Decision support in supervisory control of high-risk industrial systems (Q1097207)

The authors propose an approach to computer support of supervisory control of large industrial processes of high risk level. After a description of the decision making process and the associated decision space, they study the allocation of decision making between the designer, the operator and the automatic control system, with appropriate feedback and communication facilities between the partners. The traditional hierarchy in the control staff, in which only the designers is in charge to preanalyse all abnormal situations and to store his advice in computers, while the operators are asked to strictly follow the operating instructions, can't assure the necessary reliance: for example in case of accident the operating staff has to cope with situations which have not been analysed by the designer. The proposed new line of development is based on interactive decision making activity: the designers, by means of advanced information technology, can make available their conceptual models and tools to the operators, rather then to continue the efforts to make the preplanning more and more complete. The computer performs the online analysis of the available measured data, and the operators will choose the priorities in goals, depending on their perception of policies and available resources. Moreover, support from an expert system is provided, when several physically possible solutions may be available. Some remarks are given about the user acceptance of these new technologies.