A hierarchical fault diagnosis method using a decision support system applied to a chemical plant

Automatic control systems with sophisticated control algorithm can be very large and complex. In order to improve the automatic process control, it is important to develop fault diagnosis strategy. A hierarchical scheme of fault detection and isolation based on decision support system (DSS) is presented. For fault diagnosis, a knowledge based procedure is required. In addition to analytic symptoms, heuristic information have to be taken into account. A pattern recognition method, a unified representation of all symptoms, and a fuzzy expert system are integrated in the diagnostic tool. This approach is applied to sensor fault detection in a chemical plant, and compared with different methods.