Novel Use of Scanning Methods to Investigate the Performance of Screw Connections in Timber-Concrete Composite Structures

. Tis paper investigates the shear force capacity, stifness, and efective length of the connection screws in timber-concrete composite structures. Ten samples (six hardwood and four softwood) were fabricated with the connection screws installed at diferent angles through the interface. Te shear force capacities and the global stifness characteristics of the connections were determined directly from double shear tests. Te local characteristics of the screw connections were investigated by scanning the fnal residual screw shapes at the end of the tests for softwood specimens. Using the 2D digital scans of the screws, the screw curvatures were determined. From the curvatures, the local distribution of moment along the screw embedded within the concrete at the conclusion of the test was estimated. Te distance of the plastic hinge in the screw within the concrete from the interface between the concrete and timber (the efective length) was obtained from the maximum bending moment location calculated via this image scanning method. Empirical equations of efective screw length were developed from the test data and applied in a shear force capacity model for softwood. Tese new equations of efective length of inclined screws in connections predicted the shear force capacity of the connection better on the softwood specimens. In hardwood specimens, the screw failed in snapping. An equation of shear force capacity was developed based on the infuence of the inclination angle of the screw with the reduction factors and can predict the shear capacity of the connection in hardwood specimens.


Introduction
Timber-concrete composite (TCC) systems are used in a variety of applications including fooring systems for multistory buildings and bridge decks [1][2][3][4][5][6][7]. TCC makes use of the concrete to resist compression and the timber to carry tension [8,9]. Te stifness and strength of TCC structural systems in fexure are dependent on the performance of the shear connectors [10][11][12][13]. If these shear connectors do not perform well, large defections and deformations will manifest in the structural system and potentially cause safety and serviceability failures [14][15][16][17]. Many types of shear connections can be used, such as glued joints [18] and mechanical fasteners (e.g., screws, steel dowels, notches, and plates) [19]. For more information on TCC structures, we see the state-of-the-art review articles by Yeoh et al. [2] and De Araujo et al. [20]. Tis study investigates the global shear performance of screwed connections in TCC. Eurocode 5 (EC5) is often used as a guide for those wishing to design screw connections in TCC structures [21]. However, there are limitations in the guidance given in EC5 with respect to TCC structures, as the formulations for predicting shear force capacity in connections in single and double shear (Figure 1(a)) are based on data from timber-timber (T-T) connections, as shown in equations (1) and (2) (formulations presented in CEN [21] EC5 cl. 8.2.2). Tere is also no consideration of the efect of plastic hinges on the screw in the basic formulation of shear force capacity for the screw connection.
For a fastener in a T-T system in single shear, For a fastener in a T-T system in double shear, where F V � shear force capacity; f h,1,k � embedment strength for middle timber; f h,2,k � embedment strength for edge timber member; t 1 � thickness of middle timber; t 2 � thickness of edge timber; and d � screw diameter. We note that a small number of previous research studies have attempted to predict mechanical properties of screwbased timber-concrete composite shear connectors. Symons et al. [11] discussed a model to calculate the slip moduli of inclined screw timber-concrete composite connections, assuming that the screw behaves as a beam on a 2-D elastic foundation with negligible deformation within the concrete. In a separate study, Symons et al. [12] presented an upper bound plastic collapse predictive model for screw connections strength, assuming that the screws behave perfectly plastically, and the concrete remains undamaged. Research by Moshiri et al. [22] also introduced a predictive model for the screw connection strength in crossed or X-formation where the screws resist shear tension and shear compression stresses, while the concrete remains undamaged. In 2020, Du et al. [23] proposed the stifness model of the inclined screw connection in timber-concrete structures. Te model was developed by assuming that the concrete remained undamaged. Te screw connector within the timber and concrete was modelled as a semiinfnite beam method and included the inclination angle in the stifness model. Mirdad and Chui [24] proposed the stifness prediction of the timber-concrete composite connection with inclined screws and a gap. In their stifness prediction model, all the assumptions and the early hypotheses were focused on the local characteristics of the screws embedded within the timber. Te models were developed without any signifcant efect on the screw embedded within the concrete. Mirdad and Chui [24] also proposed analytical models of the shear strength of the screw connections within the timber to the concrete. Mirdad and Chui's models used the basis of Johansen's theory in their models to come up with the model that was able to estimate the strength of the inclined screw connection in the timber to concrete structures. Tis model also ignores any signifcant efect of the concrete on the embedded screw. In summary, EC5 and the predictive models from previous works all ignored any signifcant efect of the concrete on the connection behaviour.
Based on Gelf et al. [25], the embedment strength of the screw embedded within the concrete is infuenced by the efective length of the screw. Te efective length of the screw is the distance of the hinge in the screw from the interface between the concrete and timber. However, the model proposed by Gelf et al. [25] is limited to a screw angle of 90°. For screws at other angles, the distance of the hinge from the interface between the concrete and timber may also have a signifcant efect on the shear force capacity of the screw connection. It is also expected that the distance of the hinge from the interface between the concrete and timber is infuenced by the screw angle. To obtain the value of embedment strength for the screw embedded within the timber, the formulations from CEN [21] EC5 clause 8.3.1.1) were used as shown in equation (3) (for softwood) and 4 (for hardwood). While to obtain the embedment strength for screws embedded within the concrete (f h,c ), the formula of embedment strength from the ACI-318 standard can be used and shown in equation (5) [26]: where ρ k � density of timber and f c � concrete strength.
In predicting the shear force capacity of the screw connections with various angles, the estimation does not only rely upon force due to the embedment strength (F h ). Te strength of the shear force capacity also came from the withdrawal strength of the screw (F ax ) embedded within the concrete and timber. Te combination of embedment strength and withdrawal strength of the screw is the total 2 Advances in Civil Engineering shear force capacity of the screw connection (equation (6)). Te formula to calculate force due to embedment strength in this study is shown in equation (7) which is also based on the EC5 code [21]. Te formula for the withdrawal strength of the screw embedded within the timber can be obtained by using equation (8) (in timber) and equation (9) that are taken from the EC5 [21]. Te withdrawal strength of the screw embedded within the concrete can be determined by using equation (8) (in concrete) and equation (10) (bond stress between concrete and screw) which are adapted from Eurocode 2 (EC2) [27] F where Post-test observations of inclined screw TCC connections reported by Sebastian et al. [28] suggested that the screws can deform quite signifcantly in the concrete under longitudinal shear loading as shown in Figure 2. Terefore, in this research, the local deformation characteristics of the screws themselves were studied to better understand the global performance of TCC systems. Te data from two series of TCC shear test specimen tests, one set using hardwood and the other using softwood, are combined in this paper. A novel scanning method was then used to analyse the deformed shapes of the screws at the end of the softwood tests and determine the efective length of the screws embedded in the concrete.
Te main aim of this study was to combine data from two series of timber-concrete composite connection tests, one series using hardwood and one using softwood, and to develop a model to predict the interface capacity. Te secondary aim of the research was to determine the local bending moment distributions along the screw connectors in timber-concrete composite structures. For further information on the results of this experimental program and the experimental methodology, we need to see Ducas [19] and Bin Mohd Snin [29].

Specimen Fabrication.
In the frst series of tests, six double-shear hardwood (Bauchee) specimens were constructed. Tis was later followed by a second series of tests on four double-shear softwood (Spruce) specimens. In both cases, the specimens had a variety of screw inclination angles θ. Te screws were generally installed in an X-formation (as shown in Figure 3) to ensure tension and compression that were developed when a shear force was applied to the specimen, except for three cases where parallel fasteners (at either 45°or 90°angles) were used.
Te screw fasteners used in this work were partially threaded (PT) screws, 6 mm in diameter and 210 mm long. When installing the screws in the hardwood specimens, predrilling was required, but this was not necessary for the softwood specimens. Table 1 shows the identifcation codes for the tested specimens, the inclination angle, and the formation of the screws. Te location of the screws and geometry of the specimens are shown in Figure 3.

Material Properties.
Tensile testing was performed on a single screw according to ASTM [30]. Figure 4 shows the stress-strain curve obtained from this testing. Young's modulus of the screw fastener was calculated from the slope in the linear zone of Figure 4 as 222.9 GPa. Table 2 shows the interpolation functions ftted to the data for the "linear zone," "plastic zone," and "strain hardening zone." Tese functions were used in the subsequent analysis to calculate the stresses across the screw in both compression and tension (Section 4.2). Table 3 shows the results of the ASTM [31] tests on cylindrical specimens of the concrete used in the experimental work. Figure 5(c), were installed on the timber at the locations shown in Figure 5(b), to measure the slip between the concrete and timber. All the specimens were tested using a 500 kN capacity Dartec-2420 machine (1998 model), Figure 5(a), in displacement control (1 mm/min). Boundary conditions of a pinned support on one side of the specimen and a roller support on the other side ( Figure 5(b)) were used to help eliminate the friction between the timber and the concrete when the test was running and thus avoided any frictional efects contributing to the measured shear force capacity. Data logging was conducted using a Visha 8000_3 at a sampling rate of 10 Hz. Testing was continued until the specimens showed signifcant signs of failure (e.g., concrete cracking or a loud breaking noise).

Shear Force Capacity and Slip Modulus
BSI [32] gives the following equation for calculating the global stifness (Ks) for TCC structures: where n � 0.4 or 40 (serviceability limit state), 0.6 or 60 (ultimate limit state), and 0.8 or 80 (which is used to calculate the ductility, see: Stehn and Johansson [33], and where F v,max � maximum load and δ n � displacement of the connection at the particular point. Te results of shear force capacity against displacement for all the tests are shown in Figure 6. Te maximum shear capacities from the plots were used to investigate the local characteristics of the screws later in this paper. Table 4 summarises the test results, giving the measured stifness   Advances in Civil Engineering (using equation (11)) and shear force capacity (capacity per side) for the ten specimens. For the hardwood specimens, specimen X45 h exhibited the stifest connection at 74.4 kN/mm, 70.9 kN/mm, and 64.3 kN/mm for K s40 , K s60, and K s80, respectively. Te relatively small variation in these values indicates that there was relatively little nonlinear behaviour. For the softwood specimens, X45 s exhibited the stifest connection at 37 kN/ mm, 26.8 kN/mm, and 18.3 kN/mm for K s40 , K s60, and K s80 , respectively. Overall, when the screws were inclined at 45°,    Advances in Civil Engineering the stifness of the connection was at its greatest. For further analysis of the screws, shear force capacities were divided by two to obtain the maximum shear force capacity in a single screw. Te failure mode of the hardwood specimens was observed, and it shows that the screw was snapped as shown in Figure 7 for specimens P45 h and P90 h. From this observation, the higher concrete grade (Mix 1 : 26.5 MPa) in hardwood specimens causes the connection to fail as it snaps at the screw between the interface of concrete and timber. For this reason, a batch of the lower concrete grade (Mix 2 : 16.36 MPa) to the timber specimen was fabricated to see the signifcant efect of the concrete on the location of the plastic hinge on the screw in the specimens. Compared to the hardwood specimens, the connections in the softwood specimens failed in a ductile manner (Figure 8). During the tests, the screws engraved the timber to make the channels. At this time, the slip of the screw connections in softwood specimens increased more than in hardwood specimens. Tis situation makes the screw connections in softwood specimens behave in a ductile manner. Meanwhile, the screw embedded in the concrete remained in its original position with some deformation. Te main reason for this situation happening in the timber is due to its low density and fexural strength. Te length of the channels made by the screw depended on the screw inclination ftted to the specimen. Figure 8 also shows the diferent lengths of channels caused by the screw according to the angles of 45, 60°, 75°, and 90°. It was found that the channel length decreased when the angle of the screw decreased from 90°to 45°. Tis suggests that in practice, the 90°screw will produce more signifcant deformation compared to its other inclined counterparts.

Teoretical of the Hardwood Connection Failure Mode.
Screw connections in hardwood specimens failed as they snapped on the screw. Tis failure mode can be implicated from Figure 9 where the shear force capacity of the connection is calculated based on the shear stress of the crosssectional area of the screw. Variety in the inclination angles may infuence the total cross-sectional area of 90°to 30°s pecimens (equation (12)).
Tis shows that the infuence of the angle will change the value of d ′ as shown in equation (13). In order to determine the cross-sectional area of the screw, we utilize the oval area equation, as indicated in equation (12). For hardwood samples, the shear force capacity of the connection can be given as equation (14). Based on the RCSC [34], nominal shear strength of the bolt is expressed as in equation (15).
Normally, the shear strength of the bolt is 60% of the tensile strength. In RCSC [34], it was suggested that the shear strength of the bolt τ s is taken as 0.625 of the tensile strength of the bolt σ y(s) . Reduction factor C a (0.75) is applied on equation (14) due to nonuniform force distribution between bolts and in a long joint and the minor second-order efects such as those resulting from the action of the applied loads on the deformed structure should be accounted for through a second-order analysis of the structure RCSC [34]. Another reduction C b (0.8) is applied to the shear strength to account for the reduction in shear strength for a bolt with threads included in the shear plane but calculated with the area corresponding to the nominal bolt diameter (RCSC). Finally, the shear force capacity of the connection in a hardwood specimen can be obtained from equation (16). Tis shear force capacity indicates the strength of the connection in the hardwood-concrete specimens given as follows:

Generation 2D View of Screws from Coordinates.
Using the results from sample X45s, an example of how the bending moment distribution along the screw embedded within the concrete was established is shown as follows. Tis novel scanning method was used to determine the bending along the full length of the screws at the end of the tests. At the completion of testing, the screws were cut out of the test specimens as carefully as possible to minimize any changes or variations in the screw displacements. A portable coordinate measuring machine (CMM) scanner (a FARO Model 14000 3D scanner) was then used to measure the screw shapes at the end of the tests. Te coordinates obtained from the CMM scanner were imported into MATLAB to create the scanned images shown as a two-dimensional view in Figure 10.
Tese scans were then analysed further to extract the distorted shapes of each screw (e.g., Figure 11). Figure 10(a) shows that the tension screw in specimen X45s (the leftmost screw) has no hinges along the section embedded within the concrete (the section towards the head of the screw). However, it was found that a hinge did occur in the other tension screws, and the location of this hinge moved closer to the screw head as the screw angle increased. It was also found that only a single hinge occurred in the part of the screw embedded within the concrete. Figure 10(b) shows the scanned image for the compression screws. Te top hinge in the 45°screw was closer to the screw head compared to that in the screws in the other specimens. Te hinges in the compression screws occurred when the screw was being compressed in the concrete, and this resulted in more signifcant curvatures compared to the tension screws. Te top   Figure 9: Infuence of the inclination angle on the total cross-sectional area.
hinge was located further away from the screw head as the angle of the compression screw increased from 45°to 90°. Tis was the reverse of the change in the tension screw hinge location, which got closer to the screw head as the screw angle increased. Figure 11 shows the extracted deformation data for the compression screw in specimen X45s. Te curvature formula given in equation (17) was then used to determine the curvature at every point along the length of the screw.

Analysis of the Scanned Coordinate Data for the Compression Screw (X45s).
For the compression screw in specimen X45s, the maximum curvature was −0.002 mm −1 and the distance of the plastic hinge from the timber-concrete interface was calculated as 22.5 mm. Te value of curvature obtained is useful as it can be used to calculate the fbre strains across the screw section ( Figure 12).
By knowing the distance of the edge fbre to the middle of the screw (hsi), it is possible to fnd the extreme fbre stress (εb2) using the following equation: Unfortunately, it was not possible to measure the change in diameter of the screw when the deformation was plastic; thus, the curvatures during plastic deformation were also calculated using equation (18). For the value of max curvature in the screw within specimen X45s, the max strain was calculated as −0.006.  Te local bending moments along the screw length embedded within the concrete were also obtained from the measured curvatures. To allow the local bending moments to be calculated when the screw had reached a plastic or strain hardening phase, a plot of moment against curvature was produced as shown in Figure 13 and based on equations (19)- (21). Tis plot incorporates the nonlinear stress-strain curve ( Figure 4 and Table 2) and uses a strip analysis method. Te fgure was created by calculating the stress profle for every curvature, then for each strip across the section, using the relevant moment equations equations (19)- (21) and summing the values for all the strips to get the total section bending moment. Figure 13 was then used to convert the curvatures into moments at every point along the screw length embedded within the concrete as follows:  Figure 14 shows the local bending moment distribution for the compression screws embedded within the concrete in specimens X45s, X60s, and X75s calculated using the method mentioned previously.

Plastic Hinge Locations.
Using the scanned shape of the screws at the end of the tests, it was also possible to accurately determine the location of the plastic hinges in the screws (Figure 11 for location of plastic hinges l c and l t ) Tables 5 and 6 show computed plastic hinge locations. Te distance of the plastic hinge in concrete and timber from the timber-concrete interface is named l c(scan) and l t(scan), respectively. Tables 5 and 6 also show an increasing distance for the location of the plastic hinge from the timber-concrete interface as the angle of the compression screw reduces from 90 to 45 degrees. But with the tension screw embedded within the concrete, the location of the plastic hinge at the timber-concrete interface decreased as the angle of the screw reduced from 90 to 45 degrees. Contrary to the tension screw embedded within the timber, the distance location of the plastic hinge from the timber-concrete interface increased as the angle of the screw reduced from 90 to 45 degrees.
To calculate the shear force capacity of the interface, the embedment strength of the screw must be calculated. Based on Gelf et al. [25], the embedment strength of the screw within the concrete is the concrete bearing stress multiplied by the efective length l c . However, the model presented by Gelf et al. [25] was developed only for screws at an angle of 90°. Tese new tests extend this previous work by allowing an investigation of the efective length of the screw in the concrete and timber for diferent screw angles. Te ratio between the plastic hinge distance l c(scan) and the length of the fastener in concrete L c is defned as the plastic hinge ratio ∅ Hc (equation (22)). Te distance of plastic hinge from the timber-concrete interface is used to quantify the efective length of the screw l c . A plot of the plastic hinge ratio against the angle of the screw in concrete is shown in Figure 15 to investigate the efect of the screw angle on the plastic hinge distance. Figure 15 shows that for fasteners in compression, the plastic hinge ratio decreases when the angle increases from 45 to 90 degrees and increases for fasteners in tension when the angle increases from 45 to 90 degrees. Using a linear ft to the data presented in Figure 15, equations for plastic hinge ratio for the fasteners in tension and compression, respectively, are presented in the following equations: ∅ Hc(t) � 0.0031θ − 0.11 (screw in tension), l c can be determined, as shown in equations (25) and (26), respectively. Tese equations for l c can then be used to calculate the shear force capacity of the interface.
For tension screw (for the softwood tests), For compression screw, Te plot of the plastic hinge ratio against the angle of the screw in timber was also made and shown in Figure 16. Figure 16 shows that the plastic hinge ratio for fasteners in compression decreases when the angle increases from 45 to 90 degrees. From the plot, the equations for the plastic hinge ∅ Ht(c) � −0.0041θ + 0.43 (Screw in compression). (28) l t can be determined, as shown in equations (29) and (30), respectively. Tese equations can then be used to calculate the shear force capacity of the interface.
For the tension screw (for the softwood tests),

Advances in Civil Engineering
For the compression screw, Te degree of accuracy for equations (25), (26), (29), and (30) has been measured and shown in Tables 7 and 8. Table 7 shows the degree of accuracy for equations (25) and (26) for screws embedded within the concrete. It was found that the percentage diference between equation (26) (compression) to scanned values for all softwood specimens is less than 12%. Te values calculated by using equation (25) (tension) are also compared to the scanned values and show that the percentage diference is less than 60%. However, the values     (25), (26), (29), and (30) accuracies.

Validation of the New Efective Length of Shear Force
Capacity Equations. Te new efective length formulations (equations (25), (26), (29), and (30)) were used to calculate interface shear force capacities (using equation (6)). Equations (25), (26), (29), and (30) (generated for softwood test data) were taken to be valid for the interpretation of the softwood tests. Table 9 shows the results of shear force capacity for both approaches of l c and l t calculations. Most of the values of shear force capacities calculated using this study's efective length equations were closer to the measured shear force capacities. Tis can be seen in Figure 17, where the softwood data obtained by the new equations (efective length l c and l t ) are closer to the line of equality. Terefore, it is suggested that the efective length method proposed in this study is used when it is desirable to have a good prediction of the interface shear force capacity for the softwood specimens. Meanwhile, the hardwood data obtained by equation (16) (see Table 10) is quite overestimated from the line of equality. More reduction factors may need to be included in the proposed equation (16) to obtain more factors that infuence the shear force capacity in hardwood specimens. Tis research suggests future work to discover insight into the local behaviour of the screw connections in hardwood specimens by investigating the local bending moment, axial load, and deformation along the screw within the specimens.