Strengthening steel beams with adhesively bonded composite laminates

Licentiatavhandling, 2006

Strengthening steel structures with carbon-fibre-reinforced-polymer (CFRP) laminates bonded to the steel substrate has become an interesting alternative to the conventional strengthening methods. The current research comprises analyses of the global and local behaviour of double-symmetrical steel beams strengthened with CFRP laminate bonded with adhesive to the tension flange. The beams were subjected to loads applied in four-point bending. In this research, the global behaviour is the load-strain or load-deflection behaviour of a strengthened beam, while the local behaviour comprises the interfacial stresses obtained in the bond line. These behaviours were studied in both the serviceability and the ultimate limit state. Methods for conducting these analyses were based on analytical models, laboratory tests and FE (Finite Element) models. Furthermore, in this research, different configurations of the strengthening system were studied, where CFRP laminates and adhesives with different material and geometrical properties were used. Another aim of this research was to study the failure modes and obtain practical experience from the preparation work on the strengthened beams. The results produced by each method have been analysed and compared reciprocally. The results reveal that the load-carrying capacity could be increased by about 20% for a double-symmetrical steel beam strengthened with bonded CFRP laminate. For the analysed beams, the increase in stiffness in the elastic phase of the steel beam was negligible. However, the analytical solutions revealed that the behaviour of the strengthened beams in the elastic phase can be affected by selecting another configuration for the strengthening system, if CFRP laminate with stiffer properties could be selected. This will then affect the magnitude of the interfacial stresses, which could be critical for the strengthening system. The analyses of the interfacial stresses reveal that both the shear and peeling stresses in the adhesive exhibit large-scale variations over the width of the bond line, with their maximum values at the location closest to the web plate. In addition, the shear stress that developed as a function of the magnitude of the applied load, after the steel beam had reached yielding, was almost linear and was therefore not affected by the yielding in the area at the mid-span of the steel beam. However, the interfacial shear stresses near the mid-span of the strengthened beam were strongly affected by the yielding in the steel beam and the magnitude of the shear stress in the vicinity of this area might well exceed the shear capacity of the adhesive.