Retrofit of non-seismically designed beam-column joints by post-tensioned superelastic shape memory alloy bars

Abstract

A series of tests on three full-scale substandard exterior beam-column joints were performed to investigate the efficiency of the proposed retrofit configuration, which is the use of externally applied post-tensioned shape memory alloy (SMA) bars. A major group of structural deficiencies resulting from lack of shear reinforcement in the joint, use of low strength concrete and plain round bars were taken into account in the construction of test specimens. While the reference specimen represents the as-built subassembly, the other two were retrofitted by the post-tensioned SMA and steel bars to compare the contribution of superelastic and conventional material on the response. The specimens were exposed to quasi-static cyclic loading up to 8% drift ratio to simulate an intensive level of seismic hazard. The reference specimen underwent a brittle shear failure as excessive cracks mostly concentrated in the joint panel while there was almost no damage in the rest of the RC components. A joint failure with enhanced response quantities was observed in the specimen retrofitted by post-tensioned steel bars. The specimen incorporating the retrofit solution via post-tensioned SMA bars was capable of performing an adequate performance and promoting minimization of the damage in the joint panel, which results in more ductile behavior. The hysteretic response of the SMA retrofitted specimen was validated with a refined numerical model in ATENA Science software. Experimentally observed response was also verified by an analytical model based on fracture mechanics considering the nonlinear behavior of plain concrete under tension. Due to inherent uncertainties in material constitutive laws, the analytical model was evolved to a stochastic level to propose a more advanced model for estimating the capacity of the reference and retrofitted joint. It is found that the experimental results were within the prominent range of Probability Density Functions (i.e. mean ± 1 SD) of the estimated joint tensile stress especially for the shear damaged specimens.