Experiment on Behavior of a New Connector Used in Bamboo ( Timber ) Frame Structure under Cyclic Loading

Connection is an important part of the bamboo and timber structure, and it directly influences the overall structural performance and safety. Based on a comprehensive analysis of the mechanical performance of several wood connections, a new connector for the bamboo (timber) frame joint was proposed in this paper. /ree full-scale T-type joint specimens were designed to study the mechanical performance under cyclic loading./e thickness of the hollow steel column was different among three specimens./e specimens were loaded under displacement control with a rate of 10mm per minute until the specimens reach failure. It was observed that the failures of three specimens were caused by the buckling of flanges in the compression and that the steel of connections does not yield. /e load-displacement hysteretic curve for three specimens is relatively plump, and the stiffness of connection degenerates with the increasing of cyclic load. /e maximum rotation is 0.049 rad, and the energy dissipation coefficient is 1.77. /e thickness of the hollow steel column of the connector has significant impact on the energy dissipation capacity and the strength of the connection. A simplified moment-rotation hysteresis model for the joint was proposed.


Introduction
Timber is a natural organic material, and people easily get it from nature and use it without much processing; therefore, timber was employed as a construction material a long time ago.Timber has higher strength in tension and compression parallel to grain, light mass, and good durability, some 1000year timber buildings still stand well [1].Because of friendly environment, graceful timber texture, and simple nature, timber building will still be enormously appealing to people.For the timber frame structure, the beam-to-column connection is usually the most unsubstantial part on account of fabricated construction.e mechanical behavior of timber joints directly influences the overall timber structural performance and safety.As a result, the design of timber joints is extremely important.In some Asian countries such as China, South Korea, and Japan, mortise and tenon joints [1][2][3][4][5] are traditionally used in timber buildings from dwelling houses to palaces.In these joints, steel fasteners are not used, therefore keeping the original beauty of the timber.However, due to the slippage between the mortise and tenon, the energy dissipation capacity of the whole structure is affected under earthquake load.e whole building can even collapse due to the separation between the mortise and tenon.In addition, the mortise-to-tenon connection wakens the column cross section at the connection, which reduces the vertical loadcarrying capacity and also negatively affects the energy dissipation capacity of the column.To prevent the mortise-totenon connection from separation and improve the strength of connection, Bulleit et al. [6] used wooden pegs to fasten mortise-to-tenon connection.But the study showed that the split failure owes to prying force of the wood peg occurred along with a peg hole when there is a distance between the tenon and the sill of mortise, and either shear failure of the tenon appeared.Moreover, Hong et al. [7] employed a T-type steel plate to strengthen the mortise-to-tenon connection and not to enhance the energy dissipation capacity of the joint.
Steel plate-bolted connection is another joint commonly used to connect the beam and column in a timber structure, which often is applied in the heavy timber structure and has architecture beauty because the connecting steel plate is covered with wood.However, for these types of joints, the brittle failure mode was obvious [8][9][10].Split failure along the bolt hole in the beam is the main failure mode in this kind of joint, and it reduces the load-carrying capacity of the joint.As the beam and column are slotted to accommodate the steel plate, the vertical load-carrying capacity is reduced.Under cyclic loading, the hysteretic curve shows a clear "pinching" effect because of the gap between the bolt and wall of the hole [8][9][10].
Glued-in-rod connection is also commonly used to connect the beam and column in a timber building.e rod is embedded in timber to connect the beam and column and to transfer the load from the beam to the column.With appropriate materials and good construction quality, the glued-in-rod connection has demonstrated good structural performance [11,12].e glued-in-rod connection has been widely studied and used in various projects [13].e advantage of glued-in-rod connection is that the steel rod is embedded in timber to protect the steel rod from corrosion, and the durability is better.In order to get good construction quality, generally, the steel rod should be glued well with one timber first part in advance and then being connected with another part.us, the rod hole position on the beam and column must be precise.As timber and glue are brittle materials, the principle of energy dissipation is not obvious.Vašek [14,15] used two U-type steel connectors located at upper and bottom edges of the beam to fasten the steel rods embedded in the timber beam and column.
is new method could reduce stress concentration to the beam and column and increase the energy dissipation capacity of the joint by the deformation of U-type steel connectors.In addition, the U-type steel connectors are applied to connect the timber beam and column, and the position of the hole on the steel connector can easily be changed, which is convenient for connection construction on site.
Bolted timber-timber connection is a simple and practical connection for the timber beam-to-column joint.Only bolts are used to fasten the timber beam and column and to carry load from the girder to the post [16,17].Steel nails sometimes are used to substitute for bolts.
e different mechanical performance of connection can be obtained by changing the quantity and arrangement style of bolts.Some special materials, such as steel plate or hard wood, are inserted into the contacted surface of the beam and column to improve the performance of connection as well.
For the timber beam-to-column joint, load-carrying capacity and energy dissipation are the two primary factors to judge the performance of the joint.Huang [18] gave a better energy dissipation connector for the timber frame joint, but the connection between the beam and column is not robust, slippage happens under earthquake, and shear stiffness of the joint is less.Based on the comprehensive analysis of the abovementioned joints, a new beam-tocolumn connector was developed and presented in this paper.e joint can connect top and bottom columns and also link beam and column parts without weakening the column cross section.As the mechanical behavior of this joint under the earthquake load is unknown, an experimental study was conducted to evaluate the stiffness, strength, energy dissipation, and resilience of the joint with different thicknesses of the hollow steel column.

Joint Details and Fabrication
e joint consists of two parts.One part is beam and column members made of parallel strand bamboo (PSB) [19][20][21].e beam is 55 mm in width, 200 mm in depth, and 1050 mm in length.Two beams were assembled in parallel.At the end of the beams where a bolt hole was located, carbon fiberreinforced plastic, 8 mm in width, was used to wrap the beams to prevent it from cracking along the bolt hole.e top and bottom columns are 200 mm in width, 240 mm in depth, and 700 mm in length.e column depth is paralleled to the beam axis.e other part is a steel connector, which is the most important part of the joint.
e details of the connector are shown in Figure 1.Four steel plates are welded together to form a hollow column, and then, a horizontal steel plate is welded in the middle of the hollow column to reinforce the hollow column and to transfer load from the top column to the bottom column.e top and bottom plates (flanges) of the I-shaped steel beam were bent into L shape and welded to the hollow steel column.A steel plate web was welded between the top and bottom flanges to form an I-shaped steel beam.Two shear connection plates were welded to the hollow steel column and fastened to the web of the I-shaped steel beam with a 16 mm diameter bolt.
is is to ensure the connection between the hollow steel column and the I-shaped steel beam in case the welding between the hollow steel column and I-shaped steel beam flanges is broken.
e PSB columns were placed into the hollow steel column and connected to the hollow steel column with four 14 mm diameter bolts.Two PSB beams were connected to the steel web with four 16 mm diameter bolts.

Specimen Designing.
ree joint specimens were fabricated in site.Except for the thickness of the hollow steel column, the specimens are identical.Details of the specimens are shown in Figure 2. Holes in bamboo specimen (beam and column) and steel plate are 1.5 mm greater than those of bolts, which is easy to assemble.A detailed description of thickness of each tested steel connection plate is reported in Table 1.

Mechanical Properties of Materials.
All the steel plates are of grade Q235B in accordance with the Chinese standard (GB/T700-2006) [22].e material properties of the steel plates are determined according to EN10002-1 [23].Table 2 lists the mechanical properties of the steel plates.e 14 mm diameter bolts, which were used to fasten the hollow steel column and PSB columns, had the average yield strength of 804 MPa under tension.e 16 mm diameter bolts had the average yield strength of 2241 MPa under bending.
According to the ASTM standard D143-09 [24], the ultimate compressive and tensile strength values of PSB parallel to grain are 65 MPa and 100 MPa, respectively, and the ultimate tensile strength value of PSB perpendicular to grain is 4.4 MPa.

2
Advances in Materials Science and Engineering

Test Setup.
A schematic illustration of the test setup is shown in Figure 3. e main purpose of this research is to study the mechanical performance of the connection specimens under cyclic loading.In order to load the specimen with the existing loading frame, the column was rotated an angle of 90 °and was xed to the oor channel with four 30 mm diameter bolts.A two-way 250 mm hydraulic actuator (100 kN capacity) was used to apply the cyclic load.e center of the actuator was 1000 mm above the top surface of the PSB column, and the head of the actuator was about 150 mm away from the PSB beam.
e beam displacements were monitored by a laser displacement sensor, and the corresponding load value was measured by a load cell mounted on the actuator rod.Two laser displacement sensors #3 and #4 were located at the beam ange to measure the rotation of the connection.e other two sensors #1 and #2 were located at the hollow steel column to measure the defection of the hollow column.
For the three specimens, the sensors were located at the same locations of the specimen.

Loading Schedule.
In this study, a controlled cyclic displacement scheme was used.e maximum displacement was 10 mm in the rst two cyclic loading, and the loading rate was 10 mm per minute.e maximum displacement was then increased by 10 mm after each step until specimen failure which showed specimen lost bearing loading capacity.e loading scheme is shown in Figure 4.
It was assumed that the actuator force in compression was positive, and the beam displacement moving away from the reaction wall was positive.

Test Results and Analysis
e test results of the three specimens are shown in Table 3.For each experiment, the failure mode, the energy dissipation, and the resilience model of each specimen will be discussed in the following sections.

Failure Mode.
From Figure 3, it can be observed that one lateral beam ange was pulled and the opposite beam ange was compressed under loading.For the three specimens, the failure occurred when the beam ange buckled under compressive force.Figures 5-7 show the failure mode of the three specimens.It was noticed that the beam ange buckling occurred suddenly without yielding of the steel plate, which is a brittle failure and results in the failure of test specimens.Besides the beam ange buckling, for the specimen CJ2 in which the hollow steel column is 4 mm in thickness, the hollow steel column was also deformed around the beam ange because of tensile fore of the beam ange.e bolts and PSB members with larger cross-sectional dimensions were still in good condition after the tests.Advances in Materials Science and Engineering According to tested results, it was also observed that the three specimens with the same thickness in the beam ange lost capacity in the same location and for the same reason, but they had obvious di erence in ultimate bearing capacity; the main reason was that the boundary conditions of the compressed beam ange were not ideal situation under axial loading.

Load-Displacement Hysteretic Loop at the End of the Beam.
e load-displacement hysteretic curve of the joint re ects the overall performance of connection.e load-displacement hysteretic curves at the end of the beam are shown in Figures 8-10.As can be seen from the gures, all three specimens exhibit good performance.e strength deterioration with the progress of the cyclic load was not observed.e rotational sti ness, however, deteriorated with the increase of cyclic load.
For the specimen CJ2 (4 mm thick hollow steel column), the beam ange in compression buckled suddenly after 9 displacement cycles.For the specimen CJ1 (5 mm thick hollow steel column) and the specimen CJ3 (6 mm thick hollow steel column), the beam ange in compression buckled suddenly after 8 displacement cycles.σ y is the yield stress; σ u is the limit stress; ε y is the yield strain; ε u is the limit strain; E s is the elasticity modulus.
Advances in Materials Science and Engineering Figure 9 shows that the load in the specimen CJ2 is asymmetric. is indicates that the sti ness in positive direction is larger than that in negative direction.
Slight pinching of load-de ection hysteretic curves was observed in all three specimens.Such a phenomenon may be contributed to the following reasons: on one hand, because of the gap of the bolt connection between the shear connection plate and the bolt, slippage is inevitable under the cyclic loading; on the other hand, shear deformation is not ignored due to less e ective antishear section dimension in the bearto-column connection; of course, the deformation of the hollow column frame plate also contributes to the slippage.

Moment-Rotation Relationship Curve.
Because of the large cross section of the PSB column, the sti ness of the column is quite large.As a result, it is assumed that there is no rotation in the column.Furthermore, as the PSB columns were xed to the oor by four 30 mm diameter bolts, it was assumed that the PSB column did not slide under the cyclic loading.Based on the above assumptions, the rotation of  6 Advances in Materials Science and Engineering connection is the sum of the rotation due to the de ection of the hollow steel column, de ection of the beam ange, and the gap between the bolt and shear connection plate.In this test, the rotation of connection can be calculated as follows: where Δ 4 is the displacement of laser sensor #4 under cyclic loading and 300 is the distance from sensor #4 to the surface of the column.When calculating the rotation of the joint, the displacements from sensors #3, #4, and #5 all can be used.However, the displacement of sensor #5 includes the elastic deformation of the PSB beam, so the result from sensor #5 cannot be used to calculate the joint rotation.As for sensor #3, which was placed at 100 mm away from the beam ange where beam web ends, the velocity of reciprocating deformation of the beam ange which is lagging behind the loading velocity in the beam tip a ects the result of sensor #3; therefore, the rotation of the joint based on sensor #3 is not used.e maximum rotation for each specimen is shown in Table 4.It is noted that the rotation for the specimen CJ2 with a 4 mm thick hollow steel column is the biggest among the three specimens.e specimen CJ3 with a 6 mm thick hollow steel column has the least rotation.
Because the deformation of the column is not taken into account, the moment of the joint is obtained by multiplying the load and the length from the loading point to the surface of the column.e equation for joint moment is given as follows: where P is the load applied to the beam and L load is the distance from the load point to the upper surface of the column.show the moment-rotation hysteretic behavior of the three specimens.It shows that, for all three specimens, the hysteretic behavior is stable until the buckling of the beam ange.
e initial sti ness K ini of each specimen, which is the ratio of the maximum moment to the corresponding rotation in the rst cycle, is shown in Table 4.
e initial sti ness increases with the increasing thickness of the hollow steel column.
e relative parameters for the momentrotation curve of the three specimens are showed in Table 4.

Envelope Curve.
Envelope curve is de ned as the line which connects the maximum rotation and corresponding moment in the rst cycle of the displacement step.Figure 14 shows moment-rotation envelope curves of the three specimens.It is noted that the envelope curve of each specimen is similar.
Yielding phenomenon was not observed in the whole envelope curve, but the load increase appears to slow down with the increase of cyclic displacement.Such a phenomenon is mainly due to substantial losses in sti ness of the beam ange and hollow steel column under cyclic loading.It is evident that the thickness of the hollow steel column has a signi cant in uence on the envelope curve.
From the envelope curves in the third quadrant shown in Figure 14, the specimen CJ3 has the largest load and the specimen CJ1 has the smallest load at the same displacement.
is corresponds to the thickness of the hollow steel column of Advances in Materials Science and Engineering each specimen.is indicates that the thickness of the hollow steel column has an in uence on the strength of connection.
It is also observed that the envelope curve under compressive load is di erent from the envelope curve under tensile load. is is due to the sti ness degradation of the steel plate.

Energy Dissipation.
e energy dissipation of the connection is the enclosed area under cyclic loading, as shown in Figure 15, and the displacement in Figure 15 is the maximum beam tip displacement.It is an important index for evaluating seismic performance of the joint.
According to literature [25], the energy dissipation capacity coe cient E can be expressed in the following equation: Coe cient E for the three specimens CJ1, CJ2, and CJ3 is 1.70, 1.77, and 1.67, respectively.It can be seen that the specimen CJ2 has the biggest coe cient E, and the specimen    16.Point A is the summit point of the rst step cycle, and from point 0 to point A, the loading sti ness is expressed with secant sti ness K ini .After passing point A, loading sti ness K 2 becomes smaller. is is due to the shear deformation of the beam and bending deformation of the hollow steel column.K 3 is the unloaded sti ness from point B, which is obviously larger than the loading sti ness K 2 .At the start of unloading, only beam anges act in tension and compression, and shear deformation of the beam and bending deformation of the column plate were not happening.When the unloading reaches zero, the loading in the opposite direction starts.
When the load starts to load in the opposite direction, the shear deformation of the beam and de ection of the hollow steel column recover rst along with the close of the gap between the shear connection plate and the bolt, followed by the load increase of the joint.erefore, the hysteretic curve of connection has slight pinching, and the sti ness of the slope is called as slippage sti ness K 4 .Based on the tested results, the average sti ness values K ini , K 2 , K 3 , and K 4 of each specimen are given in Table 5, in which K 2 , K 3 , and K 4 are the slopes of regression lines.

Conclusions
In this study, the mechanical performance of a new type connector in timber frame building was investigated.
ree full-scale specimens with di erent hollow steel column thicknesses were subjected to cyclic loading up to failure, and the primary conclusions can be summarized as follows.
For the three specimens, failures were due to the compression buckling of the beam ange, and yielding in the beam ange was not observed.In addition, deformation perpendicular to the hollow steel column was observed in the specimen CJ2; however, no obvious deformation was noted in the specimens CJ1 and CJ3.Buckling of the beam ange governs the moment resistance and rotation capacity of connection.
It is also noticed that the hysteresis loops for the three specimens appear full, with slight pinching in every specimen.
e energy dissipation coe cient E for the three specimens CJ1, CJ2, and CJ3 is 1.70, 1.77, and 1.67, respectively, which indicates that the three specimens have good energy dissipation.
e hollow steel column thickness has signi cant in uence on the energy dissipation of the connection.Although the thinner hollow steel column has better energy dissipation, it has less connection sti ness.
e maximum rotations of the specimens CJ1, CJ2, and CJ3 under cyclic loading are 0.0375, 0.049, and 0.044, respectively.e thinner the hollow steel column, the larger the rotation of the connection.
Under the cyclic loading, the connection undergoes substantial losses in sti ness.
According to the envelope curve and moment-rotation curve, a hysteresis model for this type of connection under cyclic loading is presented.Employing a thick hollow steel column will have bigger rst secant sti ness K ini under cyclic loading.Advances in Materials Science and Engineering

Figure 1 :
Figure 1: Constructional detail of the joint.(a) Steel connector, (b) beams and columns, and (c) fabricated joint.

Table 1 :
Steel plate thickness of the connector (mm).

Table 2 :
Mechanical property of the steel plate.

Table 3 :
Test results of the three specimens.

Table 5 :
Model parameters of the specimens.