Fault Diagnosis of Contaminated Insulators Based on Heterologous Entropy Feature Fusion
- Resource Type
- Conference
- Authors
- Gao, Kai; Wu, Tianyi; Hu, Zhengyong; Deng, Jinyi; Ma, Danrui; Jun, Jinli
- Source
- 2022 IEEE 5th International Electrical and Energy Conference (CIEEC) Electrical and Energy Conference (CIEEC), 2022 IEEE 5th International. :2911-2916 May, 2022
- Subject
- Engineered Materials, Dielectrics and Plasmas
Power, Energy and Industry Applications
Fault diagnosis
Pollution
Fuses
Decision theory
Power transmission
Flashover
Insulators
contaminated insulator
heterologous signal
entropy feature
feature fusion
fault diagnosis
- Language
Pollution flashover of insulators is one of the common faults in the field of power transmission and transformation. In order to improve the reliability of non-contact detection of insulator contamination state, a fault diagnosis method for contaminated insulators based on the fusion of infrared image and leakage current feature level is proposed. Extract insulator infrared images and leakage current entropy features from normal to severe fault states respectively, and then feature selection according to Fisher criterion to obtain feature quantities that can effectively characterize insulator operating conditions. The Bayesian decision theory is used to fuse the heterogeneous entropy features and faults diagnosis. The experimental results show that the infrared image and the leakage current entropy feature can effectively characterize the operating state of the insulator. The accuracy rate of the fusion diagnosis method is significantly higher than that of the infrared or leakage current feature alone, which provides a new method for insulator fault diagnosis.