Study on the Detection of Single and Dual Partial Discharge Sources in Transformers Using Fiber-Optic Ultrasonic Sensors

Partial discharge is a fault that occurs at the site of insulation defects within a transformer. Dual instances of partial discharge origination discharging simultaneously embody a more intricate form of discharge, where the interaction between the discharge sources leads to more intricate and unpredictable insulation damage. Conventional piezoelectric transducers are magnetically affixed to the exterior metal tank of transformers. The ultrasonic signals emanating from partial discharge undergo deflection and reverberation upon traversing the windings, insulation paperboards, and the outer shell, resulting in signal attenuation and thus making it difficult to detect such faults. Furthermore, it is challenging to distinguish between simultaneous discharges from dual partial discharge sources and continuous discharges from a single source, often leading to missed detection and repairs of fault points, which increase the maintenance difficulty and cost of power equipment. With the advancement of MEMS (Micro-Electro-Mechanical System) technology, fiber-optic ultrasonic sensors have surfaced as an innovative technique for optically detecting partial discharges. These sensors are distinguished by their minute dimensions, heightened sensitivity, and robust immunity to electromagnetic disturbances. and excellent insulation properties, allowing for internal installation within power equipment for partial discharge monitoring. In this study, we developed an EFPI (Extrinsic Fabry Perot Interferometer) optical fiber ultrasonic sensor that can be installed inside transformers. Based on this sensor array, we also created a partial discharge ultrasonic detection system that estimates the directional information of single and dual partial discharge sources using the received signals from the sensor array. By utilizing the DOA (Direction of Arrival) as a feature recognition parameter, our system can effectively detect both simultaneous discharges from dual partial discharge sources and continuous discharges from a single source within transformer oil tanks, meeting practical application requirements. The detection methodology presented in this paper introduces an original strategy and resolution for pinpointing the types of partial discharges occurring under intricate conditions within power apparatus, effectively distinguishing between discharges from single and dual partial discharge sources.