Statistical Estimation of <i>S-N</i> Curves for High-Strength Steels Based on Their Static Mechanical Properties

Abstract

In this study, a statistical estimation of S-N curves for high-strength steels from their static mechanical properties was attempted. First, the fatigue data of high-strength steels were extracted from the "Database on fatigue strength of metallic materials" published by the Society of Materials Science, Japan (JSMS), and the S-N curves of respective steels were determined from the JSMS standard, that is, the "Standard evaluation method of fatigue reliability for metallic materials: Standard regression method of S-N curve". The correlations between the regression parameters obtained here and the static mechanical properties were investigated. Thus, significant correlations were found between the regression parameters and static mechanical properties (e.g., fatigue limit E and tensile strength s B). Based on these correlations, the S-N curves of high-strength steels were successfully predicted from their static mechanical properties, similar to structural carbon steels and aluminum alloys in previous studies. Moreover, from the distribution of the fatigue limit E, the percentile points for the estimated S-N curve were predicted. It was finally confirmed that 84 % of the S-N data series fall within the estimated interval of ±2s, where s represents the standard deviation of the fatigue limit.