|
The widely used Brinson model allows through experimental data to describe the behaviour of shape-memory-alloys (SMAs) in different condition of stress and operating temperature. The chief advantage of this one-dimensional constitutive behaviour model is in accounting for the existence of two forms of martensite within the alloy, twinned and detwinned. An internal variable approach that represents the martensitic fraction is introduced by the model and divided into the stress induced components and temperature induced component. And providing a uni-axial temperature-stress diagram clearly separating the regions of the three stable phases: austenite (A), twinned martensite (MT), and detwinned martensite (MD). The complete knowledge of the Brinson model allows to estimate the SMA behaviour in different conditions and to improve it according to specific tasks. In order to apply the model it is necessary to experimentally estimate the SMA material parameters These parameters include the Young’s moduli of each phase and critical transformation stresses, phase transformation temperatures as a function of stress level and thermal expansion coefficients of both austenite and martensite, as well as recoverable strain, both initial and after the alloy has been trained. While the Brinson model has been developed and validated for NiTi based alloys, the increasing interest aimed at Cu-based SMAs, mainly for their competitive price and relatively high performance, requires a validation of its extension to these materials. The paper focuses on a specific alloy of interest, Cu-16Al-10Mn (%at.), but results can be extended to different chemical compositions. The paper explores the opportunity to use experimental methods based on monotonic loading and unloading of the Cu-Al-Mn specimens, as well as thermal cycling under both zero and non-zero stress levels to estimate the parameters mentioned above. Furthermore, its goal was to ascertain whether any specificity intrinsic to the Cu-based alloy would require changes to the parameter estimation method or introduction of new parameters to the model to fully describe its behaviour.
|
