Experimental Study on the Behaviour of Glulam Timber Beams Bonded with Glued-in BFRP Rods

Abstract

The technique known as glued-in rods (GiR) presents an appealing option for effectively connecting structural timber members. The utilization of fibre reinforced polymer (FRP) rods offers additional advantages, such as reduced weight, improved corrosion resistance and lower thermal conductivity. Studies examining the behaviour of GiR connections with Basalt FRP (BFRP) rods are limited, and there is currently a lack of guidance for their fabrication and design. This paper presents an experimental study on the performance of glulam timber beams bonded with glued-in BFRP rods. Epoxy adhesive was used to glue the BFRP rods into the cross-sections of glulam, forming the connections. The experimental programme includes four types of connections, investigating two configurations for the glued-in rods (Designs D1 and D2) and two rod diameters (8 mm and 10 mm). Testing was conducted with four replicates for each connection type, resulting in a total of 16 tests. To assess the moment capacity of connections, a four-point bending set-up with a 2300 mm span, was applied. The response in terms of load-displacement response was monitored and the failure mechanisms were assessed. The predominant observed failure mode was bar pull-out and tensile splitting of timber. D1 connections with 10 mm bars demonstrated superior performance in both moment capacity and maximum displacement at failure.

1 Introduction

Concerns about climate change, particularly the impact of carbon dioxide emissions, have led to increased global interest in the use of timber, owing to its sustainable properties. Timber has been employed in construction since ancient civilisation, due to its attractive aesthetics appearance, favourable strength-to-weight ratio, workability and toughness. Compared to natural timber, glued laminated (glulam) members have several advantages. The utilization of glulam can result in higher strength, as defects are distributed within the layers of the beam. However, forest depletion and the variability of natural timber makes the mass production of structural timber challenging. Addressing this challenge can be achieved, to some extent, by reusing existing timber sections. Joining sections of existing timber members allows for the production of new timber elements. Traditional timber connections, such as nailed or bolted connections, demand long machining times, substantial labour and are primarily suitable for smaller loads. Conversely, Glued-in rods (GiR) [1], are an effective method to connect timber sections, while offering several other advantages.

Previous investigations on glued-in rod connections have mainly focused on steel rods as connectors in the jointing system. The moisture content of natural timber makes the use of steel rods rather unfavourable as it is prone to corrosion penetration. The utilization of FRP rods offers additional advantages, such as reduced weight, improved corrosion resistance and lower thermal conductivity [2]. Research on the GiR technique with FRPs has recently been gaining momentum, but the main focus has been limited to Glass FRPs (GFRP), primarily due to its lower cost compared with carbon. Research on other FRP variants such as Basalt FRPs (BFRP), which offer relatively equivalent or even superior performance to GFRP considering the price, are very limited. Furthermore, despite the advantages of the GiR over conventional joint techniques, its incorporation into design is presently hindered by the lack of reliable design standards. To bridge these gaps, the current research focusses on examining the performance of glulam timber specimens joined with glued-in BFRP rods. The variables considered in the present research include the impact of the design types and rod diameter on the capacity of the timber samples. The test results provide a sound basis for formulating guidelines for the design of structural timber connections with GiR.

2 Experimental Study

2.1 Materials

This study used glued laminated timber graded as GL28, with characteristic bending strength 28 N/mm². The BFRP rods used were supplied by Magmatech [3] from their RockBar range. These rods have an elastic modulus of 54,000 N/mm² and a tensile strength of 920 N/mm² [4]. Two diameters of BFRP rods were utilized, specifically 8 mm and 10 mm. Two-part epoxy adhesives, recognized for their excellent gap-filling properties, were employed. Supplied by Rotafix [5], these adhesives have a nominal shear strength of 12.5 N/mm².

2.2 Glulam Bonded Specimens

Each glulam timber specimen had a total length of 2500 mm and was fabricated by joining two 1250 mm slices using glued-in rods. The timber specimens had nominal cross-section of 90 mm × 220 mm. The connection consisted of BFRP rods embedded through a length of 20 × rod diameter on either side (see Fig. 1a) and glued-in with epoxy-adhesive of 2 mm and 3 mm glueline thicknesses for the 10 mm and 8 mm rods, respectively. The rods were embedded in two different design configurations, D1 and D2. Both configurations included two bars on the tension and two on the compression zone, but in D1 the bars were placed in the same horizontal line, whilst in D2 all four bars were in a vertical line (see Fig. 1b). For each of the rod arrangements, two rod diameters were investigated. Four replicates were examined for each connection type, leading to a total of 16 bonded specimens. The designation applied hereafter comprises the rod diameter, followed by the design configuration and finally the replicate number (see Table 1).

Fig. 1.

Specimens with GiR connections (dimensions in mm).

Table 1. List of specimens.

2.3 Testing

In order to investigate the moment capacity of the timber joints, a series of four-point bending tests with a 2300 mm span, were carried out. In Fig. 2, a photograph of the set-up and a schematic illustration are presented. Load transfer was facilitated through the use of an I-shaped steel stiffened spreader beam. Steel rollers, in conjunction with steel bearing plates, were employed at the support points. Three linear variable displacement transducers (LVDTs) were placed at the two load points and at mid-span to capture vertical displacements. To monitor the strain distribution profiles, strain gauges were affixed to the external faces of each specimen at three locations (see Fig. 2b). A data acquisition system was employed to record all necessary data during the testing process. The monotonic load was applied via displacement control with a constant rate of 3 mm/min [6].

Fig. 2.

Four-point bending test set-up.

3 Results and Discussion

3.1 Load-Vertical Displacement Curves

The curves depicting the load versus mid-span vertical displacement for all tests are recorded and presented in Fig. 3. Each graph shows the response of a group with similar characteristics. In most instances, a small “jump” in the load-displacement response, resulting in a slight alteration to the initial stiffness, was observed shortly after loading. The observed jump, corresponding to the moment a gap opening appeared on the connection interface, is attributed to the breaking of the thin layer of residual adhesive between the two timber slices (see Fig. 1a). In general, similar performance was observed from the four replicates within each group.

Fig. 3.

Load-midspan vertical displacement curves for all specimens.

3.2 Failure Loads

Throughout the testing process, the applied load (N) was continuously monitored. The highest recorded value of the load is regarded as the failure load (N_u,Exp) of each specimen. To assess the moment resistance, the equation M_u,Exp = (N_u,Exp/2) × α where α is the distance from the loading point to the support (700 mm in this case, as shown in Fig. 2b), was applied. Table 2 presents the failure loads (N_u,Exp), the moment resistance (M_u,Exp) and the corresponding mid-span vertical displacements (δ_u,Exp).

Table 2. Test results at failure.

3.3 Failure Modes

Typical cases of failure modes are shown in Fig. 4. The prevailing failure modes were bar pull-out with wood attached and tensile splitting of timber, signifying the integrity of the interface between the adhesive and the FRP rods.

Fig. 4.

Typical failure modes.

3.4 Comparison of Studied Cases

To allow visualization of the ultimate performance and comparison among the cases studied, Fig. 5 graphically depicts the average values of ultimate loads, moment resistance and corresponding displacements for each examined group of specimens. As evident, the bonded specimens with 10 mm rods and D1 exhibited the best performance both in terms of moment resistance and of corresponding displacement. Specifically, the joint with 10 mm rods and D1 attained an average ultimate load of 44.31 kN. As anticipated, for both D1 and D2, the joints with 10 mm rods surpassed their 8 mm rods counterparts. In particular, the average moment capacity for D1 increased by 64%, when comparing 10 mm to 8 mm rods. Furthermore, when comparing the two proposed designs, improved performance was observed for D1 design (i.e., four bars vertically) for both examined rods diameters (i.e., average M_u,Exp equal to 9.46 kNm and 15.51 kNm for 8-D1 and 10-D1, respectively and average M_u,Exp equal to 9.21 kNm and 10.80 kNm for 8-D2 and 10-D2, respectively).

Fig. 5.

Graphical presentations of average values for each studied group.

4 Conclusions

This paper presented an experimental study involving 16 bonded glulam beam specimens fabricated with glued-in BFRP rods. The key parameters investigated were the rod diameter (8 mm and 10 mm) and the design configurations (D1, D2). The load-displacement curves were reported and discussed. The predominant failure mode observed was pull-out of bars at tension zone and tensile splitting of timber. Across all cases, the connections with 10 mm bars demonstrated superior performance compared to those with 8 mm bars, with the D1 configuration outperforming configuration D2. Research work is currently underway to provide design guidance for the bonded glulam specimens. Future research could examine alternative design arrangements and comparison of their moment resistance with that of continuous beams and of other connection types.