Model-based fatigue identification of NiTiCu shape memory alloys under thermomechanical cyclic loading

This work investigates the use of model-based techniques to identify functional fatigue stages of SMA when used as actuators. The model-based technique enables an online identification of fatigue level of SMA actuators, in comparison with off-line conventional fatigue analysis techniques, such as SEM and TEM. The models are identified by linearisation of three equations - the model of phase transformation, the temperature dynamics described by Joules heat-convectional cooling, and a constitutive equation relating changes in stress, strain, temperature, and mole fraction. Residues are defined as a function of simulated error and used in the decision rule to identify the fatigue level of the tested wire. The tested specimens are NiTiCu SMA drawn in 0.7 millimeter diameter wires. The materials are loaded with a constant weight under various thermal cycling, producing a set of SMA wires with different fatigue levels. The proposed fatigue identification approach has been tested with the specimens being subjected to a range of stress from 76 MPa to 204 MPa. Results show that the accuracy of the proposed method has been found to be at least 93% for all loading conditions.