Affiliations: [a] College of Information, Shanghai Ocean University, Shanghai, China
| [b] Audio Visual Education Center, Shanghai Maritime University, Shanghai, China
Correspondence:
[*]
Corresponding author. Peng Lu, College of Information, Shanghai Ocean University, Shanghai 201306, China. E-mail: lplplp20051@163.com.
Abstract: This paper combines the rough set with graph theory to deal with the problems of power transformer fault diagnosis, by the graph of decision table for fault diagnosis and its partitioned adjacency matrix. In the process, the new three-ratio decision table of fault diagnosis based on graph theory and rough set is got without conflict and missing, to derive the new fault diagnosis rules. These new rules got by the partitioned core attribute of this graph can expand the fault diagnosis range of guideline IEC-60599, and improve the defect problem of three-radio fault diagnosis method. The results of experiment based on the 62 fault samples of power transformers prove the effectiveness of the new method.
Keywords: Power transformer fault diagnosis, rough set, the graph of decision table, partitioned core attribute
