IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 40, NO.

1, FEBRUARY 2025 447

Geomagnetically Induced Current Field Test on

Large Grid-Connected Power Transformers:
Analysis, Model Development, and Simulations
Adedasola A. Ademola , Member, IEEE, Roland B. Brandis , Senior Member, IEEE,
Andreas F. Schuetzinger, Member, IEEE, Bart Simons , Luc Dorpmanns, Andrea Pinceti, Member, IEEE,
Katelynn D. Vance, Member, IEEE, Ibukunoluwa O. Korede, Senior Member, IEEE,
Robert M. Orndorff , Member, IEEE, Micah J. Till, Member, IEEE, Kyle D. Hannah, Member, IEEE,
Mike Lamb, Member, IEEE, R. Matthew Gardner, Senior Member, IEEE, and Yilu Liu , Life Fellow, IEEE

Abstract—Geomagnetic-induced current (GIC) flow in power I. INTRODUCTION

grids can cause undesirable effects such as transformer overheat-
HE adverse impacts of geomagnetic-induced current (GIC)
ing, harmonics, higher reactive power demand, etc. Many sim-
ulation models have been developed to study these effects, but
real-world verification on modern transformer designs is rare. This
T flow in the power grid during extreme geomagnetic distur-
bance (GMD) events have been extensively documented [1],
paper presents the first long-duration GIC field test in the U.S. [2], [3], [4]. GIC flow in electric transmission lines is permitted
performed on high-voltage, grid-connected transformers featuring
winding clamps and tie rods instead of conventional tie bars. Field by their connection to grounded wye-connected transformers.
measurements were taken to evaluate GIC effects. These measure- When the quasi-DC GICs flow in transformers, they cause DC
ments also aided in developing and validating thermal and electro- excitation which combines with the normal AC excitation to
magnetic transient (EMT) models of the transformers. During the induce a bias in the magnetic flux inside the transformer core,
test, significant current and voltage distortions were observed along leading to half-cycle saturation. During the saturation, large
with considerable transformer reactive power losses. Analysis of
the field measurements showed that the transformers’ hottest spot spikes of asymmetric magnetizing current are drawn by the
was at the inner windings, and their k-factors were close to factory transformers. This causes harmonic-rich distortions and signifi-
test and software default values. Thermal simulations indicated cantly increases reactive power demand, which can have system-
that the transformers would not violate their thermal limits even wide implications. For example, the simultaneous consumption
for a GIC waveform that peaks at about 200 A/phase. EMT simu- of excessive reactive power by several transformers can lead
lations revealed that increased transformer loading may reduce
GIC-induced reactive power demand and harmonics in certain to abnormally low voltage levels. The even and odd harmonics
scenarios. The study also highlighted potential inaccuracies in using from the transformers can trip voltage support devices and cause
the k-factor method to calculate transformer reactive power losses. protection devices to misoperate. Power generation capabilities
can be reduced if generators are unable to maintain scheduled
Index Terms—EMTDC, geomagnetic disturbances, geomag-
netically induced currents, harmonic analysis, power quality, voltages due to excessive reactive power production [5]. Notable
power transformers, reactive power, thermal analysis. occurrences such as the 1989 Hydro-Quebec blackout and the
2003 Swedish grid collapse highlight the potential for severe
disruptions [6], [7], [8], [9].
Aside from electrical effects, GIC can also thermally impact
grid-connected electrical equipment, leading to insulation fail-
ures or permanent damage. Stray fluxes that escape the core
Received 21 March 2024; revised 18 August 2024 and 23 October 2024;
accepted 3 November 2024. Date of publication 19 November 2024; date of cur- during saturation can cause direct heating of a transformer’s
rent version 24 January 2025. Paper no. TPWRD-00467-2024. (Corresponding structural parts, while harmonics can increase eddy current
author: Adedasola A. Ademola.) losses, which in turn heat up transformer windings and genera-
Adedasola A. Ademola was with the University of Tennessee, Knoxville, TN
37996 USA. He is now with Dominion Energy Virginia, Glen Allen, VA 23060 tor rotor windings. GIC-induced harmonics can also increase
USA (e-mail: [email protected]). transmission line heating due to skin and proximity effects
Roland B. Brandis, Andreas F. Schuetzinger, Andrea Pinceti, Katelynn D. which increases with harmonic frequencies [10]. GIC also has
Vance, Ibukunoluwa O. Korede, Robert M. Orndorff, Micah J. Till, Kyle D.
Hannah, Mike Lamb, and R. Matthew Gardner are with Dominion Energy mechanical effects including transformer vibration and noise,
Virginia, Glen Allen, VA 23060 USA. which noticeably increases even at low GIC amplitudes [11]. In
Bart Simons and Luc Dorpmanns are with Royal SMIT Transformers B.V., most cases, the mechanical effects are considered less important
6531 Nijmegen, The Netherlands.
Yilu Liu is with the University of Tennessee and Oak Ridge National Labo- than the more dangerous electrical and thermal effects.
ratory, Knoxville, TN 37830 USA. Considering the potential these effects have to cause high-
Color versions of one or more figures in this article are available at impact and wide-area disruptions, GIC has generated a lot
https://doi.org/10.1109/TPWRD.2024.3502642.
Digital Object Identifier 10.1109/TPWRD.2024.3502642 of interest among academia, power systems engineers and

0885-8977 © 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.

Authorized licensed use limited to: Chaitanya Bharathi Institute of Tech - HYDERABAD. Downloaded on June 24,2025 at 03:54:51 UTC from IEEE Xplore. Restrictions apply.
 448 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 40, NO. 1, FEBRUARY 2025

government entities, leading to numerous GIC related studies

and simulation tools [12], [13], [14]. To verify these simulation
studies, there is a need for physical GIC tests on high-voltage
power transformers which are the principal sources of GIC-
related issues on the power grid. While laboratory experiments
have been performed [15], these tests are often on low-voltage
transformers with kilo volt-amp power ratings and are insuffi-
cient in scale to accurately replicate GIC effects on large high-
voltage power transformers. Model verification through factory
tests of large power transformers is possible, but such tests do not
represent actual service conditions during a GMD event when
transformers are connected to the grid and carrying load [16]. In Fig. 1. Aerial view of the single-phase autotransformers that make up TX1.
some cases, the test system voltage can collapse during factory
tests or have exaggerated distortions as the transformer core
saturates. Moreover, most laboratory or factory testing is only to develop and validate thermal and electromagnetic transient
able to test relatively low GIC (or DC) magnitudes of a few (EMT) models. The thermal model was subsequently used to
amperes due to limitations of the test equipment. According estimate the transformers’ thermal response to realistic GIC
to IEEE Std. C57.91, these tests are not of significant value to waveforms under full load conditions. The EMT model was used
understanding GIC effects on power transformers under realistic to study the electrical effects of GIC on the transformers under
operating conditions [16]. various load conditions. This approach was useful because it
Across the world, only a few large-scale GIC tests have been allowed investigation of GIC effects on transformers under re-
performed on grid-connected power transformers. The earliest alistic, load-carrying service conditions, without exposing DEV
ones occurred in the 1980s with three tests on the operating customer loads to GIC effects.
system of Minnesota Power, USA. The first was performed on It is worth noting that a pre-test study, based on EMT modeling
an unloaded 500/230 kV, 200 MVA, shell-type, single-phase and simulations, was conducted to evaluate the potential grid
autotransformer. The other two were performed on two loaded impacts of the field test. The methodology and results of this
230/115 kV, 333 MVA, three-phase transformers; one with a study were previously published in [21]. The study demonstrated
three-limb core design, and another with a shell-type core [17]. that the system-level impact of the GIC field test would be
DC injection during the three tests lasted for 2, 1, and 2 minutes, minimal and would not affect the continuity or quality of the
respectively. These short durations led to limited insights regard- power supply. This, among other factors, provided confidence
ing the thermal effects of GIC on the tested transformers due in the performance of the field test.
to the relatively longer thermal time constants of transformers The rest of the paper is organized as follows. Section II
[18]. The most recently documented large-scale GIC field test describes the GIC field test and Section III presents the analysis
on a grid-connected power transformer occurred in Finland of the measured parameters. Details about the development
before 2002 [19]. This test was simultaneously performed on two and validation of the thermal and EMT models are reported in
unloaded 400/120 kV, 400 MVA, three-phase transformers by Sections IV and V, along with simulation results. Conclusions
connecting a DC source to their common neutral points. This was are presented in Section VI.
a long-duration test involving DC injection for up to 30 minutes
until the tested transformers reached saturation equilibrium. In II. GIC FIELD TEST
the 20 years since this test was completed, transformers have
The substation where the GIC field test was performed has two
had a number of internal design modifications, and the results
500/230 kV transformers, TX1 and TX2, connected in parallel.
of the test may no longer be applicable.
The GIC test was performed on TX1 while TX2 continued
This study analyzed the measurements collected in the fall of
to serve load. Both transformers are grounded, wye-connected
2022 during a large-scale GIC field test performed by Dominion
transformers made up of banks of three single-phase autotrans-
Energy Virginia (DEV), a U.S. electric utility, in collaboration
formers, each rated at 168 MVA (for Oil-Natural-Air-Natural
with Royal SMIT Transformers, a manufacturer in The Nether-
rating) and 280 MVA (for Oil-Natural-Air-Forced rating). Fig. 1
lands. The field test, initially proposed in 2013, was performed at
shows an aerial view of the three single-phase autotransformers
a 500/230 kV substation and is the first-ever, long-duration test
that make up TX1: TX1A, TX1B and TX1C.
in the U.S. to capture both the electrical and thermal effects
of GIC on large high-voltage power transformers. The GIC
field test adds to the body of knowledge because of the unique A. GIC Field Test Setup
independent clamping design of the tested transformers’ which A back-to-back (B2B) configuration for two out of the three
avoids using tie bars in the core structure and instead utilizes single-phase autotransformers was used for the GIC field test,
winding clamps with tie rods situated farther away from the as shown in Fig. 2. This is a typical configuration used by
core. Thus, the test provided value to reveal the GIC withstand transformer manufacturers to perform no-load DC injection tests
capabilities of these transformer designs. Moreover, using mea- in the factory. This configuration allows for the simultaneous
surement data from the tested transformers, Dominion was able testing of two similar transformers because the secondary sides

Authorized licensed use limited to: Chaitanya Bharathi Institute of Tech - HYDERABAD. Downloaded on June 24,2025 at 03:54:51 UTC from IEEE Xplore. Restrictions apply.
 ADEMOLA et al.: GEOMAGNETICALLY INDUCED CURRENT FIELD TEST ON LARGE GRID-CONNECTED POWER TRANSFORMERS 449

considered as saturation equilibrium [20]. However, the long du-

ration DC injection in this test allowed the transformer structural
parts to attain thermal equilibrium so that their final temperature
in response to DC magnitudes can be measured.
For multiple DC injection tests that were performed on the
same day, the injection periods were separated by a minimum
two-hour rest period with zero DC input. This allowed the trans-
formers to effectively cool down before the next DC injection.
Thus, each injection period was considered to be independent
[21]. Measurements collected during the GIC test included trans-
former temperature, transformer sound, ambient temperature,
dissolved gas analysis data, and current and voltage measure-
ments. In the transformer, temperatures were measured at the
windings, core legs, yoke clamping plate, winding clamping
plates, top and bottom support, tie rods, and tank. The top and
bottom oil temperatures were also collected. To continuously
monitor harmonics at the test location up to the 15th order,
Fig. 2. GIC field test B2B transformer configuration. a power quality meter was installed at Bus 1. Digital fault
recorders were activated at regular intervals to capture high-
resolution current and voltage readings at selected substations
in the vicinity of the test location.
are connected in series opposition, that is, the induced AC
electromagnetic field at the secondary windings cancel out each
other. Thus, they can serve as a test circuit for the injection of III. ANALYSIS OF FIELD TEST MEASUREMENTS
DC.
This section presents the analyses of some of the measure-
HA , HB , and HC in the figure are the primary terminals and
ments including transformer hotspot temperature, reactive and
XA , XB and XC are the secondary terminals of the autotrans-
non-active power demand, current and voltage harmonics, and
formers. For the test, HB and HC were disconnected from their
voltage drop at the test substation.
primary bus, but HB was looped to HA , thereby effectively
connecting TX1A and TX1B to a single phase (Phase A) of
Bus 1. Meanwhile, XC was disconnected from its secondary bus A. Transformer Hotspot Temperature
to completely isolate TX1C. Forty winding turns of TX1A and
Measurements from the transformers’ thermal sensors were
TX1B (out of a total of 1141) were separated to make windings
used to locate the structural part with the highest temperature and
YA and YB, which formed the test circuit. Thus, the initial phase
the magnitude of temperature rise in response to the injected DC
voltage at the H-terminals was 525 kV/3 and the Y-terminals
currents. The direct mechanism of transformer heating during
was 9.78 kV. A DC voltage source was placed in the test circuit
GIC injection is by the leakage fluxes that escape the core
and its voltage was varied until the desired DC current flowed
path during saturation. These fluxes generate high eddy-current
in the test circuit. The galvanic isolation of the test circuit from
losses in any metallic material in the core vicinity, thereby caus-
the primary windings prevented DC from flowing into the grid.
ing it to heat up significantly. Typically, the metallic structural
parts closest to the core are the tie bars or flitch plates, thus,
B. GIC Injection
they usually have the highest temperature during GIC injection
Being quasi-DC, the GIC was replicated by injecting DC [22], [23]. However, the tested transformers do not have tie bars
current into the ad-hoc tertiary windings of the transformers but use winding clamps and tie rods instead as shown in Fig. 3.
during the field test, similar to the previous field test in [19]. The advantage of this transformer design is that the metallic
The DC voltage source was controlled to ramp up the DC current parts prone to hotspot heating have been situated farther away
injection at 10 Adc /s until the current reached one of five target from the core thereby reducing stray fluxes that reach them and
current magnitudes injected separately: 292 A, 584 A, 875 A, consequently reduce their temperature rise during GIC injection.
1167 A, and 1459 A. From the primary windings’ perspective, Analysis of the field test results showed that the top of the
these are roughly equivalent to 10 A, 20 A, 30 A, 40 A and inner low-voltage (LV) windings of the transformers had the
50 A based on the winding ratio between the primary and test highest temperature rise. Fig. 4 presents the temperature rise
windings. Both sets of currents will create similar DC offsets in above oil of these inner windings (as defined in IEEE C57.91-
the transformer core flux. 2011 [32]) for each injected DC magnitude. For conventional
Each level of DC was injected for 2 hours, unlike the previous transformers that have tie-bars, the most critical structural part
GIC field test in the U.S. that lasted for a maximum of 2 minutes. are the tie-bars which are directly affected by flux exiting the
Typically, the offset DC flux in the core stops increasing within core. In contrast, the core legs in the tested transformers’ design
minutes of DC injection when the magnetizing current’s DC have no tie-bars. Therefore, their most critical structural parts
component becomes equal to the injected current; this can be were the yoke plates, and to a lesser extent, the winding clamps.

Authorized licensed use limited to: Chaitanya Bharathi Institute of Tech - HYDERABAD. Downloaded on June 24,2025 at 03:54:51 UTC from IEEE Xplore. Restrictions apply.
 450 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 40, NO. 1, FEBRUARY 2025

Fig. 5. Reactive and non-active power drawn by TX1A against injected DC

current.

Fig. 3. Internal design of the transformers tested.

expected, Q, which only accounts for the displacement between

fundamental currents and voltages, was smaller than NA, which
comprised all power that was non-useful due to displacement
and distortion of the current and voltage waveforms. An obvious
linear relationship was also observed.
The general relationship between NA or Q and the DC current
magnitude was estimated using the well-known k-factor formula
given by:

N A = kN A × Idc × Vrms,ll (1)

Q = kQ × Idc × Vrms,ll (2)

Where Idc is DC current in amperes and Vrms,ll is the line-

Fig. 4. Temperature rise of the transformer inner windings and structural parts
for various magnitudes of DC injection. to-line RMS voltage in kilovolts at the same voltage level as the
DC injected.
Analysis of the measurement data showed that for the trans-
formers kNA was 1.255 and kQ was 0.782. The kNA value, pro-
For comparison, the temperature rise of these structural parts
vided by the transformer manufacturer from their DC injection
were also included in Fig. 4.
factory test, was 1.2. This shows that factory test results can pro-
The figure shows that the inner winding temperature rise
vide good estimates of the additional reactive/non-active power
was higher than that of the structural parts. This is a unique
drawn from the grid by a transformer under the influence of GIC.
observation since previous works on conventional transformers
A slight variant of the k-factor (popularly used in commercial
assert that the structural parts, such as the tie-bar or yoke plates,
software like PowerWorld) is the K-factor, which is calculated by
experience larger and faster temperature rise than windings
setting Idc and Vrms,ll to per-unit values relative to the peak base
[18], [24]. Thus, expected transformer thermal response to GIC
current and RMS base voltage, respectively [26]. For the tested
should not be generalized for transformer designs like those
transformers, KQ was found to be 1.928, which is larger than
tested. All other transformer parts including tie rods, top and
the default value of 1.8 that PowerWorld uses for transformers
bottom supports, tank wall shielding, and tank cover, did not
with similar core designs. Therefore, using the factory test’s
show any additional heating, but rather just followed the local
k-factor will underestimate the actual GIC-induced reactive and
oil temperature. Moreover, the core temperature rise caused by
non-active powers by about 4.4%, while using PowerWorld’s
the slight increases in no-load loss were negligible. Dissolved
default k-factor will lead to a 6.6% underestimation. These are
gas analysis during the test also did not show any increase.
acceptable errors, considering the difficulties of performing GIC
field tests to obtain accurate k-factors, as well as the variability
B. Reactive Power and Total Non-Active Power
of the k-factors based on transformer type, internal design, and
Since the current and voltage during the field test contained age [13].
significant harmonics, measurement data was decomposed into It is worth mentioning that the potential effect of the trans-
harmonic components using fast Fourier transforms. The funda- formers’ reactive power consumption on system voltage was a
mental frequency reactive power (Q) and total non-active power major concern prior to the GIC test. However, the maximum
(NA) were calculated as recommended in IEEE Std. 1459-2010 voltage drop observed during the test was less than 1% for the
[25]. Fig. 5 presents Q and NA drawn by TX1A during the GIC largest DC injection. This small voltage drop had no noticeable
field test. (Note that these values were similar for TX1B.) As effect on grid operation during the test.

Authorized licensed use limited to: Chaitanya Bharathi Institute of Tech - HYDERABAD. Downloaded on June 24,2025 at 03:54:51 UTC from IEEE Xplore. Restrictions apply.
 ADEMOLA et al.: GEOMAGNETICALLY INDUCED CURRENT FIELD TEST ON LARGE GRID-CONNECTED POWER TRANSFORMERS 451

TABLE I
CURRENT AND VOLTAGE DISTORTIONS AT PHASE A DURING THE GIC FIELD
TEST

Fig. 6. Current drawn by the test transformers for DC injection of 50 Adc.

Fig. 8. Neighboring substations to the test location.

Fig. 7. Frequency content of the transformers’ RMS currents and total Phase A
RMS current during one DC injection test.

C. Current and Voltage Distortion

The currents drawn by TX1A and TX1B, along with the total
current drawn from Phase A during the 50 Adc injection, are
presented in Fig. 6.
Due to the B2B configuration, the transformers saturated in
opposite directions and at different half-cycles; this observation
was previously explained in [21] and was also observed in the
previous GIC field test in Finland [19]. Thus, the peak of the total Fig. 9. Percentage point increase in THDv at nearby buses to the test location.
current drawn from Phase A is less than the sum of the current
peaks of TX1A and TX1B. Fig. 7 shows the frequency content of
the current from TX1A and TX1B and the total currents through during normal operating conditions. On the other hand, THDv is
Phase A to the grid during the same DC injection period. While the ratio of total RMS harmonic voltage and the fundamental
both transformers produced significant even and odd harmonics, frequency voltage. The table shows significant harmonics in
the total Phase A current is devoid of even harmonics. The both current and voltage during the DC injection tests. In fact,
B2B configuration made the even harmonics cancel out, as THDv for the 20 A DC injection and above exceeded the 1.5%
the transformers produced identical and symmetrical current limit recommended by IEEE Std. 519. Note that the IEEE Std.
waveforms, shifted by half-a-period. 519 limits were specified for steady-state operation, so its ap-
It is worth noting that the harmonic cancellation capabilities of plication in transient GMD scenarios is conservative. Moreover,
nearby STATCOMs were disabled to prevent them from masking background harmonics of about 0.92% THDv already existed at
the extent of harmonic propagation during the field test. Bus 1 prior to the GIC field test. Thus, the test was not solely
Table I presents the total demand distortion (TDDi) of the responsible for the observed THDv.
Phase A current and the total harmonic distortion (THDv) of Bus THDv was also monitored at seven neighboring substations
1 voltage for different DC injections. As defined in IEEE Std. at varying distances away from Bus 1 (see Fig. 8).
519, TDDi was calculated as the ratio of the total RMS harmonic Fig. 9 shows significant increases in THDv at these buses
current and the maximum load demand current at Phase A of up to 0.67 percentage points, which illustrates the ability of

Authorized licensed use limited to: Chaitanya Bharathi Institute of Tech - HYDERABAD. Downloaded on June 24,2025 at 03:54:51 UTC from IEEE Xplore. Restrictions apply.
 452 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 40, NO. 1, FEBRUARY 2025

Fig. 10. Block diagram for the ETTM software to create thermal models of
transformers (modified from [27]).

Fig. 11. Comparison of field measurements and simulation results to validate

harmonics to flow through transmission line impedances over transformer thermal model.
considerable distances. During a real GMD event where several
transformers simultaneously produce harmonics, the harmonics
may amplify themselves enough to cause severe grid voltage
distortions. Comparing Figs. 8 and 9, there is no correlation
between distance from the harmonic source and rise in harmonic
voltage. This is because harmonic propagation depends on the
harmonic impedance of the transmission circuit including its
resonance characteristics, which can lead to unexpected voltage
distortions in different portions of the grid. This underscores the
need for wide-area harmonic-flow analysis to accurately assess
propagation of harmonics during GMD scenarios.

IV. THERMAL MODEL CREATION AND SIMULATION USING

FIELD TEST MEASUREMENTS
A. Model Creation and Validation
Fig. 12. GIC waveform for supplemental GMD scenario and a scaled-up GIC.
To leverage the field measurements, EPRI Transformer Ther-
mal Modeling (ETTM) software was used to create a thermal
model. ETTM creates a thermal transfer function by using the
to the actual transformers. This may impact the accuracy of the
transformer’s thermal response to DC step functions, such as the
estimated peak temperatures for short-duration GIC peaks.
hotspot temperature measurements. The underlying concept of
ETTM’s operation is detailed in [27]. As shown in Fig. 10, the
B. Simulation of Transformer Thermal Response to GIC
inputs to the software are the transformer’s asymptotic temper-
Under Full Load
ature rise (ΔTasymp ) for DC injections of different magnitudes
(i.e., the IGIC vs. ΔTasymp data), along with the time-series data The validated thermal model was used to simulate the trans-
of the temperature rise for one DC injection (i.e., the ΔT vs. t former’s response to a GIC waveform with a peak of about
data). 85 A/phase (see Fig. 12) with a focus on the transformer’s
For this study, the measurements were obtained from the field hottest spot, that is, the inner windings. The waveform was
test and used as inputs for ETTM. The inner winding temperature obtained from the GMD model developed in [28]; but instead of
measurements for DC injections of 10 A, 20 A, and 40 A were using a peak geoelectric field value, the time-series data of the
used to create a thermal model. Then the model was used to geoelectric field (obtained from [29]) for the supplemental GMD
predict the transformer’s thermal response to DC injection of 30 scenario was used. Note that NERC created the supplemental
A and 50 A, to validate the modeling approach. GMD scenario with geoelectric field strength of 12 V/km as an
Fig. 11 compares the simulated thermal response and the field extension of the benchmark GMD scenario. The supplemental
test measurements of these injection magnitudes. The simulation GMD models localized electric field enhancement due to spatial
was able to predict with reasonable accuracy the winding tem- complexities of high-latitude geomagnetic fields [30]. A scaled-
perature rise above oil and its trajectory for unknown DC mag- up GIC waveform having about 200 A/phase (that is, 600 A)
nitudes. This result validated the thermal model’s performance peak was also considered. This GIC peak is almost twice the
for subsequent evaluation of thermal response to untested GIC highest recorded GIC reading of 330 A in the world [31].
waveforms. It is worth noting that the trajectories of the field and The thermal response calculation also assumed a conservative
simulated thermal responses are slightly different; specifically, oil temperature of 105 °C, which is the recommended limit for
the thermal model has a slightly longer time constant compared normal life expectancy loading of transformers according to

Authorized licensed use limited to: Chaitanya Bharathi Institute of Tech - HYDERABAD. Downloaded on June 24,2025 at 03:54:51 UTC from IEEE Xplore. Restrictions apply.
 ADEMOLA et al.: GEOMAGNETICALLY INDUCED CURRENT FIELD TEST ON LARGE GRID-CONNECTED POWER TRANSFORMERS 453

Fig. 14. B-H trajectory for simulated 50A/ph DC injection before and after
Fig. 13. Thermal responses of the tested transformers to GIC. tuning the transformer model.

IEEE Std. C57.91 [32]. The oil temperature is typically associ- field intensity (H) of the model, both initially and after tuning.
ated with transformer load due to the long thermal time constant The figure shows that the initial saturation properties were
of transformer oil. Thus, incorporating the oil temperature into underestimated; the tuned transformer model exhibits a more
the simulation made it possible to leverage the no-load data severe slope in the saturation region.
obtained from the field test to estimate the transformer’s thermal A reason for the inaccuracies of the initial model may be
response under full load conditions. because of the severe voltage and current distortions experienced
The simulations do not account for variability in the hotspots’ as the transformer began to saturate during the factory test where
cooling rate (due to viscosity changes of the bulk oil in the a strong transmission grid was unavailable. Moreover, the largest
transformer tank) under different loading conditions. IEEE Std. DC injected into the transformer during the factory test was 5.33
C57.91 highlights that oil cools less effectively during no-load ADC due to the temperature limitations of the power supply.
conditions, compared to full -load conditions [32]. A study in Since saturation characteristics of transformers are non-linear
[22] that considered oil viscosity effects also demonstrated that in nature, the initial V-I datapoints may not be accurate for the
temperature rise during no-load conditions could be up to 17.2% deeper levels of saturation achieved during the field test where
greater than during full-load conditions. Based on this, it was up to 50 ADC was injected. This further accentuates the need
understood that the proposed method to superimpose the no-load for this GIC field test to correctly calculate relevant transformer
temperature rise over the expected full-load top-oil temperature properties for subsequent modeling efforts.
would overestimate the actual hotspot temperature at full load. To perform EMT simulations, a reduced 11-bus equivalent
Nonetheless, in vulnerability studies, a conservative approach is of the DEV network was developed in EMTDC/PSCAD to
generally preferable. include the field test location and the portion of the system
With the several layers of conservative assumptions, the one bus away. Therefore, the network equivalent included the
slightly longer time constant of the thermal model is well com- 500 kV and 230 kV buses at the test substation along with
pensated. Fig. 13 shows that for both GIC waveforms, the hotspot neighboring substations represented by six 500 kV and three
temperature of the tested transformer will not reach the 180°C 230 kV buses. The development process and validation of the 11-
short-term emergency limit specified for transformer windings bus network-equivalent model for the test location and nearby
in IEEE Std. C57.91 [32]. These results reveal the significant buses is detailed in our previous work [21].
capability of the tested transformers to withstand GIC. It is External boundary conditions were replicated using equiva-
unlikely that these transformers will suffer thermal damage due lent-voltage sources, and the transmission lines were modelled
to GIC even during extreme GMDs. using the frequency-dependent models in PSCAD. The B2B
configuration of TX1A and TX1B were replicated in PSCAD
V. EMT MODEL CREATION AND SIMULATION USING FIELD using single-phase, unified magnetic-equivalent circuit (UMEC)
TEST MEASUREMENTS transformer models. Both transformers were modeled with the
following parameters: a leakage reactance of 0.367 p.u. for the
A. Model Creation and Validation
B2B configuration and 0.1418 p.u. for the normal load-carrying
To create an initial EMT model for our field-tested trans- configuration; no-load losses of 0.00018 p.u.; copper losses of
formers, the saturation V-I datapoints obtained from the manu- 0.00145 p.u.; core aspect ratios of 0.71 and 1.0 for yoke-to-
facturer’s factory DC injection test was used. Subsequently, the winding limb length and area, respectively.
model’s saturation V-I datapoints were tuned until the current A DC voltage source was placed in the path of the test
drawn by the transformer in the simulation environment matched windings to simulate DC injection and was controlled by a ramp
the field test measurements for the same DC injection. Fig. 14 signal of 2 kVdc /s (the same as in the field test) until the DC
compares the magnetic flux density (B) against the magnetic current in the primary winding reached the desired magnitudes.

Authorized licensed use limited to: Chaitanya Bharathi Institute of Tech - HYDERABAD. Downloaded on June 24,2025 at 03:54:51 UTC from IEEE Xplore. Restrictions apply.
 454 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 40, NO. 1, FEBRUARY 2025

Fig. 15. Comparison of field measurements and EMT simulation for trans-
Fig. 17. Comparison of current frequency content in B2B and 3Ph transformer
former current during 50Adc injection.
configurations.

conditions. DC injection was achieved by connecting a DC

voltage source to the transformer HV neutral. The voltage source
was controlled using a PI controller with a proportionality gain of
4 and an integral time constant of 0.5. At no-load, the response of
the 3-ph and B2B configurations were compared. The polarity of
the current at saturation was the same for the 3-ph transformers
since DC was injected in their common neutral winding, unlike
the B2B configuration where the direction of the DC current
flow was opposite.
Fig. 17 compares the frequency content of the HV winding
current for both configurations. The 3-ph configuration allowed
both even and odd harmonics into the grid with decreasing har-
Fig. 16. Comparison of field test and EMT simulations for reactive power and
monic magnitudes as harmonic order increases, unlike the B2B
non-active power drawn by transformers during DC injection tests. configuration which only had odd harmonics. From a testing
perspective, this is an advantage for the B2B configuration,
since odd harmonics are more common in the grid; hence,
Unlike in the field test, DC injection in the simulation was sensitive grid devices are typically designed to be more resistant
run for about 120 seconds to allow transformer saturation to to them. The 3-ph configuration also allowed DC current flow
reach an equilibrium state. The tuned currents, as well as the into the grid; because injecting current through the neutral into
non-active and reactive power consumed during the field test and the HV windings may have the adverse effect of saturating
simulation, were compared (see Figs. 15 and 16). The simulation other transformers nearby. From the grid perspective, the B2B
accurately calculated the current, reactive power, and non-active configuration caused higher odd harmonics in phase A than the
power drawn by the transformers for the same DC injection, thus, 3-ph configuration. This is because of the addition of the odd
validating the transformer EMT model. harmonics from the two transformers connected B2B to a single
phase, while only one transformer is connected to each phase in
the 3-ph configuration.
B. Simulation of Transformer Electrical Response to GIC The 3-ph transformer was loaded by connecting its secondary-
Under Loaded Conditions side terminals to the 230 kV bus at the test substation in the EMT
The validated EMT transformer model was used to investigate simulation environment. The phase angles of the voltage sources
GIC electrical effects on the tested transformers under various at the downstream 230 kV buses were varied to change the power
load conditions. Three copies of the UMEC transformers were flowing through the transformers, thereby representing various
connected in a wye-wye configuration to form a three-phase operating conditions. The transformer operates in the step-down
(3-ph) transformer in normal load-carrying configuration. The mode, i.e., direction of real power flow is from the 500 kV to
phase voltages at the secondary side of the transformers were the 230 kV side. Table II presents the power flow conditions
changed to 132.79 kV (or 230 kVll ) to connect to the 230 kV including voltage (V), real power (P) and reactive power (Q)
side of the system. The leakage reactance was also updated to at the transformer HV and LV sides before GIC injection. 50
account for the change to normal operating configuration. A/ph DC was injected through the neutral to the HV windings.
DC current of 150 A was injected into the high-voltage (HV) Table III presents the steady-state power flow conditions after
neutral. This is equivalent to 50 A/phase, assuming balanced GIC injection.

Authorized licensed use limited to: Chaitanya Bharathi Institute of Tech - HYDERABAD. Downloaded on June 24,2025 at 03:54:51 UTC from IEEE Xplore. Restrictions apply.
 ADEMOLA et al.: GEOMAGNETICALLY INDUCED CURRENT FIELD TEST ON LARGE GRID-CONNECTED POWER TRANSFORMERS 455

TABLE II
POWER FLOW CONDITIONS AT TRANSFORMER TERMINALS BEFORE GIC
INJECTION

TABLE III
POWER FLOW CONDITIONS AT TRANSFORMER TERMINALS AFTER 50 A/PH
GIC INJECTION

TABLE IV
GIC-INDUCED REACTIVE POWER LOSSES FOR DIFFERENT LOADING SCENARIOS

Fig. 18. Transformer magnetizing current during DC injection for various load
conditions.

TABLE V
TRANSFORMER FLUX DUE TO GIC-INDUCED SATURATION FOR DIFFERENT
LOAD SCENARIOS

Using Table II data, the reactive power losses due to load

before the GIC injection, Qloss,load , was calculated as the dif-
ference between the reactive power flow from the HV to the
LV buses for each load scenarios. Similarly, the transformer’s
total reactive power losses after GIC injection, Qloss,total , was
calculated using Table III data. ΔQGIC was calculated as the dif-
ference between Qloss,total and Qloss,load . This is the portion of
Qloss,total assumed to be solely dependent on the GIC injection.
Qloss,load , Qloss,total , and ΔQGIC are presented in Table IV. The
table shows that with increasing load, Qloss,load increased sig-
nificantly and the contribution of ΔQGIC to Qloss,total reduced. Consequently, the peak of the final AC flux after GIC injection
Table IV indicates that ΔQGIC does not change significantly also decreased for lower terminal voltages. Thus, the excursion
due to transformer load or power flow for the same GIC magni- of the flux peak into the transformer saturation region, which
tude. This observation has also been noted in previous literature affects the magnetizing current and ΔQGIC , was also slightly
[5], [33]. The slight reduction in ΔQGIC was attributed to the lower when terminal voltages reduced. This demonstrates that
decrease in the peak of transformer magnetizing current as ΔQGIC is a function of AC core flux and transformer voltages,
shown in Fig. 18. From Table II, the increase in reactive power which can be influenced by transformer load or power flow
demand for the considered load scenarios caused a reduction in conditions depending on transformer design, power factor, and
terminal voltages even before GIC injection. The lower voltages direction of energy flow (e.g., step-up, or step-down). These
resulted in reduced AC flux in the transformers, as shown in findings are consistent with those in [20], [34], [35].
Table V, where the peak and peak-to-peak values of the AC flux For large system GMD studies, ΔQGIC is typically calculated
before GIC injection decreased with lower terminal voltages. using the well-known k-factor method in (2) which considers

Authorized licensed use limited to: Chaitanya Bharathi Institute of Tech - HYDERABAD. Downloaded on June 24,2025 at 03:54:51 UTC from IEEE Xplore. Restrictions apply.
 456 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 40, NO. 1, FEBRUARY 2025

TABLE VII
LOAD EFFECTS ON HARMONIC PRODUCTION OF TRANSFORMERS FOR 50 Adc
INJECTION

procedures during GMD events, these results suggest that

Fig. 19. Effect of terminal voltage on transformer saturation for the same GIC there’s a significant advantage to reducing load on transformers
magnitude. in the system that are vulnerable to GIC. For instance, Table IV
shows that a load reduction for the tested transformers from
TABLE VI 75% to 50% reduced Qloss,load by 55.6% and increased ΔQgic
COMPARISON OF THE K-FACTOR METHOD TO EMT SIMULATION WHEN by merely 1.4%; the total Qloss,total reduction was 30%. This
CALCULATING REACTIVE POWER LOSS FOR A 50Adc INJECTION could be significant during an extreme GMD event. Table VII
suggests that the same load reduction may cause an increase
of 0.67% in harmonic current and 0.72% in harmonic voltage.
However, these increases are small and should not lead to
significantly higher equipment heating or protection system
misoperation caused by harmonics.

VI. CONCLUSION
This paper focused on the first-ever, long-duration GIC field
the HV-side terminal voltage and GIC magnitude. The equation test performed on high-voltage, grid-connected power trans-
assumes that, for the same GIC magnitude, the relationship formers in the U.S. and the first test globally in over 20 years.
between ΔQGIC and terminal voltage is linear with a slope equal The field test was designed to replicate GIC flow and observe
to the k-factor. However, Fig. 19 shows that the relationship its impact on transformers with a unique internal core design
between terminal voltage and ΔQgic is slightly non-linear for that uses winding clamps and tie rods instead of tie bars on
the same GIC magnitude in this study. This is because core flux the core legs. This test allowed for greater insight into how
is not completely determined by either the HV or LV terminal these transformers would operate during a GMD event, and to
voltages for a transmission transformer. This may introduce physically verify GIC impacts on the grid.
calculation errors when using the k-factor method. Moreover, While conventional transformers typically have their hotspots
a k-factor that is set using no-load GIC tests may not sufficiently at the tie bars or yoke plates, analysis of the field measurements
consider possible load effects. revealed that the hotspot for the test transformers was at the top
To verify, the Qloss,total from the EMT simulation was com- of inner LV windings. This implies that expected transformer
pared with the Qloss,total calculated using the k-factor method as thermal response to GIC should not be generalized for trans-
shown in Table VI for the same IDC value of 50 A/ph. Given that former designs like those tested. The field test also allowed for
previous analysis in Section III-B obtained a k-factor of 0.782 the determination of the transformers’ k-factor which was close
for the tested transformers, the table shows that the error could to the generic k-factors and those obtained from prior factory
be as high as 8.4%, even in a situation where the transformer’s tests. This supports the use of generic k-factors where physically
k-factor is known with high certainty. This finding supports the verified k-factors are not available, especially for large system
unproved statement in [13] that the accuracy of k-values may GMD studies with hundreds of transformers.
depend on loading conditions even for the same transformer As expected, the transformers produced even and odd harmon-
type. Moreover, a prior study showed that the k-factor method ics during the test, but the even harmonics were cancelled out on
could lead to errors as high as 13% [36]. In that case, the error the grid side because of the B2B connection of the transformers.
rate was only for different electric field/GIC values, not for Nevertheless, considerable current and voltage distortions were
varying power flow conditions as is the case here. Nevertheless, observed at the test substation above the IEEE Std. 519 har-
the use of the k-factor is presently unavoidable for large system monic limits for normal operating conditions. Analysis revealed
GMD studies, and Table VI indicates that the results are mostly that background voltage harmonics contributed to the observed
conservative. distortion levels. Higher voltage distortions were also observed
As the magnetizing current peak reduced for the higher load at locations several kilometers away from the test location. The
scenarios, harmonic production also reduced, as shown in observed harmonic propagation supports the need for wide-area
Table VII. When considering recommended operating harmonic flow analysis to reveal potential amplifications or

Authorized licensed use limited to: Chaitanya Bharathi Institute of Tech - HYDERABAD. Downloaded on June 24,2025 at 03:54:51 UTC from IEEE Xplore. Restrictions apply.
 ADEMOLA et al.: GEOMAGNETICALLY INDUCED CURRENT FIELD TEST ON LARGE GRID-CONNECTED POWER TRANSFORMERS 457

cancellations when multiple transformers are simultaneously [5] V. Albertson, J. Kappenman, N. Mohan, and G. Skarbakka, “Load-
producing harmonics during a real GMD event. flow studies in the presence of geomagnetically-induced currents,” IEEE
Trans. Power App. Syst., vol. PAS-100, no. 2, pp. 594–607, Feb. 1981,
The field measurements also provided valuable data for de- doi: 10.1109/tpas.1981.316916.
veloping a thermal model for the tested transformers. Thermal [6] S. Guillon, P. Toner, L. Gibson, and D. Boteler, “A colorful blackout:
simulations demonstrated that the hotspot temperature rise of The havoc caused by auroral electrojet generated magnetic field vari-
ations in 1989,” IEEE Power Energy Mag., vol. 14, no. 6, pp. 59–71,
the tested transformers would be lower than IEEE Std. C57.91 Nov./Dec. 2016, doi: 10.1109/mpe.2016.2591760.
limits even for a GIC waveform with 200 A/phase peak. This [7] D. H. Boteler, “A 21st century view of the march 1989 magnetic
suggests that the tested transformers have high GIC withstand storm,” Space Weather, vol. 17, no. 10, pp. 1427–1441, Oct. 2019,
doi: 10.1029/2019sw002278.
capabilities to withstand realistic GMD scenarios. The measure- [8] M. Wik, A. Viljanen, R. Pirjola, A. Pulkkinen, P. Wintoft, and H. Lundstedt,
ments also aided in fine-tuning the saturation characteristics “Calculation of geomagnetically induced currents in the 400 kV power grid
of the transformer model. Subsequent EMT simulations con- in southern Sweden,” Space Weather, vol. 6, no. 7, Jul. 2008, Art. no. 2659,
doi: 10.1029/2007sw000343.
firmed that change in power flowing through the transformer [9] M. Wik, R. Pirjola, H. Lundstedt, A. Viljanen, P. Wintoft, and A. Pulkkinen,
can vary AC core flux which will directly impact GIC-induced “Space weather events in July 1982 and October 2003 and the effects of
reactive power demand and harmonics. Analysis further showed geomagnetically induced currents on swedish technical systems,” Annales
Geophysicae, vol. 27, no. 4, pp. 1775–1787, Apr. 2009, doi: 10.5194/an-
that transformer load reduction during extreme GMD events is geo-27-1775-2009.
beneficial considering the relatively large contribution of load [10] “3002017707 - Assessment guide: Geomagnetic disturbance harmonic im-
to the total reactive power losses. pacts and asset withstand capabilities,” Electric Power Research Institute,
Palo Alto, CA, USA, Tech. Rep. 3002017707, 2019.
It is worth mentioning that the tested transformers have [11] X. Wu et al., “Diagnosis of DC bias in power transformers using vibration
been successfully recommissioned, returned to service, and are feature extraction and a pattern recognition method,” Energies, vol. 11,
currently operating without issues. This paper underscores the no. 7, Jul. 2018, Art. no. 1775, doi: 10.3390/en11071775.
[12] “3002018769 - Geomagnetically induced current harmonic tool
benefits of the GIC field test to enhance model development and (GICharm): GICharm version 2.0,” Electric Power Research Institute, Palo
validation, while fostering better understanding of GIC effects Alto, CA, USA, Tech. Rep. 3002017447, 2020.
on contemporary power transformer designs. [13] K. Shetye and T. Overbye, “Modeling and analysis of GMD effects
on power systems: An overview of the impact on large-scale power
systems,” IEEE Electrific. Mag., vol. 3, no. 4, pp. 13–21, Dec. 2015,
ACKNOWLEDGMENT doi: 10.1109/mele.2015.2480356.
[14] A. Rezaei-Zare, J. Mahseredjian, and A. Haddadi, “3002014059 - Ge-
Special thanks to Dominion Energy’s Electric Transmission omagnetic disturbance (GMD) transformer thermal analysis tool: EPRI
division, encompassing our dedicated engineers, field techni- transformer thermal model (ETTM),” ResearchGate, Berlin, Germany,
Tech. Rep. 3002014059, 2018, doi: 10.13140/RG.2.2.26826.62400.
cians, and operators, whose unwavering support was instrumen- [15] H. K. Chisepo, C. T. Gaunt, and L. D. Borrill, “Measurement and FEM
tal in bringing this field test to fruition. We appreciate PJM, analysis of DC/GIC effects on transformer magnetization parameters,”
our reliability coordinator and transmission planner, for their in Proc. 2019 IEEE Milan PowerTech, Milan, Italy, 2019, pp. 1–6,
doi: 10.1109/ptc.2019.8810423.
collaborative efforts both before and during the test period. [16] IEEE Guide for Establishing Power Transformer Capability while under
Royal SMIT Transformers B.V. and their proficient engineers Geomagnetic Disturbances, IEEE Standard C57.163-2015, 2015.
also have our gratitude for generously sharing their expertise [17] J. Kappenman, Mitigation of Geomagnetically Induced and DC Stray
Currents. Duluth, MN, USA: Minnesota Power and Light, 1983.
along with pertinent data and models that were crucial in plan- [18] “3002017708 - Transformer thermal impact assessments for DC withstand
ning and executing the field test. We are thankful to Resilient capability: Examining the impacts of geomagnetically induced current
Power System LLC for providing the essential DC test set and for (GIC) on transformer thermal performance,” Electric Power Research
Institute, Palo Alto, CA, USA, Tech. Rep. 3002017708, 2019.
their invaluable onsite assistance. Our sincere appreciation goes [19] M. Lahtinen and J. Elovaara, “GIC occurrences and GIC test for 400 kV
to Schweitzer Engineering Laboratories for their contribution of system transformer,” IEEE Trans. Power Del., vol. 17, no. 2, pp. 555–561,
equipment and expertise, enabling the seamless high-speed data Apr. 2002, doi: 10.1109/61.997938.
[20] E. E. Bernabeu, “Single-phase transformer harmonics produced during ge-
acquisition of transformer voltage and currents. omagnetic disturbances: Theory, modeling, and monitoring,” IEEE Trans.
Power Del., vol. 30, no. 3, pp. 1323–1330, Jun. 2015, doi: 10.1109/tp-
REFERENCES wrd.2014.2371927.
[21] A. A. Ademola, A. Pinceti, K. A. Vance, K. D. Jones, R. M. Gardner,
[1] NERC, “2012 special reliability assessment interim report: Effects and Y. Liu, “System impact assessment of a proposed geomagnetically-
of geomagnetic disturbances on the bulk power system,” Feb. 2012. induced current field test at a dominion energy virginia substation,” in Proc.
[Online]. Available: https://files.givewell.org/files/shallow/geomagnetic/ Grid Future Symp., Chicago, IL, 2022, Art. no. 15. [Online]. Available:
Geomagnetic_Disturbance_Task_Force_2012.pdf https://cigre-usnc.org/wp-content/uploads/2022/10/4D-1_Ademola.pdf
[2] Office of Risk Management and Analysis United States Department of [22] L. Marti, A. Rezaei-Zare, and A. Narang, “Simulation of transformer
Homeland Security, “Geomagnetic storms,” OECD/IFP Futures Project hotspot heating due to geomagnetically induced currents,” IEEE Trans.
on Future Global Shocks, Jan. 2011. [Online]. Available: https://www. Power Del., vol. 28, no. 1, pp. 320–327, Jan. 2013, doi: 10.1109/tp-
oecd.org/gov/risk/46891645.pdf wrd.2012.2224674.
[3] Government of Canada - Office of Critical Infrastructure Protection [23] R. S. Girgis and K. B. Vedante, “Impact of GICs on power transformers:
and Emergency Preparedness, “Threat analysis: Geomagnetic storms Overheating is not the real issue.,” IEEE Electrific. Mag., vol. 3, no. 4,
- reducing the threat to critical infrastructure in Canada,” Apr. 2002. pp. 8–12, Dec. 2015, doi: 10.1109/mele.2015.2480355.
[Online]. Available: http://www.epc-pcc.gc.ca/emergencies/other/TA02- [24] M. Akbari, M. Mostafaei, and A. Rezaei-Zare, “Estimation of hot-spot
001_E.html heating in OIP transformer bushings due to geomagnetically induced
[4] I. A. Erinmez, J. G. Kappenman, and W. A. Radasky, “Man- current,” IEEE Trans. Power Del., vol. 38, no. 2, pp. 1277–1285, Apr. 2023,
agement of the geomagnetically induced current risks on the na- doi: 10.1109/TPWRD.2022.3212322.
tional grid company’s electric power transmission system,” J. At- [25] IEEE Standard Definitions for the Measurement of Electric Power Quan-
mospheric Sol.-Terr. Phys., vol. 64, no. 5, pp. 743–756, Mar. 2002, tities Under Sinusoidal, Nonsinusoidal, Balanced, or Unbalanced Condi-
doi: https://doi.org/10.1016/S1364-6826(02)00036-6. tions, IEEE Standard 1459-2000, 2010.

Authorized licensed use limited to: Chaitanya Bharathi Institute of Tech - HYDERABAD. Downloaded on June 24,2025 at 03:54:51 UTC from IEEE Xplore. Restrictions apply.
 458 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 40, NO. 1, FEBRUARY 2025

[26] PowerWorld, “PowerWorld simulator GIC,” in Proc. PowerWorld Client [31] R. Pirjola and D. Boteler, “Geomagnetically induced currents in European
Conf., IL, 2014, pp. 15–17. high-voltage power systems,” in Proc. IEEE 2006 Can. Conf. Elect.
[27] A. Rezaei-Zare, A. Haddadi, and J. Mahseredjian, “Geomagnetic distur- Comput. Eng., 2006, pp. 1263–1266, doi: 10.1109/ccece.2006.277540.
bance (GMD) transformer thermal analysis tool: EPRI transformer thermal [32] IEEE Guide for Loading Mineral-Oil-Immersed Transformers and Step-
model (ETTM),” in “Technical Update” Electric Power Research Insti- Voltage Regulators, IEEE Standard C57.91-2011, 2011.
tute, 2018. [Online]. Available: https://www.researchgate.net/publication/ [33] J. Kappenman, V. Albertson, and N. Mohan, “Current transformer and
331245789 relay performance in the presence of geomagnetically-induced currents,”
[28] A. A. Ademola et al., “Sensitivities of geomagnetically induced currents IEEE Trans. Power App. Syst., vol. PAS- vol. 100, no. 3, pp. 1078–1088,
in dominion energy Virginia to the neighboring grids and transformer Mar. 1981, doi: 10.1109/tpas.1981.316574.
blocking schemes,” in Proc. 2023 IEEE Power Energy Soc. Gen. Meeting, [34] R. A. Walling and A. N. Khan, “Characteristics of transformer exciting-
Orlando, FL, 2023, pp. 1–5. current during geomagnetic disturbances,” IEEE Trans. Power Del., vol. 6,
[29] PowerWorld, “Transformer time-series for NERC benchmark GMD no. 4, pp. 1707–1714, Jan. 1991, doi: 10.1109/61.97710.
event and supplemental GMD event,” Oct. 2014. Accessed: May [35] J. Chen, C. Liu, M. Wang, and T. Wang, “Power system responses to geo-
25, 2023. [Online]. Available: https://www.powerworld.com/knowledge- magnetic disturbances recognized using phasor measurement recordings,”
base/transformer-time-series-for-nerc-benchmark-gmd-event Elect. Power Energy Syst., vol. 113, pp. 932–940, 2019.
[30] NERC, “Supplemental geomagnetic disturbance event description,” [36] A. Haddadi, R. Hassani, J. Mahseredjian, L. Gerin-Lajoie, and A. Rezaei-
Atlanta, GA, 2017, Accessed: Jan. 09, 2022. [Online]. Available: https:// Zare, “Evaluation of simulation methods for analysis of geomagnetic
www.nerc.com/pa/Stand/Project201303GeomagneticDisturbanceMitig- disturbance system impacts,” IEEE Trans. Power Del., vol. 36, no. 3,
ation/Supplemental_GMD_Event_Description_June_2017.pdf pp. 1509–1516, Jun. 1, 2021, doi: 10.1109/tpwrd.2020.3010195.

Authorized licensed use limited to: Chaitanya Bharathi Institute of Tech - HYDERABAD. Downloaded on June 24,2025 at 03:54:51 UTC from IEEE Xplore. Restrictions apply.