Long-term behavior of connections for glubam-concrete composite beams

Glubam-concrete composite (BCC) structure is a construction technique where glue-laminated bamboo beam is connected to an upper concrete slab using different types of connectors. The long-term behavior and creep mechanism of BCC system are very complex due to different time-dependent behavior of three components. This paper performed a series long-term push-out tests on four types of connections under static load. The tests lasted for a period of seven months in uncontrolled sheltered indoor condition. Important results such as the relative slip-time curves are presented in order to characterize time-dependent behavior of connectors. Furthermore creep coefficient constitutive models are provided for predicting the long-term behavior of connections.


Introduction
Glubam, a new type of glue-laminated bamboo, was developed by Xiao et al. [1]. Through designing with special laminate structure of bamboo fibers, glubam exhibits relatively higher mechanical properties compared with typical engineered lumber product, glulam ( Fig.1) [2,3]. However, for glubam-only flexural members, the main drawback of mechanical properties is of relatively low stiffness which may lead to increased deflection compared with reinforced concrete structure or steel structure. Improving the bending performance of glubam flexural members is a key issue for the modern bamboo structures. The glubam-concrete composite (BCC) structure is a construction technique where an upper concrete slab is connected to a lower glubam beam by shear connectors (Fig.2), allowing the best properties of both materials to be exploited, combining the high compression resistance of concrete with the greater capacity of glubam to withstand bending and tension.
The mechanical performance of two materials can be utilized. Several experiments on BCC connections and beams have been performed to date [1][2][3][4][5][6][7] Based on the short-term performance study on connectors [4], the following four connectors were selected for long-term push-out test, as follows: (1) A continuous steel mesh shear connector (SM) with 2mm thick ×400 mm length, was made with Q235 steel plate and was actually cut from a punching-hole mesh. The steel mesh was 100 mm wide of which one half is glued into a slot in glubam while the other half is embedded into concrete, as shown in Fig. 2a .
(2) The notched connector (NC) with 100 mm length and 50 mm depth , was used by combining with a screw in the glubam beam. The screw is same as SC connector that was glued in glubam beam but reached 50 mm into concrete, as shown in Fig. 2b . connection types. The nut was tightened by a torque wrench 28 days after curing concrete , as shown in Fig.   2d.
All push-out specimens were composed of two 100 mm×400 mm×400 mm concrete slab and a 112 mm×140 mm×400 mm glubam beam, as shown in Fig.2. The mechanical properties of the above four shear connectors were obtained by short-term push-out test, as shown in Table 2.

Test setup
According to some long-term push-out tests on TCC system, every shear specimens was subjected to a sustained load of approximately 0.3F u , which represents the quasi-permanent part of the serviceability design load [11], where F u is the average ultimate load determined with the short-term shear tests on symmetrical specimen.
It should be noted that the total load applied on the specimens is equal to 0.6F u considering two connectors in each specimen.
In order to conduct the long-term test in limited room, authors has developed a new long-term test setup, as shown in Fig. 3. and Fig. 4. The experimental device utilizes leverage principle (Fig.5): F 1 ×L 1 =F 2 ×L 2 , L 1 =0.31m, L 2 =1m, F 2 =0.31F 1. The load F 2 obtains by compressing the spring with a jack. After that, the fixing nut is tightened to keep the load constant. The details of long-term sustained load are shown in Table 3.  In the interface between glubam beam and concrete slab, 4 mechanical displacement gauges were installed to measure the relative slip. At almost same time every day, the relatively slip, as well as temperature and relative humidity of room, were record.

Experimental results
Through the analysis of experiment data for 210 days, the average relative slip-time curves of the four connectors shown in Fig.6 were obtained. It can be seen from the relative slip-time curve: (1) All specimens exhibit initial slip  el at the moment of loading, and the quantity of slip is relate to the slip stiffness.
(2) The relative slip increases rapidly in the beginning of 100 days, and then its increment decreases gradually and tended to be stable.
(3) All curves are not smooth caused by varying temperature and relative humidity. Generally, this fluctuation is larger in the earlier than in the later period.
The creep coefficient is used to describe the time-dependent behavior of shear connectors in BCC system, and it is defined as following, where, (t) denotes the relative slip of the connection at time t ,  el denotes the initial relative slip at beginning of load application.
Considering the actual situation, most of the long-term tests will not last for service life (50 years). Therefore, some relatively accurate long term mechanical models or creep constitutive models need be established for predicting the relatively slip over a long period on account of stable experimental data [10][11][12].
For SC, NC and PNC connectors, Kelvin rheological model are selected to analyze the long-term slip, and it is expressed as following, where, t 0 and t, signify the loading time and the final time, respectively. The J i and τ i is a constant coefficient related to material properties.
On the other hand, the slip curve of SM specimen is different from other connectors that relatively slip where, A and m, are regression parameters coefficients.
The fitting formulas and calculating curves are shown in Table 4 and Fig. 6. Table 5 summarizes the slip modulus, initial elastic slip, the average slip, and fitting error for the tested connections. There are the large difference between the analytical fitted curve and experimental curve of specimen SM before 100 days, after that a good correspondence can be discovered.
Since creep constitutive models are used to predict the long-term slip of connectors at end of service life (50 years), all tested results show that the fitting errors after 100 days is less than 3%, indicating that these creep models can be used to evaluate the long-term performance of connectors. Therefore, the final slip of each connector is predicted at end of service life (50 years) and the results are also presented in Table 5. It can be observed that the predicting value under the long-term load of 0.3F max is significantly lower than its ultimate slip obtained from the short-term push-out test.

Conclusions
This paper investigated the long-term slip on four types of connectors for glubam-concrete composite beam.
Based on the results, the following conclusions were drawn: (1) The relatively slip of the four types of connector increases rapidly in early stage are large, and then the increment decreases gradually and becomes stable at end of test.
(2) Kelvin rheological model and power function model are used to analyze the long-term slip of these four connectors. Compared with the test data, the proposed models are found to accurately predict the slip development within the test period.
(3) The predicting values at end of service life are significantly lower than the corresponding ultimate slips obtained from the push-out test under the long-term load of 0.3F max .