Structured fault-detection and diagnosis using finite-state automaton

Fault detection and diagnosis in large, complex dynamic systems is important for safe operation. Faults may occur in the process, the sensors, the actuators and the measuring instruments independently or simultaneously. Some faults can be detected by direct measurement, but many can only be inferred through indirect means. We propose to address the problem of fault-detection and diagnosis using finite-state automaton models (FAM). These models partition the state-space into finite regions and contain information on system trajectory within these regions. A framework for modelling faults using FAM is discussed. A fault-detection engine is then developed to detect and isolate faults.