Distribution network insulation fault risk warning method based on LSTM and FCM

Statistics show that most power outages are caused by distribution network insulation failure. It is difficult to detect the insulation deterioration of distribution network in time, and the failure to effectively warn the insulation fault brings difficulties to the inspection work of distribution n...

Full description

Saved in:
Bibliographic Details
Published in2024 4th International Conference on Neural Networks, Information and Communication (NNICE) pp. 1539 - 1543
Main Authors Zhao, Yu, Bai, Hao, Mao, Xing, Li, Ping, Yang, Qian, Nie, Ding
Format Conference Proceeding
LanguageEnglish
Published IEEE 19.01.2024
Subjects
Clustering algorithms
distribution grid
Distribution networks
Fault diagnosis
fault warning
FCM
Inspection
Insulation
insulation fault
LSTM
Transportation
Urban areas
Online AccessGet full text

Cover

More Information
Summary:Statistics show that most power outages are caused by distribution network insulation failure. It is difficult to detect the insulation deterioration of distribution network in time, and the failure to effectively warn the insulation fault brings difficulties to the inspection work of distribution network. In order to assist the distribution and transportation inspection department to take effective prevention and control measures to improve the reliability of distribution network, a distribution network insulation fault early warning model based on long short-term memory network (LSTM) and fuzzy C-means (FCM) clustering algorithm is proposed. First, by classifying the operation data of the power grid, 25 characteristic variables of the distribution network are selected, and the principal component analysis method is used to fuse the characteristic variables to form low-dimensional characteristic variables that can characterize the operation state of the distribution network. Then, based on the fused feature variables, FCM clustering algorithm is used to identify the boundary conditions for the distinction between the insulated fault state and the non-insulated fault state of the distribution network, and LSTM is used to predict the characteristic quantity of the distribution network in the short term, and the fault occurrence time in the short term is predicted by combining the characteristic quantity prediction results and the obtained state boundary conditions. Realize insulation fault early warning; Finally, taking the operation data of 39 feeder lines in a prefecture-level city as an example, the effectiveness of the proposed method is verified.
DOI:10.1109/NNICE61279.2024.10498803