Metal-Oxide Arrester On-Line Processing and Alarm Threshold Selecting Method

Abstract

Metal-Oxide Arrester (MOA) On-line monitoring reflects the state of the equipment, especially when it comes to leakage current which shows the Insulation strength. As the differences in the operating environment and the influence of temperature and moisture on leakage current testing, each MOA has its own typical value. However, nowadays the MOA alarm threshold is only one, lack consideration of different operation environments. It will cause overhaul or lack of repair. This paper proposes an individualized MOA On-line Alarm Threshold Selecting Method. First, aggregate MOA on-line monitoring data, including 1 year's MOA monitoring data. Then, preprocessing detection data, including identification the singular value by 3 deltas rule and smooth the noise by low pass filter. For the problem of monitoring data distortion, identification the singular value by 3 deltas rules and smooth the noise by Yang Hui triangle. Then sort the data on multiple criteria and calculate the leakage threshold based on the confident, which correspond to the fault ratio. By this means, the MOA targeted alarm threshold has been selected. Field applications shows that the targeted alarm threshold decrease the false alarm rate greatly. It illustrated that the method is valid and economical, and worth popularization.