You are on page 1of 8

CIGR-719 2015 CIGR Canada Conference

21, rue dArtois, F-75008 PARIS


http : //www.cigre.org Winnipeg, Manitoba, August 31-September 2, 2015

Performance of single-phase core-type power transformers under influence of


geomagnetically induced currents analysis and factory testing

W.Ziomek, D.Boyd, K.Vijayan T.Kalicki


CG Power Systems Canada Inc Hydro One Networks Inc
Canada Canada

SUMMARY

The impact of geomagnetically induced currents on operation of power grid and power
apparatus especially transformers is in the recent years of increasing interest. Power
system regulatory organizations, power utilities, and equipment manufacturers are deeply
involved in collecting data, analyzing performance of the system and equipment.
This paper will describe fundamentals of geomagnetic disturbance, generation of GICs
and their impact on in-service transformers.
A calculation procedure used to define the performance of single phase power
transformers under GIC will be reviewed on example of large power transformer and a high
voltage autotransformer. The analysis allows calculating the part-cycle core saturation,
resulting current pulse train, related reactive power requirement, current harmonics, thermal
effects in core and windings, and sound level. These characteristics allow generating the
loading guide for a specific transformer under GIC.
The results of factory testing performed on two back-to-back connected units will be
described and compared to the analytical results.

KEYWORDS

Power transformer, geomagnetic disturbance, geomagnetically induced currents, transformer


performance characteristics

Main authors email: waldemar.ziomek@cgglobal.com


INTRODUCTION

The Earth's magnetic field forms the magnetosphere which protects us from most of the particles the
Sun emits with the solar wind and other solar disturbances. Geomagnetic disturbance (GMD), or
geomagnetic storm is an interval of time when an intense and long-lasting interplanetary convection
electric field leads, through a substantial energization in the magnetosphere-ionosphere system, to an
intensified Earth ring current which is sufficiently strong to rapidly decrease the Earth magnetic field,
driving it into high levels of negative polarity. Primary causes of the enhanced interplanetary electric
fields are: (i) Interplanetary Coronal Mass Injections (CME) and (ii) Corotating Interaction Regions
(CIR) produced by interaction between the fast solar-wind streams, emanating from coronal holes,
with the slow streams and characterized by high magnetic fields. The frequency of geomagnetic
storms increases and decreases with the sunspot cycle: CME driven storms are more common during
the maximum of the solar cycle while CIR driven storms are more common during the minimum of
the solar cycle [1-5]. A geomagnetic disturbance of sufficient intensity and duration changes the
Earths magnetic field thereby inducing atypical electric fields in the Earth. The horizontal field near
the surface of the Earth, typically measured in V/km and characterized by the Earth surface potentials,
induces the flow of geomagnetically induced currents (GIC). When a CME strikes the magnetosphere,
the displacement of the geomagnetic field relative to the conductive ionosphere increases the
magnitude of currents flowing 100-150 km above the Earths surface in the electrojets, to the order of
millions of amperes. These electrojet currents induce quasi-dc potentials in series with the
transmission lines, which in turn, drive the flow of GIC wherever there is a path for them to flow. The
GICs that flows in the transmission lines in the loops formed by transmission lines, grounded-wye
transformers, and the paths through the ground between the transformer neutrals affect the power
grids operation [6].
When the geomagnetically induced voltage appears in the loop formed by a transmission line,
transformer, and the ground - almost all of this voltage appears initially across the unsaturated
magnetizing inductances of the transformers as these inductances are higher than that of the
transmission line. Because the induced voltage has very-low frequency (quasi-dc, at frequency
typically lower than 1 Hz), it causes an increase in the dc flux density offset, allowing initially only a
low dc current to flow until the superimposed dc flux, dc, and the ac peak of the transformer core
flux density, ac, reach the transformer core saturation level (see Fig.1.). The dc voltage drops across
the circuit resistances and causes a decrease in the quasi-dc voltage applied to the transformer and
therefore the rate of increase of the dc flux density offset slows down. The dc flux density offset stops
increasing when the dc component of the magnetizing current equals the dc induced voltage divided
by the circuit resistance. This dc current is the steady-state GIC which gives rise to the eventual dc
flux density shift in the transformer core.

Figure 1 Magnetizing characteristic with dc flux causing the part-cycle saturation of the
excitation current

2
The magnitude of this flux density shift depends on the magnitude of the GIC current, the number of
turns in the windings carrying this current, and reluctance of the path of the dc flux induced by GIC.
The result is that the dc flux adds to the ac flux in one half-cycle and subtracts from the ac flux in the
other half-cycle. When the dc flux is high enough, the peak flux density in the magnetic core reaches
the pre-saturation range in one half of the cycle resulting in core saturation for a small part of a cycle.
This is referred to as a part-cycle saturation. Fig. 1 shows the -I characteristics which represents the
B-H curve of core material, as the magnetic field intensity H is proportional to the current I and the
magnetic flux is proportional to the magnetic flux density B. The BH characteristics of the
transformer core materials is very non-linear and for higher magnitudes of DC, the core provides a
much higher reluctance to the DC ampere-turns and therefore results in a smaller incremental increase
in the flux density shift and a higher peak magnetizing current pulse im depending on the air core
inductance Lair-core of the circuit.
For a given GIC profile, the waveform of the magnetizing current in a part-cycle saturation can be
determined. Following this, an equivalent sinusoidal excitation current determined based on empirical
data. This allows for calculation of the reactive power consumption of the transformer under the
specified DC loading profile. The harmonic content of the magnetizing current under a part-cycle
saturation may also be determined through the Fourier transform. Determination of the harmonic
content allows for calculation of a corresponding harmonic factor to account for increased eddy loss
and associated heating of constructional parts. As the magnetizing current is typically on order of the
rated transformer current during a part-cycle saturation, this, in conjunction with increased eddy loss,
are used to determine heating in all components and provide a basis to generate a loading guide for the
transformer.
Three-phase, three-limb cores with core form construction provide a high reluctance path to DC flux.
This is because this core type offers an order of magnitude higher magnetic reluctance to the DC
ampere-turns in the core tank magnetic circuit. The DC flux has to pass through the very high
reluctance path from the top yoke of the core to the tank cover, through the tank walls, and return to
the bottom yoke through, the high reluctance path from the tank bottom. Therefore, this core type is
less susceptible to part-cycle core saturation; however it is susceptible to high magnitudes of
magnetizing current. Single-phase, shell-form, and five-legged core-form transformers present much
lower reluctance to the DC flux within the core. Therefore, they are more susceptible to part-cycle
core saturation at lower levels of DC. Also, while the flux density of the core is an important factor in
determining the final value of the peak of the sum of the AC and DC flux density in three-phase core
form transformers with a threeleg core, this is not the case for transformers with other core types [7-
12].

OBJECTIVE AND TEST SETUP


The goal of the experiment was the determination of the self-protective capability of a single phase,
transformer with two wound core legs, with regards to Geomagnetically Induced Currents (GICs).
GICs was simulated through the injection of DC current into the transformer HV neutral. The
collected experimental data were used to determine a safe operational limit for the unit under GIC
event. Testing proceeded by increasing DC stress upon the units until such point where the units
and/or test equipment are considered endangered or the limits of capability of the test equipment have
been reached.
The test setup used in this experiment is a modification of the Loading Back Method as described in
IEEE Std C57.12.90. The general test setup is depicted in Figure 1. The two transformers utilized in
the experiment, SVC1 and SVC2, were single phase, 60/80/100 MVA, 230kVY 19.65kV units of
identical design. AC excitation for the units is provided from the mains via the manufactures loader
and regulator transformers. Due to limitations of the test facilities available, it was not possible to
supply both the necessary excitation current and circulate the equivalent load current the units will see
in operation. As such the units were tested in a no load condition and tested values were compared to
the calculated transformer characteristics under GIC.

3
a. b.
Figure 1 a. Test Setup Schematic; b. Photograph of Actual Setup

As GIC currents are quasi-DC in nature, these were simulated with a DC source comprised of a battery
and resistor bank in various configurations so as to allow the injection of various DC levels. A 2.362
resistor from the bank is utilized as a shunt so as to allow the monitoring of the AC current in the
HV circuit. A reactor (not depicted) can be considered in the HV circuit in order to limit the possible
AC circulating current. The necessity of this device was to be determined experimentally by
measuring the circulating current in the HV circuit at rated excitation and 0 DC input to the test setup.
In order to prevent the overheating of winding or constructional parts during the testing, fiber optic
probes and thermocouples were installed within the unit. Probes are located within the windings, on
transformer core, tie-plate, and clamping system in the area of the core window. The probes were
monitored throughout the testing process. Additionally, the units cooling systems were run at full
capacity and the additional water cooling system was connected to both units to further ensure no
overheating occurs.

PROCEDURE
The first step of the process was the determination of whether or not a reactor would be required to
limit the AC circulating current in the HV circuit. The units were excited at rated voltage and a total
circulating current of 150 mA was measured in the HV circuit. As such a reactor was deemed
unnecessary.
Before testing was to begin in earnest, DC was slowly injected into the test setup to determine whether
the response from all equipment involved was acceptable. The DC source was set to 2.6A and AC
excitation slowly ramped up. At approximately 95% excitation, the HV side overcurrent protection on
the regulator tripped. The ratio was increased to allow for a higher threshold. The DC source was
then disconnected and the units demagnetized. It was noted that this took a substantial amount of
time, over 5 minutes.
The following procedure was utilized:
The required DC current for the test was injected into the transformer neutrals
The AC excitation is brought up to 90% and the response of the test system examined
for any possible indications of danger to the system
AC excitation is brought up to 110%
Measurements of the input power, excitation current magnitude, harmonics and wave
shape, sound level, and fiber optic readings are taken
The AC excitation is reduced in 5% and the above repeated until a 90% excitation
reading has been taken.
The unit is demagnetized in preparation for the next DC level

Testing was halted at a level of 10A DC injected into the transformer neutrals due to a limitation in the
capacity of the testing equipment. If AC excitation levels exceeding 100% were attempted protective

4
circuitry in the testing equipment was tripped. Various configurations of capacitors were attempted to
compensate for the extreme reactive loading but were unsuccessful.

EXPERIMENTAL RESULTS AND ANALYSIS OF DATA

Using the oscilloscope, the current was measured off the shunt in the HV circuit, and the part-cycle
saturation was proven be achieved. A sample waveform has been provided in Figure 3.

Figure 3 - HV Shunt Voltage at 110% AC and 5A DC

Both the exciting current waveshape and harmonic component data were measured as a function of the
DC bias level. The exciting power per unit of core weight for the range of DC bias and AC excitation
levels tested is presented in Figure 4. Also presented is the units normal excitation current as seen in
Figure 5 and the comparative excitation current at various AC excitation levels when a DC bias is
applied as shown in Figure 6. An anomaly appears to have occurred during the 5A DC injection test.
The values measured during the 2.5A DC and 10A DC injection conform to anticipated trends as does
the 110% AC measure at 5A DC injection. In general however, an upward trend relative both to
increasing AC excitation and increasing DC current injection is apparent. This variation is believed to
arise from the sequence of DC application. In an effort to protect the units, the highest DC current,
and as such related stresses, were applied initially and subsequent DC levels were less strenuous. This
was believed to have magnetized the core to an excessive level for the subsequent voltage applications
and as such skew the data.

90 140

80
120

70

100
60
Exciting Current (A rms)
Exciting Power (VA/kg)

50 80
90 Normal Excitation
100 2ADC Excitation
5ADC Excitation
110
40 10A DC
60

30

40

20

20
10

0 0
0 2 4 6 8 10 12 80 85 90 95 100 105 110 115 120 125
DC Bias (A) AC Excitation (%)

Figure 4 - Exciting Power vs DC current Figure 5 - Excitation Current vs ac voltage

5
90

80

70

60

Exciting Current (A rms)


90
50
95
100
105
40
110

30

20

10

0
0 2 4 6 8 10 12
DC Bias (A)

Figure 6 - Per Unit RMS Excitation Current Relative to Excitation Current without DC Bias

The harmonic content of the measured excitation current during the GIC test was reasonably close to
calculated values for such conditions. Due to limitations with the test equipment, only the odd
harmonics present within the system are measured. Even harmonics are expected however and
calculated values for both are presented compared to the measured odd harmonic values in Figure 7
and Figure 8. Observed discrepancies between calculated and measured values are believed to arise
from assumptions of ideality utilized in the model as well as unaccounted for interactions between the
units, test equipment, and AC source.

The harmonic content of the excitation current is presented in Figure 8. It is apparent that substantive
harmonic components are present within the signal, on the order of the primary harmonic. A common
trend throughout is an increase in harmonic magnitude at lower levels of DC bias with a marked
decline at the 10A DC bias for 3rd, 5th and 7th harmonics.

90

80

70

60
Excitation Current (A rms)

3
50 5
7
9
40 11
13

30

20

10

0
0 2 4 6 8 10 12
DC Bias (A)

Figure 7 - Calculated vs Measured Harmonics


Referenced to the Fundamental Frequency Figure 8 - Harmonic Content of Excitation
Current at 100% AC

Sound level measurements were taken at selected points around the transformer tanks. The average of
the measurements was then taken and plotted below in Figure 9. From this we can see both the AC
excitation and DC bias levels have a large impact on the units sound level.
The data collected during the experiment were compared to the calculated values and allowed for
simulation of different levels of loading under specific GIC conditions.

6
91.0

90.0

89.0

88.0
Total Sound Level [dBA]

2.6 A DC
87.0 4.75A DC
9.75A DC

86.0

85.0

84.0

83.0
80 85 90 95 100 105 110 115
AC Excitation [%]

Figure 9 - Total Sound Level

CONCLUSIONS

The tests performed in this research allowed to confirm the accuracy of analytical methods in
predicting the behavior of a single-phase unit under GIC conditions.

Based upon the experimental data collected and analytical calculations, it was recommended to the
user that the unit is operated within the limits presented below in Figure 10. If the GIC events last
longer, the unit needs to be observed for hotspots and abnormal noise. At full load, the maximum GIC
current that unit can withstand is 200 A in the neutral for a maximum time of around 8.5 seconds.
The unit in operation will withstand the above levels of GIC currents. However, it has to be
emphasized that the flow of GIC currents through a transformer will result in core saturation,
development of internal hotspots, significantly increased noise level and vibrations. Considering these
detrimental effects it is recommended that whenever possible the unit is switched off when GIC
currents exceeding the recommended levels or durations are detected.

Figure 10 - GIC Withstand Curve for Currents up to 200A DC

7
BIBLIOGRAPHY

[1] W.D.Gonzales et al What is a geomagnetic storm, Journal of Geophysical Research, Vol.90,


1994
[2] G.Rostoker et al Physics of magnetic storms, Magnetic Storms, Geophysical monograph,
1998
[3] Y.Kamide et al Magnetic storms: current understanding and outstanding questions, Magnetic
Storms, Geophysical monograph, 1998
[4] B.T.Tsurutani The interplanetary causes of magnetic storms: A review, Magnetic Storms,
Geophysical monograph, 1998
[5] N.U. Crooker Solar and heliospheric geoeffective disturbances, Journal of Atmosphere and
Solar-Terrestrial Physics 62, 2000
[6] IEEE Standard C57.163 Draft Guide for Establishing Power Transformer Capability while
under Geomagnetic Disturbances under development
[7] John Kappenman, Geomagnetic Storms and Their Impacts on the U.S. Power Grid, Report #
Meta-R-319 by Metatech Corporation submitted to Oak Ridge National Laboratory, 2010
[8] Ramsis Girgis & Chung-Duck Ko, Calculation Techniques And Results Of Effects Of GIC
Currents as Applied to Two Large Power Transformers, IEEE Transactions on Power Delivery,
Vol. 7, No. 2, April 1992
[9] P. Picher, L. Bolduc et al, Study of the acceptable DC current limit in core-form Power
Transformers, IEEE Transactions on Power Delivery, Vol. 12, No. 1, January 1997
[10] Matti Lahtinen & Jarmo Elovaara, GIC Occurrences and GIC Test for 400 kV System
Transformer, IEEE Transactions on Power Delivery, Vol. 17, No. 2, April 2002
[11] Nobuo Takasu et al, An experimental analysis of DC excitation of Transformers by Geo -
magnetically Induced Currents, IEEE Transactions on Power Delivery, Vol. 9, No. 2, April
1994
[12] R. Girgis, K. Vedante Effects of GIC on Power Transformers and Power Systems,
Transmission and Distribution Conference and Exposition (T&D) IEEE PES, 2012

You might also like