The objective of this study is to evaluate the effect that mechanical training has on the properties of NiTi based shape memory alloys. The unique mechanical behavior of shape memory alloys, which allows the material to undergo large deformations while returning to their original undeformed shape through either the shape memory effect or superelastic effect, has shown potential for use in seismic design and retrofit applications for civil engineering structures. However, cyclic loading has been shown to degrade the energy dissipation capacity and decrease the recentering capability of the material due to fatigue effects. It has been recommended that mechanical training of superelastic shape memory alloys prior to use in applications can limit these fatigue effects. A factorial experimental design is employed to explore the optimal number of mechanical training cycles, strain level of training, and the effect of the loading rate after training in order to minimize the degradation in the loading plateau stress, residual strain, and equivalent viscous damping properties. The results presented can serve as a guide to optimizing the properties of NiTi shape memory alloys for seismic applications. The ability to obtain stable properties of shape memory alloys under a specified training schedule further supports the eventual implementation of the material into actual building and bridge systems as seismic design and retrofit devices.
The cyclic behavior of shape memory alloys (SMAs) in their austenitic form is studied to determine the most appropriate method of modeling in terms of both accuracy and ease of implementation. Four different models for SMA behavior are evaluated: (a) a simple nonlinear elastic model, (b) a trigger-line model, (c) a one-dimensional thermomechanical model, and (d) a one-dimensional thermomechanical model which accounts for the behavior of SMAs under cyclic loading. Using a two degree-of-freedom bridge model with SMA restrainers and a single degree-of-freedom building model with SMA cross-braces, the effect of using the different models on the seismic response of the bridge and building is evaluated. Using a suite of nine earthquake ground motions, the displacement response histories with the four different models are compared. The results illustrate that although the models show quite different behaviors for the SMAs, the resulting responses of the bridge and building are insensitive to the type of model used. For most of the ground motion records used, the difference in the maximum displacement for the four models was less than 15%. This study lends support to the use of more simplified models when evaluating the effectiveness of the SMAs for seismic response modification.
This study evaluates the properties of superelastic shape memory alloys under cyclical loading to asses their potential for applications in seismic resistant design and retrofit of civil engineering structures. Shape memory alloy bars are tested to evaluate the effect of bar size (diameter) and loading history on the strength, equivalent viscous damping, and recentering properties of the shape memory alloys in superelastic form. The bars are tested under both quasi-static and dynamic loading. The results show nearly ideal superelastic properties can be obtained in large diameter shape memory alloy bars. However, comparing these results to previous studies, the more common wire form of the shape memory alloys show higher strength and damping properties compared with the large bars. The recentering capabilities (based on residual strains) are not affected by the section size of the bar. Overall, the damping potential of superelastic shape memory alloys is low for large diameter bars, typically less than 7% equivalent viscous damping. Degradation of the superelastic properties of the shape memory alloys occurs for cyclical strain greater than 6%, leading to increased residual strains and reduction in energy dissipated. Finally, strain rate effects are evaluated by subjecting the shape memory alloys to loading rates representative of typical seismic loadings. The results show that increased loading rates lead to slight decreases in the equivalent damping, but have negligible effect on the recentering of the shape memory alloys.