Parametric Identification of CVTs via Frequency Response Analysis: A Comparison Between Levenberg-Marquardt and PSO Methods

Authors

  • Mario E. Borba
  • Ulisses Netto
  • Diego Issicaba
  • Antonio F. Aquino

Keywords:

Capacitive Voltage Transformer, Parametric Identification, Levenberg-Marquardt, Particle Swarm Optimization, Frequency Response Analysis

Abstract

This paper compares the Levenberg-Marquardt (LM) and Particle Swarm Optimiza- tion (PSO) methods for parametric identification of the equivalent circuit model of Capacitive Voltage Transformers (CVTs) in two stages. In the first, both methods are validated using reference parameters extracted from the literature, with considerations regarding the strong sensitivity of LM to initial conditions and the robustness of PSO across different initializations. In the second, PSO is applied to identify the parameters of a CVT installed in the Brazilian National Interconnected System (NIS), using a frequency response curve obtained from field measurements, achieving a satisfactory fit of the normalized correction factor. The results indicate that PSO is better suited for this application, given the absence of a priori information about the equipment parameters in field conditions.

Downloads

Published

2026-10-05