Infrared Image-Based Condition Assessment of 33kV SIR Housed ZnO Surge Arresters

Lightning surge arresters are key components for the protection and reliability of power systems. During service, surge arresters are aged by deteriorating their properties. It is important to monitor conditions periodically in order to avoid failures of power apparatus. This study proposes an online-based condition monitoring method for surge arresters using principal component analysis (PCA), support vector machine (SVM) and thermal image processing of infrared images captured from 46 energized 33kV lightning surge arresters belonging to Ceylon Electricity Board (CEB). Field-energized aged arresters (18 samples), laboratory-tested aged arresters (27 samples) and a brand-new arrester were taken for the investigation. The main focus was to build a model to identify the healthiness of the lightning surge arresters based on their temperature profiles. Leakage current, insulation resistance and finite element method (FEM) based simulations were considered to validate the healthiness of the arresters. Using this model, it can be visualized whether the surge arrester condition is acceptable, only using a single thermal image of the arrester as an input. Even though, an arrester islocated in good condition region of the proposed model, the distance to the decision boundary indicates how close that arrester to be in unhealthy region.