Abstract
A separable glued-laminated timber (GLT, Larix kaempferi Carr.)-steel beam system is presented in this work for easy recycling at the time of disposal. The minimum thickness of steel required to induce compressive GLT failure was assembled with GLT by inclined screws. In a total of 8 GLTs, 3 GLTs were not reinforced (control group), and 5 GLTs were reinforced with steel plates (comparison group). In the GLT in the comparison group, a steel plate (SPHC, yield strength: 227 MPa, modulus of elasticity 166.33 GPa) was installed with screws (∅9x160mm, 45°). The deflection and load of specimens were measured by a third-point bending test to derive their bending stiffness and load-carrying capacities. All specimens in the control group showed brittle tensile failure, but all specimens in the comparison group showed ductile behavior and maintained a load-carrying capacity of about 30 kN. After the compression failure of the GLT, there was no damage to the screw connection, while the steel plate was extended. Based on the behavior of the steel, a GLT-steel beam prediction model was developed, similar to the structural design method for reinforced concrete.
Download PDF
Full Article
Bending Behavior of Separable Glued-Laminated Timber (GLT)-Steel Beam Combined with Inclined Screws
Sung-Jun Pang,a and Jung-Kwon Oh a,b,*
A separable glued-laminated timber (GLT, Larix kaempferi Carr.)-steel beam system is presented in this work for easy recycling at the time of disposal. The minimum thickness of steel required to induce compressive GLT failure was assembled with GLT by inclined screws. In a total of 8 GLTs, 3 GLTs were not reinforced (control group), and 5 GLTs were reinforced with steel plates (comparison group). In the GLT in the comparison group, a steel plate (SPHC, yield strength: 227 MPa, modulus of elasticity 166.33 GPa) was installed with screws (∅9x160mm, 45°). The deflection and load of specimens were measured by a third-point bending test to derive their bending stiffness and load-carrying capacities. All specimens in the control group showed brittle tensile failure, but all specimens in the comparison group showed ductile behavior and maintained a load-carrying capacity of about 30 kN. After the compression failure of the GLT, there was no damage to the screw connection, while the steel plate was extended. Based on the behavior of the steel, a GLT-steel beam prediction model was developed, similar to the structural design method for reinforced concrete.
DOI: 10.15376/biores.18.2.3838-3855
Keywords: Glued-laminated timber; Steel; Composite beam; Bending performance; Reinforcement
Contact information: a: Department of Agriculture, Forestry and Bioresources, Seoul National University, Seoul, Republic of Korea; b: Research Institute of Agriculture and Life Sciences, Seoul National University, Seoul, Republic of Korea; * Corresponding author: jungoh@snu.ac.kr
INTRODUCTION
Carbon emission reduction has become a key topic in the construction industry. Timber construction has a low-carbon emission due to the light weight and easy cutting of the timber elements (Gerilla et al. 2007; Yan et al. 2010; Hafner and Schäfer 2018; Sandanayake et al. 2018; Li et al. 2019; Röck et al. 2020). In addition, timber retains carbon absorbed as the tree grows. Structural timber within a building will store the carbon as long as the building is maintained. Thus, the use of timber as a structural element contributes to the atmospheric carbon emissions reduction of buildings (Petersen Raymer 2006; Resch et al. 2021; Pang et al. 2022).
Structural timber is used as a beam due to its high bending strength compared to its weight. The design value of structural timber is based on the 5% lower limit, which is the 5th value from the lowest strength when 100 destructive tests are conducted (Pang et al. 2013, 2020). This is because a high safety factor is required for timber due to brittle failure. If the timber is subjected to ductile failure rather than brittle failure, then the safety factor can be lowered. Steel is a typical structural material with good ductility. The combination of steel and glued-laminated timber (GLT) is used in dry construction and the components are recyclable (Tsai and Le 2018). In addition, both materials are cut or drilled in the factory using CNC cutting machines (Lee et al. 2021). Thus, the machining precision is high and the machining error is small (Tolerances: ±1mm range). This means that the two materials have many advantages that can be used as a hybrid.
Several researchers tried to develop timber-steel hybrid beams (Moritani et al. 2021). André (2006) showed that when the GLT is reinforced by steel, the bending performance of the GLT beam is improved and the deviation is reduced. Riola Parada (2016) showed that the creep of a GLT beam reinforced by steel was around 85% (after one year) and 50% (after the fifth year) of the GLT beam. Loss et al. (2016a, b) used an I-shape steel beam to avoid the brittle failure of the timber beam and timber floor in the fabrication of modern buildings. Tsai and Le (2018) reinforced the web of an I-shaped steel beam with timber using screws, and the strength of the composite beam was improved by about 15%. Wu et al. (2021) developed an I-shaped wood-steel composite beam and reported that an elastic-plastic phase appeared as a result of the bending performance test. As mentioned above, researchers revealed the advantages of timber-steel composite beams. However, the structural behavior of the hybrid beam was designed and analyzed in the elastic range.
When a GLT beam is glued with other materials (steel or fiber-reinforced polymer (FRP)) by an adhesive, the risk of glueline delamination has been reported (Metelli et al. 2016; Schober et al. 2015), and it is hard to expect good durability. Tomasi et al. (2009) tried to develop a ductile bending behavior of GLT beams by inserting steel bars. They showed that steel reinforcement provides a simple and reliable solution for robust timber structures. Wang et al. (2021) developed a glulam beam mechanically reinforced by a steel rod. They showed the possibility of ductile failure due to the yielding of glulam in compression parallel to the grain or yielding of the steel rod.
As mentioned above, most GLT-steel beams were developed by gluing GLT and steel. However, this study developed a separable glued-laminated timber (GLT)-steel beam without gluing as an eco-friendly solution to building construction. The GLT-steel beam was assembled using inclined screws without gluing, which enabled the easy separation of the materials for recycling purposes. The minimum thickness of steel for inducing compressive GLT failure was investigated, and the reinforcing effect and bending behavior of the developed GLT-steel beam were analyzed. Based on the experimental test, a prediction model for the load-carrying capacity of the GLT-steel beam has been developed.
DESIGN OF GLT-STEEL BEAM
Figure 1 shows schematically the concept of the separable GLT-steel beam. In this study, the GLT-steel beam is designed to fail by a universal test machine with a load capacity of 100 kN, based on the previous study (Pang et al. 2018). The load-carrying capacity of the GLT-steel composite beam was designed by applying the composite beam theory, under the assumption that GLT and steel are completely combined and linear elasticity materials. Figure 2 shows the load-carrying capacity, neutral axis position, and total thickness according to the steel plate thickness when a GLT (80 mm (width) × 120 mm (thickness) × 2450 mm (length)) is reinforced with steel plates. As the thickness of the steel plate increases, the thickness of the composite beam increases and the neutral axis goes down toward the bottom.
In the case of load-carrying capacity, as the thickness of the steel plate increased, the load that the composite beam could support also increased, but the increase rate of the load decreased when the thickness was 4 mm or more. When GLT was reinforced with a steel thickness of 3 mm or less, the maximum load of the beam was governed by the tensile failure of the steel plate. However, when GLT was reinforced with a steel thickness of 4 mm or more, the maximum load of the beam was governed by compression failure of GLT. Therefore, in this study, the GLT was reinforced with a 4 mm thick steel plate to check whether compression fracture of GLT occurred.
Fig. 1. Separable glued-laminated timber-steel beam
Fig. 2. Load-carrying capacity, neutral axis position, and total thickness of beam according to the steel plate thickness
EXPERIMENTAL
Materials
Figure 3 shows the cross section of GLT and the combination of lamina. One lamina grade (E9) was used, and the modulus of elasticity (MOE) of that grade was greater than 9 GPa. The GLTs were manufactured using four layers of larch species (Larix kaempferi Carr., the density of 550 kg/m3) lamina according to the KS F3021 standard (KS F3021 2013). The size of GLT was 80 mm (width) × 120 mm (thickness) × 2450 mm (length). The GLT specimens were stored for one week in a bending test laboratory (20 °C, 65% R.H.). After the bending test, the moisture content of GLT was measured with a moisture tester (HM-530, Kett Electric Laboratory, Tokyo, Japan) and was about 6 ± 2%.
Table 1 shows the dimensions and mechanical properties of GLT and steel plate. Eight GLTs were prepared, and all of the actual bending stiffness of the GLTs were measured. GLTs were divided into two groups to have similar bending stiffness. One group (3 out of 8) was used to measure the bending strength of GLT, and the other group (5 out of 8) was reinforced with a steel plate.
Fig. 3. Cross section of GLT and GLT-steel beam
Table 1. Dimensions and Mechanical Properties of GLT and Steel Plate
1) Glued-laminated timber (KS F3021 2013)
2) Experimental value in this study
3) Steel Plate Hot Commercial (KS D3501 2018)
4) Experimental value from Suh et al. (Chang-Min and Hyun-Chul 2009)
5) Experimental value from mill test certificate
To reinforce the GLT, a 4 mm thick steel plate (Steel Plate Hot Commercial, SPHC (KS D3501 2018)) was installed at the bottom of the GLT. The steel plate and GLT were fixed with self-tapping screws (VGS ∅9 × 160mm (ETA-11/0030 2019)). Axial slip moduli of screws are significantly higher compared to slip moduli in shear (Dietsch and Brandner 2015). Thus, the screw was installed at an angle of 45° using a washer (HUS945, Rothoblaas) to minimize the slip between the GLT and the steel plate.
When the GLT is reinforced with a 4 mm steel plate, the maximum load of the composite beam is about 20 kN, and the shear force generated between the GLT and steel is 84.7 kN. The design value of the inclined screw used is 11.25 kN (Rotho Blaas Srl 2020) and 8 screws can support the shear force generated between the GLT and steel plate. However, 10 screws were installed at intervals of 60 mm to prevent the screw connector failure and to observe the bending failure of the GLT-steel beam. The axis of the screw was rotated 20 degrees toward the center of the GLT to prevent the GLT from being cracked by the screws, as shown in Fig. 4.
Fig. 4. Screw installation (45° inclined, 20° rotation)
Experimental Tests
The deflection and load resistance of specimens were measured by a third-point bending test to derive their bending stiffness and load-carrying capacities (Fig. 5). The load was applied using a universal test machine (Zwick GmbH & Co., Ltd., Ulm, Germany) according to ASTM D198 (2010). A yoke was installed on the neutral axis of the specimen to measure the pure deflection.
Fig. 5. Configuration of third-point loading test for the specimens
The actual displacement was measured at the center of the specimen using the linear variable displacement transducers (LVDT). The test span and the loading speed were 2,280 mm and 10 mm/min, respectively. The bending stiffness and bending moment of the specimens were calculated using Eq. 1 (BS EN 408:2010+A1:2012 2012; Pang et al. 2019; Pang and Jeong 2019) and Eq. 2, respectively,
(1)
(2)
where (EI)test is the experimental bending stiffness of the specimen (N·mm2), Le is thedistance between the load position and support positions (mm), L is thespan of the specimen (mm), P1 and P2 are theloads corresponding to 10% and 40% of the maximum load, Pmax, respectively (kN), d1 and d2 are thedeflections corresponding to P1 and P2, respectively (mm), and Mmeasured is the measured bending moment capacity (kN·m).
RESULTS
GLT Beam
Figure 6 shows the failure modes of the GLT beam. All specimens were broken around the knot in the tensile zone. At the time of failure, the load resistance dropped sharply, as shown in Fig. 7. This brittle failure is a typical failure mode of GLT or wood in bending tests (Pang et al. 2011, 2018, 2021; Pang and Jeong 2019).
Table 2 shows the bending properties of the GLT specimens. The average bending stiffness of the GLT in the failure test was 0.168 × 1012 N∙mm2, which was similar to the stiffness in the grading test. The average load-carrying capacity of the GLT was 28.2 kN, and the bending moment capacity was 10.7 kN∙m.
(a) Tensile failure of the bottom lamina
(b) Tensile failure around knot at the bottom lamina
Fig. 6. Failure mode of the GLT beam
Fig. 7. Load-displacement curve of the GLT beam
Table 2. Bending Properties and Failure Modes of GLT and Reinforced GLT
1) Effective bending stiffness by Eq. 1
2) Measured in grading test to divide the specimens into two groups.
3) Measured in failure test to evaluate the maximum load-carrying capacity of the specimen
4) (EIeff in grading test / EIeff in failure test) × 100
5) Maximum load
6) Bending moment resistance calculated by Eq. 2
7) Coefficient of variation
GLT-steel beam
Figures 8 through 10 show the failure mechanism of the GLT-steel beam. In two of the five specimens, tensile failure occurred first at the finger joint (Fig. 8a) or knot (Fig. 9a) of GLT. The finger-joint and knots are the main factors causing the strength reduction (Pang et al. 2011, 2018, 2021; Pang and Jeong 2019). Unlike the GLT beam, compression failure occurred at the top of the GLT (Fig. 8b and Fig. 9b) and large deformation occurred (Fig. 8c and Fig. 9c).
In three of the five specimens, compressive failure occurred first at the top of the GLT (Fig. 10a). After that, tensile failure occurred at the knot, but the tensile failure of the specimens did not lead to rapid brittle failure. The specimens showed ductile behavior due to the elongation of the steel plate. As described in the material section, the number of screws was calculated and installed to sufficiently support the expected shear force between GLT and steel plate. During the experiment, no screw was pulled out, and no damage was found between the screw and the steel plate. Therefore, the elongation of the steel plate shows that the screw connection sufficiently supported the shear force between GLT and the steel plate.
Figure 11 shows the load-displacement curve of the GLT-steel beams. Unlike the GLT beam, the GLT-steel beam did not lose its load-carrying capacity significantly after the initial failure and maintained a load-carrying capacity of approximately 30 kN. This may be because the steel plate is in a plastic state in which the stress does not increase even when the strain increases. The maximum load-carrying capacity was higher in the case where the compression failure occurred first (48 kN to 51 kN) than in the case where the tensile failure occurred first (32 to 38 kN). This shows that the tensile part (bottom of GLT) contributed to the load-carrying capacity of the GLT-steel beam until the tensile failure occurred.
(a) Tensile failure at finger joint of GLT
(b) Compression failure on top of GLT
(c) Large deformation
Fig. 8. Fracture mechanism of GLT-steel beam in which tension failure (finger joint) of GLT occurs first (RGLT-1)
(a) Tensile failure at knot of GLT
(b) Compression failure on top of GLT
(c) Large deformation
Fig. 9. Fracture mechanism of GLT-steel beam in which tension failure (knot) of GLT occurs first (RGLT-2)
(a) Compression failure on top of GLT
(b) Tension failure on bottom of GLT
(c) Large deformation
Fig. 10. Fracture mechanism of GLT-steel beam in which compression failure of GLT occurs first