FORM SENSITIVITIES TO CORRELATION: APPLICATION TO FATIGUE CRACK PROPAGATION BASED ON VIRKLER DATA

Abstract eng:
Sensitivities are important information in a structural reliability analysis. They help in quantifying the role of various input parameters in the failure probability and make therefore the basis of any sound structural reliability analysis. This paper addresses the sensitivities to correlation between random variables, which are rarely studied in the structural reliability field except in a very few references, see e.g. reference [1]. The methodology for deriving these sensitivities is detailed in the first part of this paper, based on accurate solutions obtained by numerical integration of Nataf model-derived equations. These quantities are obtained as standard by-products of the FORM analysis, without extra calls to the structural model. The methodology has been validated over some reference examples. It is then applied to stochastic crack growth fatigue in the second part of the paper, for which a high sensitivity of the failure probability to correlation between parameters of the Paris-Erdogan equation has been observed. The examples investigated in this second part illustrate the strong role correlation could have on reliability results and they also points out the important effects which could result from neglecting some slight bias between the structural model and reference experimental data.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket