Exploring transformer fault detection using RFID technology

Abstract

Real-time monitoring and fault diagnosis of transformers are essential for the stable power system operation. This paper presents an RFID-based transformer fault feature extraction and classification algorithm. Experiments show that monitored current signals are stable while the temperature peak is 356°C. Hilbert decomposition reveals regular current and voltage patterns that can be used as fault indicators. Signal strength classification accuracy reached 80% . At rated load, the transformer temperature soared to 186°C, indicating overheating issues. The monitoring during a sample day showed that overload events were concentrated from 16:00-20:00, which required attention. The approach helps accurately identify transformer fault types from real-time RFID data for proactive maintenance. Compared to reactive repairs after failures, this not only improves employee productivity but also reduces costs. Based on customized RFID deployment, the algorithm contributes to the stability and economy of power infrastructure.