PSI - Issue 38

Di Song et al. / Procedia Structural Integrity 38 (2022) 546–553 Di Song and Chao Yu/ Structural Integrity Procedia 00 (2021) 000–000

547

2

1. Introduction NiTi shape memory alloys attracted great attention in the last decades mainly on their excellent superelastic and shape memory properties. The superelasticity of NiTi SMAs is usually applied in the fasteners and connecters of aerospace structures since this material presents more than 8% recoverable strain in tension loadings instead of 1% for other kinds of metals and steels. Such components usually endure complex cyclic loadings in the applications, leading to the cyclic deformation and fatigue failure are key issues that should be solved, especially for the non proportional multiaxial loading cases. The uniaxial cyclic experiments are relatively widely performed by many researchers up to now, where the deformation properties of the gradual accumulation of the residual strains, the degradation of the transformation start stresses, and the decrease of the dissipation energy per cycle have been observed and analyzed in a series of literature （ Gall et al. (2001,2002), Sehitoglu et al. (2001), Zhang et al. (2008), Dayananda and Rao (2008), Liu et al. (1999), Eggeler et al. (2004), Strnadel et al. (1995a, b), Sehitoglu et al. (2001a, b, 2006), Nemat-Nasser and Guo (2006), Predki et al. (2006), Wang et al. (2010) , Saleeb et al. (2013), and Song et al. 2014a, b). The performances of the special cyclic loading behaviors are defined as the “transformation ratcheting” of NiTi shape memory alloys by Kang et al.(2009). Moreover, the uniaxial fatigue tests are also reported by many researchers ((Kang et al.(2012), Dunand-Châtellet and Moumni (2012), Maletta et al.(2014) and Song et al. (2015a) ). In these tests, the “Z shaped” fatigue lives with the increasing loading levels is regarded as the most special feature which distinguished with other metals and steels. The mechanism of such performance is illuminated by Song et al. (2014a, 2015b, 2016, 2017a) as the transformation extent of the loadings greatly influences the fatigue lives of NiTi SMAs, the loading cases with lower mean stresses but higher transformation extent may present much shorter fatigue lives. Song et al.(2015b) has also proposed a damage-based fatigue prediction model for NiTi SMAs, which includes the damage derived from the initiation and the propagation of the micro-cracks, as well as that from the martensite transformation and its reverse. The proposed fatigue model can give a reasonable prediction of the damage evolution of each loading case, and the fatigue lives predicted are all located in 1.5 times error bands. The multiaxial cyclic and fatigue experiments for NiTi SMAs are also gradually reported in recent years, where the whole-life transformation ratcheting in five types of non-proportional multiaxial paths are performed and compared by Song et al.(2014a, 2015a, c, 2017a, 2019) and the fatigue lives in the loading cases with various applied stress levels and loading paths are also analyzed. It is found in the studies that the loading path can greatly affect the transformation ratcheting and fatigue lives of NiTi shape memory alloys, leading to a more complex situation to accurately predict the fatigue life in each loading case. Mahtabi et al. (2015) has demonstrated that nearly all the classic fatigue models are not appropriate for the life-prediction for NiTi SMAs due to the 20-100 times of errors. Thus, Song et al. (2017b) extend their uniaxial fatigue model to the multiaxial region, where the damage derived from the martensite re-orientation is also taken into consideration therein. The model mentioned highly increased the prediction accuracy of multiaxial loadings up to twice error bands. Moreover, Song et al. also proposed a simplified multiaxial life-prediction model for NiTi SMAs, which can give predictions with only four pre-conducted loading cases. That model greatly decreased the required data in the predictions, and the predicted accuracy can be maintained in three times error bands. Meanwhile, the above-mentioned damage evolution and fatigue models appropriate for NiTi SMAs all adopt dissipation energy as the damage parameter, which means that the evolution of the residual strains during the cyclic loadings cannot be predicted by such kinds of fatigue models (indeed, they can be predicted by constitutive models). Thus, in this paper, a model which can predict the accumulation of residual strains with various non-proportional multiaxial loading cases is proposed, and the multiaxial fatigue lives in these loadings are also predicted. The results can provide a new method for the prediction of the multiaxial transformation ratcheting and fatigue life for NiTi shape memory alloys. 2. Experimental methods The near-equiatomic NiTi SMA micro-tubes from Jiangyin Fasten-PLT Materials Science Co., Ltd, China are adopted in the fatigue tests. The outer diameter of the micro-tubular specimen is 2.5 mm, and the inner one is 2.2 mm. The austenite finish temperature Af is 302 K, measured by a differential scanning calorimetry (DSC), which