This paper presents a methodology for anticipating failures in a component up to the end of its life cycle. Often, feedback data is not sufficient and must be complemented by the analysis of expert judgment. The methodology developed aims at anticipating the degradation mechanisms responsible for aging, and evaluating their relevance and related uncertainties. This is necessary information for risk analysis related to the operating of a component up to the end of its life cycle. Lastly, the methodology is applied to a nuclear component.