Mechanical behavior of a novel precast beam-to-column connection with U-shaped bars and engineered cementitious composites

Abstract

This article investigates a novel precast connection, with U-shaped bars extending from precast column to connect with the longitudinal bars in precast beams. To improve the seismic behavior of the connection, engineered cementitious composites, one kind of highly ductile concrete, were introduced into the core area of the connection, which also act as the cast-in-place material in the beam top and end. Prior to the test, finite element modeling was conducted to determine the proper splice length between U-shaped bars and beam reinforcements and also to evaluate the bonding performance of the proposed connection. The experimental program was then carried out on a monolithic connection, a precast connection with normal concrete as well as a precast connection with engineered cementitious composite, after which the seismic behaviors of the connections including their failure mode, hysteresis characteristic, stiffness degradation, ductility, and energy dissipation were analyzed. All three types of connections underwent typical flexural failure where the joint area remained intact. The negative carrying capacity, ductility, and energy dissipation were slightly lower for the connection with concrete, while the connection with engineered cementitious composite exhibited satisfactory behavior comparable to monolithic specimens. The latter connection with engineered cementitious composite is therefore suggested to be applied in highly seismic region.

Keywords

cyclic loading engineered cementitious composite finite element precast connection U-shaped bars

Introduction

Precast concrete structures have shown various advantages over conventional cast-in-place concrete structures, such as better quality control and improved efficiency in time and cost. However, precast structure failure by earthquake is mostly caused by connection failure between precast elements. Design of connection details is therefore crucial for the safety of the whole structure.

For precast connections, the beam and column are normally connected by cast-in-place concrete, and the cast-in-place location is either in the column or in the beam end. There have been several studies on connections with the cast-in-place location in the column. Xue and Yang (2010) conducted a test of four full-scale precast concrete connections which involved cast-in-place concrete columns. In the test, a large casting area is needed for the columns, which significantly reduces the construction efficiency. In view of this, Ozturan et al. (2006) came up with a new connection with cast-in-place column, where the columns were precast with a gap in the middle. Within the gap, the upper and lower columns were only connected by vertical reinforcements and U-shaped bars were installed as flexural reinforcement in the precast beam for anchorage. Although this type of connections can certainly save casting area and construction time, a gap in the column is not desirable since it forms the weakest point in the column before casting. To provide adequate strength and stability for the column during the installation process, Parastesh et al. (2014) developed a moment-resisting connection, where the columns were precast continuously with diagonal bracing bars and shear links in the gap. It should be pointed out that while this connection provided good structural integrity and exhibited decent ductility and energy dissipation, the overly complicated reinforcements in the joint zone will make the concrete casting extremely difficult.

Compared to the connections with the cast-in-place column, the connections with the cast-in-place beam end can save construction time, maintain column integrity, and avoid complicated diagonal bracing bars. Moreover, such practice perfectly follows the principle of “strong column–weak beam” in seismic design theories. Nevertheless, as there is no reserved gap in the column, the longitudinal bars in the beam must be cut apart at the joint. The stress transfer at the joint is then a big concern that may cause severe failure at the beam end during earthquakes. In such a case, the need to ensure effective stress transfer at the connections while maintaining structural integrity is urgent.

In this study, improvements are made in two ways: In the first way, to ensure force transmission, additional U-shaped bars at the connection are employed as connecting elements to lap with longitudinal bars in the beam. Compared with straight spliced bars, the application of U-shaped bars can effectively reduce the straight splice length and minimize the fieldwork area in the beam end. In addition, reinforcing bars are used in the connection region to improve the carrying capacity of the beam end. In the second way, substitution of concrete with engineered cementitious composites (ECCs), one kind of high-performance fiber-reinforced composites (FRCs), is adopted in the joint and beam end to improve the seismic performance of the connection.

The use of FRCs in precast connections has been shown to be effective to improve their seismic behavior. Vasconez et al. (1998) showed that the FRC-based connection design can remarkably increase the plastic hinge length in the beam end, and their strength, energy dissipation capacity, and stiffness have been improved significantly. ECC is well known for being superductile, accompanied with the multiple-cracking and strain-hardening behaviors. Its ultimate tensile strain can reach over 6%, while the crack width is controlled to below 80 µm. During compression, the strength of ECC is similar to normal concrete but the corresponding strain is almost twice (Billington and Kesner, 2003). Moreover, ECC shows good deformation compatibility with steel reinforcements which can help avoid bond splitting cracks (Fischer and Li, 2002). Owing to these properties of ECC, members cast with ECC show various advantages over the ones cast with normal concrete, such as improved load-carrying capacity, residual stiffness, and energy absorption (Fischer, 2003; Hou et al., 2014; Parramontesinos and Canbolat, 2005; Yuan et al., 2013). Choi et al. (2013) studied the connections for precast specimens with ECC where the specimen behaved in a ductile manner and showed higher strength than monolithic concrete connections. Therefore, introducing ECC into the connection is expected to help achieve improved strength, ductility, and energy dissipation. Meanwhile, ECC in the lap splice region can ensure the stress transfer efficiency between spliced bars and eliminate bond deterioration due to its deformation compatibility with reinforcements.

In this study, a novel precast connection with U-shaped bars and reinforcing bars in the joint region will be investigated. ECC materials are introduced to prevent bond deterioration and improve seismic resistance of the connection. It should be mentioned that many investigations have observed significant bond deterioration in the connection regions due to insufficient splice length of steel reinforcements (Shariatmadar and Beydokhti, 2014). To avoid incident bond deterioration between beam longitudinal bars and U-shaped bars and also to reveal the stress transfer mechanism of the proposed connection, finite element analysis is to be conducted with the software ATENA to acquire a proper splice length between longitudinal bars and U-shaped bars. After this, the structural behaviors of two precast connections with different cast-in-place materials (concrete and ECC) were tested and compared with conventional monolithic connections. The seismic behaviors of the connections including failure mode, hysteresis characteristic, stiffness degradation, ductility, and energy dissipation were analyzed. The findings of this study are expected to provide insights into the design of novel precast connections incorporating novel materials.

Details of the proposed beam-to-column connection

In the proposed connection, the precast beams are supported on the column with cast-in-place concrete/ECC in the beam end and top (Figure 1). The beams are prefabricated with bared longitudinal reinforcements and bared stirrups. Longitudinal reinforcements in the vicinity of the joint are shaped into a 90-degree hooked anchorage. The precast column is continuous at the joint zone with reserved holes for the U-shaped bars, reinforcing bars, and continuous longitudinal bars at the top of the beam. U-shaped bars are designed as connecting elements and reinforcing bars are provided as flexural reinforcement at the connection region. High-performance grouts are used between the steel bars and reserved holes to ensure good bonding between them. Moreover, higher strength concrete or ECC is used in the precast joint region and at the beam top and end.

Figure 1.

Schematic diagram of the precast connection.

Finite element simulation and discussion

To estimate the proper splice length of U-shaped bars and reinforcement, and also to verify the feasibility of the proposed connection, finite element modeling with ATENA was performed prior to the test. Figure 2 shows the finite element model and sectional reinforcements of the connection, where the clear span of the beam is 1700 mm and the height of the column is 2450 mm. Brick solid elements and truss elements are adopted to model the concrete and steel bars, respectively, and spring elements are adopted to simulate the bond behavior between steel bars and concrete (the interfacial bond–slip relationship follows Comite Euro-International Du Beton (1993)). A bilinear model with strain-hardening behavior is used to model the steel bars, where the yielding strength, ultimate strength, and elastic modulus of the steel bars are 412.8, 578, and 208 GPa, respectively. The constitutive relationship of concrete is adopted from Comite Euro-International Du Beton (1993), where the compressive strength of concrete is 34 MPa and the compressive strain at peak stress and crushing strain are 0.002 and 0.0038, respectively.

Figure 2.

Illustration of the finite element model: (a) finite element model and sketch of the connection; (b) schematic illustration of steel reinforcements.

Concerning the boundary conditions, the bottom displacement of the column is restrained in all three directions and the top displacement is only restrained in the X and Y directions. An axial compressive force is applied at the top of the column, and asymmetric loading is applied at the end of the beams by displacement control. The model is meshed with the size of 100 mm and the “Newton–Raphson” iterative procedure is adopted as the solution method. Both displacement and residual convergence criteria are adopted in the computation with the tolerance of 0.01.

To examine the validity of the proposed finite element model, simulation on a concrete connection test from the reference (Guan et al., 2016) was performed and compared with the test results. The comparison has proven that the boundary conditions, meshing method, and calculation method adopted in this study have been selected properly. The details of the comparison can be found in Appendix 1.

Analysis of splice length of steel reinforcements

To determine the proper splice length between U-shaped bars and beam longitudinal reinforcements, parametric studies were conducted with the proposed model by varying the splice length from 200 to 500 mm. Figure 3 shows the load–displacement curves for the left end of the beam. It is found that the load-carrying capacity and ductility increase evidently with increasing splice length. However, further increase of splice length leads to a very small increase in carrying capacity and ductility after the splice length reaches 400 mm. It can thus be deduced that 400 mm is the optimal splice length for stress transfer.

Figure 3.

Load–displacement curves at different splice lengths.

In order to clarify the force transfer mechanism between U-shaped bars and longitudinal bars, some monitoring points are set along reinforcements in specimen FE-400. Figure 4 shows the stress–displacement curves of steel reinforcements at different monitoring points. It can be observed that the stress of CL (monitoring points in the left and right side of the connecting (U-shaped) bars) first increases linearly with increasing deflection and then remains nearly constant at a high value close to that of UR. In contrast, the stress for BL is extremely low, indicating that external load has been transferred successfully from the longitudinal bars to the U-shaped bars. Figure 5 shows the stress distributions along the U-shaped bars and longitudinal bars at the different load levels of 40.49, 89.36, and 133 kN. No significant difference can be seen in the U-shaped bars at 40.49 kN, while the stress is very high in the range from −300 to −100 mm for the loads of 89.36 and 133 kN. Then, the stress decreases substantially outside the joint region and reduces to about 60 MPa at the position of −600 mm. However, for the bottom reinforcing bars of the left beam, the stress reduces considerably at the beam end from −300 to −200 mm and then increases linearly with larger distance from the joint. The mechanism behind the above observations is stated as follows: Two bars placed side by side forming a lapped splice can be regarded as two anchored bars and the transfer of force from one bar to another is achieved with the development of bond stress in the surrounding concrete along the splice length. For the U-shaped bar, the length of the portion between −600 and −200 mm can be regarded as the anchorage length, over which the stress of the U-shaped bars decreases substantially accompanied with the bond–slip behavior at the interface of bars/concrete. For the longitudinal bars, the position of BL can be regarded as the anchorage end of the bottom bars. The stress is therefore extremely low at BL but the bar soon regains increasing stress proportionally to the distance from the joint. This proves again that the stress can be transferred effectively from the longitudinal bars to the U-shaped bars over the splice length of 400 mm.

Figure 4.

Stress–displacement curves of steel reinforcements at different monitoring points. Cl and CR: monitoring points in the left and right side of the connecting (U-shaped) bars. UL, UR, BL, and BR represents the monitoring points in the up and bottom longitudinal bars at the left and right side, respectively.

Figure 5.

Stress distributions along the steel reinforcements at different load levels.

It should be highlighted that the splice length, 400 mm, is only determined for the specimen specified in this study. To determine splice length that applies to connections with various parameters, a comprehensive parametric study is included in Appendix 2.

Analysis of bond strength between U-shaped bars and concrete

To enhance the bond strength between steel reinforcements and concrete, high-performance grouts are used between the U-shaped bars and reserved holes. Six kinds of bond–slip relationship are defined in the finite element models (Figure 6), where FE-G and FE-P represent the good and poor bond–slip relationship, respectively; FE-0.5P, FE-0.3P, and FE-0.1P represent 0.5, 0.3, and 0.1 times bond strength of FE-P, respectively; FE-0 represents the negligible bond strength between steel reinforcements and concrete.

Figure 6.

Various bond–slip models between steel reinforcements and concrete.

Figure 7 shows the load–displacement curves on the left beam end with different bond–slip models. Based on the simulation results of FE-0, FE-0.1P, FE-0.3P, and FE-0.5P, both the carrying capacity and initial stiffness increase significantly with increasing bond strength. However, for specimens above FE-0.5P, the increase of bond strength results in a negligible change in carrying capacity and initial stiffness, indicating that FE-0.5P is optimal for force transfer between steel reinforcements and concrete. Note that the bond strength of the grout in real construction is approximately four times larger than FE-0.5P, so it should be sufficient for the proposed connection and the assembling methods are feasible.

Figure 7.

Load–displacement curves with different bond–slip models.

Experimental program

Specimen details and loading configuration

To evaluate the seismic behavior of the proposed connection, specimens with concrete (JC) and ECC (JE) in the cast-in-place zone were prepared and a monolithic connection JMO was also tested for comparison purposes. Based on the prior finite element analysis, a splice length of 400 mm was employed. Dimensional details of the subassembly are shown in Figure 8.

Figure 8.

Configurations and reinforcement details of the test connections (unit: mm): (a) specimen JMO and (b) specimens JC and JE.

Figure 9 shows the schematic diagram of the test setup. The precast concrete column was pinned at the base and top to simulate the real situation where the moment was zero at the middle of the column. Two hydraulic jacks were used at the free ends of the beam to apply reverse cyclic loading, and a hydraulic jack was used to simulate the axial load in the column. Two load cells were installed to monitor the loads applied at the beam ends, and two linear variable differential transformers (LVDTs) were placed at the loading points to acquire the vertical displacement of the cantilever beams.

Figure 9.

Schematic experimental test setup.

An axial compressive ratio of 0.3 was applied on the column to simulate the external load transferred from the upper floors. Then, asymmetric loadings (P₁ and P₂) were applied on the beam ends by displacement control. The maximum displacement in each cycle was increased with a predetermined increment Δ_y, where two loading cycles were conducted for each displacement increment. After the eighth cycle, the maximum displacement increment was adjusted to 2Δ_y to accelerate the loading process. The loading history is shown in Figure 10. The test was stopped after the applied vertical load had reduced to 85% of the ultimate load due to limitations of machines.

Figure 10.

Cyclic loading history for each specimen.

Material properties

For the ECC materials, fly ash and cement were used as the binders and polyvinyl alcohol (PVA) fiber was used as the reinforcement. Figure 11 shows the tensile stress–strain curves for the ECC material by direct tensile tests where the ultimate tensile strain exceeded 4% and the tensile strength exceeded 5 MPa. A number of cubic specimens (150 mm on each side), including concrete and ECC specimens, were cast and tested under compression. The results are shown in Table 1. It should be mentioned that the strength of ECC was lower than that of concrete due to lack of coarse aggregates of ECC. The mechanical properties of steel reinforcements have been tested and are shown in Table 2.

Figure 11.

Tensile stress–strain relationship of ECC.

Table 1.

Material properties of concrete and ECC.

Specimen	Strength of concrete (MPa)	Strength of ECC (MPa)
Precast beam	37.1	–
Precast column	37.1	40.5
Cast-in-place zone	41.2	35.5

ECC: engineered cementitious composite.

Table 2.

Material properties of the steel reinforcements.

Diameter (mm)	Area (mm²)	Yield strength (MPa)	Ultimate strength (MPa)	Modulus of elasticity (GPa)
8	50.3	501.2	639.4	208
10	78.5	495.0	643.2	208
16	201.1	462.5	622.9	208
18	254.3	459.9	612.1	208
25	490.6	431.1	591.5	208

Experimental results and discussion

Cracking pattern and failure mode

Figure 12 shows the crack propagation patterns and failure modes of the tested connections. The comparison will be described as follows.

Specimen JMO. After the first cracking, several flexural cracks distributed uniformly along the beam and some diagonal cracks appeared at the joint. When the displacement reached 60 mm, longitudinal splitting cracks were first observed at the beam bottom. Existing beam cracks then extended throughout the depth of the beam and concrete spalling occurred at the plastic hinge of the beam. The specimen failed in flexural mode.

Specimen JC. At the displacement of 10 mm, first flexural cracks were formed at the beam. More cracks then concentrated in the beam at 200 mm away from the beam-to-column interface. This indicates that the reinforcing bars in the beam end can effectively prevent the development of cracks near the interface. Afterwards, diagonal shear cracks were observed in the connection region and no bond deterioration was observed near the U-shaped bars before the displacement of 60 mm, indicating that the splice length of 400 mm was sufficient for stress transfer. At the final failure, an interfacial crack appeared at the beam-to-column interface and longitudinal splitting cracks were then observed along the longitudinal bars. Concrete crushing finally took place at the beam plastic hinge, leading to the failure of the specimen.

Specimen JE. When the displacement reached 20 mm, multiple diagonal cracks appeared in the joint, and flexural cracks distributed uniformly along the entire length of the beam. Different from concrete elements, the widths of the existing cracks remained constant, while many fine hairline cracks occurred with increasing external loading. This can be attributed to the excellent crack width control ability of ECC. When the displacement reached 40 mm, an interfacial crack occurred at the beam-to-column interface and then flexural cracks appeared in the beam with many micro-cracks. The test was stopped when the vertical load had reduced to 85% of the ultimate load due to limitations of the machine. During the whole test, the specimen maintained good integrity and no debonding was observed between the steel bars and ECC.

Figure 12.

Crack patterns and failure modes: (a) specimen JMO, (b) specimen JC, and (c) specimen JE.

Hysteretic behavior

Figure 13 shows the hysteretic and envelope curves for the three connections. The beam displacement at 1000 mm away from the connection region is used as the displacement in the hysteretic curve. As expected, the specimen JMO showed full hysteretic curves with relatively little pinching due to the integrity and strong bond between the reinforcements and concrete. For the specimen JC, the hysteretic curves were basically at the elastic state in the first two cycles and the residual deformation was very small, while the pinching effect of the hysteresis loops was apparent after the displacement of 40 mm. This is because more longitudinal bars were used in the beam ends for the precast specimens, which inevitably caused severer longitudinal splitting cracks. These longitudinal splitting cracks, along with the interfacial cracks formed at the beam-to-column interface for precast connections, cause the pinching phenomenon. Compared with JC, the hysteretic curves of JE were slightly fuller, and the load-bearing capacity in the negative direction was 19% higher than that of specimen JC given the fact that the compressive strength of ECC was lower than that of in situ concrete. This can be attributed to the compatible deformation between the longitudinal reinforcements and the ECC material. Since ECC is highly ductile with the ultimate strain capacity even higher than that of steel, splitting of their interfacial bonding or spalling of ECC can be avoided under external loading, which significantly improves the load-bearing capacity of ECC specimen under the negative loading. Along with the multiple-cracking feature of the ECC specimen, its hysteretic curve is therefore fuller than that of the concrete specimen which normally fails by one localized large crack. The excellent ductility and damage tolerance of ECC effectively enhance the load-carrying capacity and energy dissipation ability of the connection.

Figure 13.

Hysteretic and envelope curves for the three connections: (a) specimen JMO, (b) specimen JC, and (c) specimen JE.

Stiffness degradation

The secant column stiffness (K) is defined as the slope of the line between the maximum load and the corresponding displacement for the positive and negative directions during a loading cycle. K is calculated as

K = \frac{| + F_{i} | + | - F_{i} |}{| + Δ_{i} | + | - Δ_{i} |}

(1)

where +F_i and −F_i are the positive and negative peak loads of the first hysteresis loop at each load level, respectively, and +Δ_i and −Δ_i are the corresponding displacements.

The secant stiffness for the three specimens is computerized and plotted in Figure 14. As can be seen, the stiffness decreased continuously as the displacement levels increased, due to the increasing cumulative damage in the tested specimens, and each specimen experienced severe stiffness degradation at the end of the test. It can be observed that there was no significant difference in stiffness degradation between the precast connections and the monolithic connection. The initial stiffness for the specimen JC was slightly higher, which was mainly attributed to the reinforcing bars in the joint zone. However, no evident difference can be seen in its latter stiffness. Although ECC had a lower elastic modulus than concrete, the stiffness degradation trend of the specimen JE was very similar to that of concrete connections, especially after the displacement of 30 mm. It can be reasoned that the multiple fine cracks formed in ECC components lead to a high effective moment of inertia, which compensates for the lower elastic modulus.

Figure 14.

Stiffness degradation of the three specimens.

Energy dissipation

The energy dissipated during a cycle is calculated as the area enclosed by the hysteresis load–displacement loops in that cycle, and the cumulative energy dissipated is calculated as the sum of energy dissipated in each loop until the specified cycle. Figure 15 shows the comparison of cumulative energy dissipation of the tested specimens. Generally, all the specimens experienced a similar pattern of energy dissipation that energy dissipation increased with the increasing load level. For the specimen JC, a 22.4% decrease in cumulative energy dissipation was observed in the final displacement compared with the specimen JMO. This is attributed to the weaker interface and severer longitudinal splitting cracks caused by more longitudinal bars placed in the plastic hinge. What is more, configuration of more longitudinal bars in the plastic hinge for the specimen JC hindered the steel bars from deforming completely and thus reduced the energy absorption capacity. However, smaller difference can be found between the specimens JE and JMO. The crack width for the connection of the specimen JE was much finer than that of concrete specimens due to the fact that the fibers bridged and restrained the cracks of ECC materials. These fine hairline cracks effectively dissipated energy. Besides, ECC with high compressive strain can ensure that the bars are fully strained, by which the reinforcement can contribute more to energy dissipation. Overall, substitution of concrete with ECC in the connection region can significantly enhance the energy dissipation ability of the connections.

Figure 15.

Cumulative energy dissipation curves of the three specimens.

Displacement ductility

The concept of ductility is a key element to evaluate the seismic performance of a structure, which is calculated by the envelope curve in both the positive and negative directions following the calculation method of concrete. The ductility coefficient is defined as the ratio of the ultimate displacement to yield displacement. The ultimate displacement is determined as the displacement when the vertical load falls to 85% of the ultimate strength and the yield displacement is determined according to the criteria for equivalent elasto-plastic energy absorption (Park, 1989). As shown in Figure 16, when the area of S₁ is equal to the area of S₂, the position of point B is determined. The value of horizontal coordinate of point B can be defined as the yield displacement.

Figure 16.

Method used to define yield displacement.

As summarized in Table 3, the ductility coefficient of the specimen JMO was 3.05–3.55 and that of the specimen JC was 3.25–3.07, which illustrated that the proposed connection behaved in a ductile manner. The ductility coefficient of the precast specimen JC in the negative direction was slightly lower than that of the monolithic connection. This resulted from the reinforcing bars placed in the joint zone which leads to a higher yield displacement. Nevertheless, for the positive direction of the specimen JC, where the same reinforcements as the monolithic specimen are arranged, a similar yielding displacement as the monolithic one was also achieved.

Table 3.

Comparison of displacement ductility factors.

Specimen	Loading direction	Yielding load		Peak load		Final load		Ductility coefficient	Failure mode
Specimen	Loading direction	P_y (kN)	Δ_y (mm)	P_max (kN)	Δ_max (mm)	P_0.85 (kN)	Δ_0.85 (mm)	Ductility coefficient	Failure mode
JMO (WEST)	Positive	132.76	25.89	154.58	60.13	131.39	78.97	3.05	FF
JMO (WEST)	Negative	105.57	21.47	112.33	29.86	95.48	76.13	3.55	FF
JC (WEST)	Positive	135.42	25.79	154.26	40.13	131.12	83.69	3.25	FF
JC (WEST)	Negative	91.45	24.27	100.28	29.98	85.24	74.4	3.07	FF
JE (WEST)	Positive	126.77	26.40	144.72	61.90	123.01	89.90	–	No failure
JE (WEST)	Negative	105.27	26.13	119.51	41.05	101.56	75.13	–	No failure

FF: flexural failure.

As tabulated in Table 3, the specimen JE tended to have a slightly larger yield displacement than the specimen JC owing to small elastic modulus or better damage resistance of ECC. For the ultimate displacement, Fischer and Li (2002) considered a 25%–50% strength drop as the failure state for reinforced ECC components. However, due to limited permissible displacement of the hydraulic actuator, the test was stopped when the external load reduced to 85% of the ultimate load. As a result, the specimen JE had not reached final failure when the test was terminated and the ductility coefficient of ECC cannot be calculated from this test. Nevertheless, the final displacement for the specimen JE in the negative direction was similar to that of specimen JMO, while the displacement in the positive direction was much larger when the applied load fell to 85% of the ultimate strength. It can therefore be deduced that the specimen JE would exhibit a larger ductility coefficient at failure if the test was not terminated.

Conclusion

A novel precast connection was proposed and investigated in this article. Experimental program was carried out on monolithic connection, precast connection with normal concrete as well as precast connection with ECC, and the following conclusions can be drawn:

Based on the discussion of finite element analysis, a splice length of 400 mm was determined to be the optimal length to transfer the load between spliced bars. The use of high-performance grouts can provide the bond strength between U-shaped bars and concrete. Therefore, the assembling method proposed in this article is feasible.

All precast connections showed typical flexural failure but the specimen with ECC exhibited fuller hysteretic curves because of its superior ductility and damage tolerance.

Both the precast connections with concrete and ECC exhibited similar stiffness degradation as the monolithic connection. The cumulative energy dissipation of precast specimen with concrete is 22.4% lower than that of the monolithic specimen due to the weak interface between the precast and cast-in-place concrete. However, the application of ECC can offset this defect and improve energy dissipation significantly.

The performance of the precast connection with ECC is found to be much better than the one with concrete in most of the aspects and even comparable to the monolithic specimen, making it a promising application in highly seismic region.

Footnotes

Appendix 1

Appendix 2 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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of China under Grant Nos 51778131 and 51708109, and the Distinguished Young Scholar Foundation of Jiangsu Province under Grant No. BK20160027.

References

Billington

Kesner

(2003) Cyclic response of highly ductile fiber-reinforced cement-based composites. ACI Materials Journal 100(5): 381–390.

Choi

(2013) Development and testing of precast concrete beam-to-column connections. Engineering Structures 56(6): 1820–1835.

Comite Euro-International Du Beton (1993) CEB-FIP Model Code 1990: Design Code. Lausanne: Thomas Telford Ltd.

Fischer

(2003) Intrinsic response control of moment-resisting frames utilizing advanced composite materials and structural elements. ACI Structural Journal 100(2): 166–176.

Fischer

(2002) Influence of matrix ductility on tension-stiffening behavior of steel reinforced engineered cementitious composites (ECC). ACI Structural Journal 99(1): 104–111.

Guan

Guo

Xiao

et al . (2016) Experimental study of a new beam-to-column connection for precast concrete frames under reversal cyclic loading. Advances in Structural Engineering 19(3): 529–545.

Hou

Zhu

et al . (2014) Calcium silicate hydrate from dry to saturated state: structure, dynamics and mechanical properties. Acta Materialia 67: 81–94.

Ozturan

Ozden

Ertas

(2006) Ductile connections in precast concrete moment resisting frames. PCI Journal 51: 66–76.

Parastesh

Hajirasouliha

Ramezani

(2014) A new ductile moment-resisting connection for precast concrete frames in seismic regions: an experimental investigation. Engineering Structures 70(9): 144–157.

10.

Park

(1989) Evaluation of ductility of structures and structural assemblages from laboratory testing. Bulletin of the New Zealand National Society for Earthquake Engineering 22(3): 155–166.

11.

Parramontesinos

Canbolat

(2005) Experimental study on seismic behavior of high-performance fiber-reinforced cement composite coupling beams. ACI Structural Journal 102(1): 159–166.

12.

Shariatmadar

Beydokhti

(2014) An investigation of seismic response of precast concrete beam to column connections: experimental study. Asian Journal of Civil Engineering 15(6): 849–867.

13.

Vasconez

Naaman

Wight

(1998) Behavior of HPFRC connections for precast concrete frames under reversed cyclic loading. PCI Journal 43(6): 58–71.

14.

Wallace

(1996) Behaviour of Beam Lap Splices Under Seismic Loading. Christchurch, New Zealand: University of Canterbury.

15.

Xue

Yang

(2010) Seismic tests of precast concrete, moment resisting frames and connections. PCI Journal 55(3): 102–121.

16.

Yuan

Pan

et al . (2013) A comparison of engineered cementitious composites versus normal concrete in beam-column joints under reversed cyclic loading. Materials and Structures 46(1): 145–159.