Seismic behaviour of fibre-reinforced-polymer- and steel-strengthened timber connections

Abstract

In the event of a seismic attack, the structural integrity of moment-resisting timber frames in the joint region may become compromised and hence the joint may not be able to transfer bending moment around the frame. Often, replacement of a damaged joint is not an option and hence efficient but effective strengthening and repair schemes for such joints are necessary. This article reports the results of 15 tests on 10 metal dowel-type moment-resisting timber connections subjected to monotonic or cyclic loading. Joints are either strengthened or repaired with epoxy, fibre-reinforced polymer composites or steel plates. The ability of the test joints to resist the imposed cyclic loading is presented in the context of hysteresis responses. Recommendations for strengthening and repair interventions are made based on strength, stiffness, ductility, energy dissipation and damping characteristics of the test joints.

Keywords

cyclic loading experimentation fibre-reinforced polymer joints timber

Introduction

For moment-resisting timber frames, joints play an important role in determining the strength and stiffness of the structure. Dowel-type timber joints are popular in modern timber structures due to their relatively high strength and stiffness, flexibility in application, as well as ductile behaviour provided that they are properly designed. The load is transferred between the joint components via timber bearing as well as dowel action arising from the fasteners (Porteous and Kermani, 2007). Common fasteners include nails, screws, bolts and steel dowels. The joint region is, however, most vulnerable in cases of overload especially caused by earthquake attacks. In order to improve or at least reinstate the overall performance of timber frame structures, it is necessary to develop strengthening or repair schemes for the joint regions.

Existing studies have reported the development of various methods and techniques to improve the performance of dowel-type timber joints. In order to increase the local bearing capacity of the timber, materials are applied on the surface of the timber member, such as bonded steel plates (Leijten, 1988), punched nail plates (Blaß et al., 2000; Blaß and Schädle, 2011), densified veneer wood (DVW; Leijten, 1991, 1999; Rodd and Leijten, 2003) and glass-fibre-reinforced polymer (GFRP) materials (Chen, 1999; Haller et al., 2006; Haller and Wehsener, 1999; Kasal et al., 2002, 2004a, 2004b). Other strengthening approaches include the insertion of lag screws laterally into the timber members in order to reduce the risk of timber splitting and to carry part of the bolt bearing load (Blaß and Schädle, 2011; Lam et al., 2010). A resin injection technique has also been developed to date to strengthen joints (Davis and Claisse, 2001; Rodd et al., 1989). The injected resin filled the gap between the bolt and the hole in order to achieve a full bond between the bolt and the timber. There are, however, few studies on strengthening timber joints with carbon-fibre-reinforced polymer (CFRP) composites. Also, there is a lack of detailed cyclic loading tests on steel-reinforced timber joints.

The aim of the study reported herein is to investigate the performance of various strengthening and repair strategies for cyclically loaded dowel-type moment-resisting timber joints. The results reported form part of a larger study on timber joints and moment-resisting frames (Yang, 2013). Four strengthening schemes and one repair scheme are proposed, namely, (1) externally bonded wet lay-up CFRP plates, (2) near-surface mounted CFRP rods, (3) externally bonded steel plates, (4) externally bonded pultruded CFRP plates and (5) epoxy repair. The strength, stiffness, ductility, energy dissipation and damping results of 14 cyclic loading joint tests are reported and the results are analysed in accordance with BS EN 12512:2001 (2001). The performances of the joints as well as the effectiveness of the strengthening schemes are compared and evaluated with each other as well as that of a monotonically loaded control joint.

Experimental setup

Timber joint specimens

L-shaped double shear glued laminated timber (glulam) joints were constructed and the test setup is shown in Figure 1. The glulam members were fabricated in Shenzhen, China, and they were made of yellow pine that was sourced from the United States. The glulam sections were composed of seven layers of timber that resulted in a section height of 245 mm (Figure 2) with an average tested density of 587 kg/m³. One central member (135 mm wide) and two side members (65 mm wide) were joined via four 16-mm mild steel dowels for each test joint. The diameter of the dowel holes in the timber was 17 mm and the holes were spaced at 100-mm centres. The length of the connecting rods was 350 mm and they contained 50-mm threaded ends. The distance between the dowels and the nearer timber ends was 120 mm. This positioning satisfied Eurocode (BS EN 1995-1-1:1994, 1994) requirements for the prevention of end splitting failure. In addition, two steel plates were connected to the ends of each glulam member via five 12-mm threaded rods. The rods were inserted into 13-mm-diameter holes. The steel plates were connected to the test frame or actuator by a 30-mm pin at each end.

Figure 1.

(a) Test setup and (b) schematic of the test setup.

Figure 2.

Details of glulam members (unit: mm): (a) elevation and (b) cross-section.

Considering the size of the test frame and the range of the actuator, the steel plates were connected at an axis oriented 45° to that of the glulam members. Assuming that the glulam members and steel plates were rigid bodies, deformation was assumed to occur only within the joint area. Based on the available ram displacement of the actuator, the maximum possible rotational deformation was calculated to be +0.227 rad (joint opening) and −0.192 rad (joint closing). The actual measured maximum rotation was about +0.19 and −0.17 rad, respectively. This was due to the clearance within the whole system, fabrication variability and bending deformation of the glulam members. Finally, the actuator was laterally constrained so only in-plane movement of the test joints was permitted.

Details of strengthening and repair schemes

Test programme

Four strengthening schemes and one repair scheme were investigated, namely, (1) externally bonded wet lay-up CFRP plates (herein EB scheme), (2) near-surface mounted CFRP rods (herein NSM scheme), (3) externally bonded steel plates (herein SP scheme), (4) externally bonded pultruded CFRP plates (herein FP scheme) and (5) epoxy repair (herein ER scheme).

The test joints were divided into two main categories, namely, (1) strengthening and (2) repair. Table 1 shows these two broad categories as well as a detailed account of the joint parameters within. Joints in the strengthening category (apart from the control joints J1, J2, J8–J10) contained strengthening interventions for joints that had not previously undergone loading (i.e. joints that are referred to as undamaged). Joints in the repair category contained repair interventions for joints that had undergone loading (i.e. damaged joints). Plain joint J1 was tested under monotonic loading as a preliminary trial test to obtain the yield displacement in order to determine the cyclic loading parameters for the remainder of all test joints. Four plain joints (J2, J8–J10) were tested under cyclic loading as control specimens and they also served as damaged joints that were subsequently repaired in the repair category as, respectively, joints J2R and J8R–J10R. Joint J1 was also repaired as joint J1R. Undamaged joints J3 and J4 were strengthened with different EB schemes and repaired joint J8R was strengthened with an EB scheme. Undamaged joint J5 and repaired joint J10R were strengthened with an NSM scheme. In addition, undamaged joint J6 and repaired joint J9R were strengthened with an SP scheme. Finally, undamaged joint J7 was strengthened with an FP scheme. All joints in the repair category were subjected to epoxy repair around the dowel region.

Table 1.

Details of test specimens.

Strengthened joints		Repaired joints
ID	Strengthening scheme	ID	Repair scheme
J1	Nil (plain joint)^a	J1R	Epoxy^b
J2	Nil (plain joint)	J2R	Epoxy
J3	EB90
J4	EB45
J5	NSM
J6	SP
J7	FP
J8	Nil (plain joint)	J8R	Epoxy + EB45
J9	Nil (plain joint)	J9R	Epoxy + SP
J10	Nil (plain joint)	J10R	Epoxy + NSM

J: joint; EB: externally bonded FRP; NSM: near-surface mounted FRP; SP: steel plate; FP: FRP plate; R: joint repaired with epoxy; EB90: fibres oriented 45° to grain; EB45: fibres oriented 0, 90° to grain.

Monotonic load. Remaining joints subjected to cyclic loading (Figure 8).

Joints J1, J2 and J8–J10 were repaired with epoxy adjacent to the dowels to produce joints J1R, J2R and J8R–J10R, respectively.

Externally bonded wet lay-up CFRP plate (EB scheme)

Two layers of carbon fibre sheet of 30 mm width were bonded adjacent to the dowel holes in a wet lay-up manner relative to the longitudinal axis of the side members. Two configurations were investigated, namely, (1) perpendicular FRP strips (herein EB90) and (2) ±45° strips (herein EB45). Since the bearing forces acted about the rotational centre of the four dowels, EB90 was ±45° to the grain direction. The fibre direction was therefore parallel and perpendicular to the timber grain for the EB45 scheme. The epoxy used in this scheme was Sikadur 300. Prior to bonding, the surface area to be strengthened was sanded with an electric sander containing 80CW sandpaper and then wiped clean with pure alcohol. This grade of sandpaper was coarse enough to create a reasonably rough bond surface without excessively damaging the timber. Fibre sheets were bonded onto both sides of each glulam member in a wet lay-up manner. Typical bonded EB90 and EB45 schemes are shown in Figure 3(a) and (b), respectively.

Figure 3.

Repair and strengthening schemes: (a) EB90 specimen; (b) EB45 specimen; (c) NSM specimen – application of epoxy into grooves; (d) NSM specimen – insertion of rolled fibre sheets into grooves; (e) SP specimen – application of epoxy on steel plate; (f) SP specimen; (g) FR specimen – FRP plate segment; (h) FP specimen; (i) repair specimen – reinstated bolt hole; and (j) repair specimen details.

Near-surface mounted CFRP rod (NSM scheme)

FRP rods used in the NSM scheme were made from rolled fibre sheets of 54 mm width (i.e. 10 fibre bundles) in a wet lay-up manner. The rods were mounted into grooves cut into the timber perpendicular to the grain direction but adjacent to the dowel holes (Figure 3(c)). The grooves were made with a 6.35-mm (1/4 in)-diameter drill bit and the grove depth was 5 mm. The surface of grooves was sanded with 80CW sandpaper and cleaned with alcohol. Following the application of epoxy to the grooves (Figure 3(c)), the saturated rolled fibre sheet was mounted (Figure 3(d)). The epoxy used in this scheme was Sikadur 300.

Externally bonded steel plate (SP scheme)

Mild steel plates of 2.3 mm nominal thickness and the dimensions of 245 mm × 200 mm were utilised. Holes of 176 mm diameter were also drilled into the plate and Araldite 420 was chosen as the bonding agent. The viscosity of this epoxy prevented it from excessive leaking after the steel plate was bonded onto the timber. The timber surface was sanded and cleaned in the same manner as the EB scheme. The oxidised surface layer of the bonded side of the steel plates was removed with the aid of an angle grinder. Epoxy was applied to both the timber and steel plate surfaces and this scheme was applied to both sides of the outer timber members (Figure 3(e) and (f)).

Externally bonded pultruded CFRP plate (FP scheme)

Bidirectional pultruded CFRP plate was utilised. The nominal thickness of the plate was of 1/8 in (3.175 mm) and the width was 4 in (101.6 mm). The plate was cut into 245-mm-long segments and two 17-mm holes were drilled into each plate segment (Figure 3(g)). The plates were then bonded with the longer plate dimension oriented perpendicular to the timber grain direction. The two bonded CFRP plates therefore formed a reinforced area of 245 mm × 203.2 mm (Figure 3(h)). This is a similar area to the SP scheme (i.e. 245 mm × 200 mm). The shear planes of the joints (i.e. both sides of central member and one side on each side member) were strengthened. Araldite 420 was used in this scheme. The timber and CFRP plate surfaces were sanded and cleaned in the same manner as the EB scheme.

Timber reinstatement with filled-in epoxy (repair scheme)

Following the application of monotonic or cyclic load to selected plain joints (i.e. J1, J2, J8–J10), the timber adjacent to the steel dowels was found to be crushed. The original geometry of the timber was then repaired with Sikadur 300 epoxy to produce the respective joints J1R, J2R and J8R–J10R. The low viscosity of this epoxy, which limited the introduction of air bubbles, was a convenient material to work with. The repair involved several steps, namely, (1) seal one side of the timber member with tape at each damaged hole location, (2) fill the entire hole with epoxy and (3) remove the tape after 7 days of air curing and drill 17-mm-diameter holes to form the dowel holes. The final repair is shown in Figure 3(i) and the epoxy wedges that filled the originally crushed timber are evident. Finally, any splitting or cracking of the timber member present after the initial loading was filled with epoxy that was applied with a syringe (Figure 3(j)). During this process, the timber was clamped for a period of 7 days. Following epoxy repair, the aforementioned strengthening schemes were then applied as per the configurations specified in Table 1.

Instrumentation and test procedures

The load was applied by a 250-kN MTS actuator in a ram displacement control manner. Joint J1 was initially subjected to monotonic loading at a rate of 0.5 mm/min. The yield displacement (V_y) according to the actuator displacement was +21 mm. The loading protocol for the cyclic test is shown in Figure 4 and it was determined in accordance with BS EN 12512:2001 (2001). Based on the monotonic load test result, the estimated yield displacement (V_y,est) was assumed to be 20 mm (note that the tested value was 21 mm). The amplitudes of the first two cycles were based on 0.25V_y,est and 0.5V_y,est, respectively. The following cycles were then set in accordance with 0.75V_y,est, V_y,est, 2V_y,est, 4V_y,est and 6V_y,est. The load rate for the cyclic tests was adjusted so that each cycle could be completed within 30 min. In addition, each load cycle was applied three times for each cyclically loaded joint.

Figure 4.

Cyclic loading procedure.

Joint rotation was calculated from two linear variable differential transformers (LVDTs) that were spaced at 110 mm as shown in Figure 5(a). The LVDTs were connected to an aluminium frame that was attached to the end of the central member. An aluminium plate with surface parallel to the grain direction of one side member was fixed to the rotational centre of the joint (i.e. the centre position of the four dowels). These two LVDTs moved relative to this plate. The rotation between the central member and side members (i.e. joint rotation, φ) was calculated as follows

φ = \tan^{- 1} [\frac{(v_{1} - v_{2})}{l}]

(1)

where v₁ and v₂ are the measured displacements of the two LVDTs and l is the distance between the two LVDTs of 110 mm. Joint opening (φ > 0) occurred when the angle between the central member and side members was greater than 90° and joint closing (φ < 0) occurred when the same angle was less than 90°.

Figure 5.

Joint measurement instrumentation. (a) Measurement of joint rotation: (i) schematic drawing and (ii) test picture of specimen J4. (b) Measurement of joint translational deformation.

Joint translation was measured by two linear displacement transformers (LDTs) as shown in Figure 5(b). A threaded rod was positioned perpendicular to the joint plane that was attached to the centre of the joint between the four dowels. The two LDTs were secured to the load frame and measured the threaded rod movement. Joint moment, M, was then calculated from

M = F (v_{h} + H)

(2)

where F is the actuator load, v_h is the horizontal movement of the rotational centre of joint and H is the original horizontal distance between the rotation centre of joint and the axis of load (i.e. H = 601 mm as per Figure 1(b)).

Material properties

Timber

The compressive strengths and elastic moduli were obtained from compression tests in accordance with BS EN 408:2003 (2003). Seven specimens were tested parallel to the grain direction and an additional seven specimens were tested perpendicular to the grain direction. The test specimens were randomly selected from the same batch of glulam material. The parallel-to-grain specimen consisted of cross-sectional dimensions of 30 mm × 50 mm and a height of 180 mm. The perpendicular-to-grain specimen consisted of cross-sectional dimensions of 50 mm × 70 mm and a height of 90 mm. The elastic moduli were calculated from four longitudinally positioned LVDTs located each side of the test specimens. The arising elastic modulus and compressive strength properties in the parallel-to-grain direction were 6630 (1673) and 42.86 (5.98) MPa, respectively, where the values in the brackets are standard deviations (SDs). The corresponding properties in the perpendicular-to-grain direction were 388 (83) and 4.63 (0.61) MPa.

FRP plate (wet lay-up)

The properties of the CFRP plate were obtained from tensile tests on five flat rectangular coupons tested in accordance with ACI 440.3R-04:2004 (2004). The test coupons were made from two layers of carbon fibre sheet of 35 mm width in a wet lay-up procedure. The same role of 0.166 mm nominally thick carbon fibre sheet was used for all tests. The elastic modulus was 231 (14) GPa, the tensile strength was 2768 (178) MPa and the ultimate strain was 11,982 (889) με, where the results in brackets are standard deviations.

FRP plate (pultruded)

The properties of the pultruded FRP plate were provided by the manufacturer (Strongwell, n.d.). The tensile strength and elastic modulus were 852 MPa and 62.19 GPa, respectively, while the clamped bearing strength was 351 MPa (loaded via a torqued fastener), and the unclamped bearing strength was 214 MPa (loaded via a normal pin).

Epoxy

The properties of two types of epoxy were obtained from tensile testing of dog-bone-shaped coupons in accordance with BS EN 527:1999 (1999). There were three Sikadur 300 coupons and five Araldite 420 coupons tested. The elastic modulus, tensile strength and ultimate strain results of the Sikadur 300 were 3433 (31) MPa, 48.03 (6.23) MPa and 17,324 (2978) με, respectively, where the values in the brackets are standard deviations. The corresponding results for the Araldite 420 were 1719 (27) MPa, 21.52 (0.46) MPa and 18,356 (1525 με).

Steel rod

The steel rods used as dowel-type fasteners were delivered in two batches. The first batch was used in joints J1, J2, J1R and J2R, while the second batch was used for the remainder of the test specimens. The material properties were obtained from tensile testing of the steel rod coupons with a diameter of 16 mm and a length of 600 mm in accordance with BS EN 10002-1:2001 (2001). Four coupons were tested for each batch and the results are summarised in Table 2.

Table 2.

Material properties of steel.

Material	Elastic modulus, GPa (SD)	Yield strength, MPa (SD)	Ultimate strength, MPa (SD)
Steel rod (batch 1)	216 (8)	359 (7)	403 (8)
Steel rod (batch 2)	209 (8)	355 (17)	391 (13)
Steel plate	211 (12)	306 (18)	391 (30)

SD: standard deviation.

Steel plate

Four steel plate coupons were tested in accordance with BS EN 10002-1:2001 (2001). The steel plate coupons of 2.3 mm nominal thickness had a width of 50 mm and a length of 600 mm. The properties arising are presented in Table 2.

Experimental results

Behaviour and failure modes

General failure modes

All joints failed by mode ‘j’ as categorised by BS EN 1995-1-1:1994 (1994) for double shear timber-to-timber connections. The components of this failure mode are shown schematically in Figure 6(a) where each steel dowel contained two plastic hinges (Figure 6(b)). The timber adjacent to the dowel failed in bearing and this deformation was irrecoverable (Figure 6(c)). Splitting cracks were observed on the side members and such cracks always propagated along the grain direction as shown schematically in Figure 6(d). For the joint opening cases (i.e. positive rotation), the bearing force of dowel 1 in Figure 6(d) pointed towards the free end of the side members. In this case, tension was induced perpendicular to grain that leads to cracking (herein joint opening crack). For the joint closing cases (i.e. negative rotation), cracking propagated from dowel 2 in Figure 6(d) (herein joint closing crack).

Figure 6.

General failure modes: (a) failure mechanism; (b) failure of steel dowel; (c) local deformation of timber and (d) definition of joint opening crack and joint closing crack.

In the following sections, the behaviour of the joints is described individually. Descriptions are made with reference to the south side and north side members as per Figure 1(a).

Plain joints

All the plain joints (i.e. J1, J2, J8, J9 and J10) exhibited mode ‘j’ failure. Large deformation occurred in the joints and the maximum tensile strength was reached in the side members which led to splitting. Typical behaviour involved joint opening and closing cracks occurring on the north and south side members, respectively, of joint J2 during the 15th load cycle. The maximum amplitude was also first achieved in this cycle.

Repair scheme

The repaired joints without FRP or steel plate strengthening (i.e. joints J1R and J2R) failed in a more brittle manner than the plain joints. For joint J1R, the joint opening crack formed first on the south side member during the 12th cycle. Then, during the 14th load cycle a joint opening crack appeared on the north side member. A joint closing crack also occurred on the south side member at this load level. For joint J2R, a joint opening crack formed on the north side member during the 15th load cycle. During the negative rotation phase of the same load cycle, cracks were found to form on both side members. In the 16th load cycle, the south side member cracked when subjected to joint opening.

In these two joints, the steel dowels were pushed onto the epoxy wedges at large deformation. The epoxy wedges were therefore pushed into the adjacent timber which led to permanent damage in the timber. There was, however, no crushing or damage found in the wedges themselves. A typical failure mode is shown in Figure 7(a).

Figure 7.

Failure modes of strengthened joints: (a) repaired joint, (b) EB90, (c) EB45, (d) NSM, (e) SP, (f) dowel in SP joints and (g) FP.

EB scheme

No cracks were observed after the tests in the timber members strengthened with EB schemes (i.e. joints J3, J4 and J8R). For joint J3 that contained the EB90 scheme, the steel dowels pressed into the FRP perpendicular to the fibre direction (Figure 7(b)). These fibres bent and broke; however, the fibres located further away from the bearing zone were not visually affected and no debonding was observed. For joints J4 and J8R that contained EB45 schemes, the dowel bore into the intersection position of the orthogonally placed FRP (Figure 7(c)). These fibres crushed and broke although there was no noticeable damage to the remainder of the FRP. The epoxy wedges in all three joints did not crush although they were pushed into the adjacent timber.

NSM scheme

No cracking was observed in the timber for joints strengthened with NSM schemes (i.e. joints J5 and J10R). The FRP rods in the vicinity of the dowels were deformed (Figure 7(d)) although the damage to the FRP was local. For joint J10R, the epoxy wedge was pushed into the adjacent timber at high joint deformation.

SP scheme

The steel plates were found to have failed in bearing adjacent to the dowels (Figure 7(e)) where the steel locally buckled and yielded (joints J6 and J9R). There was also shear deformation of the steel dowels at the shear planes as the stiffness and strength of the steel plates were significantly higher compared to the timber (Figure 7(f)). Some of the steel plates debonded during the test and then strength degradation of the joint resulted. The north side steel plate partially debonded during the 15th load cycle of the positive rotation phase. A crack was also found between the steel plate and the timber substrate. The joint strength reduced in subsequent load cycles following debonding. Debonding occurred in the positive rotation phase of the 12th load cycle for the repaired joint J9R. Partial debonding occurred on the steel plates for the north side member as well as the north side of the central member. For the negative rotation phase of this load cycle, these two plates then fully debonded. Afterwards, joint strength decreased significantly. During the 15th load cycle, joint opening and closing cracks were observed to occur on the north side member.

After test, poor quality of bond between the steel plates and the timber substrate in places was visually detected. This occurred because the thin steel plates were easily bent during the cutting and drilling stage of preparation. This caused the plate and the timber to not be fully in contact with each other. Nonetheless, the SP scheme did increase the strength of the joint by about three times the strength of the plain timber joints.

FP scheme

The pultruded FRP plates locally crushed due to dowel bearing (Figure 7(g)) for joint J7. As the FRP plate was not deformed during the preparation stage, and hence remained reasonably flat, there was adequate bond between the plate and the timber. Finally, no cracking was observed in any of the timber members.

Moment–rotation responses

Monotonic loading

The moment–rotation response of the monotonically loaded joint J1 is shown in Figure 8(a). At the initial loading stage (φ < 0.015 rad), the moment and tangential stiffness increased only gradually with rotation due to clearance between the steel dowels and the adjacent walls of the timber holes. Once the steel dowels came into contact with the sides of their hole, the load increased in an approximately linear manner with rotation (0.015 rad < φ < 0.030 rad). With further increase in rotation (0.030 rad < φ < 0.060 rad), the stiffness of the joint decreased gradually and this represents joint yield. Following this transition stage, the moment increased in a reasonably linear manner with joint rotation albeit with reduced stiffness. This test stopped when the actuator ram had displaced 100 mm. The joint rotation at this endpoint was 0.173 rad. According to BS EN 12512:2001 (2001), the yield displacement (V_y) was +21 mm by actuator displacement.

Figure 8.

Moment–rotation responses: (a) J1 – plain joint; (b) J2 – plain joint; (c) J8 – plain joint; (d) J9 – plain joint; (e) J10 – plain joint; (f) J1R – epoxy repair scheme; (g) J2R – epoxy repair scheme; (h) J3 – EB90 scheme; (i) J4 – EB45 scheme; (j) J8R – epoxy repair + EB45 scheme; (k) J5 – NSM scheme; (l) J10R – epoxy repair + NSM scheme; (m) J6 – SP scheme; (n) J9R – epoxy repair + SP scheme and (o) J7 – FP scheme.

Cyclic loading

The hysteresis curves arising from the joints subjected to cyclic loading are presented in Figures 8(b) to (o). It can be observed that the envelopes of the hysteresis curves (i.e. backbone curves) exhibited similar shapes when compared to the moment–rotation curve of the monotonic loading test on joint J1. There were two basic types of hysteresis behaviours however. For brittle failure arising from timber splitting or plate debonding, the backbone curves showed distinct strength degradation. This type of behaviour was observed in joints J1R, J2R, J6 and J9R, and the curves are plotted in Figure 8(f), (g), (m) and (n), respectively. Even though the strength was reduced under repeated load for the remaining test joints (herein impairment of strength, as to be defined in the following sections), there was no brittle joint failure. For this class of behaviour, the backbone curves increased in moment with rotation, except for occasional load drops due to strength impairments.

Not all hysteresis curves were symmetric for the first few cycles (i.e. amplitudes = 5, 10 and 15 mm) due to variabilities in manufacturing and test setup. The moments were, however, still small. For the following discussions related to energy dissipation and equivalent viscous damping ratio, the data arising from larger amplitude cycles (i.e. amplitudes = 20, 40, 80 and 120 mm) were used as the results were more stable.

All hysteresis responses exhibited pinched behaviour and such behaviour is common for dowel-type timber joints. The curves were pinched in the first and third quadrants when the load in the unloading stage was small. The moment carried by the joint arises from three mechanisms, namely, (1) bearing strength of the timber (and the FRP or steel strengthening materials), (2) bending of steel dowels and (3) friction between the steel dowels and the timber holes. All three mechanisms contributed to joint deformation in the first load cycle after yield. When the joints were unloaded or reloaded, the dowels slid in the holes and did not bear into the timber. Therefore, only bending (or reversed bending) of the dowels as well as friction force acted at this stage.

Table 3 provides a summary of the hysteresis curves at the amplitudes V_y,est, 2V_y,est, 4V_y,est and 6V_y,est (i.e. C6–C17). Both the positive and negative rotation phases for each cycle are reported. The positive rotation phase is equal to the half cycle when φ > 0 (i.e. the angle between the central member and the side members is greater than 90°). The negative rotation phase is equal to the half cycle when φ < 0 (i.e. angle between central member and side members is less than 90°). The peak rotation of each cycle is denoted by φ_peak (i.e. maximum rotation in the positive rotation phase and minimum rotation in the negative rotation phase; unit: radians). M_peak is the moment corresponding to φ_peak (unit: kN m). In addition in this table, ΔM is the impairment of strength value defined in accordance with BS EN 12512:2001 (2001) (unit: kN m). This term refers to the loss of strength under repeated cyclic loading and it is equal to the decrease of strength arising from three repeated load cycles at the same amplitude. The definitions of φ_peak, M_peak and ΔM are shown graphically in Figure 9(a). For all joints except J6 and J9R, M_peak coincides with the maximum moment (i.e. minimum in the negative rotation phase) achieved for each set of the three cycles per amplitude. For joints J6 and J9R, M_peak is influenced by plate debonding and strength degradation. For C15 of joint J6, the maximum moment was 24.85 kN m and the minimum moment was 20.49 kN m. For C12 of joint J9R, the maximum moment was 24.40 kN m and the minimum moment was 20.49 kN m.

Table 3.

Details of hysteresis loops.

Amplitude			V_y,est			2V_y,est			4V_y,est			6V_y,est
Scheme	ID		M_peak	φ_peak	ΔM	M_peak	φ_peak	ΔM	M_peak	φ_peak	ΔM	M_peak	φ_peak	ΔM
Plain	J2	+	3.88	0.034	0.23	5.17	0.068	0.40	6.77	0.141	0.89	7.41	0.219	1.53
		−	−1.32	−0.033	−0.02	−5.23	−0.065	0.09	−7.46	−0.127	0.54	−7.85	−0.186	0.67
	J8	+	3.62	0.023	0.25	5.85	0.054	0.43	7.63	0.122	1.04	8.64	0.183	1.21
		−	−1.48	−0.027	0.08	−5.34	−0.055	0.23	−7.05	−0.113	0.66	−7.59	−0.162	0.70
	J9	+	3.54	0.034	0.22	5.38	0.068	0.34	6.89	0.140	0.81	7.77	0.208	1.03
		−	−1.03	−0.033	−0.01	−4.60	−0.065	0.18	−6.16	−0.127	0.49	−6.65	−0.179	0.56
	J10	+	0.84	0.022	0.02	4.66	0.053	0.18	7.01	0.120	0.75	7.74	0.178	1.05
		−	−3.59	−0.027	0.18	−5.34	−0.057	0.35	−6.68	−0.116	0.54	−7.12	−0.162	0.82
Repair	J1R	+	2.86	0.025	−0.01	7.69	0.052	0.25	9.76	0.118	2.49	7.67	0.197	1.08
		−	−5.12	−0.027	0.20	−7.08	−0.058	0.50	−7.73	−0.118	1.88	−6.39	−0.148	0.81
	J2R	+	3.87	0.023	0.05	7.65	0.051	0.39	9.64	0.118	1.12	9.02	0.169	2.35
		−	−2.89	−0.023	0.10	−6.75	−0.051	0.32	−8.55	−0.109	0.88	−7.38	−0.147	1.87
EB	J3	+	4.66	0.026	0.27	6.57	0.059	0.62	8.59	0.128	0.92	9.51	0.187	1.04
		−	−0.69	−0.025	−0.04	−5.18	−0.051	0.31	−8.05	−0.110	0.79	−9.11	−0.157	0.89
	J4	+	1.14	0.023	−0.08	5.93	0.053	0.43	8.65	0.120	0.90	10.14	0.178	1.15
		−	−4.30	−0.024	0.18	−6.67	−0.052	0.43	−8.76	−0.110	0.84	−9.77	−0.158	0.97
	J8R	+	3.07	0.025	0.02	8.09	0.054	0.44	10.64	0.120	0.85	11.94	0.190	1.10
		−	−3.60	−0.023	0.04	−7.84	−0.051	0.43	−10.24	−0.109	0.84	−11.36	−0.164	1.00
NSM	J5	+	1.23	0.025	−0.12	6.03	0.053	0.29	9.54	0.118	0.90	11.14	0.177	0.95
		−	−4.60	−0.025	0.23	−7.23	−0.053	0.55	−9.73	−0.110	0.77	−11.07	−0.158	0.90
	J10R	+	3.31	0.024	0.11	7.66	0.054	0.42	10.66	0.121	0.79	12.55	0.192	1.24
		−	−2.63	−0.024	0.10	−7.27	−0.052	0.41	−10.59	−0.111	0.67	−12.10	−0.169	0.98
SP	J6	+	7.39	0.021	0.31	17.74	0.046	1.25	24.73	0.110	3.60	20.65	0.185	3.16
		−	−5.16	−0.022	0.19	−16.59	−0.044	1.01	−24.26	−0.098	2.78	−20.49	−0.163	2.67
	J9R	+	8.40	0.021	0.15	19.54	0.047	1.50	19.83	0.111	7.92	14.60	0.191	2.21
		−	−3.87	−0.024	0.11	−15.93	−0.046	1.10	−12.66	−0.104	0.81	−13.19	−0.164	1.51
FP	J7	+	4.09	0.024	0.26	8.80	0.051	1.03	11.64	0.116	2.30	12.80	0.186	2.72
		−	−3.30	−0.023	0.12	−8.17	−0.050	0.70	−11.10	−0.108	2.02	−11.97	−0.166	2.54

EB: externally bonded FRP; NSM: near-surface mounted FRP; SP: steel plate; FP: FRP plate.

V_y,est is the estimated yield displacement by actuator movement, V_y,est = 20 mm; M_peak is the moment at maximum (positive) or minimum (negative) rotation (unit: kN m); φ_peak is the peak rotation, that is, maximum rotation in the positive rotation phase or minimum rotation in the negative rotation phase (unit: radians); ΔM is the impairment of strength (unit: kN m).

+represents the positive rotation phase (φ > 0) and – represents the negative rotation phase (φ < 0).

Figure 9.

(a) Details of hysteresis loops and (b) definitions of strength and stiffness.

Strength, stiffness and ductility

The strength, stiffness and ductility are plotted graphically in Figure 9(b) and are defined in accordance with BS EN 12512:2001 (2001). These quantities are extracted from the backbone curves of the moment–rotation responses. There are two backbone curves for each joint to consider, namely, positive rotation phase and negative rotation phase. The results arising from both of these phases are presented in Table 4.

Table 4.

Strength, stiffness and ductility results.

Scheme	ID	Positive rotation phase (φ > 0)						Negative rotation phase (φ < 0)
Scheme	ID	K	M_y	M_u	φ_y,mod	φ_u,mod	D	K	M_y	M_u	φ_y,mod	φ_u,mod	D
Plain	J1	180	5.43	8.51	0.030	0.162	5.38	–	–	–	–	–	–
	J2	131	4.40	7.41	0.034	0.220	6.54	157	−6.19	−7.85	−0.039	−0.164	4.15
	J8	175	5.23	8.64	0.030	0.180	6.03	150	−5.48	−7.59	−0.037	−0.152	4.15
	J9	131	4.83	7.77	0.037	0.203	5.49	181	−4.45	−6.65	−0.025	−0.153	6.24
	J10	146	5.42	7.74	0.037	0.162	4.36	151	−4.82	−7.12	−0.032	−0.159	4.98
Repair	J1R	212	8.07	9.76	0.038	0.106	2.80	225	−6.06	−7.73	−0.027	−0.116	4.30
	J2R	247	7.26	9.64	0.029	0.110	3.74	203	−6.76	−8.55	−0.033	−0.102	3.06
EB	J3	185	5.86	9.51	0.032	0.187	5.91	209	−5.93	−9.11	−0.028	−0.136	4.79
	J4	231	5.95	10.14	0.026	0.160	6.20	206	−6.17	−9.77	−0.030	−0.155	5.18
	J8R	340	7.33	11.94	0.022	0.173	8.04	378	−6.57	−11.36	−0.017	−0.152	8.72
NSM	J5	250	6.59	11.14	0.026	0.157	5.97	209	−7.19	−11.07	−0.034	−0.156	4.54
	J10R	243	7.76	12.55	0.032	0.181	5.66	255	−7.93	−12.10	−0.031	−0.156	5.02
SP	J6	696	17.73	24.85	0.025	0.098	3.84	807	−17.15	−24.26	−0.021	−0.082	3.85
	J9R	805	18.62	24.40	0.023	0.072	3.11	920	−15.77	−17.83	−0.017	−0.060	3.52
FP	J7	288	8.26	12.84	0.029	0.176	6.16	303	−7.81	−11.97	−0.026	−0.154	5.96

EB: externally bonded FRP; NSM: near-surface mounted FRP; SP: steel plate; FP: FRP plate.

K is the stiffness (unit: kN m/rad); M_y is the yield moment (unit: kN m); M_u is the ultimate moment (unit: kN m); φ_y,mod is the modified yield rotation (unit: radians); φ_u,mod is the modified ultimate rotation (unit: radians); D is the ductility factor (dimensionless).

The ultimate moment, M_u, that the joint achieved during the test is used to refer to the strength. For the positive rotation phase, M_u refers to the maximum moment resisted by the joint. For the negative rotation phase, M_u refers to the minimum moment.

The stiffness is defined as the slope of a straight line of constant slope that passes through coordinates (φ₁₀, M₁₀) and (φ₄₀, M₄₀) on the moment–rotation curve. More specifically

K = \frac{(M_{40} - M_{10})}{(φ_{40} - φ_{10})}

(3)

where K is the stiffness of the moment-resisting joint; M₁₀ and M₄₀ are 10% and 40% of the ultimate moment, respectively; and φ₁₀ and φ₄₀ are the rotation corresponding to M₁₀ and M₄₀, respectively. The yield point (φ_y, M_y) is defined as the intersection of two straight lines. The first line passes through coordinates (φ₁₀, M₁₀) and (φ₄₀, M₄₀), while the second line is the tangent arising from the moment–rotation envelope with a slope 1/6th of the stiffness.

The ductility, D, is defined in accordance with BS EN 12512:2001 (2001) as per

D = \frac{φ_{u}}{φ_{y}}

(4)

where φ_u is the ultimate rotation and φ_y is the yield rotation. Note that φ_u is the rotation corresponding to the yield moment M_u. To eliminate the influence of variability in manufacturing and the clearance between the steel dowel and the adjacent walls of the timber hole at the beginning of the test, the modified yield rotation, φ_y,mod, and the modified ultimate rotation, φ_u,mod, are defined as

φ_{y, \mod} = φ_{y} - \frac{(4 φ_{10} - φ_{40})}{3}

(5)

φ_{u, \mod} = φ_{u} - \frac{(4 φ_{10} - φ_{40})}{3}

(6)

The definitions of φ_y,mod and φ_u,mod are shown graphically in Figure 9(b) and the joint ductility is therefore calculated as follows

D = \frac{φ_{u, \mod}}{φ_{y, \mod}}

(7)

For the joints that exhibited strength degradation (i.e. splitting in joints J1R and J2R, and debonding in joints J6 and J9R), M_u was reached before the maximum rotation was achieved. For all other joints, M_u refers to the moment at maximum rotation. Based on the moment–rotation curves for these cases, the moment increases with rotation. In addition, the actual ductility may be greater than the calculated value. The value of D is therefore the minimum value of the real ductility of these joints.

The average yield strength (in absolute terms) of the plain joints was found to be 5.14 kN m and the yield values ranged from 4.82 to 5.43 kN m. The average ultimate strength was found to be 7.70 kN m, and the values ranged from 6.65 to 8.51 kN m.

Inspection of the results reveals the following trends. For epoxy repaired joints (J1R and J2R), both the strength and stiffness were increased as compared to the undamaged joints (J1 and J2). The strength and stiffness were increased up to 30% and 89%, respectively. The joint ductility was, however, limited because the repaired joints cracked early in the test. The results were also influenced by joint opening and closing cracks that occurred on the side members.

There was no significant difference observed between the joints strengthened with the EB90 and EB45 schemes. In comparison to the mean value of the plain joints, the yield and ultimate strengths of the strengthened joints increased up to 20% and 32%, respectively. Comparing the result of joint J8R with those of joints J1R and J2R, it can be seen that the EB scheme prevented cracking of the timber and also enhanced joint ductility. The stiffness was also increased. The application of the EB45 scheme for the epoxy repaired joint increased the stiffness by up to 70% with respect to the average stiffness of joints J1R and J2R. In addition, the joint ductility was increased over 2.3 times compared with the average of joints J1R and J2R.

For undamaged and repaired joints, the strength of the joints containing the NSM scheme was higher than those of joints containing the EB schemes. The stiffnesses between the undamaged and repaired joints were similar and they increased about 53% compared to the plain joints. The NSM scheme also produced an acceptable level of ductility since no timber members cracked.

Joints containing the SP schemes (J6 and J9R) performed in a superior manner in both strength and stiffness due to the high bearing strength and stiffness of the steel plates. The ultimate strength of these two joints was more than three times that of the plain joints, while the stiffness was increased more than four times. The ductility of these joints was low due to plate debonding. Nonetheless, the moment-carrying capacity was higher than that of all the other joints in the post-debonding cycles.

The strength and stiffness of the joint containing the FP scheme (joint J7) were lower than those of the joints containing the SP scheme. The former was, however, higher than the remaining strengthened joints (with the exception of stiffness for joint J8R). The behaviour of joint J7 was ductile since there was no observed debonding or cracking.

Energy dissipation and equivalent viscous damping ratio

Energy dissipation results are presented in Table 5 for the amplitudes V_y,est, 2V_y,est, 3V_y,est and 4V_y,est. These levels were chosen to coincide with the post-yield joint behaviour. The dissipated energy for each load cycle is the area enclosed by the corresponding hysteresis loop (unit: kN m).

Table 5.

Energy dissipation results.

Amplitude		V_y,est			2V_y,est			4V_y,est			6V_y,est			Total^a
Scheme	ID	C6	C7	C8	C9	C10	C11	C12	C13	C14	C15	C16	C17	Total^a
Plain	J2	40	22	19	266	160	143	1065	689	622	1716	1172	1077	7067
	J8	21	12	10	245	143	125	1034	653	582	1502	1089	1000	6445
	J9	27	18	16	249	146	130	1007	652	585	1494	1093	1009	6468
	J10	21	11	9	204	130	114	946	608	547	1336	986	902	5845
Repair	J1R	41	26	24	318	220	199	1314	828	674	1380	1019	913	7015
	J2R	17	10	10	264	181	166	1273	799	723	1464	991	833	6741
EB	J3	34	18	16	265	144	125	1113	650	581	1557	1067	978	6601
	J4	24	14	12	252	148	129	1142	664	591	1639	1131	1030	6810
	J8R	11	9	9	283	183	167	1389	819	736	2229	1491	1341	8685
NSM	J5	28	15	13	267	167	148	1196	706	634	1778	1206	1099	7299
	J10R	15	9	8	272	182	166	1365	816	741	2308	1541	1393	8829
SP	J6	39	15	14	572	123	106	2688	781	690	3826	1976	1630	12,490
	J9R	31	9	8	568	140	120	2551	1010	823	2737	1848	1638	11,499
FP	J7	26	16	14	333	160	139	1442	771	659	2135	1289	1157	8175

EB: externally bonded FRP; NSM: near-surface mounted FRP; SP: steel plate; FP: FRP plate.

V_y,est is the estimated yield displacement by actuator movement, V_y,est = 20 mm.

Total energy dissipated during test, from C1 to C17.

All values are in kN m.

The average total dissipated energy of the plain joint (J2) under cyclic loading was 6456 kN m. Upon consideration of the results of joints containing the strengthening and repair schemes, the SP joints had the highest energy dissipation. This was 78% higher than the average result of the plain joints. The FP scheme was better performing than the NSM schemes for the undamaged joints. In addition, the EB schemes did not noticeably increase the energy dissipation capacity. Even though the repaired joints J1R and J2R had higher stiffness and strength than the plain joints, there was not much improvement in energy dissipation. This was due to poor joint ductility. It should be noted though that the EB and NSM schemes enhanced the ductility of the repaired joints. The energy dissipation of joints J8R and J10R was at least 24% higher than that of the repaired joints that did not contain strengthening.

Upon consideration of each set of three load cycles for the same amplitude, the dissipated energy of the first cycle was recorded to be higher than those of the second and third cycles. In addition, the energy dissipated in the second cycle was higher than that in the third cycle. This result demonstrates the pinching effect and joint degradation with increased deformation. For each three-cycle set, the dissipated energy in the third cycle was about 40%–60% of that in the first cycle. For the joints containing the SP scheme, the energy of the third cycle was less than 35% of that in the first cycle and this was largely achieved before the plate had debonded. This result indicates that the SP scheme cannot increase the energy dissipation capacity significantly during the repeated loading cycles.

According to BS EN 12512:2001 (2001), the equivalent viscous damping ratio, ζ_eq, is calculated as follows

ζ_{eq} = \frac{E_{d}}{2 π E_{p}}

(8)

where E_d is the energy dissipated per half cycle and E_p is the available potential energy. The definitions of E_d and E_p are in turn shown diagrammatically in Figure 10. The ζ_eq values extracted from C6–C17 (i.e. Cycle No. 6–Cycle No. 17) for each joint test specimen are presented in Table 6.

Figure 10.

Definition of equivalent viscous damping ratio.

Table 6.

Equivalent viscous damping ratio results.

Amplitude		V_y,est			2V_y,est			4V_y,est			6V_y,est
Scheme	ID	C6	C7	C8	C9	C10	C11	C12	C13	C14	C15	C16	C17
Plain	J2	0.092	0.051	0.045	0.170	0.101	0.090	0.203	0.145	0.136	0.197	0.157	0.151
	J8	0.062	0.033	0.029	0.140	0.086	0.075	0.192	0.138	0.128	0.170	0.136	0.131
	J9	0.061	0.038	0.035	0.133	0.078	0.070	0.183	0.131	0.121	0.167	0.135	0.129
	J10	0.024	0.021	0.021	0.079	0.070	0.064	0.177	0.127	0.117	0.166	0.134	0.129
Repair	J1R	0.040	0.034	0.033	0.087	0.085	0.079	0.198	0.149	0.125	0.179	0.143	0.131
	J2R	0.026	0.017	0.016	0.107	0.081	0.075	0.192	0.134	0.126	0.177	0.144	0.135
EB	J3	0.083	0.043	0.038	0.155	0.081	0.071	0.187	0.117	0.108	0.160	0.118	0.111
	J4	0.037	0.031	0.030	0.094	0.068	0.063	0.178	0.115	0.106	0.156	0.117	0.111
	J8R	0.018	0.015	0.015	0.096	0.071	0.065	0.184	0.117	0.108	0.171	0.122	0.114
NSM	J5	0.028	0.022	0.020	0.088	0.074	0.066	0.168	0.111	0.102	0.151	0.109	0.103
	J10R	0.034	0.021	0.020	0.108	0.079	0.072	0.179	0.114	0.107	0.165	0.119	0.111
SP	J6	0.053	0.018	0.017	0.123	0.026	0.023	0.168	0.056	0.052	0.177	0.105	0.090
	J9R	0.044	0.011	0.009	0.119	0.029	0.025	0.211	0.124	0.102	0.169	0.128	0.118
FP	J7	0.051	0.031	0.028	0.125	0.067	0.060	0.180	0.114	0.104	0.157	0.112	0.106

EB: externally bonded FRP; NSM: near-surface mounted FRP; SP: steel plate; FP: FRP plate.

V_y,est is the estimated yield displacement by actuator movement, V_y,est = 20 mm.

Note that the equivalent viscous damping ratio is dimensionless.

The equivalent viscous damping ratio is a non-dimensional value and it changes from cycle to cycle. For smaller amplitude cycles (i.e. C6–C8), ζ_eq was small and this suggests that the hysteresis behaviour was not fully developed. As the amplitude increased though and the non-linear behaviour was well developed, ζ_eq increased and became relatively stable. The maximum value of ζ_eq occurred in C12 for most of the joints and on average the value was 0.186. In the last set of three cycles (C15–C17), ζ_eq decreased. In addition, among the sets of three load cycles within the same amplitude, the equivalent viscous damping ratio of the first cycle was higher. The ratio then decreased and was similar in the following two cycles. This showed that the performance of the joint degraded with repeated loading cycles. The reduction of ζ_eq was on average 38% of the first cycle for the same amplitude. The decrease in ζ_eq was most obvious in the joints containing SP schemes, where a reduction of more than 70% was observed before the plates debonded. Even after the plates debonded, this reduction was still higher than most of the other results. This indicates that although the bonded steel plates enhanced the strength and stiffness significantly for the first cycle in each set of three cycles, they were not very effective in the subsequent repeated loading cycles.

Conclusion

The results of a series of tests on the effectiveness of FRP and steel strengthening and repair schemes for moment-resisting timber joints have been reported. A total of 15 tests have been performed on 10 joint specimens, while 14 of the joints were subjected to cyclic loading and the remaining one was subjected to monotonic loading. The test results were analysed largely in accordance with BS EN 12512:2001 (2001), of which the strength, stiffness, ductility, and characteristics of the hysteresis responses have been reported. The following conclusions can be drawn:

The use of epoxy repair to damaged joints can increase the strength and stiffness of the joints. The joint ductility was, however, poor if no additional strengthening was applied to the joint.

The EB schemes improved joint ductility due to the prevention of timber cracking before an earthquake event. There was, however, no strong evidence to show an effective enhancement in joint strength and stiffness. With application of the EB scheme to the repaired joint, both the strength and stiffness were enhanced significantly. This was due to avoidance of brittle failure arising from cracking or debonding.

The NSM scheme increased both the joint strength and stiffness. The NSM scheme was, however, less effective in strengthening the repaired joint in comparison to the EB strengthening. No brittle failure was found to occur for any of the NSM joints.

The SP scheme was found to increase both the strength and stiffness of the joints significantly. Joint ductility was, however, limited due to plate debonding. In addition, the energy dissipation capacity and equivalent viscous damping ratio were observed to decrease during repeated load cycles.

The strength of the FP-strengthened joint was higher than those of the all other strengthened joints besides the joints containing the SP scheme. The FP scheme also provided the second highest stiffness of the strengthened undamaged joints. No debonding or other brittle failure modes were observed during this test.

The eventual method to be adopted by engineers will depend on the joint and the method for repair. NSM will be particularly suited to the restoration of historical structures. The external bonding of FRP plates, however, is preferred if there are no restrictions on the visibility of the repair or strengthening intervention.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

ORCID iD

Scott T Smith

References

ACI 440.3R-04:2004 (2004) Guide test methods for fiber-reinforced polymer (FRPs) for reinforcing or strengthening concrete structures.

Blaß

Schädle

(2011) Ductility aspects of reinforced and non-reinforced timber joints. Engineering Structures 33(11): 3018–3026.

Blaß

Schmid

Litze

et al . (2000) Nail plate reinforced joints with dowel-type fasteners (Paper 8.6.4.). In: Proceedings of the WCTE 2000-world conference on timber engineering, Vancouver, BC, Canada, 31 July–3 August.

BS EN 10002-1:2001 (2001) Metallic materials – tensile testing – part 1: methods of test at ambient temperature.

BS EN 12512:2001 (2001) Timber structures-test methods-cyclic testing of joints made with mechanical fasteners.

BS EN 1995-1-1 (1994). Eurocode 5: design of timber structures – part 1-1: general rules and rules for buildings.

BS EN 408:2003 (2003) Timber structures. Structural timber glued laminated timber. Determination of some physical mechanical properties.

BS EN 527:1999 (1999) Plastics – determination of tensile properties.

Chen

(1999) Mechanical behavior of fiberglass reinforced timber joints. PhD Thesis, École Polytechnique Fédérale de Lausanne, Lausanne.

10.

Davis

Claisse

(2001) Resin-injected dowel joints in glulam and structural timber composites. Construction and Building Materials 15(4): 157–167.

11.

Haller

Wehsener

(1999) Use of technical textiles and densified wood for timber joints. In: Proceedings of the 1st international RILEM symposium on timber engineering, Stockholm, 13–14 September, pp. 717–726.

12.

Haller

Birk

Offermann

et al . (2006) Fully fashioned biaxial weft knitted and stitch bonded textile reinforcements for wood connections. Composites Part B: Engineering 37(4–5): 278–285.

13.

Kasal

Heiduschke

Haller

et al . (2002) Fiber-reinforced beam-to-column connections for seismic applications. In: Proceedings of the CIB W18 meeting, Kyoto, Japan, 16–19 September, pp. 1–11.

14.

Kasal

Heiduschke

Kadla

et al . (2004a) Laminated timber frames with composite fibre-reinforced connections. Progress in Structural Engineering and Materials 6(5): 84–93.

15.

Kasal

Pospisil

Jirovsky

et al . (2004b) Seismic performance of laminated timber frames with fiber-reinforced joints. Earthquake Engineering and Structural Dynamics 33(5): 633–646.

16.

Lam

Gehloff

Closen

et al . (2010) Moment-resisting bolted timber connections. Proceedings of the Institution of Civil Engineers – Structures and Buildings 163(4): 267–274.

17.

Leijten

AJM

(1988) Steel reinforced joints with dowels bolts. In: Proceedings of the international conference on timber engineering, Seattle, WA, 19–22 September, vol. 2, pp. 474–488.

18.

Leijten

AJM

(1991) Development of reinforced joints with densified veneer wood. In: Proceedings of the international timber engineering conference, London, 2–5 September, pp. 3.36–3.42.

19.

Leijten

AJM

(1999) Locally reinforced timber joints with expanded tube fasteners. HERON 44(3): 131–161.

20.

Porteous

Kermani

(2007) Structural Timber Design to Eurocode 5. Oxford: Blackwell.

21.

Rodd

Leijten

AJM

(2003) High-performance dowel-type joints for timber structures. Progress in Structural Engineering Materials 5(2): 77–89.

22.

Rodd

Hilson

Spriggs

et al . (1989) Resin injected mechanically fastened timber joints. In: Proceedings of the 2nd Pacific timber engineering conference, Auckland, New Zealand, 28–31 August, pp. 350–357. Auckland, New Zealand: The University of Auckland.

23.

Strongwell (n.d.). Available at: http://www.strongwell.com/wp-content/uploads/2013/04/SAFSTRIP-Brochure.pdf (accessed August 2017).

24.

Yang

(2013) Influence of strengthening repair schemes on dowel type timber joints moment resisting frames. PhD Dissertation, Department of Civil Engineering, The University of Hong Kong, Pokfulam, Hong Kong.