PSI - Issue 39

Matteo Benedetti et al. / Procedia Structural Integrity 39 (2022) 65–70 Author name / Structural Integrity Procedia 00 (2019) 000–000

69 5

It is now clear that the fatigue damage mechanisms that prevail in plain and notched components are different. This could hinder the applicability of the theory of critical distances, which postulates that the critical distance L can be inferred from a plain and a notched or cracked specimen configuration. Specifically, we term "plain&threshold" the approach based on the determination of L starting from plain fatigue strength and crack threshold. With "plain¬ched" we denote the inverse search method recently derived by us in this paper and based on the use of a plain sample and a notched samples with optimized shape to minimize the sensitivity to experimental uncertainty. To overcome the above discussed shortcoming, we explore in this work the possibility of inferring L from two notched geometries with different notch severity. The fatigue characteristics determined in this way are representative of the fatigue damage mechanism that rules the neighborhood of the tip of the notch. In this way, it is possible to deduce, along with the critical length L * , an intrinsic plain fatigue strength ∆ ∗ , which ideally represents the fatigue strength measured using miniaturized plain samples extracted from the notch tip. The mathematical formalism can be found in Benedetti et al. (2021).

Table 1. Results of the critical distance inversion methods. Plain&treshold Specimen geometry (a)&(f) Plain&sharp Specimen geometry (a)&(b)

Blunt&sharp Specimen geometry (c)&(b)

∆

/2 ( ) 158

∆ /2 ( ) 158

∆ ∗ /2 ( ) 269

L th (mm)

L (mm)

* (mm)

L

0.625 0.136 Table 1 lists the plain fatigue strength and the critical distance evaluated according to these three possible approaches. To apply the first approach "plain&threshold", the crack growth threshold was determined from fatigue crack growth experiments performed on a M(T) specimen. It can be noted that the two approaches based on the plain fatigue limit predict a very large value of the critical distance, above 0.5 mm, well above the values typically reported in the literature for structural metallic materials. On the contrary, the approach "Blunt&Sharp" based on two notched specimen geometries leads to an estimate of a much shorter critical distance L * , approximately 0.14 mm, and an intrinsic plain fatigue strength ∆ ∗ /2, (about 270 MPa) significantly higher than that displayed by the plain samples. Table 2. Prediction of the notch fatigue strength and crack growth threshold of inde- pendent geometries not used in the calibration of the critical distance method. 0.530

Specimen geometry Notch 60° R0.2 (b) Notch 60° R1 (c) Notch 90° d14 (d) Notch 90° d6.5 (e)

Exp. (MPa/ MPam 0.5 )

Plain&treshold

Plain&sharp

Blunt&sharp

Pred.

Err. (%)

Pred.

Err. (%)

Pred.

Err. (%)

94.6

102

7.5

-

-

-

-

118

105

-11

99.3

-16

-

-

112

104

-7.2

97.5

-13

105

-6.4

123

143

16

133

8.4

130

6.1

M(T) (f)

14.0

-

-

12.9

-7.9

11.1

-20

The predictions listed in Table 2 of the fatigue strength of independent notched variants permit assessing the suitability of these three critical distance approaches to predict the notch fatigue resistance of DCI. A systematic comparison is possible only for variants (d) and (e), which are used in neither of these approaches. Interestingly, the "Blunt&Sharp" approach is the only one able to keep the absolute relative error well below 10%. Conversely, the "Plain&Threshold" and "Plain&Sharp" predictions are affected by larger errors, since these latter two approaches are influenced by fatigue damage mechanisms occurring in plain and M(T) specimens that are scarcely representative of those taking place in notched coupons.