Application of a qualitative reasoning method to water treatment plant fault diagnosis

This paper deals with the signed directed graph (SDG) method applied to the diagnosis of a moisture extraction box clogging of a rotary dryer in a waste water treatment plant. This method is based on the causal graphs and the membership functions of the fuzzy set theory. It permits to detect the variations of the process variables and then discover the fault variable using the cause/effect relationships between these variables. The interesting point of this method is the possibility to use both quantitative and qualitative knowledge when it is not possible, from the quantitative information available on the process, to access to reliable mathematical models for diagnosis with analytical models.