Stochastic Updating of Probabilistic Life Models for Rotorcraft Dynamic Components

Probabilistic life models for rotary wing structures enable the development of maintenance programs that limit risk to acceptable levels. However, uncertainty in distributions of these models' parameters leads to overly conservative predictions of structural component lifetimes. Probabilistic modeling uncertainty can be reduced by updating distributions in the probabilistic model with information contained in maintenance data. The additional information permits more accurate statements on remaining life of structural components to be made, allowing condition-based maintenance and forecasting. A framework is developed to update probabilistic rotorcraft structural life models with maintenance data. The hierarchical Bayesian approach is adopted where the current parameter distributions serve as prior distributions, and maintenance data are included through a likelihood function. Updated distributions accounting for the inspection data are obtained using Bayes' rule. An example analysis of maintenance data is presented, where a probabilistic safe-life fatigue model is updated with crack detection and corrosion inspection findings.