Value of analyzing operating events

If one had no uncertainty regarding scenarios, frequencies, or consequences associated with a hazardous operation, then from a simplistic quantitative point of view, there would be nothing more to be learned by analyzing operating experience. However, this is not generally the case; several kinds of reasons can be offered for analyzing operating experience in the context of risk management. Even if the scenario set is complete, there may be uncertainty regarding the level of performance at a facility (whether its programs are suitably maintaining equipment reliability). Alternatively, even if performance is satisfactory, there may be risk contributors that are significant but not modelled, or modelled inadequately. Although some catastrophic events may occur with no operational precursors, in other cases, the potential for certain risk contributions may be revealed by analysis of experience. Methods for extracting insights of this kind from analysis of operating experience are still evolving, especially the problem of deciding which events to focus on. The paper will examine the usefulness of measures of "surprise" in this regard.