HRA, Human Reliability Analysis: the HRA project examines the reliability of human actions, taking account of various boundary conditions. The state of the art is the determination of failure probabilities for actions that are required during an incident, but are not performed. On the other hand, there is no efficient method for the systematic identification and quantification of unplanned actions which negatively impact the progression of an incident – known as EOC (Errors of Commission). Work was undertaken on the following points in 2009:

  • Development of a method for the systematic identification and quantification of EOC, and application of this method to a Swiss nuclear power plant: the CESA (Commission Errors Search and Assessment) method for identifying and evaluating EOC that was developed during project phase HRA-II was reviewed in the light of international experience, and was refined where necessary. For this purpose, use was also made of simulator data obtained under the auspices of the OECD Halden Reactor Project. One of the aspects examined was whether the known HRA methods correctly predict and evaluate the relevant factors of influence on the reliability of operators’ actions. The results show that the CESA method provides good predictions of these factors. The CESA method was also tested for a second Swiss nuclear power plant in 2009. Use of the method made it possible to identify and quantify a total of six actions with “EOC potential”. According to this data, a number of these actions are of comparable importance to some of the actions that were already taken into account in the plant-specific probabilistic safety analysis. Nevertheless, the influence of all identified EOCs on core damage frequency is slight. The analysis of these EOC cases did not uncover any reasons to improve the incident regulations.
  • Development of dynamic models that map the interactions between people and plant in order to support HRA: the objective is to support HRA with a dynamic model for plant and staff behaviour. This model enables the presentation of the time progression for plant parameters in interaction with the operators’ actions, which should – in particular – improve the modelling bases for operators’ diagnostic processes. In 2009, the work focused on simulating incident scenarios and evaluating them for HRA purposes. Another important aspect was the development of a method for the systematic evaluation of the simulations.