Real time outlier monitoring for power transformer fault diagnosis based on isolated forest. (January 2020)

• Record Type: Journal Article
• Authors: Shen, Li; Du, Hongjun; Liu, Shuji; Chen, Shuo; Qiao, Lin; Liu, Sai; Liu, Jiahua; Li, Kexin; Li, Jing
• Is Part Of: IOP conference series. Volume 715(2020)
• Publication Date: 2020-01
• Subjects: Materials science -- Periodicals; 620.1105
• DOI: 10.1088/1757-899X/715/1/012033
• Languages: English
• ISSNs: 1757-8981

Abstract

Abstract: In order to improve the accuracy and efficiency of transformer fault detection, this paper uses the isolated forest algorithm combined with the historical transformer characteristic gas data to establish the characteristic gas outlier recognition model, and then uses the uncoded ratio method to establish the abnormal event strategy and abnormal event library for the transformer historical fault information. Finally, the state of the transformer is diagnosed based on the established outlier detection model and the exception event library. The experimental results show that the proposed method has a great improvement in the detection efficiency and stability of outliers, and is more accurate in transformer fault diagnosis combine with abnormal event database.