Rough Set Theory for Data Mining for Fault Diagnosis on Power Transformer
Wenqing Zhao, Yongli Zhu · 2006
Because the testing data of power transformers for their condition evaluation have uncertainty characteristic, and there is little knowledge and human experience on transformer fault diagnosis, the rough set theory is used to diagnose transformers. This paper presents a new type fault decision model based on rough set theory. The traditional dissolved gas-in-oil analysis (DGA) for power transformers' condition evaluation and rough set theory based fault diagnosis are combined in the diagnostic model. The results of using the proposed model to analyze some known samples of testing data of faulty transformers shows that the model possesses strong solving ability to deal with uncertain facts. Moreover, by comparing with popular diagnosis methods like Naive Bayes classification, there is less fault data discriminated by the rough set model and the accuracy for power transformer fault diagnosis is improved using our proposed model