Abstract
This study elucidated the mechanical performance of different connections used in orthogonal ribbed beams made of poplar laminated veneer lumber (LVL). Three types of connections (namely, U-, T-, and L-shape) were fabricated and used to connect the second beam and the middle part of a main beam to form an orthogonal rib beam. A concentrated load was applied to the intersection between the main beam and the second beam. The results showed: (1) all three types of ribbed beam connection specimens showed good connecting performance and the ductile failure, including the pull-out of partial self-tapping screws, crushing of LVL at the end of the ribbed beam, and connection failure; (2) the variation of load capacity of JD2-type specimens was approximately 3.2%, which was smaller than that of the JD1- and JD3-type specimens, slightly less than 10%; (3) the ultimate bearing capacity of a connection specimen was proportional to the number of self-tapping screws; and (4) the mechanical performance of the U-shaped connection specimens was worse than that of the L-and T-shaped connection specimens. The L-shaped connection was recommended due to its better mechanical performance, simpler configuration, and more convenient fabrication.
Download PDF
Full Article
Mechanical Performance of Three Types of Connections Used in Orthogonal Ribbed Beams Made of Poplar Laminated Veneer Lumber
Yan Liu,a,* Jing Chen,a Hongwei Ma,a,* Meng Gong,b and Linfeng Zhang a
This study elucidated the mechanical performance of different connections used in orthogonal ribbed beams made of poplar laminated veneer lumber (LVL). Three types of connections (namely, U-, T-, and L-shape) were fabricated and used to connect the second beam and the middle part of a main beam to form an orthogonal rib beam. A concentrated load was applied to the intersection between the main beam and the second beam. The results showed: (1) all three types of ribbed beam connection specimens showed good connecting performance and the ductile failure, including the pull-out of partial self-tapping screws, crushing of LVL at the end of the ribbed beam, and connection failure; (2) the variation of load capacity of JD2-type specimens was approximately 3.2%, which was smaller than that of the JD1- and JD3-type specimens, slightly less than 10%; (3) the ultimate bearing capacity of a connection specimen was proportional to the number of self-tapping screws; and (4) the mechanical performance of the U-shaped connection specimens was worse than that of the L-and T-shaped connection specimens. The L-shaped connection was recommended due to its better mechanical performance, simpler configuration, and more convenient fabrication.
DOI: 10.15376/biores.17.3.4638-4655
Keywords: Poplar LVL; Connection; Orthogonal ribbed beam; Stiffness; Bearing capacity; Failure mode; Load distribution
Contact information: a: College of Civil Engineering, Yangzhou University, Yangzhou, Jiangsu 225127 P. R. China; b: Wood Science and Technology Center, University of New Brunswick, Fredericton, New Brunswick E3C 2G6 Canada; *Corresponding authors: hwma@yzu.edu.cn; liuyan@yzu.edu.cn
INTRODUCTION
Laminated veneer lumber (LVL) is a new type of building material made of fast-growing poplar wood veneer and adhesives (Liu et al. 2013; Xiong et al. 2018). With the proper selection of the wood veneer, layups, adhesive type, and hot-press bonding parameters, adhesively laminated wood products can be designed and manufactured to meet the needs of human beings. Many scholars have carried out research on the physical and mechanical properties of poplar LVL. Liu et al. (2017) measured the tensile strength, compressive strength, and other physical and mechanical properties of poplar LVL, and found that poplar LVL had relatively uniform material properties and small variation, making it a suitable load-bearing material for building structures. Aydın et al. (2004) conducted physical and mechanical tests on poplar LVL made of two wood species (eucalyptus and beech) and two adhesives (urea formaldehyde and polyvinyl acetate) and found that the type of adhesives resulted in different strength characteristics of the LVL panels. Ido et al. (2010) tested the strength of LVL made from five wood species and examined the lamination direction of veneer. It was found that the strength of the LVL material was related to the direction of the wood texture during assembly, except for radiata pine. Cui et al. (2016) found that the bending property of the LVL was considerably improved when the fiber line was implanted into the fast-growing poplar LVL. Meng et al. (2016) conducted mechanical tests on LVL made of poplar and eucalyptus with different veneer thicknesses and found that the modulus of elasticity (MOE), modulus of rupture (MOR), and horizontal shear of the LVL increased as the veneer thickness decreased. The experimental results showed that the poplar LVL, as an engineering wood material, exhibited stable material structure, good mechanical properties, high reliability, and low variability.
With the development of modern wood structures, the traditional connection methods used for the construction of wood structures may not meet modern performance requirements. To address this, researchers have, over years, conducted many studies on the mechanical properties of wood connections. He and Sun (2008) and Xu et al. (2011a) tested light wood truss joints connected by serrated plates, because the tooth plate connections could be the weak link of the whole truss. They analyzed the ultimate bearing capacity, the ultimate sliding capacity, and the ultimate tensile capacity of the serrated plates, and discovered that this kind of joint exhibits high ductility and low variability. Xu et al. (2011b) studied the bearing capacity of single bolted and group bolted connections in glulam under tension parallel to the grain direction of the wood and examined their failure modes and failure mechanisms. They found that the mechanical properties of a glulam-to-glulam single bolted connection was related to the thickness ratio of the glulam, and the mechanical properties of the glulam-to-glulam bolted connections were related to the number of bolted columns and the number of bolts in each column. Lu et al. (2016) designed and fabricated 9 glulam beam-column bolted joints, 7 of which were strengthened with self-tapping screws. They conducted monotonic and low cyclic loading tests and found that glulam beam-column bolted joints easily caused wood tearing damage, and the connections strengthened with self-tapping screws could effectively avoid cracking in the wood. Xu et al. (2009) experimentally and numerically studied the behaviour of dowel-type steel-to-timber joints tension loaded perpendicular to the grain. They found two primary types of failures, i.e., splitting and embedding. They used the experimental results to validate a three-dimensional (3D) non-linear finite element model and found that the numerical results were in good agreement with the experimental results. Bader et al. (2016) investigated the behavior of connections made of slotted-in steels in LVL members bonded by single dowels with diameters of 12 and 20 mm with an aim to elucidate and subsequently simulate the load–displacement behavior of steel-dowel-LVL connections. They found that a pronounced nonlinear behavior of the single-dowel connections was observed for all load-to-grain directions. Sebastian et al. (2018) carried out double-shear tests on beech LVL-concrete composite connections based on the screw connectors singly inclined at either 45° or 90°. It was found that in the specimens with the same screw orientation, the longitudinal shear force was applied either in forward or reverse, because, in practice, concrete shrinkage, moisture-induced timber expansion and oscillatory, e.g., seismic, or dynamic loads could induce the reversal force on the connection. Li (2005) studied the performance of the composite floor composed of pre-stressed, laminated slab and steel beam under uniform load bearing. The results showed that the composite floor can work together, and the rigid connectors can share most of the tensile stress in the concrete flange plate. Lu et al. (2016) conducted monotonic and low-cycle repeated loading tests on glulam beam-column-bolted joint specimens. They found that glulam beam-column-bolted joints were prone to wood tearing damage, and applied self-tapping screws to reinforce the joint. The effect of the screw bite force effectively inhibited the cracking of the wood texture. Pathak and Charney (2008) simulated the lateral force of the floor and found that the nail connection was the primary source contributing to the in-plane flexibility, which was independent of the aspect ratio of the floor. Hassanieh and Valipour (2020) conducted the shear tests on three groups of wooden frames and discovered that the oriented strand board (OSB) panel thickness and nail size considerably affected the stiffness and strength. In summary, it could be seen from the above review that there have been limited studies on steel connections reinforced with self-tapping screws.
To fill this gap, nine full-scale connection specimens with three groups of different types of connectors were designed and fabricated. They were assembled and tested under concentrated loading, which was used to study the effect of the different connectors on the flexural performance, failure mode, bearing capacity, and stiffness. The results could provide a reference for the design and application of this type of connection in modern wooden structures.
EXPERIMENTAL
Design and Fabrication of the Specimens
Three types of poplar LVL-ribbed beam connections were designed and manufactured, each of which had 3 replicates, generating a total of 9 specimens. The design parameters of each type of connection are detailed in Table 1. Each specimen had a width of 1200 mm and a depth of 1200 mm. The inner part was composed of a primary beam and two secondary beams connected by connections to form an orthogonal ribbed beam joint. The section sizes of the three ribbed beams were 40 mm by 235 mm and the section size of the four ribbed beams of side fixing frame were 80 mm by 235 mm, which is illustrated in Fig. 1. The ribbed beams at the connection area were connected with three types of connections, i.e., U-shape, T-shape, and L shape, as shown in Fig. 2, all of which were made of Q235B steel. The mechanical performance of the Q235B material was measured by Ma et al. (2018, 2019). According to relevant Chinese standards GB/T 50708-2012 (2012) “Technical code of glued laminated timber structures” and GB50005-2017 (2017) “Standard for design of timber structures”, the screwing patterns and numbers of self-tapping screws were selected. The self-tapping screws were made of 304 stainless steel countersunk head screws with a specification of M4×20, with a yield strength of 205MPa and a tensile strength of 520MP. The sizes of the specimens are given in Fig. 2.
Table 1. Design Parameters of the 3 Types of Specimens
Fig. 1. Schematic of a connection specimen
Fig. 2. Dimensions and construction of the connectors
Fig. 3. Schematic of the experimental setup (Note: 1- Reaction frame; 2- Pressure jack; 3-Connector; 4-Pressure sensor; 5- H-Steel support; and 6-Four ribbed beam specimen)
Material Properties of Poplar Laminated Veneer Lumber (LVL)
The physical and mechanical properties of the poplar LVL were determined in a previous study by the author (Liu et al. 2017). The average moisture content of the poplar LVL was 13.1% and the average density was 0.57 g/cm3. The mechanical properties of the poplar LVL are listed in Table 2. These quantities were previously tested by the authors as well (Liu et al. 2017).
Table 2. Material Properties of Poplar Laminated Veneer Lumber (LVL)
Test Setup and Loading Scheme
A hydraulic jack was used to apply a concentrated load at the mid-span of a ribbed beam connection. The testing procedure referred to the methods for testing timber structures stipulated in the Chinese standard GB/T50329-2012 (2012) and ASTM E2322 (2015). The loading setup is shown in Fig. 3. First, the ultimate load of a specimen was estimated according to the formulas for calculating the shear bearing capacity and compression failure bearing capacity of a pin fastener provided in the Chinese standard GB50005-2017 (2017). A preload was applied to the specimen before actually starting the loading. The load was kept for 5 min when the load was added to 5% of the ultimate load and then unloaded. All measuring equipment was zeroed prior to loading. A hierarchical loading scheme was followed, i.e., the first stage was loaded with approximately 10% of the ultimate load, and the second stage was loaded with approximately 5% of the ultimate load until it was damaged. The holding time for each loading stage was 5 min.
Instrumentation
In order to measure the changes of the connection deflections and rib strains of each specimen during the loading process, the displacement meters and strain gauges were arranged at the bottom of the connection rib beam, and load cells were placed at the end of the rib beam and the corner of a specimen to measure the load distribution of a connection rib specimen. The arrangement of the measuring points is shown in Fig. 4, among which W1 through W7 represent the displacement meter numbers, C1 through C7 represent the strain gauge numbers (note: the JD2 specimens were not measured at the measurement points C2, C5, and C6 due to the connection construction), and N1 through N8 represent the load cell number.
Fig. 4. Arrangement of the measuring locations
GENERAL BEHAVIOR AND FAILURE PATTERN
JD1 specimen (U-shaped type)
When loaded to 21 kN, the specimen emitted a slight squeezing sound. When the load reached 33 kN, a crack began to appear between the plug and the slot of a connector, and the crack continued to grow as the load increased.
Fig. 5. The failure morphology of specimen JD1
When the load reached 51 kN, the self-tapping screws between Beam 2 and the slot were slightly pulled out. When the load reached 55 kN, the end 2 of the beam slid down, and the subsidence distance reached 4 mm, as shown in Fig. 5a. When the load reached 60 kN, the wood in contact with the load cell was crushed, as shown in Fig. 5b. At this time, the end of Beam B-2 was separated from the 3-axis side beam, and the clearance reached 10 mm, as shown in Fig. 5c. When the final load reached approximately 68 kN, the self-tapping screws between Beam 2 and the slot was pulled out, and the connection separation gap reached approximately 14 mm, making the specimen fully fail, as shown in Fig. 5d.
JD2 specimen (T-shaped type)
Before the load was loaded to 42 kN, the specimen showed no obvious failure. When the load reached approximately 52 kN, the specimen emitted a slight squeezing sound. When the load reached approximately 55 kN, the two T-type connectors at the connection began to separate. As the load continued to increase, the end 2 of the beam connected to the axis side of Beam A appeared to slide down, as shown in Fig. 6a.
Fig. 6. The failure morphology of specimen JD2
At this time, the right end of Beam B-2 appeared to sag 5 mm at the connection along the axis side of Beam 3, as shown in Fig. 6b. When the load reached approximately 65 kN, the gap between the end of Beam B-1 and Beam B-2 at the connection and Beam 2 was approximately 10 mm, as shown in Fig. 6c, and the spacing between the two connectors reached approximately 7 mm, as shown in Fig. 6d. Finally, the specimen failed.
When the load reached approximately 12 kN, the specimen produced a slight sound due to the crushing of the wood. When the load reached approximately 27 kN, Beam 2 produced a fiber tearing sound in the wood, and the right end of Beam B-1 and the left end of Beam B-2 began to sag. When the load reached approximately 36 kN, obvious deflection occurred at the midpoint of Beam 2. When the load reached approximately 42 kN, the tearing sound of the wood fibers became quite loud, and the self-tapping screws at the lower part of the connector at the 2 connections of the beam were pulled out, as shown in Fig. 7a. As the load continued to increase, the support at the end of Beam 2 began to slide down. When the load reached approximately 50 kN, oblique cracks appeared at the right upper end of Beam B-2 where it met the connector, and the wood was pulled apart, as shown in Fig. 7b. When the final load reached approximately 60 kN, the screws at the lower end of Beam 2 were pulled out, and the gap between Beam 2 and Beam B-1 was approximately 10 mm, as shown in Fig. 7c. Meanwhile, the end of Beam 2 connected to the axis side beam of A was compressed approximately 3 mm, and the wood at the upper end of Beam 2 was torn, as shown in Fig. 7d. The specimen finally failed.
Fig. 7. The failure morphology of specimen JD3
Analysis of failure
According to the above description on the failure of each specimen, the overall initiation and propagation of failure in a specimen was as follows: (1) pull-out of the self-tapping screws at the lower part of the specimen; (2) separation of the connection and the ribbed beam; (3) crushing of the wood at the end of the ribbed beam; and (4) failure. In summary (as shown in Fig. 8), first of all, the screws connected on both sides of the lower end of Beam 2 at the connection were pulled out. Some of the screws began to fracture, the bearing capacity of the connection began to decrease, the deflection at the connection increased, the end of each rib beam was stressed, the end of the beam slid down, and finally the end wood in contact with the support was crushed.
Fig. 8. Failure evolution of a nodal specimen
ANALYSIS AND DISCUSSION
Load-displacement Response
The load-displacement curve of each specimen was plotted (Fig. 9). The displacement linearly increased as the load increased at the initial stage of loading. When the load reached 80% of the ultimate load, the displacement growth rate of the specimen was accelerated, and the load-displacement curve becomes nonlinear. When the load continues to increase, the displacement growth rate was further accelerated due to the pull-out of some screws in the specimens. Before the failure of a specimen, the ductile behavior appears obvious. Throughout the whole loading process, it is not surprising that the midpoint of a specimen exhibited the maximum displacement. During the loading process, some screws in the mid-span connection area were pulled out, weakening the connection performance of the specimen, and reducing the load transfer from the 2-axis ribbed beam to the B axis ribbed beam. During the course of the loading process, some screws were pulled out from the connection, reducing its load capacity. Therefore, the displacements at the supports of Beam B-1 (point W4) and Beam B-2 (point W7) were smaller than the displacements at the support of Beam 2 (points W1 and W3). For specimens JD1 and JD3, the displacements of the ribbed ends in two directions were 6.10 mm and 5.49 mm, respectively, while for type JD2, the displacement of the ribbed ends in two directions was only 1.87 mm. This was because of the presence of the T-plate at the bottom of specimen type JD2, which increased the overall deformation performance of the connections.
Fig. 9. Load-displacement curves of three types of specimens tested
Load-strain curves
The load-strain curve at each measuring point is shown in Fig. 10. It can be found that these curves almost linearly increased throughout the whole loading stage. Under load, the strain change at the middle point of Beam 2, i.e., point C2, was the largest, and as the load increased, the strain change at point C2 began to show a nonlinear increase. For Beam B-1 and Beam B-2, the strains at their midpoints C4 and C7 and the end points C5 and C6 of the beam were very small. This could be because the self-tapping screws at these points were pulled out, and the bottom end of Beams B-1 and B-2 were detached from Beam 2 near the connection points.