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.
Random - weighted Monte Carlo Simulation Method for Structural Reliability Analysis
