Estimation of fatigue S-N curves of welded joints using advanced probabilistic approach
This paper provides a new advanced probabilistic approach for reliable estimation of the fatigue characteristic S-N curves of welded joints both for constant amplitude (CA) and variable amplitude (VA) loading conditions. The presented approach, which is referred to as the ML-MCS approach, combines Maximum Likelihood method (ML) and Monte-Carlo Simulations (MCS) method to estimate true p-quantiles of CA and VA S-N curves by using complete experimental data-sets. The ML-MCS approach includes a linearization method for use of S-N curves in combination with linear damage accumulation rule as well as for direct comparison with current standards. Application of the ML-MCS approach on two study cases and comparison with current standards shows that the use of the ML-MCS approach may have a significant impact in re-definition of CA and VA S-N curves of current standards and in particular of the CAFL, of the S-N curve second slope and of the critical value of accumulated damage at failure. The last section of the paper provides accurate guidelines for future experimental tests needed for re-definition of current standards. (C) 2016 Elsevier Ltd. All rights reserved.
Manuscript_revised-accepted.pdf
openaccess
716.9 KB
Adobe PDF
9e8d80a9c3dbd94c0666db16d84eecfa