PSI - Issue 19

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

235

5

3. Results and discussion

Numerical values of the specimen lifetimes obtained in the individual tests are summarized in Tab. 1. From the plot in Fig. 4 it may be easily observed that the specimens tested under anisothermal conditions exhibit in most of the cases significantly lower lifetimes then the specimens subjected to isothermal conditions. Only the in-phase “cold - start” tests are between the 20 °C and 200 °C isothermal LCF curves. Moreover, the in-phase test conditions appear to be less damaging than the out-of-phase loading for this type of material. To verify quality of predictions provided by the DOA and to check if the connection of the selected material constitutive model along with this TMF damage model is capable of providing good lifetime predictions, numerical simulation of the material thermo-mechanical response due to the experimental conditions has been performed. Stress-strain response of the specimens subjected to the thermo-mechanical loading conditions was simulated by the finite element method (FEM) and compared to the response acquired experimentally. Plots in Fig. 5 show the experimentally and numerically obtained mechanical response representing the “cold - start” test No. 1 and No. 2 as was denoted in Tab. 1. The Chaboche elastic-plastic non-linear kinematic hardening model was initially identified by using the CSSCs measured by the LCF isothermal tests. By optimizing the model parameters, a good match between experiment and simulation may be achieved as is shown in Fig. 5. Comparison of prediction by the DOA and measured specimen lifetimes is in Fig. 6. Isothermal and anisothermal raw data is used, showing maximum scatter bands around the ideal match represented by the 1st quadrant axis. As may be observed from the diagram, the maximum ratio / =̇ 5 . In the case of anisothermal test conditions, the DOA provides non-conservative predictions for all experiments. A more significant prediction error is associated with the “hot - start” tests (450-600 °C temperature range). In other words, DOA calibrated by these isothermal data set and applied to predicting the TMF damage due to the stress-strain response simulated numerically by employing the Chaboche elastic-plastic material model provides better lifetime predictions at higher temperature ranges. From the obtained results it was also observed that the method can well distinguish different damaging effect of the in phase and out-of-phase loading.

Fig. 4. Overall results of the isothermal and anisothermal fatigue tests.