Winding deformation identification using a particle swarm optimiser with passive congregation for power transformers

This paper presents a new approach to identify distributed parameters of a lumped-element model for a power transformer. A simplified circuit of a lumped-element model is developed to calculate frequency responses of transformer windings in a wide range of frequency domain. In order to seek optimal parameters of the simplified circuit of the transformer, an intelligent learning technique, based on particle swarm optimiser with passive congregation (PSOPC), is developed to identify model parameters using frequency response measurements. Simulations and discussions are presented to explore potentials of the developed approach. Index Terms – Transformer modelling, winding deformation, particle swarm optimiser, passive congregation.