Knowledge-Based Approach to Real-Time Supervisory Control

Increasing complexity of industrial plants necessitates the usage of advanced programming techniques in process control. Artificial Intelligence programming offers new opportunities in constructing complex software systems. This paper describes an approach which has been used to build Intelligent Process Control System (IPCS). IPCS expand the conventional process control systems with an additional, "knowledge-based layer", where declarative programming methods are used extensively. The knowledge-based components represent a model of the process from multiple viewpoints and the simulator, monitor, control, diagnostics and operator interface subsystems are coupled to a symbolic process database. A special architecture, the Multigraph Architecture has been used as a general framework for integrating symbolic and numeric programming techniques in a distributed environment. The knowledge-based components have been implemented as "Autonomous Communicating Objects". The application of concurrent programming techniques made possible the introduction of new approaches in the design of the monitoring and diagnostic system. A prototype IPCS has been implemented on a network of VAX computers.