A COMPARATION OF MODEL AND SIGNAL BASED FAULT DETECTION METHODS

In industrial processes, physical components or automatic control components, sensors, and/or actuators are often affected by un-permitted or un-expected deviations from their normal operation behavior. The fault detection task consists in determination of the fault occurrence and time of detection. The diagnostics procedures are based on the observed analytical and heuristical symptoms and the heuristic knowledge of the process. In this paper a comparison between different methods of fault detection and an overview of the fault diagnosis procedure for technical system is provided. For some classes of processes the structure and some parameters are well known but for others only rough models are available. Therefore the methods for fault detection and diagnosis are mainly different and several examples are explained.