PSI - Issue 5

M. Dabiri et al. / Procedia Structural Integrity 5 (2017) 385–392 M. Dabir et al. / Structural Integrity Procedia 00 (2017) 000 – 000

391

7

strain curve. Using the modified version of the SED method resulted in almost the same life estimations by the elastic plastic finite element model. The linear rule, alternately, is only suggested when a simple method is desired for a quick solution in the absence of computational resources. As seen from Fig. 3, the FE and TCD yielded the most accurate estimations compared to the experimental values. In the case of FE and other approximation methods, it has been shown that the level of conservatism increases with sharper notches (Zeng & Fatemi, 2001). The same issue was seen in this study when comparing the estimated lives of two specimens with different radii. The TCD method could be considered the most advanced technique for the stress/strain analysis of notched components and fatigue life estimations. Although supported by the same philosophy of critical distance as noted by Peterson and Neuber, this approach is developed for a unified method originally inspired by Tanaka (Tanaka, 1983) and developed by Taylor (Taylor, 1999) (Susmel & Taylor, 2010) (Taylor, 2007) (Taylor, 2001) to account for both crack and notch analysis by showing their dependency on a material characteristic length. A few advantages of this technique (in point method form, as used in this study) are as follows:  The same numerical model used for elastic-plastic finite element analysis can be used without any further modifications, and the required results can be obtained through a post-processing stage.  The material characteristic length can be considered a material constant regardless of notch configurations and the number of cycles in a low cycle regime (although not supported by any theoretical grounds). This allows analysis of the specimens with different notch radii using the reference value obtained by the calibration through the test sample.  The high level of conservatism in life estimations, especially in sharp notches observed in other approximations, is minimized. All of the fatigue life predictions by TCD for the specimens with a sharper notch are fairly close to the experimental values and on the conservative side. Life estimations of the specimens with a blunter notch, however, were mainly located on the non-conservative side but fell within an error factor of 2, as shown by the dashed line in Fig. 3b. It can be concluded that the TCD method can be considered a promising technique in notch analysis under constant amplitude loading. The interested reader is referred to the sources provided in the reference list to verify the abilities of this method in the analysis of specimens experiencing variable amplitude (Susmel & Taylor, 2015) and multiaxial loading (Gates & Fatemi, 2016).

Fig. 3. Comparison of estimated and experimental lives for notch specimens (a) R = 0.5 mm and (b) = 1.5 mm.

Conclusions

The results of this study are highlighted as follow:  Methods such as Neuber’s rule, SED and the linear rule are just rough approximat ions, resulting in highly conservative estimations.