전력용 변압기 결함 진단을 위한 Aging Index 적용 알고리즘 연구

다국어 초록 (Multilingual Abstract)

The increase in electric power demand has led to a growth in the number of installed power transformers. As the initially installed transformers approach their operational lifespan, the necessity for technologies to diagnose their condition has emerged. The Dissolved Gas Analysis (DGA) method is a representative technique for diagnosing transformer condition by analyzing the gases generated due to internal transformer faults. In this paper, to evaluate the thermal and electrical degradation of transformers based on DGA data, the gas concentrations were normalized. Subsequently, an Aging Index was proposed by calculating
and applying weights using Grid Search, OPTUNA, and Genetic Algorithm. Furthermore, index-based pattern analysis was conducted, and the long-term variability and prediction uncertainty of the index were evaluated by applying LSTM and Monte Carlo simulation.The proposed weight-based index can quantitatively reflect the trends for each fault type, complementing the limitations of existing diagnostic methods, and is expected to be effectively utilized in selecting priorities
for the maintenance and replacement decision-making of transformers.