Controlled pre-magnetization method for inrush current reduction in electric arc furnace transformers
DOI:
https://doi.org/10.24425/aee.2026.6352Abstract
This paper investigates transformer inrush currents during the energisation of arc furnace transformers in electric arc furnace (EAF) power supply systems. The study focuses on an industrial DSP-100 installation operating with harmonic filters and reactive-power compensation equipment. A nonlinear reduced-order lumped-parameter model is formulated using explicit phase-domain relations for the supplying-system equivalent, the furnace-transformer branch, the connected filter/compensation branches, and the nonlinear magnetic circuit. The surviving DSP-100 oscillograms are used as an experimental anchor and for a waveform-level consistency assessment of peak current, inter-phase peak-current variation, and transient decay. Because the original high-resolution recorder export, complete channel documentation, and recorder characteristics are no longer available, this comparison is not presented as a unique experimental validation of the nonlinear saturation law or of the detailed nonsinusoidal inrush waveform. A model-level evaluation framework for controlled pre-magnetization is then formulated using peak inrush current, transient decay time, and an energy-based transient severity index. Relative peak-current error and RMSE are retained as waveform-consistency metrics, while the practical interpretation of the severity index is linked to equipment-specific I²t withstand limits. The experimentally anchored model provides a basis for comparative and sensitivity assessment of controlled pre-magnetization in the DSP-100 operating environment. Prototype-level confirmation of controller performance, nonlinear-waveform fidelity, and the influence on associated compensation and filtering equipment remains a subject of further industrial validation.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Archives of Electrical Engineering

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.