Abnormal Events Management in Complex Process Plants: Challenges and Opportunities in Intelligent Supervisory Control

One of the most important challenges facing control system engineers is the design and implementation of next-generation control systems that can assist operators in making supervisory control decisions such as in abnormal events management (AEM), start up and shut down, controller performance assessment and so on. Operator failure to exercise the appropriate supervisory control decisions often have an adverse effect on product quality, process safety, occupational health and environmental impact. The economic impact of such abnormal situations is enormous; an estimated $20 billion/year in losses in the petrochemical industries alone in the US. Furthermore, process safety, occupational health and environmental hazards are ever increasing in importance in response to heightening public concern and the resultant tightening of regulations. Thus, there exist considerable incentives in developing intelligent control systems that can provide automated operator assistance for supervisory control situations for complex process plants. People in the process industries view this as the next major challenge in control systems research, design and application. Since fault detection and diagnosis is an important first step in AEM, I start with an overview of the various approaches to fault diagnosis, before discussing the challenges and the encouraging emerging trends. Recent progress in this area has promising implications on the use of intelligent systems for inherently safer design, operator training, abnormal events management, and process hazards analysis.