Experience in PD defect detection and location using smart sensors with AI deployed in distribution grid

Over the past 10 years, EDP Redes España has been using online partial discharge (PD) monitoring to diagnose the insulation condition of high voltage (HV) and medium voltage (MV) cables, power transformers, and switchgears in the distribution grid. Initially, spot monitoring was conducted every 3 years on critical assets using portable devices and PD diagnosis experts. The results demonstrated the effectiveness of this method in anticipating insulation faults, with the ultimate goal of applying predictive maintenance across all assets to detect and locate any PD defect before failure. Since 2021, EDP Redes España, in collaboration with Ampacimon, has developed a permanent monitoring smart sensor capable of automatically detecting and locating any PD defect in the distribution grid. Critical requirements such as sensor sensitivity and coverage, as well as data communication between the smart sensor and the central server, have been analysed. Artificial intelligence (AI) algorithms and data models have been developed to reduce human interaction and operational costs. The new smart sensor has been field-validated and compared with manual diagnosis by PD experts. The paper describes relevant cases of PD defect detection and location.