Reconstruction of the material probabilistic S-N relations with given confidence

Because of the huge cost involved in data acquisition, probabilistic S-N relations should be given on a wide applicable sense to perform the reliability analysis at arbitrary survival probability-confidence (P-C) level. However, some existent fatigue databases give the material probabilistic S-N relations only with several fixed P-C levels. To realize the reliability analysis on a sense of arbitrary P-C level, a Monte Carlo simulation method is presented for reconstruction of the relations. Test data are re-gotten by a method under the new simulation policy that matching production practice according to original statistical parameters. Details are given with respect to the possible cases of known conditions. The relations are finally determined by maximum likelihood estimation on a general model to realize the analysis at arbitrary P-C level. Reconstruction of the relations for 60Si2Mn spring steel has indicated the availability and feasibility of present method.