Method for analysis of operating events occuring in French nuclear power plants

All the serious industrial accidents which have occurred over the last few years have brought to the forefront the significant involvement of human factors. If we consider only our own specialized area, the revealing experience of Three Mile Island comes immediately to mind. Here human participation in the event was fully illustrated, justifying the attention which was subsequently to be focussed on this aspect. TMI set things moving on an international scale. Then came Tchernobyl, demonstrating that improvements in the human factor field can never be considered as definitive, but require our constant efforts and cooperation. At the present time, human factor specialists in the nuclear field have at their disposal two main categories of data: . those derived from inquiries and event analyses in real nuclear installations . those obtained on simulators, based on observation. These categories have notable specific differences: . data from real plants are obviously much wider in scope, more diversified and realistic, since they cover all the equipment, activities and specialized branches which are encountered in a power plant and which combine to create hazard. . the simulator is closer to the laboratory. The data collected are probably easier to quantify, since they result from reproducible experiments, based on very well defined postulates. The ability to study serious operating events is, of course, a vital advantage. However, it must be borne in mind that only situations in a control room are analyzed and that these are only a fraction of those contributing to the total risk incurred. Finally, it is perfectly obvious that a simulated situation is not real and probably modifies human behaviour. We feel that it would be judicious to have a corrosponding corrective factor integrated where necessary in the numerous simulator analyses presently available. In the present paper , I shall discuss data derived from operating feedback, that is to say from real events. (author)