Information Prioritization for Diagnostic Assistance in Complex Industrial Processes

Industrial processing today shows a trend toward increasing complexity and automtion. This often implies for the human operator a role mainly as supervisor over a large set of monitoring instruments. A fundamental difficulty arises in emergency situations: the amount of information that the monitoring instruments may present to the operator can be overwhelming. There may be hundreds of plant variables competing for the operator's attention. As a result, the performance of the human operator can become degraded in emergency situations. Automatic prioritization of the information is needd to assist the operator. This paper discusses sophisticated approaches for accomplishing such prioritization.