In 2014, the Integrated Life Cycle Management (ILCM) computer code was developed through collaboration between the Electric Power Research Institute (EPRI) and Électricité de France (EDF) to provide a standard methodology to support effective decision making for the long-term management of selected station assets. In order for the ILCM program to become a standard tool in the industry, additional work was needed in the development of a Likelihood of Replacement (LoR) calculator. The LoR calculator estimates the likelihood that a component will have to be replaced due to failure or reasons other than failure, such as high maintenance cost, inability to maintain, obsolescence, and other similar reasons. Expert elicitation was chosen as the method of gathering data and opinions on component replacement probabilities.
The majority of expert elicitation techniques consist of experts giving opinions on the probability of replacement at various points in time, from which a reliability curve can be calculated. Furthermore, any failure or replacement data is subjectively incorporated in to the expert’s opinion.
The present work uses Bayesian analysis to provide an objective method for statistically combining expert opinion with failure and replacement data. This paper also describes the process of extracting a Weibull LoR curve from expert’s opinions and reported failures and replacements. The expert’s work history and answer confidence is used to assign uncertainty in their answers and calculate 5th, 50th, and 95th percentile credibility Weibull curves for the probability of replacement.
