13th International Conference on Fracture June 16–21, 2013, Beijing, China -7- propagating trans-granular. This has been confirmed in an light-optical investigation of a cross section. For the samples tested under creep-fatigue loading (see Fig. 7) the crack behaviour was found to be independent on the holding time similarly like specimens tested under creep loading (Fig. 6 a, b). Around the crack tip an area of high cavity density was identified and the crack path under creep-fatigue loading is comparable with the crack path under pure creep condition. Also along the crack path creep cavities were observed. Figure 7: Specimen after CFCG-test at 580°C with a), b) HT=6 min ( b) detail view from a)), c), d) HT=60 min (d) detail view from c)), at 600°C with e), f) HT=6 min (f) detail view from e)), g), h) HT=60 min (h) detail view from g)) 3.3. Creep-fatigue interactions The crack propagation rate da / dN under creep-fatigue loading is influenced by two failure mechanisms, and therefore depends on the load level, the mean stress, the stress ratio R and the temperature. Generally, there are three different areas [10]. At high frequencies, the crack growth per cycle is independent of frequency, because the cycle duration is so short that no time remains for creep crack growth. With decreasing frequency, the crack growth per cycle da / dN is greater. The time between two load cycles is then sufficiently large, so that fatigue crack growth and creep crack growth can overlap. With further decreasing frequency of the fatigue crack propagation loses its meaning. The crack propagation is now determined only by creep crack growth. In literature [11, 12] different relationships exist to describe the crack growth under creep-fatigue loading, for example the following equation recommended for creep-ductile materials (see also [4]): C ( K) C ((C ) ) dt dN da q h t 0 t avg 1 n0 0 = ∆ + (9) According to [11] the final equation can be: h *m 2 1 m h 2m 1 n0 0C ( K) CK t C C t dN da + = ∆ + − (10) with th as holding time. The first term is a pure-fatigue contribution reflecting no effect of hold time and corresponding to crack-growth behaviour at short hold times and high frequencies. This term can be evaluated if the Paris-law coefficients are available. For further evaluation, the coefficients from Table 2 will be used. The second term shows a nonlinear power-law dependence of e) g) h) a) Crack tip Crack tip d) b) f) c) Crack tip Crack tip Crack propagation direction
RkJQdWJsaXNoZXIy MjM0NDE=