Anticipating Cleaning Schedules for 22 KV Porcelain Insulators through Leakage Current Analysis with Machine Learning

In 22 KV Electricity Distribution networks, the Pin insulators are susceptible to flashovers due to contamination layers formed on the surfaces in industrial and coastal areas. Contamination layers can result in the formation of dry bands, leading to partial discharge in the surrounding air. This cascade effect ultimately culminates in flashovers, which not only degrade the lifespan of pin insulators but can also cause punctures, resulting in power supply interruptions. In this study, leakage current variations are measured with respect to different contamination levels. This data is used to create an AI model. The model is able to predict the contamination level for any leakage current. The model has the capability to forecast contamination levels based on leakage current, offering insight into the degradation of insulation strength. This degradation correlates directly with the lifespan of insulators. By accurately assessing insulator maintenance needs, the model facilitates prolonging their life cycle.