Composites for Building Assembly

Structural members made from fibre reinforced polymer (FRP) composites are alternatives for construction of durable and lightweight structures. Generally, FRP composites consist of a polymer matrix reinforced by high-strength and stiffness fibres. The way that the fibres are arranged in the polymer matrix forms an anisotropic composite. Compared to conventional construction materials, the key advantages of FRP composites are lightness in weight, resistance to corrosion, and embrace to design for manufacturing and assembly, with possible other features from specific constituents and tailored formations. The success of FRP composites as the primary structural members in building construction requires proper designs of such structural members and connections, in consideration of the anisotropic and brittle characteristics of the composite materials. This work therefore aims to introduce a series of FRP structures and connections developed in this way dedicated for prefabrication and design for manufacturing and assembly. It is supported by a range of research activities, in order to advance member and connection solutions and to provide structural stiffness in compensation of low elastic modulus for glass fibre reinforced polymer (GFRP) composites and to provide ductility in contrast to their material brittleness. With the experience gained from these researches, large scale FRP structures built using the developed structural components and connections are also introduced.

An adhesively bonded modular GFRP web-flange sandwich system for use in building floor construction is described in this chapter. Sandwich units are developed by incorporating standard pultruded GFRP box (i.e. square hollow section) or I-profiles between two GFRP flat panels to form built-up modular sections with considerable improvement of bending stiffness. These modular sections may then be assembled in the transverse direction to form a one-way spanning slab system. Sandwich specimens with different span-to-depth ratios and core configurations were prepared via adhesively bonding the component profiles, and were then tested under four-point bending. It can be found that the span-to-depth ratio greatly influenced the failure mode, and that inserting foam into the core of the sandwich significantly improved the load-carrying capacity. Also, adhesive bonding was able to provide full composite action at both serviceability and ultimate loads, depending on the quality of the bond. Finally, structural theory was used to estimate the bending stiffness and load-carrying capacity of the sandwich specimens and good agreement with the experimental results was found.

This chapter presents a modular assembly of sandwich structures using glass fibre reinforced polymer (GFRP) components for two-way slab applications. The sandwich assemblies were built-up sections made from pultruded web-core box (i.e. square hollow section) profiles incorporated between two flat panels, connected via adhesive bonding or novel blind bolts. Two different pultrusion orientations were achieved and examined i.e. flat panels with pultrusion directions either parallel (unidirectional orientation) or perpendicular (bidirectional orientation) to the web-core box profiles. The effects of pultrusion orientation, shear connection and the presence of foam core materials on strength and stiffness were investigated by experimental testing of two-way slab specimens until failure. Sandwich slabs with unidirectional orientation showed premature cracking of the upper panel between fibres, whereas those with bidirectional orientation showed local out-of-plane buckling of the upper flat panel. The unidirectional slab also showed greater bending stiffness than bidirectional slabs due to the weak in-plane shear stiffness provided by the web-core profiles, which introduced partial composite action between the upper and lower flat panels of the bidirectional slab. Finite element models and analytical techniques were developed to estimate deformation, failure loads and the degree of composite action, and these showed reasonable agreement with the experimental results.

This chapter presents an experimental and modelling investigation into modular composite beam structures using web-flange fibre reinforced polymer (FRP) and steel for building floor construction. The modular FRP slabs are formed from adhesively bonding pultruded box profiles (i.e. square hollow sections) sandwiched between two flat panels. They are then connected via adhesive or one-sided bolted connections to steel beams to form a composite system. Two different fibre (pultrusion) configurations are investigated in this chapter: flat panel pultrusion with direction either parallel or perpendicular to the box profiles. Composite beams were tested under four-point bending and evaluated for bending stiffness, load-carrying capacity, and the degree of composite action within the FRP web-flange sandwich slab and that provided by the shear connections. All the composite beams showed ductile load–deflection responses, with yielding of the composite beam commencing prior to failure of the FRP slabs. Furthermore, adhesive bonding provided full composite action, but the novel bolted connections with a certain spacing provided either full or partial composite action, dependent on the pultrusion configuration of the FRP slab. An analytical procedure is also developed to evaluate the bending stiffness and load-carrying capacity of the composite beams. Finite element analysis was further employed in this chapter, showing good comparisons to the experimental results.

The shear stiffness of proposed blind bolts was experimentally investigated in this chapter through steel-FRP joints loaded in tension, for shear connections within the steel-fibre reinforced polymer (FRP) composite beam systems. The number of bolt rows (either one or two bolt rows) and the effect of the pultrusion orientation of the FRP web-flange sandwich slab on the joint stiffness and joint capacity were examined. It was found that joint capacity and failure modes were dependent on the pultrusion configuration of the FRP slab and the number of bolt rows. A unidirectional configuration consisting of an FRP slab with box-profiles (i.e. square hollow sections) parallel to flat panels exhibited shear-out failure, whereas a bidirectional orientation consisting of an FRP slab with box-profiles perpendicular to flat panels exhibited both shear-out and net-tension failure in the FRP component. The experimentally derived shear connector stiffness was then used in a proposed design formulation to predict the bending stiffness of modular steel-FRP composite beam systems, considering two kinds of partial composite actions. These were the composite action provided by the blind bolt shear connector and the composite action provided by transversely-oriented webs within the slab. Agreement was observed between the experimental results and the proposed design formulation.

This chapter focuses on the mechanical performance of fibre reinforced polymer (FRP) columns with square hollow sections (SHS) in axial compression. Width-thickness ratio (b/t) is an important geometric parameter for the local buckling of plates and therefore also for such SHS columns. The effects of b/t on the failure modes and load-carrying capacities of pultruded glass fibre reinforced polymer (GFRP) SHS columns are investigated in this chapter. Two SHS with different b/t values of 10.7 and 15.9 respectively are examined under axial compression. Experimental results reveal that local buckling occurs in section B (b/t = 15.9) but not in section A (b/t = 10.7). From a theoretical analysis, a formulation of critical b/t values is established at the boundaries between the failure modes of such GFRP SHS columns under compression, considering the different boundary conditions of the SHS side plates. It is commonly understood that global buckling occurs in columns with higher non-dimensional slenderness λ. This is only true when the width-thickness ratio b/t is less than the derived critical value. Experimental results from this study and previous literature are consistent with the developed theoretical estimations of failure modes and load-carrying capacities for GFRP SHS columns, considering the effects of both non-dimensional slenderness, λ, and width-thickness ratio, b/t.

Web-flange sandwich structures were built up by two glass fibre reinforced polymer (GFRP) panels and square hollow sections (SHS) in between through adhesive bonding or mechanical bolting. Experiments in compression were conducted in order to understand the failure modes including global and local buckling, load-bearing capacities, load–displacement curves and load-strain responses. Accordingly the effects of different connection methods and different spacing values between the SHS were clarified. Sudden debonding failure between GFRP panels and inner SHS columns was found on adhesively bonded specimens; while mechanically bolted specimens showed evident lateral deformation and progressive failure until the ultimate junction separation failure on the GFRP SHS columns. Local buckling was found on GFRP panels of specimens with a larger spacing between the two SHS. Finite element analysis and analytical modelling were performed to estimate the load–displacement curves and the critical stress for the local buckling on GFRP panels, where consistent agreements with experimental results were received.

Timber and steel studs or posts are commonly used in wall constructions for buildings. In this context and with the results from previous chapters, pultruded glass fibre reinforced polymer (GFRP) studs may provide an alternative solution considering their light weight and improved durability. However, integrating the GFRP wall studs to a steel frame structure is challenging, as proper connection methods are required. A sleeve connection was proposed and examined in this chapter for wall studs to steel beams. Pultruded GFRP stud was fastened to the sleeve connector by one of three methods: ordinary bolt, one-sided bolt and adhesive bond. The connector was then fastened to the steel beam through ordinary bolts. Connections with conventional steel angles were also prepared for comparison purpose. A series of moment-rotation experiments were conducted on these stud-to-beam connections. In addition, two stud lengths were designed in order to study the connection behaviour under shear force dominant loading and moment dominant loading conditions. Experimental results were obtained including failure mode, moment-rotation response, shear-rotation response, joint rotational stiffness and capacity. It was found that the bonded sleeve connection outperformed all the other connections and was classified as a rigid and partial strength connection.

This chapter numerically investigates the proposed bonded sleeve connection for joining tubular glass fibre reinforced polymer (GFRP) composites and steel members. Experimental results focused on mechanical responses of such specimens using bonded sleeve connections and conventional steel angle connections were introduced in previous chapter. These results are used to set the benchmark for detailed finite element (FE) modelling in this chapter. In the detailed FE analysis, bolt geometry including head, shank and washer were accurately modelled. Paired contact elements were used for simulating the contact and slip behaviour between bolt shanks and holes, washers and steel or GFRP. The pretension force in the bolts was also taken into account by implementing pretension elements. The FE models developed were first validated against the experimental results in terms of failure mode, moment-rotation curves and strain responses. Parametric studies were then undertaken to investigate the structural behaviour of the bonded sleeve connections considering the effects of major design parameters such as endplate thickness, bonding length, number of bolts, etc. It was found that the endplate thickness dominates the initial stiffness and the elastic moment capacity of the bonded sleeve connection and the presence of central one-sided bolts may improve the elastic moment capacity of the bonded sleeve connection.

This chapter presents the cyclic performance of bonded sleeve connections for joining tubular glass fibre reinforced polymer (GFRP) beams and columns. Specimens with different endplate thickness and number of bolts are examined under cyclic loading. The hysteretic moment-rotation responses of specimens, including rotational stiffness, ultimate moment and rotation capacity, and local strain responses are experimentally obtained and comparatively investigated. The cyclic performance of beam-column specimens is also characterized in terms of their ductility and energy dissipation capacity. Excellent ductility and energy dissipation capacity can be achieved through yielding of the steel endplate prior to the final connection failure. Detailed finite element analysis is also performed to describe the cyclic performance of beam-column specimens with bonded sleeve connections. Numerical and experimental results agree well in terms of hysteretic moment-rotation responses, ductility and energy dissipation capacity. Further parametric study of the endplate thickness provides evidence that reduction in endplate thickness may decrease the moment capacity with satisfactory ductility and energy dissipation capacity through the full development of steel yielding.

Bonded sleeve joints formed by telescoping a steel tube connector for bolt-fastening have been shown as effective means for assembling tubular fibre reinforced polymer (FRP) members into more complex structures such as planar or space frames. A theoretical formulation is developed in this chapter to estimate the capacity of such joints in axial loading for circular tube sections and the predictions are validated by experimental results covering various section sizes and bond lengths. The formulation is based on the bilinear bond-slip constitutive relationship considering elastic, softening and debonding behaviour at the adhesive bonding region. Finite element (FE) analysis is also conducted to estimate the joint capacity and to describe shear stress distribution in the adhesive layer, validating the theoretic results. The theoretical formulation is therefore further used to study the effects of design parameters including bond length and adherend stiffness ratio, again validated by FE results. An effective bond length can be calculated by the theoretical formulation for the joint capacity at both the elastic limit and the ultimate state. Given a bond length, an optimal adherend stiffness ratio can also be identified to achieve the maximum joint capacity at the elastic limit or the ultimate state.

A splice connection is introduced in this chapter for connecting tubular fibre reinforced polymer (FRP) members. This connection consists of a steel bolted flange joint (BFJ) and two steel-FRP bonded sleeve joints (BSJs). The BFJ connects two steel hollow sections, each of which is telescoped into the targeted tubular FRP member through adhesive bond, forming a BSJ. To evaluate the performance of the proposed splice connection under axial loadings, BSJs of four different bond lengths and BFJs of two bolt configurations are tested individually. Finite element (FE) models are developed which feature a bilinear bond-slip relation, contact behaviours and bolt pre-tensioning. Comparisons are made between experimental and FE results in terms of load–displacement behaviours, ultimate capacities and strain responses. Besides being capable of identifying an effective bond length for the BSJ and modelling the yielding process of the BFJ, FE analysis provides insight into the distribution of adhesive shear stress over the bond area of the BSJs, and the steel yield line pattern on the flange-plate of the BFJs. Verified by experimental results, the FE modelling technique is then utilised to understand the integrated axial behaviours of a complete splice connection.

This study investigates the cyclic performance of splice connections developed for hollow section fibre reinforced polymer (FRP) members. Splice connection specimens, each consisting of a steel bolted flange joint between two hollow section steel-FRP bonded sleeve joints, are prepared in three configurations with difference in bolt arrangement or bond length. Correspondingly, detailed finite element (FE) models are constructed with consideration of yielding of the steel components, damage in the adhesive bond, pre-tensioning of the bolts and contact between the bolt-fastened parts. Tested under a cyclic flexural loading, the specimens experience different levels of yielding in the steel flange-plates before ultimate failure in the FRP member or in the steel flange-plate. Ductility and energy dissipation capacity are demonstrated in a specimen where plastic deformation of the steel flange-plates is fully developed. The strain responses are also analysed to identify damage in the adhesive bond and yielding in the flange-plates. The FE modelling agrees with the experimental results in terms of moment-rotation and load-strain responses, and can also predict the initiation of the ultimate failure in the FRP using the Tsai-Wu failure criterion.

Multicellular web-flange composite structures were assembled using glass fibre reinforced polymer (GFRP) box sections as web sections and plates as face sheets, further with glass magnesium (GM) or gypsum plaster (GP) panels on the surface for fire protection. The structures were subjected to ISO 834 fire curve from underside where the GM or GP panels were installed and a constant load on top to introduce bending during fire exposure. Experimental results showed that the fire endurance times of the structures before failure were extended from 54 min without protective panels to 83 min or 103 min by a single layer of GP or GM panel. When double layers were used, the fire endurance time increased to 113 min for GP panels and 158 min for GM panels. Numerical modelling was further established to estimate the temperature distribution in the specimens. Effects of the GM and GP layers on the thermal and mechanical performances of loaded specimens in fire can be clarified. The fire performance of GFRP multicellular web-flange composite structures enhanced by fire resistance panels may also be well demonstrated experimentally and numerically.

With the success of applications in aerospace, marine, electrical, automotive transportation industries, fiber reinforced polymer (FRP) composites have also found their places in civil engineering. The use of glass fibers (therefore GFRP) further reduces material costs and incorporates beneficial environmental aspects such as low energy consumption and low carbon dioxide emissions. Pultrusion is an automated and economical manufacturing method for continuous production of constant cross-section structural profiles with consistent material properties. However, in comparison to steel, pultruded GFRP composites are highly orthotropic materials and associated with lower elastic modulus and shear strength and lack of ductility at material level. In the previous chapters, a range of design concepts and structural solutions have been developed with experimental demonstrations for structural members and connections using pultruded GFRP composites, in consideration of their distinct material features. Experimental and modelling investigations have shown results of their mechanical performance of the developed structural members and connections in terms of stiffness, strength or even ductility. In the last chapter of this work, large scale structural applications are introduced with the use of developed structural members and connections. Further highlights of such structural applications are their ways of construction where design for manufacturing and assembly (DfMA) is practiced.