The prediction of the time-dependent failure rate has been studied, taking into account the operational history of a component used in applications such as system modeling in a probabilistic safety analysis in order to evaluate the impact of equipment aging and maintenance strategies on the risk measures considered. We have selected a time-dependent model for the failure rate which is based on the Weibull distribution and the principles of proportional age reduction by equipment overhauls. Estimation of the parameters that determine the failure rate is considered, including the definition of the operational history model and likelihood function for the Bayesian analysis of parameters for normally operating repairable components. The operational history is provided as a time axis with defined times of overhauls and failures. An example for demonstration is described with prediction of the future behavior for seven different operational histories.

No abstract

No abstract

Nuclear power plants produce electricity with relatively low operational costs and low impacts on the environment, but are also associated with the fear of severe accidents. These are extremely rare, but once they occur, they put an immense pressure on the accident management team. In this paper, we report about an ongoing development of a decision support system called Severa, aimed at supporting the decision-making team during the course of an accident or a training exercise. The software is being developed in the context of the EU H2020 project NARSIS. The system assesses the plant damage state, predicts possible accident progressions and assesses available management actions and their consequences. Two already implemented modules are presented: a monitoring and radioactive-release assessment module. The methodological approach primarily relies on qualitative rule-based multi-criteria models, but also includes other techniques: data analysis, probabilistic safety assessment and eventtree modelling.

KT_kt-110039-19.8.09/stmm ediaköthen Z.Sˇimicetal.:P rediction of time dependents tandbyfailureratesforperiodicallyt ested components 74(2009) 5-67