Simultaneous uncertainty quantification of fracture mechanics based life prediction model parameters

A Bayesian technique is developed for simultaneous estimation of the equivalent initial flaw size (EIFS) and crack growth rate distributions in fracture mechanics based life prediction models from inspection data. The inference method separately quantifies the different sources of variability to obtain a data set independent EIFS distribution without the requirement for prior knowledge of the uncertainties in other parameters. The ability to simultaneously and accurately predict the EIFS distribution and crack growth rate noise from a small, highly-scattered data set is demonstrated in a numerical example.