超長寿命疲労試験データの評価法に関する研究

抄録

In this paper, a statistical method for the evaluation of the data related to the super-long life fatigue region is proposed. According to a research paper published by the “Research Group on the Statistical Aspects of Material Strength”, failures are classified into two modes. That is, the failures due to cracks originated on the surface and those due to crack resulting from internal material defects. Therefore, to evaluate the fatigue properties in this region, the experimental data should be separated according to these two modes.
To decide the optimum separation between these two modes, we make a first separation based on the number of cycles. Then we draw the two fatigue strength plots on probability papers, and we calculate the two coefficients of correlation. The separation line is shifted and the previous procedure is repeated. The optimum separation is the one for which the sum of these coefficients reaches its maximum.
Using this proposed method, a computational system to decide the S-N curve to be used in the design has been developed. To validate the proposed method, we applied the computational system to the fatigue test data of different materials. The agreement of this separation with that of obtained by SEM observation allow us to judge positively our method and to confirm the accuracy of the obtained S-N curves.