부산대학교 도서관

The trend toward intellectualization is pervasive in the field of power equipment fault diagnosis. However, due to the sample database's limitations, intelligent diagnosis of transformer insulation aging states near the internal hotspot regions remains a challenge. To address this issue, an optimization-design-based simulation model is proposed for the development of a frequency domain spectroscopy (FDS) database for diagnosing nonuniform aging states of power transformer. The simulation model is developed in the current work using distributed relaxation circuits and fast Fourier transforms, and the electric parameters used are identified using adaptive iteration particle swarm optimization. Finally, verification established that the extended simulation model was capable of simulating the FDS of a system with nonuniformly assigned aging states. In this regard, the contribution will aid in the establishment of a trustworthy FDS database for the purpose of advancing intelligent diagnosis of the aging state of transformer oil–paper insulation at hotspots.