Avaliação do efeito de escala nos valores de carga de colapso de interface entre concreto e polímero reforçado com fibra de carbono

Abstract

This research is a study about the debonding phenomenon at interface between concrete and carbon fiber reinforced polymer (FRP) by way of performance and analysis of experimental tests. The objective of this study is to investigate the size effect in concrete and FRP debonding. The influence of variation of size concrete specimen on interface failure was analyzed when the same FRP geometry was adhered. Two concrete cylinders specimens are superimposed on plane of cross section, and the specimens are connected by three FRP pieces symmetrically adhered in the longitudinal perimeter. The lower part is fixed and the top is pulled. This induces appearance of shear stresses at interface. Different cylinders specimens made and core samples were used: 5cm x 10cm, 10cm x 20cm and 15cm x 30cm. Different thirteen geometries of the composite material were adhered. The failure happened in concrete in some experiments because concrete reached the normal strength before the shear stress required for the debonding, especially in smaller specimens because the tensile load carrying capacity for these cases is lower. An influence of specimen size was not noticed in the maximum load and shear stresses for debonding, and stiffness. Although experiments with smaller specimens generate good results, the reason that limits their use is the fact that the concrete often reaches the normal stress capacity. In addition, a very small FRP geometry was used in tests of debonding with smaller specimens. The use of a small FRP geometry resulted in a high shear stress, and with this use, it is necessary to use a high correction factor. The extraction process can cause damage to concrete such as micro cracking, especially in smaller specimens. It was not conclude about the influence of use of core samples in interface failures, but a change of the load versus displacement graph was noted. The influence of use of core samples happened differently in each test. An debonding analysis in tests with core samples may not be appropriate.