Sequential fuzzy diagnosis method and identification method of membership function by probability and possibility theories

When building up a fuzzy diagnosis system for machinery diagnosis, fuzzy relation between failure symptoms and failure categories must be defined for fuzzy inference. However, it is not easy to search out the failure symptoms by which all failure categories can be distinguished perfectly and automatically. In order to resolve the problem, we proposed (1) a new type of symptom parameter function called off-group type of symptom parameter (OGSP); (2) the identification method of the OGSP; (3) the identification method of the membership function of OGSP; (4) the algorithm of sequential fuzzy inference by using the OGSP and its membership function for diagnosis. The efficiency of the above methods has been verified by applying them to the ball bearing diagnosis system