Towards Unique Circuit Synthesis of Power Transformer Winding Using Gradient and Population Based Methods

Abstract

The aim of the paper is to synthesize a nearly unique, physically realizable and mutually coupled ladder circuit representation of a two-winding transformer by identifying the most accurate and reliable optimization technique such that prior knowledge in selecting the initial guess and search space is avoided. To this end, magnitude and phase of the driving-point impedance function are captured by performing sweep frequency response analysis (SFRA). Then, two widely used population-based algorithms namely, particle swarm optimization (PSO) and artificial bee colony (ABC) and one gradient-based sequential quadratic programming (SQP) algorithm are implemented. Two case studies are considered namely, i) 15 MVA, 66/11.55 kV transformer and, ii) 111 kVA, 7.33/1.22 kV transformer. The performance of the aforementioned algorithms is compared using three evaluation parameters namely, repeatability, accuracy and reliability