Reliability Analysis for Inspections of CANDU Components and Systems

Abstract

Condition assessment of complex systems requires more than inspection data for the individual elements of the system. At the component level, knowledge of the probability of detection (POD) and sizing accuracy for defects is needed to properly interpret results. At the system level, probabilistic and deterministic methods guide safety assessments and life management strategies. The field of reliability analysis provides a framework for quantitative application of inspection results to both levels. Quantitative knowledge of inspection reliability can reduce the requirement to use conservative values for every quantity. A sound analysis of inspection reliability can produce a more targeted inspection scope, leading to fewer but more effective inspections and, consequently, lower radiation dose. Restart cases can be based on correct detection probability distributions rather than worst-case scenarios, and inspection intervals and expected life estimates can be optimized based on measured sizing uncertainties. This paper illustrates applications of inspection reliability analysis to optimize the inspect-repair-operate cycle for major CANDU primary heat transport elements, such as steam generator tubing, feeder piping, and fuel channels.