Assessment methodology for actions taken after events

The safety of nuclear installations is based on staggered safety levels. These safety levels comprise technical measures as well as organisational and administrative measures. For technical measures detailed regulations exist which command requirements like necessary redundancy levels, diversity and inservice inspections. The assessment of these various measures is performed in deterministic and probabilistic analyses. Compared to the technical measures there is no comparable systematic description of the measures foreseen by human actions. The safety significance of a specific human action within the staggered safety concept can be assessed only in detailed and time consuming analyses. There has been no fast and simple method to cover technical, organisational, and administrative measures in one assessment. Therefore, GRS has developed such a method on behalf of the German Ministry for Environment, Nature Conservation and Reactor Safety (BMU). This presentation is based on the findings of this research project 1. The method was developed and tested in co-operation with two German NPP. The method is therefore applicable for German NPP or NPP with similar design in power operation modes. There may be limits in the application for, shut-down states of operation and for NPP of different design. Main basis of the method developed are the International Nuclear Event Scale (INES) and probabilistic approaches. The INES has four categories in the aspect 'defense- in-depth', taking into account the level 0. The experience shows that these four categories are sufficient to communicate the safety significance of an event to the public. INES has the goal to assess quickly the safety significance of an event in nuclear installations. But, the human factors area is not covered in detail. In the method developed the human factor aspects are considered taking into account PSA approaches. The barriers based on administrative and organisational measures are assessed due to their effectiveness which is estimated on data used in PSA. During the development of the method it was taken care the method was compatible to the INES aspect 'defence-in-depth'. This should avoid difficulties for the acceptance of the method, because INES is well established. It was shown that the classification of the safety significance into 4 classes was sufficient for practice use. The method developed by GRS is intended to: - assess the potential significance of a postulated human error (proactive analyses), - assess the actual safety significance of a specific human action within an event (reactive analysis), - assess the actions taken after events to prevent recurrence