13th International Conference on Fracture June 16-21, 2013, Beijing, China measured values of Qfor each specimen were whin 15% with respect to the mean value. Thanks to the precise experiment set-up, the dissipated energy per cycle Em d1 are with 5% of the mean value in these fatigue tests. The constant stress amplitude fatigue tests (12 stress levels, 30 specimens in Tab.2) were performed to obtain the S-N curve by the traditional procedure (Fig.7). Meanwhile, Fig.8 shows that the relationship between the dissipated energy Em d1 and fatigue lifetime Nf is very similar to the S-N curve. This curve was defined as Em d1 -N curve in this work. The scatter bands of the two curves are given for 10% and 90% survival probability. The linear fitting equation of Em d1 -N curve is written as, log(Em d1 )=−0.77log(Nf)+8.33 (4) where the related coefficient R2 Em d1−N between Em d1 and Nf equals to 0.94. The linear fitting equation of S-N curve is written as, log(σa)=−0.13log(Nf)+2.85 (5) where the related coefficient R2 S−N between σa and Nf equals to 0.89. 10,000 100,000 1,000,000 1E7 80 120 160 200 240 280 Constant fatigue test A (MPa) A (MPa) N f (cycle) stop without failure max R2=0.89 log( a )=-0.13log(Nf )+2.85 200 300 400 500 600 700 a Figure 7: S-N curve by traditional staircase procedure of 316L material (Rσ=0.2, f=14Hz). As R2 Em d1−N>R2 S−N , Em d1 -N curve by dissipated energy measurements shows more accurately than S-N curve by traditional fatigue test method. In such a case, it’s possible to extrapolate the Em d1 -N curve by dissipated energy measurements to predict the residual fatigue lifetime during various stress amplitude fatigue test under traction-traction cyclic loadings (i.e. Rσ=0.2). 8
RkJQdWJsaXNoZXIy MjM0NDE=