• CSCD核心库收录期刊
  • 中文核心期刊
  • 中国科技核心期刊

ELECTRIC POWER CONSTRUCTION ›› 2022, Vol. 43 ›› Issue (3): 50-57.doi: 10.12204/j.issn.1000-7229.2022.03.006

• Application of Artificial Intelligence in Power Grid Fault Diagnosis and Location ·Hosted by Professor WANG Xiaojun, Associate Professor LUO Guomin and Associate Professor SHI Fang· • Previous Articles     Next Articles

Transformer Health Status Evaluation Based onRough Set G1 Combined Weighting

YUAN Wanling, CUI Zixuan, YU Hongbo, ZOU Xiaosong, XIONG Wei, YUAN Xufeng   

  1. The Electrical Engineering College, Guizhou University, Guiyang 550025, China
  • Received:2021-09-29 Online:2022-03-01 Published:2022-03-24
  • Contact: ZOU Xiaosong
  • Supported by:
    National Natural Science Foundation of China(52067004);National Natural Science Foundation of China(51867007);Guizhou Provincial Science and Technology Fund Project([2019] 1128)

Abstract:

In view of the imperfection of the index system and singleness and one sidedness of the index weight determination method in the construction of the power transformer health index model, a transformer health status evaluation method based on the optimal combination of rough set-G1 (order relationship analysis method) is proposed in this paper. The comprehensive index system model of transformer health status is established by integrating transformer aging, electrical test, oil color spectrum test, oil quality test, furfural test and accessories. The objective weight is obtained by rough set and the subjective weight is obtained by G1 method. The objective and subjective optimal combination method is used to scientifically weight each index. Finally, through the overall health index model of power transformer, the health indices of 7 transformers with different voltage levels are calculated to evaluate their health status. Compared with the traditional index system and weight method, it is verified that the index system proposed in this paper is more in line with the evaluation practice, the index weight is more reasonable, and the accuracy and reliability of transformer health status evaluation are improved.

Key words: rough set, G1 method, optimize combination and weighting, health index, transformer condition evaluation

CLC Number: