Stored-energy-based fatigue criterion for shape memory alloys

This paper presents a new stored-energy-based model for structural fatigue of shape memory alloys (SMAs) where the conversion of hysteresis work into dissipation and stored energy is discussed in detail. The results show that during cyclic pseudoelastic process, while part of the hysteresis work is dissipated into heat, the remainder is stored in dislocations and in residual martensite variants. At macroscopic scale, during first few cycles, the stored energy is large and strongly influences the thermomechanical behavior of the SMA, it then gradually decreases at each cycle and approaches zero at the shakedown state. At microscopic scale, however, the stored energy continues to accumulate in the micro-structure and when it reaches a critical value, fatigue cracks initiate and propagate. In this study, a stored-energy-based criterion is proposed and its validation with experimental data shows that the stored energy is the relevant parameter to predict the fatigue of SMAs.