Investigations on the Effectiveness of Self-tapping Screws in Reinforcing Bolted Timber Connections

Transcription

1 Investigations on the Effectiveness of Self-tapping Screws in Reinforcing Bolted Timber Connections Mohammad Mohammad Group Leader, Building Systems Forintek Canada Corp, Eastern Division Quebec, Canada Alexander Salenikovich Assistant Professor Faculty of Forestry & Geomatics, Laval University Quebec, Canada Pierre Quenneville Professor and head of Civil Engineering Department Royal Military College of Canada Kingston, Canada Summary Timber connections require particular consideration to prevent brittle failures. Various reinforcing techniques have been proposed over the years for bolted timber connections such as truss plates and fibre-reinforced composite (FRC) or internally reinforced using threaded rods and glued-in rods. Self-tapping screws have high withdrawal resistance and can easily be installed in timber members. This paper reports the results of a research program carried out to investigate the effectiveness of self-tapping screws in reinforcing bolted timber connections when loaded parallel to grain or perpendicular to grain. Various series of bolted glulam specimens of different configurations were tested with and without reinforcement, and their strength, stiffness, ductility, and failure modes were compared. 1. Introduction Connections in heavy timber structures play an essential role in providing stability and stiffness to the structure. The interruption of continuity in the timber structure caused by the presence of joints results in a decrease in the overall strength and stiffness of the structure which in turn implies an increase in the dimension of the assembled timber elements. In many cases, timber connection details govern the design of the various structural members. The efficient design and construction of such connections systems often determine the level of success of wooden buildings when compared to other buildings made of steel or concrete. Innovative timber connection systems have been developed over the years; however, those connection systems have proven to be either too expensive or difficult to fabricate. Bolted connections remain quite popular in timber structure designs due to their cost-effectiveness and ease of fabrication. However, certain configurations may produce brittle failures of the joints, which are not predictable in a rational form in the current design documents. Typical brittle failure modes in bolted timber connections are: shear plug, group tear-out and splitting due to excessive tensile

2 stresses perpendicular-to-grain (as the bolt wedges its way through the wood). Preventing these types of failure would enhance bolted connection capacity and maintain the overall integrity of a timber structure, thus, enabling wood to compete favorably with other engineering materials. Various reinforcing techniques have been proposed over the years for bolted timber connections. Examples of such reinforcing techniques include externally reinforced connections using truss metal plates [1], and fibre-reinforced composite (FRC) [2], [3] or internally reinforced using gluedin rods and threaded rods [4], [5]. Results showed that the reinforcement could significantly increase ductility and strength of the connection. Reinforcement methods using self-tapping screws have been investigated. Some of these studies focused on the reinforcement of notched beam supports [6]. Others looked at bolted wood-to-wood connections with side wooden members reinforced with self-tapping screws and developed a model to calculate the load in the reinforcing screws [7]. In all cases, results showed that reinforced specimens exhibited a very ductile failure mode whereas the non-reinforced specimens failed in a brittle manner. Investigations on the mechanical properties of self-tapping screws in wood were also carried out [8], where embedment strength, withdrawal strength, pull-through and bending capacity of self-tapping screws were determined for modeling purposes. Self-tapping screws have high withdrawal resistance and can easily be installed in timber members compared to other means of reinforcement which require surface preparations and drilling; therefore, they could become a good solution for reinforcing new connections or repair existing ones that are showing signs of cracking. This is particularly important for timber connections that are designed for high loads, because their details govern the overall design of other structural members. However, technical information on the effectiveness of self-tapping screws in the reinforcement of bolted timber connections and in comparison to other reinforcing techniques is not available at the moment. This paper reports some test results of various connection configurations with reduced end distances, with and without reinforcement. The goal of the study was to provide relevant information on use of self-tapping screws as means of reinforcement for bolted timber connections and/or as means of designing more compact joints. Results of this research will assist industry initiatives to promote the use of wood products in structural applications, especially in the commercial and industrial sectors where heavy timber connections are quite common. The results are also expected to serve as a useful benchmark for structural design professionals and code officials in pursuing the idea of using timber in structural systems for non-residential buildings. 2.0 Experimental Investigation 2.1 Materials All tested connections were steel-wood-steel type with steel side plates. The main members were from Spruce-Pine (SP) or Douglas-fir (DF) glued laminated timber (glulam), 20f-EX stress grade with a cross-section of 80x190 mm (4x8 in.) for the specimens loaded parallel to grain and 20f-E stress grade, 80x304 mm for the specimens loaded perpendicular to grain. In an effort to randomize any manufacturing/material flaws and to equally distribute the potential density variations amongst specimens, test groups were created by selecting pieces from several billets. For all test groups SAE Grade 2 bolts, 19 mm (3/4 in.) in diameter, were utilized (F y avg = 448 MPa, F u avg = 544 MPa). The steel side plates were of 300W grade, 9.5 mm (⅜ in.) thick. Two types of self-tapping screws were

3 used in the parallel to grain study. Type A screws were Simpson Strong-Drive wood screws, 6x152 mm in size, with a hex washer (trade name: SDS 1/4x6 ). Type B screws were produced by Screw Products Inc. These were 8x152 mm (5/16x6 in.) in size with ceramic-coated Star Drive flat head. Truss plates were 51x153 mm, SK-20 manufactured and provided by Jager Building Systems. For the perpendicular to grain tests, WURTH screws, 8x200 mm were used. All wood members were conditioned to attain 12% equilibrium moisture content prior to testing. 2.2 Test set-up Testing procedures outlined in [9] were followed. All connections were loaded until one of the ends of the joint failed. The failure was attained when one of the following conditions was reached: when the load dropped with no recovery, or due to excessive deformation. The failure was generally attained within 10 min. except with joints reinforced with truss plates, which failed at a very large slip, and it took up to 20 min. to reach failure. Load-slip data were measured with an electronic data acquisition system. A total of four displacement transducers (LVDT s) were utilised to record displacements. Typical specimens are shown in figure 1. End distance LVDT 4 Glulam LVDT 3 Self-tapping screw Self-tapping screws Parallel to grain Loading Self-tapping screws LVDT 2 LVDT 1 Self-tapping screws Perpendicular to grain Loading Fig 1 Test set up and specimen configurations 2.3 Specimens Specimen configurations have been selected in such a way to trigger brittle failure; i.e., row shearout or splitting for parallel to grain loading or splitting for perpendicular to grain loading. In total, 20 configurations were tested with ten specimens in each series. Table 1 gives details of specimen configurations and materials for the parallel to grain loading and Table 2 gives the details for the perpendicular to grain ones.

4 The Canadian design provisions for bolted timber connections [10] stipulates that the minimum end distance should not be lower than seven times the diameter of the bolt or 50 mm for a loaded end. For unloaded end, it should not be lower than four times the diameter of the bolt or 50 mm. A reduced loaded end distance (5d) was used in this study for most of specimens in order to investigate the impact of reinforcement on improving the ductility of connections with smaller end distances than those specified in the design norm. Table 1 Configurations of tested series parallel to grain loading Series Material No. of End No. of Bolt Row rows distance bolts/row spacing spacing Type of reinforcement 1 SP 80x d 1 N/A N/A No Reinforcement 2 SP 80x d 1 N/A N/A Screws type A 3 SP 80x d 1 N/A N/A Truss plate 4 DF 80x d 1 N/A N/A No Reinforcement 5 DF 80x d 1 N/A N/A Screws type A 6 SP 80x d 2 4d N/A No Reinforcement 7 SP 80x d 2 4d N/A Screws type A 8 SP 80x d 1 N/A 5d No Reinforcement 9 SP 80x d 1 N/A 5d Screws type A 10 SP 80x d 1 N/A 5d Truss plate 11 SP 80x d 1 N/A 5d No Reinforcement 12 SP 80x d 1 N/A 5d Screws type A 13 SP 80x d 1 N/A 5d Splits & screws type A 14 SP 80x d 1 N/A 5d Screws type B 15 SP 80x d 1 N/A 5d Two screws type A The configurations of the specimens loaded perpendicular to grain were selected to verify the effect of end distance on the splitting resistance and further, to verify if screws could be used to avoid potential brittle failures. For this paper, only the effect of screw reinforcement on the connection resistance will be reported. Table 2 Configurations of tested series perpendicular to grain loading Series Material No. of End No. of Bolt Loaded rows distance bolts/row spacing edge Type of reinforcement 16 SP 80x d 2 3d 4d No Reinforcement 17 SP 80x d 2 3d 4d 16 - Re-tested with screw 18 SP 80x d 2 3d 4d Screw 19 SP 80x d 2 3d 4d No Reinforcement 20 SP 80x d 2 3d 4d 19 - Re-tested with screw 3. Results and Discussion 3.1 Failure modes

5 The predominant failure mode in specimens loaded parallel to grain, with 5d end distance was row shear-out, except for those reinforced with metal truss plates (Series 3 and 10), in which considerable crushing of wood at the bolts was observed. Splitting was also prominent in many specimens with 5d end distance, especially, in those reinforced with self-tapping screws. Specimens with 7d end distance, with and without reinforcement, failed mainly in splitting. For Series 4 and 5 made with D-fir, the main type of failure was row shear-out. Even though, self-tapping screws did not seem to affect the failure mode in DF specimens with single bolts, they reduced the row shear failure in DF specimens with double bolts which failed in splitting. In Series 13, where the specimens with artificial splits were reinforced with self-tapping screws, the splitting propagated along the artificial splits under one or both bolts. Little evidence of wood crushing was observed. Generally, the crack propagation in non-reinforced joints was more pronounced than in the reinforced ones. In the specimens loaded perpendicular to grain, all failure modes involved splitting. For the ones reinforced with screws, head pull-through was observed. 3.2 Analysis of the load-slip profile The load-slip curve of the failed end was used to determine the following characteristic points: proportional limit (P PL, PL ), point of maximum resistance (P max, max ), and failure point (P fail, fail ), as shown in Figure 2. Load P PL PL Initial slope P max max Pfail fail Fig 2 Characterization of performance parameters of tested connections Initial slope range Slip According to [11], the yield point is determined as the intersection point between the load-slip curve and the straight line, which is offset 5% of the bolt diameter from the line of the initial slope. This approach was not suitable to characterize the joints tested in this project, because non-reinforced joints were brittle and did not exhibit a yield point, whereas in reinforced joints the 5% diameter offset was found deep in the nonlinear portion of the curve. For this reason, the transition to the plastic yielding was better characterized by the proportional limit (PL). To determine PL, a tangent line was fit to the load-slip curve for every ten data points above the initial slope range. The PL was determined as the point on the load-slip curve where the tangent line deviated at least 15% from the initial slope. 3.3 Ultimate loads As it can be seen in figure 3, the reinforcement with Type A screws did not affect the maximum load sustained by the joints for most series tested, regardless of the joint configuration. Neither Type B screws (Series 14) added significantly to the load capacity of the joints. No significant

6 differences in P max were observed between DF and SP specimens (Series 4 and 5 vs. Series 1 and 2). However, comparing Series 8 with Series 13, where the specimens with artificial splits were reinforced with screws, it was evident that the reinforcement of the damaged joint provided as much as 70% of the original load capacity assuming that the specimens with artificial splits had no strength at all. This is a significant contribution, especially for the repair of already split structural members. Table 3 Summary of test results parallel to grain loading Series P max fail max K ini ave P PL PL fail / PL Type of reinforcement kn mm mm kn/ mm kn mm No Reinforcement Screws type A Truss plate No Reinforcement Screws type A No Reinforcement Screws type A No Reinforcement Screws type A Truss plate No Reinforcement Screws type A Splits & screws type A Screws type B Two Screws type A Table 4 Summary of test results perpendicular to grain loading Series P max (kn) COV Type of reinforcement % No Reinforcement % Group 16 re-tested with screw % Screw % No Reinforcement % Group 19 re-tested with screw Reinforcement of single-bolt connections with metal truss plates (Series 3 vs. Series 1) did not affect the load capacity of the joint; however, the reinforced two-bolt connections (Series 10 vs. Series 8) showed on average 18% increase of the load capacity (see figure 3). One possible explanation for such an increase is that the truss plates resisted tensile stresses in the direction perpendicular to grain after some deformation had already occurred. With a single-bolt arrangement, row shear-out was predominant, and truss plates could not contribute much to prevent the formation of shear planes parallel to the load direction. With regards to connections tested perpendicular to grain, no significant statistical differences were observed between un-reinforced and reinforced groups. The only observation one could make is

7 that the mode of failure of the reinforced ones was screw head pull-through and this, forced the COV to be relatively lower. It is interesting to note that no differences existed between reinforced connections that were previously tested and then repaired and reinforced one that were previously un-tested. This again, leads one to confirm that the head pull-through controlled the ultimate load and that the tension strength perpendicular to grain of the member did not contribute to the connection resistance at ultimate. Pmax (kn) N/R Screw s type A Pmax (kn) N/R Screw s type A Truss plate Series Series (a) (b) Fig 3 Mean values of P max of tested series with (a) and without reinforcement (b) 3.4 Ductility Test results in Table 3 show that the ductility of joints is increased due to reinforcements. Ductility ratio N/R Screw s type A Truss plate Series Fig 4 Comparison of ductility ratio of specimens with and without reinforcements parallel to grain. 4. Conclusions Figure 4 shows that, on average, SP specimens reinforced with truss plates exhibited the highest ductility ratio, although DF specimens with screws (Series 5) were also highly ductile. Truss plates improved the residual strength of the tested connections due to crushing of wood prior to failure, and it was reflected by the large deformations on the load-slip curves. This improvement in the residual strength translates into more ductile connections, which are capable of absorbing more energy than non-reinforced ones. For critical joints in timber structures, this may be a valuable benefit during any seismic events. This study dealt with the experimental investigation of the effectiveness of self-tapping screws as means of reinforcing bolted timber connections loaded either parallel or perpendicular to grain. Twenty series of double-shear steel-wood-steel connections with one or two bolts per where tested. The specimens were reinforced with self-tapping screws or truss plates. Some of the specimens had artificial splits made with a band saw and were reinforced prior to testing.

8 Test results showed that reinforcement had positive effects on the joint ductility, whereas the load capacity and initial stiffness were not significantly affected. This was observed for parallel and perpendicular to grain loading. Reinforcement with either self-tapping screws or truss plates altered the failure mode in certain tests. A ductile failure mode with considerable crushing of wood was evident in connections reinforced with truss plates and for the perpendicular to grain specimens, screw head pull-through controlled the ultimate load. 5. References [1] Hockey, B. and F. Lam, and H.G.L. Prion Truss plate reinforced bolted connections in parallel strand lumber, Canadian Journal of Civil Engineering. 27: [2] Soltis, L. A., R.J. Ross, and D.F. Windorski Fiberglass-reinforced bolted wood connections, Forest Products Journal. 48(9): [3] Chen, C.J Mechanical behaviour of fiberglass reinforced timber joints, Ph.D. Thesis, No. 1940, EPFL. Lausanne, Switzerland. [4] Quenneville, J.H.P. and Mohammad, M Anti-check bolts as means of repair for damaged split ring connections, Proceedings of the World Conference on Timber Engineering. July 31-August 3. Whisler, BC, Canada. [5] Mastschuch, R Reinforced Multiple Bolt Timber Connections. M.Sc. Thesis, University of British Columbia, Vancouver, BC, Canada. [6] Blass, H.J. and Bejtka, I Reinforcement perpendicular to the grain using self-tapping screws, Proceedings of the 8th World Conference on Timber Engineering, Volume I, Lahti, Finland, [7] Blass, H.J. and M. Schmid. Self-tapping screws as reinforcement perpendicular to the grain in timber connections, In: Joints in Timber Structures. Proceedings of the International RILEM Symposium. RILEM Publications S.A.R.L. Stuttgart, Germany, [8] Hansen, K.F Mechanical properties of self-tapping screws and nails in wood, Canadian Journal of Civil Engineering. 29: [9] American Society for Testing and Materials D1761. Standard test methods for mechanical fasteners in wood. Annual Book of ASTM Standards. West Conshohocken, PA. [10] Canadian Standard Association Engineering Design in Wood (Limit States Design), O86-01, Rexdale, Ontario. [11] ASTM D5652. Standard test methods for bolted connections in wood and wood-based products. Annual Book of ASTM Standards. West Conshohocken, PA.