Benefits of strengthening timber with fibre-reinforced polymers

Doktorsavhandling, 2011

As a material, timber offers a wide range of alternatives in terms of structural applications. The natural defects present in timber are, however, the source of large variations in mechanical properties. This drawback has been partially counteracted by using engineered wood products, EWP, (glued-laminated timber (glulam), composite I-beams, parallel strand timber and so on), instead of solid wood. A few decades ago, fibre/polymer composite materials made their entrance in the civil engineering arena. They are generally used as strengthening devices. The content of this thesis is related to the benefits of strengthening timber with fibre-reinforced polymers (FRP), in particular for members loaded in tension perpendicular to the grain and in compression parallel to the grain. In the first part of this thesis, the research focused on the experimental and numerical study of strengthening glulam members loaded in tension perpendicular to the grain. The weak mechanical properties of wood in tension perpendicular to the grain are often the origin of catastrophic brittle failure. In order to enhance the tension strength perpendicular to the grain and achieve more ductile failure, flax fibre and glass fibre-reinforced polymer composites were used to strengthen glulam specimens. Three series of specimens of glulam timber (flax fibre reinforced, glass fibre reinforced and unreinforced) were tested in tension perpendicular to the grain. For an approximate amount of FRP reinforcement of 1.2% in volume (thickness of ~ 0.7 mm), an increase in tensile strength of up to 74% was shown, with a stiffness increase of up to 41%. For all reinforced specimens, semi-ductile failures were observed. A parametric study was carried out using both the Monte Carlo (MC) method and the first-order second-moment (FOSM) method. The variation in the mechanical properties of the flax fibres appeared to be the driving parameter for the strength of the system. A numerical analysis was made to model unreinforced and flax fibre composite reinforced glulam specimens loaded in tension in the direction perpendicular to the grain. Two-dimensional models were used to study the elastic response and the softening of the specimens. Damage initiation and propagation was modelled based on the “fictitious crack model”. Cohesive elements, together with a traction separation law, were used. The glulam model, where high tensile stresses perpendicular to the grain are expected, considered cylindrical orthotropy (annual rings). The tensile stresses perpendicular to the grain obtained with the numerical model are comparable to those obtained from experiments. Cohesive interface elements have been used successfully to model the crack formation and propagation in the glulam under tension perpendicular to the grain. The limiting factor when it comes to designing timber structures is often the stiffness properties of timber products. The stiffness requirements in the serviceability-limit state, both short-term and final deformation, especially in horizontal members, are a factor that often makes it necessary to increase the dimensions of the member. This causes an increase in material use and thereby higher production costs. This negative feature could be reduced by strengthening with carbon fibre-reinforced polymer (CFRP). In the second part of the thesis, the research focused on the study of strengthening glulam members loaded in compression parallel to the grain. In the experimental investigation, the effect of strengthening glulam beams loaded in bending with different amounts of laminate on the tension and compression side was investigated. A simple analytical model for wood, based on a compression test on small-scale timber specimens and taking account of the plasticity of wood under compression, was developed to analyse the behaviour of the beam under flexural loading. All types of reinforcement resulted in higher stiffness, as well as higher ultimate moment capacity. It is shown that compression tests on small-scale timber specimens do not provide accurate material data that can be used for beam analysis. Explanations based on theory and experiments are discussed in order to understand this discrepancy. Partial agreement between experimental results and analytical results is, however, reached in terms of stiffness and moment capacity by modifying the wood material data for compression. Further experimental investigations are needed in order to strengthen the findings reported in this thesis.