Knowledge Property of Rotating Machine Fault Diagnosis and Vague Sets Expression for Frequency-domain Knowledge

摘要

This essay is to discuss what function and nature domain expert knowledge demonstrates in fault diagnosis system. Then come up with conclusions that the particularity and apriority of knowledge guarantee reliability of knowledge base, and realization principles of knowledge base, i.e. pervasive knowledge and particular knowledge constitute basic resources for knowledge base, and the initialization of knowledge base is more important than self-learning ability of the system. Analyze the one-dimensional particularity that Fuzzy sets of knowledge shows within the existing fault diagnosis system based on expressions of uncertainty within rotating machinery fault diagnosis knowledge. Thus propose 3 membership grades as positive, negative and hesitant grade for expressions of vague sets in fault diagnosis system knowledge or rules. By combining with frequency-domain characteristic of high-speed rotating machinery, we establish vague sets that associating frequency-domain characteristic with rotating machinery vibration fault types.