Solutions on fuzzy diagnosis by inverse process with natural language input

Several kinds of diagnosis techniques based on the kinetic or statistical model of a nuclear plant have already been utilized in operation. Rule-based approaches have also been proposed to realize a flexible diagnostic method by using fuzzy sets, which consider the rules describing the relationship between the cause and symptom of the failures. These might be able to infer the failure by modus tollens using implications to represent the relation between the cause and symptom. In the first step of a diagnosis, it needs a broad outline of the failure from slight symptoms. The paper attempts to express solutions of fuzzy diagnosis by an inverse process, and the natural language input for fuzzy diagnosis with a dialog box on Visual C++.