Procedural guidance levels in computerized plant operation

The computerization of the operation of nuclear plants tends, in some cases, to put emphasis on some problems that existed already in conventional hardwired panel-based control rooms. Among these, the search for an optimal balance between operator initiative and the prescriptive guidance provided by the operating procedures has always been a subject of concern, which is coming to the foreground with the computerization of these operating procedures.In this case, the main concern in to balance the excessive rigidity of the computerized tool that often reflects only theoretical prescriptions, by providing for some human adaptation to the actual operating conditions and to the actual operator activity. On the other hand, a minimum level of operator guidance must be maintained to avoid operating errors and missing of important constraints which may in some cases not appear clearly in the logical unfolding of the task. Thus the problem is to search for computerized guiding means realizing a good compromise between total computer responsibility (task automation) and unbounded operator initiative.The Reactor Operation Assistance Laboratory (LAER) of the CEA has launched a reflexion about this problem related to the share of decisional responsibility between the operators and the computerized operation systems. This paper resumes the basic aspects of procedural guidance, insisting upon the distinction between the strategical level (tasks structuration accounting for the plant physical constraints) and the tactical level (task carrying out respecting local constraints on plant systems).The presentation focuses on the tactical aspects, for which several guidance alternatives, (following automation level scales proposed in the literature), are analyzed on the basis of a generic system start-up task illustration. A computer prototype has been realized to put in a concrete form these various alternatives.This paper concludes with a discussion about the potential benefit of extended comparative tests on a representative testing bench (author) (ml)