This report contains a specification and implementation description for the statistical tool (desktop software program) which will allow detailed reporting and flexible ranking of operational events. The program is named Operational Events Ranking Tool (OPERATE). The initiating idea was that detail statistical reporting with a robust and consistent way of selecting most important events for more detailed investigation is very important. This provides better insight and saves the need for complete detail investigation which is neither feasible nor useful. It allows also to identify events trends and patterns across large databases.The foundation for the statistical reporting and ranking is defined by the events parameters which are the base for reporting, grouping and ranking. This requires that each event should have characterisation which allows use of selected parameters for grouping and ranking. The set of parameters could be in general user-defined, but for initial OPERATE implementation a fixed set of parameters is defined based on the WANO operating experience programme and the IRS coding schemes. Five parameters are used for events grouping: Activity, Direct cause, Systems, Components, and Root cause & Causal factor. Seven parameters are used for ranking: Frequency, Trend, Extension, Multiple, Safety, Category and Consequences. Some other parameters are used for statistics reports and future criteria (e.g. Vendor, Reactor type, Time of event, Status and Group of staff).An important part of the implemented method is the way how ranking indexes are calculated. This is done in a consistent way based on previous experience and new research. OPERATE allows for each ranking index to be valued on the basis of relative importance determined by the experts' judgement evaluated with analytical hierarchy process (AHP). In this way weighting is easily determined and the importance of different parameters could be consistent. The tool is capable of analysing ranking for all groupings based on the expert and AHP determined relative importance. In order to investigate the ranking confidence (uncertainty and sensitivity), the tool can analyse ranking for user-selected variations of parameters importance (e.g. ±30% weighting for all ranking indexes etc.). This is useful to determine the sensitivity of the results, which could expose the criticality of certain ranking parameters and respected weights. In addition to ranking results OPERATE is able to produce nineteen different reports about events (e.g. frequencies for all parameters values on yearly and average base, etc.) which could help further focus on experts investigation. In parallel to the program development five years of US NRC LERs events (2007-2011) has been characterised in order to demonstrate and verify the tool. These results are presented in the report. OPERATE is a very flexible tool with potential for further use and development.
展开▼
