Operator aids and expert systems in user computer interfaces

Recent events have demonstrated the potential for catastrophic accidents at process control facilities, resulting in severe economic damages or loss of human life. Human operators play a central role in the outcome of accidents in process control plants, because of their responsibility to make decisions regarding the appropriate corrective actions needed to control the event. In recent years, researchers have been attempting to apply Artificial Intelligence (AI) methods for developing computer-based decision aids for process control operators. Much of this research activity has taken place within the nuclear industry. In order to assess the implications of expert systems for nuclear reactor operators, the United States Nuclear Regulatory Commission (USNRC) has sponsored a research program at the Idaho National Engineering Laboratory (INEL). Included in this program have been the development of a prototype expert system for nuclear reactor operators, as well as two experiments to measure the effects of the expert system on operator performance in simulated accident conditions. This paper briefly summarizes the experience gained during this research program and assesses the potential future of expert system decision aids for process control operators