Abstract

This study aims to the understanding of the impact of uncertainties related to constitutive model parameters on the cyclic response of CuAlBe Shape Memory Alloy (SMA) strands, and to define an adequate methodology for constructing the strands, thereby reducing uncertainty in their behavior. Employing a Bayesian framework and utilizing experimental data from tensile tests, the prior knowledge of mechanical properties at the wire level is updated. Various scenarios are explored, including model parameter updating for wires tested at different strain levels and from the same and different heat treatment batches. The updated model parameters are then used in a strand model, considering two scenarios: the strand is constructed from pieces cut from a single wire, post-heat treatment; or from pieces cut from wires from different heat treatment batches. The study concludes proposing a practical methodology to define test conditions for CuAlBe SMA wires and guidelines for the utilization of experimental data to predict bounds in the cyclic response of CuAlBe SMA strands.