Research on diagnosis rule acquisition for turbo-generator unit rotor misalignment fault based on rough set

摘要

Aiming at solving the problem of feature information redundancy and its different degree of importance for turbo-generator unit based on rough set, this paper uses an attribute reduction method to get the misalignment fault diagnosis rules. By fuzzifying up and discretizing the misalignment fault characteristic parameters, this article constructs the fault decision table for reduction, and meanwhile removes the redundant information to extract diagnosis rules of misalignment fault from incomplete and inaccurate data by using decision table. The application shows that this method is not only an effective way to obtain the fault diagnosis rules, but also can effectively diagnose the rotor misalignment fault.