145 Discussion on Modeling Errors of Commision and Omission (PSAM-0296)

The aim of the paper is to discuss errors of commission and omission from the point of view of human reliability assessment (HRA) and from the point of view of the operator. The paper covers the use of power plant simulators to determine circumstances under which Errors of Commission (EOCs) and Errors of Omission (EOOs) can occur. In order to successfully use simulators for these purposes a necessary requirement is to have a omprehensive data collection and analysis system. The paper makes reference to the number of simulator data collection studies that have been completed over the years, but there still is no universal acceptance of the need by the Training community. They do not have any appreciation of the real need for such systems to support training and lead to better trained operators and safer and more economically run power plants.

The paper shows that EOCs and EOOs results from the same basic issue the effect of context on the crews' performance. The context under which the crews' operate derives from their training, operating procedures, human-system interface (HIS), team and management leadership and workload. Results from simulator experimental studies indicate that crews make errors attributable to the operating context. However, many of the errors are corrected by the operators during the course of accidents. Some errors are not corrected but have no impact on the accident progression. Other errors do lead to correct actions not being taken. Yet others lead to a worsening of the consequences of the accident. For well-trained crews these later circumstances are rare, but insights from simulator studies can guide the HRA analysts to estimate both the probabilities and consequences of these states. The paper introduces one possible approach called the Holistic Decision Tree method to relate context to error probability and shows how analysis of simulator data can be used to assess the effects of context for particular accident scenarios and hence arrive at an estimate of whether the act is either an EOC or an EOO and the associated human error probability. The method can be used to indicate what steps organizations can take to reduce human error probability by effecting change in the context.