PSI - Issue 2_B

A. Giertler et al. / Procedia Structural Integrity 2 (2016) 1207–1212 Author name / Structural Integrity Procedia 00 (2016) 000–000

1210

4

The observed behavior is typical of smooth specimens made of highly tempered materials. It is reported that, if the annealing temperature is reduced, the frequency effect on smooth specimens decreases because the athermal part of the critical shear strength exceed the Peierls stress contribution [Takeuchi et al. (2008)]. To prove this, it is planned to increase the strength of the material by reducing the annealing temperature and afterwards the same experiments will be repeated. The given results show the importance of microstructure, local plastic deformation and testing conditions on the fatigue behavior. To understand the relationship between local plastic deformation and the macroscopic fatigue behavior, load increase tests have been carried out in the resonance fatigue testing machine, cf. Fig. 3 a. The load increase test starts with a stress amplitude of σ a = 360MPa. This value is far below the yield strength of the material. The stress amplitude is stepwise increased during the test by 6MPa every 100.000 cycles until fracture of the specimen. The surface of the fatigue specimens has been observed by a thermographic camera and every thousand cycles a thermogram has been stored. Additionally, the change in resonance frequency as an indicating value for fatigue damage is plotted, cf. Fig. 3a.

a)

b)

Figure 3: (a) Stepwise load increase test: the specimen response plotted as the change in temperature and resonance frequency; (b) Graphical determination of the fatigue limit

The first significant change in resonance frequency can be observed after N =2.4·10 6 cycles at a stress amplitude of 500MPa. The test was stopped until a change of frequency of Δ f =-0.1Hz has been reached, which correlates with a surface crack of more than 3mm in length. In comparison to that, the temperature measured by the thermographic camera provides a more sensitive signal out of the microstructure. The first response of the temperature up to N =100000 cycles can be attributed to a heating up process of the fatigue setup with the beginning of the test. Followed by a state of equilibrium, the continuous increase of stress amplitude leads to a negative change in specimen temperature. This behavior is due to the thermoelastic effect [Yang et al. (2004)], where a tension load applied to a metallic specimen causes a negative change in temperature. A continuous increase of specimen temperature can be observed from N =700000 cycles by local plastic deformation. The corresponding stress amplitude at N =700000 cycles was σ a =396MPa. This value corresponds well with the cyclic yield strength of σ cyc =400MPa obtained from the incremental step test. The strong increase in temperature beginning at 2.4·10 6 cycles is caused by fatigue crack propagation. A graphical approach to determine the fatigue limit is given in Figure 3b, by plotting the change in temperature vs. the corresponding stress amplitude. The range between 350MPa and 500MPa behaves almost linearly until above 500MPa the temperature increases exponentially. The linear temperature rise in the first region between 350MPa and 500MPa is dominated by local plastic deformation and the second range above 500MPa by fatigue crack propagation. Using a linear fit through each region gives an intersection of the two lines at a stress amplitude of σ a =483MPa. This value is in very good agreement with the fatigue strength of σ FL = 490MPa, which has been determined by tests under constant amplitude loading, cf. Fig 2a. An example of the resolution of the thermographic camera is given in Figure 4a. The label “C1“ corresponds to the crack initiation site. The respective temperature development vs. number of cycles at position C1 forms the data base for the diagram shown above in Fig. 3 a and b. In comparison, “C2“ marks an area in which no significant temperature