Are Events with Different Significance correlated? : Quantitative Analysis of the Correlation between Events with Different Safety Significance

摘要

Events from operating experience are collected in order to serve as source of lessons to be learned. The underlining assumption is that there is causality between events and risk. While this is obvious when accidents happen it is less clear how precursor and other less risk relevant events are connected to causes of accidents. Confirmation of the causation assumption is important to optimize operating experience feedback programs and to better maintain safety. This paper is analyzing correlation between the rates and numbers of events with different safety significance. Events rates from two nuclear energy related databases are considered. Preliminary findings are that correlation exists but not as large as expected. Future work is to restrict correlation analysis only between events which have similar causes and investigate potential to use results to improve operating experience program.