PSI - Issue 19

Martin Nesládek et al. / Procedia Structural Integrity 19 (2019) 231–237 Nesládek et al./ Structural Integrity Procedia 00 (2019) 000 – 000

232

2

components should perform well under cyclic thermal conditions, whereas mechanical testing at elevated cyclic temperatures presents a number of technical challenges. The mechanism of thermo-mechanical fatigue (TMF) combines together three basic interacting phenomena: (i) mechanical fatigue, (ii) creep and (iii) oxidation due to environmental conditions. Furthermore, generally recognized effects in mechanical fatigue are the notch, mean stress and phase shift effects, surface quality and many others. In the TMF problem, these basic effects are combined in some way together with the mentioned environmental influence and material viscoplasticity. The key effort in TMF investigation is correct identification of the most pronounced effects and application of a model that well describes them with respect to their mutual contribution to total material damage. Although it may be considered as controversial, the additive law of damage is usually assumed to hold among the mechanical, creep and environmental damages (Sehitoglu and Boismier, 1990) due to its simplicity. Among the approaches utilized in constructing the damage models suitable for low-cycle fatigue, the energetic concept is often applied. One of the early works is by Morrow (1964) who proposed an energy based criterion applicable to isothermal cases of fatigue loading. Ostergren (1976) developed an energy-based approach considering a portion of the dissipated strain energy above a threshold tensile stress to predict fatigue damage under various load regimes with stress/strain rate effects. Skelton (1991) observed that the strain energy at saturation cumulated for all cycles until failure may be considered as constant. The energy expended per cycle may be applied to correlate the fatigue life also in high temperature conditions. Energy-based methods can potentially be powerful tools since the dissipated strain energy quantifies the entire stress-strain cycle response and connection to the calibration data from basic tests appears to be straightforward. However, considering the basic TMF in-phase and out-of-phase strain controlled cycles, for instance, strain energy by itself would not distinguish these two cases clearly even if it is often reported that the damage induced may differ substantially (Neu and Sehitoglu, 1989). From this perspective, correct assessment of the stress/strain-temperature phase shift and the mean stress effects is still an open issue and is subject of the work by Ostergren (1976), Neu and Sehitoglu (1989) and Delprete and Sesana (2019), for instance. Recently, Nagode et al. (2009-2014) have published a numerical procedure that handles the thermo-mechanical material response continuously considering all history time points no matter if the individual input signal is reversed in a given time instant or not. The method treats the mean stress effect by evaluating a damage parameter in a procedure based on rainflow that associates the actual mean and amplitude of stress or strain with each load history time point. Continuous damage parameter evaluation, which is the Smith-Watson-Topper - SWT (Smith et al., 1970) in this case, may consequently be performed. Hysteresis operator approach is used subsequently to evaluate the mechanical damage from the Manson-Coffin curves parametrized by temperature. The method is named as Damage Operator Approach (DOA) and it is utilized in the work discussed in this paper. TMF tests that combine cyclic mechanical and thermal conditions are needed to prove quality of prediction by a selected model. However, several technical challenges are associated with them if compared to isothermal tests. To achieve a predefined temperature and mechanical load phase shift, a feedback control by both the thermocouple and the extensometer has to be ensured. Thermal expansion has to be continuously controlled and separated from the total mechanical strain provided by the signal from extensometer. The applied heating has to ensure homogeneous temperature field within the specimen gauge length not only in axial, but also in radial direction to prevent spatial deviations in thermo-mechanical response and additional un-controlled loading due to constrained thermal expansion. Therefore, thin-walled tubular specimens appear to be a proper geometry for this kind of tests. TMF tests are subject of the papers by Hyde et al. (2012), Minichmayr et al. (2008) and Okazaki and Sakaguchi (2008), for instance. The reasons for practical application of a TMF approach addressed in this paper emerged from the increasing demands for ST flexibility that alters design strategies applied until now. In the actual situation, non-stationary mechanical and thermal load cycles applied to ST components have to be assumed if their lifetime is to be assessed. In this regard, fatigue assessment techniques applied until now have to be reconsidered or new approaches have to be proposed to ensure that the mechanisms contributing to material damage are properly handled. Joint effort of the authors was to propose a methodology for TMF assessment applicable to ST shafts that would work well under various combinations of mechanical and thermal loading. To prove this capability, extensive TMF experimental campaign had to be performed. TMF experimental techniques applied to a CrMo rotor steel and the results that have been achieved until now are introduced by this paper.