Analysis of Power Transformers under DC/GIC Bias

Dennis Albert

Research output: ThesisDoctoral Thesis

Abstract

Power transformers are key components in the interconnected bulk power transmission grid. Moreover,
to ensure the reliable and stable operation of the power grid, the interaction of the transformers and
the power grid during normal and abnormal operation conditions were studied. To study abnormal
operation conditions of power transformers it may be necessary to include the non-linear hysteresis
characteristics of the transformer cores in electromagnetic transient studies. The modelling of the
hysteresis characteristics of the transformer cores requires detailed information about the transformer
core design and material. If this information is not available, it is challenging to establish an adequate
electromagnetic transformer model. Especially during deep saturation conditions, typically near two
Tesla for gain-oriented steels, an accurate modelling of the hysteresis characteristics can be essential
for the calculated phase currents. Such saturation conditions could be caused by geomagnetically
induced currents (GICs) or direct current (DC) bias caused e. g. by power electronic devices. This
work is a follow-up investigation, motivated by increased transformer sound, which could be traced
back to GICs in the high and extra high-voltage transmission grid.
This work presents a measurement based modelling approach to establish electromagnetic topology
models of power transformers, including the transformer’s core hysteresis characteristics. First the
AC saturation test was developed with the idea to saturate the outer two limbs of a three-phase
transformer core by two elevated 180◦ phase-shifted single-phase voltages. The AC saturation test
was successfully used to parametrise the hysteresis model of two transformer topology models, using
the inductance-reluctance and the capacitance-permeance analogy. Because the AC saturation test
requires a sufficiently large power source, it was further developed to the DC hysteresis test. Instead
of using a 50/60 Hz sinusoidal voltage, a DC with reversal polarity was used. The DC hysteresis test
was also successfully used to parametrise the transformer hysteresis models. The implementation of
the DC hysteresis test in a portable transformer test allows to conduct this test in the laboratory and
in the field. Together with the principle of variable core gap inductance the transformer topology
models of a 50 kVA reveal a high accuracy of the calculated and measured current waveforms during
the AC saturation and the standard no-load test, as well as the corresponding power demand.
For the measurement of transformer neutral point currents, including geomagnetically induced
currents (GICs), an existing measurement system was further developed to minimise the constraints
of the monitoring system on grid operations. The utilisation of a split-core current transducer
around the earthing switch, together with a software-supported correction of the offset drift, reveals
a low long-term offset drift of the measured transformer neutral point current. In addition to the
measurement of the transformer neutral point current, the measurement system was extended
to monitor a direct current compensation (DCC) system, installed in several transformers in the
transmission grid. The analysis of the DCC measurements, which allows a calculation of the DC per
phase, reveals an equal distribution of the DC between the high-voltage phases and the capability of
the system to minimise the effects of GICs in transformers.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Graz University of Technology (90000)
Supervisors/Advisors
  • Renner, Herwig, Supervisor
  • Høidalen, Hans Kristian, Supervisor, External person
Publication statusPublished - 10 Oct 2022

ASJC Scopus subject areas

  • Engineering(all)
  • Energy(all)

Fields of Expertise

  • Sustainable Systems

Fingerprint

Dive into the research topics of 'Analysis of Power Transformers under DC/GIC Bias'. Together they form a unique fingerprint.

Cite this