An intelligent architecture for integrated control and asset management for industrial processes

Abnormal event management (AEM) in large manufacturing plants has evolved as a higher and increasingly vital function of process control. In this paper, an intelligent information management and control system is introduced. The different computational agents (i.e., modules) of the system are embodied in a three-layered cognitive hierarchy, which offers intelligent behavior at the system level, as well as at the level of specialized task agents. At the lower level, agents generate goal-seeking reactive behavior. Three different fault detection and isolation agents (i.e., three complementary techniques) are embedded to generate three different assessments and to enhance the fault isolation process. Other utility agents are also incorporated to address topics such as process model identification and optimization. At the middle layer, agents enable decision making, planning, and deliberative behavior. Two case-based reasoning agents are incorporated, the first manages the system in normal operation, while the other handles faulty process situations. A meta-management agent at the highest level monitors and coordinates other agents so as to make the whole system performance more robust and coherent.