A Leakage Current Forecast of the Polymeric Insulator Using ANFIS Method Based on LabView Pre-processed Thermal Image

Abstract

This paper reports estimates of leaky currents based on Thermal (Infrared) images of polymer insulators using LabVIEW as preliminary data from the ANFIS method for the purpose of condition-based non-contact monitoring. In this study, images of very high contaminated polymer insulators were taken using the FLIR-A600 thermal camera series. The laboratory pollution performance test is carried out in accordance with the standards of IEC 60507 with an 18 kV AC voltage. The pollution severity is indicated because the ESDD value is very high contaminated and the relative humidity conditions are maintained at the range of 85% RH to 95% RH and temperature at 28°C. Humidity and temperature are maintained through a mist-producing container that is connected to a controlled polymer isolator test container using an Arduino microcontroller with temperature and humidity sensors. LabVIEW is used to calculate the percentage of RGB color from a thermal image related to a certain measured leakage current. This RGB data is training data for ANFIS to predict leaky currents. The results show that this method is able to predict current leakage without spending time to reach RMSE 0.00008. By adding a calculation to determine the condition of the insulator in the category of Safe, Maintenance or Hazard so that the Field Technician can decide whether the insulator must be replaced or not. With a non-contact method that only through thermal analysis of polymer insulators can accelerate the process of checking the condition of the insulator.