Near-equiatomic NiTi shape memory alloy (SMA) torsional tube actuators were trained for two-way shape memory effect (TWSME) by repeated thermal cycling under an isobaric load. Performance of the trained actuators was assessed by thermally cycling through a complete phase transformation under a range of isobaric loads that varied from negative to positive and included loads near zero. To assess the actuation performance of the trained SMA components, extended isobaric thermal cycling and cycling under varying loads to constant strain limits was performed. Additionally, isothermal loading was applied in the fully martensitic state prior to and following training.
Results show stable TWSME when cycling under significant isobaric loading in the trained direction, however at low or negative loads (loads applied opposite to the training direction) a degradation of TWSME occurred during thermal cycling. Isothermal loading showed that martensite variant reorientation and detwinning was redistributed in the trained actuator when compared to an untrained actuator. Thermal cycling against constant strain limits was shown to have a negligible effect on the stability of the TWSME and overall performance of the trained actuator. Various combinations of isothermal and isobaric loading were shown to expand the operating range at low and negative loads. Additionally, load paths were identified which limit the degradation of TWSME over extended cycling.
The aforementioned results are discussed in the context of the correlation between uniaxial isobaric and isothermal loading and texture measurements obtained by in-situ neutron diffraction at stress and temperature. These results show various thermomechanical combinations of heating and loading sequences that yield the same final martensite texture in SMA, which highlights the ability to take different paths yet still obtain the desired actuator response while minimizing irrecoverable deformation mechanisms. The implications of extending these uniaxial results to the design and fabrication and ultimately improving the performance of torsional SMA actuators are discussed.