Estimating learning effects on nuclear power plants: A count data model with a varying time interval as a duration variable

Abstract

This paper analyses learning effects in nuclear power plant operation, which are reflected on the reliability growth and operating cycle decision. To consider two aspects of learning, a count data model with a varying time interval as a duration variable is suggested. A possibly complicated model canbe estimatedrelatively easily by a Gibbs sampling technique. The estimates are then used to investigate the effects of three types of operating experiences on the plant operation.