Abstract
Cross-laminated timber (CLT), a wood product with excellent shear resistance, is often used in modern timber constructions. Using the standards ASTM D1761-12 (2020) and NDS-2012 (2012), this study investigated the connection properties of shear bolts and screws in CLT panels. The specimens were made from spruce-pine-fir lumber and installed on a test platform using one high-strength bolt or eight screws, and then an upward load was applied to the top of the specimen. The results showed that the bolt connection provided a higher ultimate bearing capacity and elastic stiffness. The bolt exhibited virtually no deformation, and the CLT panel did not noticeably deteriorate when the connection was damaged. The distance between the bolt hole and the bottom of the CLT specimen and the angle between the outer-layer grain direction of the CLT panel and the load direction were both measured. Changes in the ductility coefficient value had an obvious effect on the connection performance of the shear bolts when the outer-layer grain direction of the CLT panel was consistent with the load direction. Contrastingly, when the outer-layer grain direction of the CLT panel was perpendicular to the load direction, the effect was negligible, and the yield load was nearly unchanged.
Download PDF
Full Article
Comparative Study on Connection Properties of Shear Bolt and Screw of Thin Cross-laminated Timber Panel
Daiyuan Zhang,a Liming Shen,a,* Xudong Zhu,a,b Sujun Zhang,b Meng Gong,c and Yuewen Gao b
Cross-laminated timber (CLT), a wood product with excellent shear resistance, is often used in modern timber constructions. Using the standards ASTM D1761-12 (2020) and NDS-2012 (2012), this study investigated the connection properties of shear bolts and screws in CLT panels. The specimens were made from spruce-pine-fir lumber and installed on a test platform using one high-strength bolt or eight screws, and then an upward load was applied to the top of the specimen. The results showed that the bolt connection provided a higher ultimate bearing capacity and elastic stiffness. The bolt exhibited virtually no deformation, and the CLT panel did not noticeably deteriorate when the connection was damaged. The distance between the bolt hole and the bottom of the CLT specimen and the angle between the outer-layer grain direction of the CLT panel and the load direction were both measured. Changes in the ductility coefficient value had an obvious effect on the connection performance of the shear bolts when the outer-layer grain direction of the CLT panel was consistent with the load direction. Contrastingly, when the outer-layer grain direction of the CLT panel was perpendicular to the load direction, the effect was negligible, and the yield load was nearly unchanged.
DOI: 10.15376/biores.18.1.827-843
Keywords: Cross-laminated timber; Connection properties; Bolt; Screw
Contact information: a: College of Science and Engineering, Nanjing Forestry University, Nanjing 210037 China; b: Yangzhou Polytechnic Institute, Jiangsu 225200 China; c: Wood Science and Technology Centre, University of New Brunswick, 1350 Regent St., Fredericton E3C 2G6 Canada;
* Corresponding author: shenlimingda@hotmail.com
INTRODUCTION
At present, reducing carbon emissions is a major objective of the international community. The construction industry accounts for up to 30% of the world’s annual greenhouse gas emissions (IEA 2019). Therefore, reducing this sector’s carbon footprint is crucial for mitigating climate change. One solution is building houses from wood because wood is a natural building material that stores carbon. However, it is difficult for traditional light timber structure buildings to meet the requirements of high-rise buildings in terms of fire performance and structural strength. Cross-laminated timber (CLT) is a novel engineered wood material that was originally developed by scientists in Germany and Austria (Schickhofer and Guggenberger 1994). This material is made of multiple mutually perpendicular layers of wood. Timber structure buildings with CLT panels as the main structural components have good fire performance, structural strength, and lateral stiffness. Many countries are currently promoting CLT mid- and high-rise timber structure buildings on a large scale (Frangi et al. 2008; Frangi et al. 2009; Klippel et al. 2014; Wang et al. 2015; Brandner et al. 2016; Sikora et al. 2016). Cross-laminated timber structures constitute a new type of building suitable for assembly construction. Therefore, the connection mode in CLT structures is an important research topic. In common CLT buildings, CLT wall panels and floors are mainly connected using hold-down fasteners and wood screws. Although this connection mode is convenient for assembly, it is susceptible to considerable nail connection damage under transverse loads, which drastically reduces the overall lateral stiffness of CLT wall systems; this is mainly manifested by the bending or pulling out of wood screws under lateral loads, as well as the sliding displacement of the overall wall panel (Varoḡlu et al. 2007; Wei et al. 2015). In general, the performance of traditional light-frame timber shear walls is primarily determined by the response of the nail connection rather than the properties of the timber members (Dinehart and Shenton, III 2000; Richard et al. 2002; Li et al. 2012, 2015; Verdret et al. 2015). Compared with light-frame timber walls, CLT walls have a better lateral stiffness. Therefore, it is necessary to study better connection methods for CLT shear wall to avoid the overall stiffness reduction of wall systems because of improper connection methods. Cross-laminated timber panel and floor connection systems provide better overall stiffness by employing both side-angle steel and bolt penetrations through CLT panels.
The stiffness of bolted CLT wall systems is related to the bearing capacity of bolted joints. Current research on the bearing capacity of bolted joints in wood is based on the yield theory proposed by Johanson in 1941. Daudeville et al. (1999) conducted pin groove bearing tests to study the influence of the bolt diameter, edge distance, and end distance on the pin groove bearing strength. When the ratio of the wood thickness to the bolt diameter is small, the failure mode of a single-bolt or two-bolt connection tends to be brittle failure. Kharouf et al. (2005) conducted numerical simulations based on a nonlinear numerical model to study this failure mode. In their simulations, a compression plastic constitutive model was used in a certain area around the bolt. The resulting plastic deformation was in good agreement with experimental results obtained using glulam steel splint bolt connections (Kharouf et al. 2005). Considering the incompleteness of the bolt failure mode, especially the failure mode of group bolts, Moses and Prion (2003) proposed an anisotropic plastic model for wood and an ideal elastic-plastic model for bolts in view of the Canadian wood structure design standard CSA O86-14 (2019). In addition, the Weibull weakest link theory has been used to predict failure at given levels of probability (Moses and Prion 2003).
To investigate the application of screw and bolt connections in thin CLT panels in terms of the difference in shear performance, this study conducted shear tests involving four groups with two connection methods. The parameters studied included the section size, connection strength, shear stiffness, and failure mode of the specimens.
EXPERIMENTAL
Materials and Test Equipment
The spruce-pine-fir (SPF) lumber used in this study was purchased from Suzhou Jingxiu Construction Technology Co., Ltd. (Suzhou, China). It is a broad-grained wood, and the proportion of latewood is 25.7%. The average timber density is 0.48 g/cm3. The timber was air-dried, after which its average moisture content was 9.65% at room temperature and humidity (23 °C and relative humidity of 65%). The SPF boards, 4000 mm in length (L), 89 mm in width (W), and 38 mm in thickness (T), were fabricated into 800 mm × 80 mm × 16 mm standard planks. Then, the planks were assembled via a cold press machine (XLB; Qingdao Guangyue Rubber Machinery Manufacturing Co., Ltd., Qingdao, China), and resin with a density of 200 g/m2 was added between the layers. The resin was a common thermosetting one-component polyurethane (PUR) adhesive (PURBOND HB S709; Harbin Chengfeng Adhesive Co., Ltd., Harbin, China).
Fig. 1. Fabrication processes of the CLT samples
Table 1. Dimensions of Bolt Connection CLT Specimens
Table 2. Dimensions of Screw Connection CLT Specimens
The pressure of the cold press was maintained at 20 MPa for 8 to 12 h at a temperature of 15 °C. Subsequently, 1200 mm × 1200 mm × 48 mm three-layer CLT panels were obtained and trimmed into 300 mm × 80 mm × 48 mm specimens. The fabrication processes of the specimens are shown in Fig. 1, and the dimensions of the eight types of CLT specimens are listed in Tables 1 and 2.
Figure 2 shows all the metal connectors and fasteners used in this study. The tests were conducted at the mechanics laboratory of the Yangzhou Institute of Industry and Technology (Yangzhou, China). The bolt shear failure test was conducted using a WDW-200 micro-controlled electronic universal testing machine developed by Nanjing UP Electronic Technology Co., Ltd. (Nanjing, China). The maximum loading force of the horizontal load was 200 kN. As shown in Fig. 3, the bolt connection CLT specimens were connected to angle brackets by one 8.8 grade high-strength bolt with a diameter of 10 mm and length of 100 mm. Meanwhile, the screw connection CLT specimens were connected to two steel plates with eight screws (four on each side).
Fig. 2. Metal connectors and fasteners used in the tests
Fig. 3. Experimental setup for a shear test of CLT specimens: (a) setup for bolt connection specimens and (b) setup for screw connection specimens
The steel plates were connected to angle brackets by two 8.8 grade high-strength bolts. The angle bracket and the underlying steel beam were connected by one 8.8 grade high-strength bolt with a diameter of 20 mm and length of 60 mm. All the nuts were finger-tight. After the specimen was installed, the loading head was moved slightly upward to tighten the specimen.
Methods
The bolt shear test was conducted as per ASTM D1761-12 (2020). After the specimen was installed, the loading head was moved slightly upward to eliminate the gap error of the mounting hole. After the test started, the loading head was moved upward at a rate of 5 mm/min and stopped when the bearing capacity dropped to 80% of the ultimate load or obvious damage occurred (ASTM D1761-12 2020).
Fig. 4. Typical load-displacement curve for bolt and screw connection
The elastic stiffness K, ultimate bearing capacity Pmax, corresponding displacement Δmax, yield displacement Δy, and ductility coefficient D can be obtained based on the load-displacement curves from the test, . In particular, Pmax and Δmax were taken from the peak point of the load-displacement curve. According to ASTM D1761-12 (2020), the slope of the linear part of the load-displacement curve represents the K value, obtained using a two-point line of 10% Pmax and 40% Pmax of the load-displacement curve. According to the American NDS-2012 standard (2012), the linear part of the initial curve is offset by 5% of the bolt or screw diameter, and the intersection of the line and the curve represents the yield load Py.
RESULTS AND DISCUSSION
Failure Modes
Figure 5 shows the bolts after the test. Almost no bending occurred in the bolts. Figure 6 shows the deformation and failure of the screws after the test.
Fig. 5. Deformation of the bolts after the test: (a) bolts of type 1 specimens, (b) bolts of type 2 specimens, (c) bolts of type 3 specimens, and (d) bolts of type 4 specimens
Considering the specimen types in Table 2, when the screw length was 60 mm, the screw deformation of type 6 specimens, whose outer-layer wood grain direction was perpendicular to the load direction, was larger, and bending occurred in the middle of the screws. The bending deformation of screws in the type 6 specimens, whose outer-layer wood grain direction was parallel to the load direction, occurred near the screw head. When the screw length was 40 mm, the specimens whose outer-layer wood grain direction was parallel to the load direction underwent obvious bending deformation, and all specimens exhibited shear failure. In contrast, the specimens whose outer-layer wood grain direction was perpendicular to the load direction exhibited nearly no deformation when the specimens were destroyed.
Fig. 6. Deformation and failure of the screws after the test: (a) screws of type 5, (b) screws of type 6, (c) screws of type 7, and (d) screws of type 8
Figure 7 shows the failure modes of bolt holes of the bolt connection CLT specimens. As shown in Figs. 7(a) and (b), when the outer-layer grain direction of the CLT panel was parallel to the load direction and the distance between the bolt hole and the bottom of the specimen was 40 mm, the destruction of the CLT panel around the bolt hole was wood crushing and block shear (see red arrow in the figure). Furthermore, there were no obvious signs of damage in the middle layer of the CLT panel.
When the bolt hole was 70 mm away from the bottom (type 2 specimen), a large collapse failure occurred below the bolt hole, and the block shear of the middle layer of the CLT panel was observed from the side of the specimen. When the outer-layer grain direction of the CLT panel was perpendicular to the load direction, the transverse layers on both sides exhibited obvious transverse cracking failure.
The damage degree of the type 3 specimen, whose bolt hole was 40 mm from the bottom, was noticeably greater than that of the specimen whose bolt hole was 70 mm from the bottom. In addition, when the bolt hole margin was 40 mm, obvious block shear occurred at the bottom of the middle layer of the CLT panel. The red circle in Fig. 7(c) denotes the block shear.
Fig. 7. Typical failure appearances of the bolt connection CLT specimens during the shear tests: (a) type 1; (b) type 2; (c) type 3; (d) type 4
Figure 8 shows the failure modes of the screw connection CLT specimens. Figures 8(a) and (b) show the failure modes of specimens that employed 60-mm screws. The specimens with an outer-layer wood grain direction parallel to the load direction exhibited wood crushing under the screw holes and splitting in the vertical direction. When viewed from the side, the middle layers of some specimens had microscopic cracks. When the outer-layer wood grain direction was perpendicular to the load direction, obvious horizontal cracks were observed in the positions of the screw holes, and the size of the screw holes did not increase noticeably. Further, many cracks were observed on the sides of the specimens. Figures 8(c) and (d) show the failure modes of specimens with 40-mm screws. When the outer-layer wood grain direction of the specimens was parallel to the load direction, the failure mode was similar to that of the specimens with 60-mm screws. When the outer-layer wood grain direction was perpendicular to the load direction, the main failure mode was horizontal cracking around the screw holes, and the cracking degree was lower than that of the specimens with 60-mm screws. When viewed from the side, the adhesive layer between the outer and middle layers around the screw holes was damaged, which eventually caused the outer layer to fall off.
The failure modes of the bolt and screw connection CLT specimens are comprehensively compared as follows. First, from the perspective of deformation and failure of the metal connector, the bolts had no obvious deformation, whereas the screws had obvious deformation, except for the screws in the type 8 specimens; most of the specimens exhibited shear failure in one to three screws. Second, the failure of the specimens connected by screws was more severe. When the outer-layer wood grain direction was parallel to the load direction, the failure of the bolt connection specimens manifested as wood collapse below the bolt hole, while the screw connection specimens had a certain probability of cracking along the wood grain direction when the wood collapse occurred below the screw hole. When the outer-layer wood grain direction was perpendicular to the load direction, both the bolt hole and the screw hole produced transverse cracks along the grain. The cracks in the screw connection specimens were longer and were often transverse throughout the specimens. In addition, the outer-layer wood cracks in the screw connection specimens were more severe than those in the bolt connection specimens when viewed from the side. The outer layer of the specimens connected by 40-mm screws eventually fell off. In general, the bolt connection CLT panels were more likely to be repaired after breakage and had a higher rate of reuse than the screw connection CLT panels.
By comprehensively comparing the failure modes of metal connectors and wood in the two connections, it can be found that the bolts had no obvious failure when bolt connection was used, and the wood failure area was smaller and concentrated, which indicated that the connection part of the bolt connection CLT shear wall had better repairability.
Fig. 8. Typical failure appearances of the screw connection CLT specimens during the shear tests: (a) type 5; (b) type 6; (c) type 7; and (d) type 8
Mechanical Properties
Table 3 summarizes the test data and calculated performance indexes of all 24 specimens. Figure 9 shows the average ultimate bearing capacity (Pmax) and yield load (Py) of the eight types of CLT specimens. When the specimens were connected by bolts, the Pmax of types 1 and 2 specimens, whose outer-layer wood grain direction was parallel to the load direction, was noticeably greater than that of type 3 and type 4 specimens, whose outer-layer wood grain direction was perpendicular to the load direction. When the distance (d) between the bolt hole and the bottom of the specimen was 40 mm, the Pmax of type 1 increased 13.9% while Py increased 19.3% compared with type 3. When d was 70 mm, the Pmax and Py of type 2 were 31.6% and 51.8% greater than those of type 4, respectively. The increase was more obvious when d was 70 mm because the rest of the wood under the bolt hole was more prone to shear failure after being crushed when d was 40 mm. When d was changed from 40 mm to 70 mm, the Pmax and Py of specimens whose outer-layer wood grain direction was parallel to the load direction increased more obviously, up to 28.2% and 21.8%, respectively. However, the Pmax of specimens whose outer-layer wood grain direction was perpendicular to the load direction increased only 10.9%, whereas Py decreased 4.2%. This indicated that the failure of the specimens when the outer-layer wood grain direction was perpendicular to the load direction depended on the tensile strength of the transverse grain on both sides of the bolt hole.
When the specimens were connected by screws, the Pmax of the specimens with 60 mm screws was noticeably greater than that of the specimens with 40-mm screws. When the outer-layer wood grain direction was parallel to the load direction, the Pmax of the specimens with 60-mm screws increased 50.9% compared with that of the specimens with 40-mm screws. When the outer-layer wood grain direction was perpendicular to the load direction, this value increased 133.9%. Thus, when the CLT panels were connected with the screws, the depth of the screw penetration into the panels had a noticeably impact on the shear performance of the connection. When the 60-mm screw was used, the direction of the outer-layer wood grain did not have a noticeable effect on the test results, and the Pmax and Py of type 5 and type 6 specimens increased 4.6% and 8.3%, respectively. However, when the 40-mm screw was used, the values for the specimens whose outer-layer wood grain direction was parallel to the load direction were obviously greater than those for the specimens whose outer-layer wood grain direction was perpendicular to the load direction; Pmax and Py increased 62.1% and 38.3%, respectively. Considering the failure mode of the screw connection specimens in Fig. 8, in the specimens that employed 40 mm screws with the outer-layer wood grain direction perpendicular to the load direction, the depth of the screw penetration into the specimen was small. Therefore, the pull force of the screw acted on the outer layer of wood, resulting in separation and failure of the outer layer at the position of the screw hole.
By comparing type 2 specimens with the best performance in bolt connection and type 5 specimens with the best performance in screw connection, it can be found that the Pmax and Py of type 5 specimens were 12% and 11% higher than that of type 2 specimens, respectively. Although the bearing capacity of test piece No. 5 was higher, the total amount of steel used by 8 screws was about 1.3 times that of 1 bolt. Therefore, the bolt connection is reliable, easy to disassemble and saves material.
Table 3. Mechanical Properties of All CLT Specimens
Fig. 9. Peak load (Pmax) and yield load (Py) of the eight types of CLT specimens
Figure 10 shows the average elastic stiffness K and ductility coefficient D of all the specimens. Figure 13 shows the load-displacement curves of the bolt and screw specimens of each type after averaging. First, a comparison of the K and D values of the bolt and screw connection specimens shows that the K-mean and D-mean of the bolt connection specimens were 2.8 and 2.2, respectively. Further, the K-mean and D-mean of the screw connection specimens were 1.9 and 1.3, respectively. Thus, although the bolt connection CLT panels showed better elastic stiffness, they mainly suffered from brittle failure; hence, the ductility coefficient of the bolt connection specimens was relatively low. From observing the separate analysis of the bolt connection specimens, when the CLT panels were connected by bolts, the specimens whose outer-layer wood grain direction was parallel to the load direction had higher elastic stiffness and a larger ductility coefficient. Meanwhile, the elastic stiffness and ductility coefficient of specimens with a d value of 70 mm were noticeably higher than those with a d value of 40 mm. Among the four types of specimens connected by screws, the elastic stiffness of specimens with 60-mm screws was noticeably higher than that of specimens with 40-mm screws; however, the screw length had no noticeable effect on the ductility coefficient of the specimens (1.3 on average). This may have been due to the brittle failure of the CLT specimens connected by screws when they reached the ultimate load.