Reliability Assessment of Shape Memory Alloy Parts Based on a Probabilistic Approach

In this paper, we present a fatigue reliability assessment of shape memory alloy (SMA) parts by adopting a recent probabilistic approach based on the extension of the Dang Van (DV) high cycle fatigue (HCF) criterion to SMAs. The adopted approach considers the probabilistic aspect of the DV criterion and the reliability assessment is established using uniaxial loading experimentations taken from the literature. The aleatory uncertainty quantification of the material parameters and the loading path parameters have been taken into account. In this work, the sampling technique has been modified. Indeed, the fatigue reliability is computed by using the Latin hypercube sampling (LHS) and it is compared with the Monte Carlo Simulation (MSC) technique. The cyclic loading is accomplished in the fully martensitic phase state with a stabilized cyclic response known as elastic shake-down. In this work, the HCF reliability prediction is improved by reducing the computational time compared and considering rare events. Moreover, iso-probabilistic Dang Van zones (ISPDV) are obtained by fixing a desired reliability.