A novel method for prediction of washing cycles of electrical insulators in high pollution environments

Abstract Electrical substations located closer to coasts and industrial zones are subjected to chemical substances and marine salts that increase pollution. Therefore, to mitigate pollution's effect on insulator performance, utility operators should frequently wash insulator chains. However, this process is not an easy task and brings extra costs due to the conventional maintenance planning that defines the washing cycles. This conventional procedure does not take into account the actual insulator pollution, but rather human experience. Thus, washing cycles may occur either when the pollution is not extensive or when it already affects the reliability of the system. Hence, the paper presents a novel method to predict the adequate time to wash the insulators based on the characteristics of leakage current in a substation on the north coast of Colombia. The method then uses the Exponentially Weighted Moving Average (EWMA) control chart to determine the appropriate days to wash insulator chains. Results show that the proposed method obtains a Mean Absolute Percentage Error (MAPE) of the leakage current two orders of magnitude smaller than the actual method. Hence, the proposed method is a useful tool for maintenance planning in utilities.