An Integrated Approach to Fault Detection and Diagnosis in Power Plant (original) (raw)

This paper deals with a computer diagnostic system for a power plant simulator. The system is designed to help the operator in fault detection and isolation during operation of the power plant. A two-part knowledge base was created to increase the efficiency and flexibility of knowledge-based diagnostic system. The base consists of heuristic knowledge (set of rules and principles of system behavior) and analytical knowledge (set of mathematical models , Kalman filter algorithms and tests of random sequences). The design methodology is a hierarchical knowledge structure with compiled knowledge at higher levels of abstraction and analytical redundancy at lower ones. Such an integrated approach reduces the computational complexity associated with functionally-redundant schemes and increases the effectiveness of the knowledge-based approach .