A fuzzy fault diagnosis method applied to a steam circuit

This paper presents a hierarchical scheme of a fault detection and isolation (FDI) procedure based on a decision support system (DSS). Knowledge based procedure, analytic symptoms and heuristic information are combined in order to achieve the fault diagnosis task. Then, using a pattern recognition method, a unified representation of all symptoms and a fuzzy expert system are integrated in the diagnostic tool. The results obtained with this approach on the steam circuit in a waste water treatment plant are compared with those obtained with structured residuals using a modified Hamming's distance.