Electrical and temperature correlation to monitor fault condition of ZnO surge arrester

Third harmonic resistive leakage current and thermal imaging was commonly used as tool for monitoring and assessing the zinc oxide surge arrester condition. However, the measuring of leakage current on-site is highly influenced by electromagnetic interference (EMI) and caused difficulty to get accurate measurement. Thermography technique can be used to measure temperature of ZnO arrester without interfered by the EMI. However, temperature images resulted from thermal camera were difficult to analyze due to lack of additional information. This paper proposes thermal image temperature correlation with third harmonic resistive leakage current to monitor condition of ZnO arrester. The condition of ZnO arrester classifies as normal, suspicious and faulty based on its third harmonic resistive leakage current. Statistical features of thermal image histogram together with the third harmonic resistive leakage current were used. The multilayer perceptron (MLP) neural network was used to classify the condition of ZnO arrester. The result shows that accuracy of classification is approximately 97 %, which is highly encouraging.