Keywords
dissolved gas analysis, transformer fault diagnosis, bayesian network, normality test, fault classification
Abstract
Accurate diagnosis of power transformer faults is essential to the reliable operation of the power grid. To achieve this goal, this paper proposes a new multiclass probabilistic diagnostic model based on bayesian network and hypothesis testing of dissolved gas analysis. The bayesian network model can embed expert knowledge, learn data patterns from data and infer uncertainties related to diagnosis results, and improve the data selection process through hypothesis testing. Based on the IEC TC10 data set, this paper compares three traditional diagnostic methods to perform diagnostic experiments to verify the effectiveness of the proposed model. The results show that the maximum diagnostic accuracy of the proposed diagnostic model is 88.9%, which is greatly improved compared to traditional diagnostic methods.
DOI
10.19781/j.issn.1673-9140.2021.06.003
First Page
20
Last Page
27
Recommended Citation
HE, Ninghui; ZHU, Hongbo; LI, Xiuguang; PAN, Liangliang; ZHOU, Xiu; and NI, Hui
(2022)
"Transformer fault diagnosis based on bayesian network and hypothesis testing,"
Journal of Electric Power Science and Technology: Vol. 36:
Iss.
6, Article 3.
DOI: 10.19781/j.issn.1673-9140.2021.06.003
Available at:
https://jepst.researchcommons.org/journal/vol36/iss6/3