The role of artificial intelligence in fault-tolerant process-control systems

Abstract

This paper discusses potential applications of Artificial Intelligence in real-time process control systems from the perspective of fault tolerance. We emphasize two areas of fault tolerance, namely, fault recovery and reconfiguration. For fault recovery, we present telescopic replication technique using a frame based knowledge representation method for facilitating real-time recovery. The impact of knowledge representation on fault recovery is addressed. For reconfiguration, we discuss how to generate contingency plans to cope with unexpected situations via planning and learning by using a lattice structure. The application of these techniques to a chemical batch process-control system is shown.