Use of PSA tools and techniques in life cycle management of maintenance and operation of complex system

This paper briefly discusses the uses of probabilistic safety analysis for life cycle optimization of maintenance and operation. The focus is on developing a time-dependent model for conditional failure intensity of component with incorporated effects of maintenance (major overhauls) that could be used in fault tree evaluations simulating various time points in the past or future. Likelihood functions and Bayesian inference process for estimating model´s parameters based on operational history records are described for normally operating and periodically tested standby components. Finally, examples of Bayesian inference and modulation of reliability predictions are provided for each of the two types of components based on assumed operational history records in order to demonstrate that the model produces meaningful results.