Intelligent system for process supervision and fault diagnosis in dynamic physical systems

In recent years, the increasing complexity of process plants and other engineered systems has extended the scope of interest in control engineering, which was previously focused on the development of controllers for specified performance criteria such as stability and precision. Modern industrial systems require a higher demand of system reliability, safety, and low-cost operation, which in turn call for sophisticated and elegant fault-detection and isolation algorithms. This paper develops an intelligent supervisory coordinator (ISC) for process supervision and fault diagnosis in dynamic physical systems. A qualitative bond graph modeling scheme, integrating artificial-intelligence techniques with control engineering, is used to construct the knowledge base of the ISC. A supervisor provided by the ISC utilizes the knowledge in the knowledge base to classify various system behaviors, coordinates different control tasks (e.g., fault diagnosis), and communicates system states to human operators. The ISC provides a robust semiautonomous system to assist human operators in managing dynamic physical systems. The proposed ISC has been successfully applied to supervise a laboratory-scale servo-tank liquid process rig.