Aging Analysis of Transformer Insulation at Weakest Region: Dielectric Parameters Extraction via Immune Optimization

The conventional method is insufficient for determining the worst aging state of traction transformer oil-paper insulation due to the nonuniform aging. To address this issue, an extended relaxation distribution model based on multipopulation immune algorithm (ERD-MPIA) is proposed. This work collects frequency-domain spectroscopy (FDS) on the oil-paper system in a variety of nonuniform aging phases. Then, the MPIA is used to optimize the ERD model to establish the appropriate relaxation parameters for characterizing the worst aging stages. The verification results demonstrate that ERD-MPIA is capable of accurately decoupling the FDS of oil-paper systems subjected to nonuniform aging, allowing for the determination of the ideal relaxation parameters. Additionally, a distinct association is discussed between the relaxation parameters and the worst aging states. In that regard, the findings are expected to evaluate and detect the worst aging states of the traction transformer insulation system.