Accurate interpretation of frequency response analysis (FRA) data is crucial for power transformer windings. To address this challenge, this study suggests a new auto-synthesis method to derivate a high-frequency Equivalent Ladder Network Circuit Model (ELNCM) aimed at investigating and interpreting the transformer winding (TRW) characteristics.
The precise ELNCM is attained using the transfer function (TF) extracted from the measured FRA data. From where, five winding parameters are recited, mainly, the total ground capacitance , the equivalent inductance and capacitance , the voltage distribution factor and series capacitance . After that, the precise self and mutual inductances () of the proposed ELNCM identification are carried out using artificial intelligence (AI). As the self and mutual inductances cannot be measured directly, this paper presents an AI approach, namely, Logistic Chaotic Archimedes Optimization Algorithm (LCAOA) for improved precision.
The physical parameters ( are directly extracted from the FRA data, which is very useful in practical studies. Through the obtained results, actual identification is ensured by faithfully seeing the FRA curves created from TF-AI-ELNCM.
Going further, unlike previous studies, the proposed method eliminates the need for geometrical data and not necessitate any specific arrangement on the transformer winding. The methodology presented offers substantial practical benefits for transformer winding diagnosis, it eliminates the need for lookup tables or specific configuration models.
