Remaining life prediction methods using operating data and knowledge on mechanisms

It is important to be able to predict the remaining life of components and structures in a power plant, both for nuclear and fossil units. The information needed can be obtained from controlled laboratory experiments and the plant operating data. On materials degradation, large amounts of data from both sources are available. However, it is essential to formulate the best methodology to utilize this information so that our needs can be met. In this paper, the methods currently used for remaining life prediction are discussed with typical results. For components with multi-subcomponents such as globe valves, it is not possible to predict the remaining life using operating data, even though useful information related to aging characteristics can be obtained. On the other hand, operating data can be used to predict the remaining life of a steam generator (SG), since the degradation of the tubes is the life limiting factor for a SG. It is shown in this paper that Weibull statistics is a useful tool in predicting the remaining life of a component utilizing the operating data. Mechanistic approaches are also necessary in formulating the working hypothesis, which can be validated by experiments and in some cases by utilizing the operating data.