See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/344172368

Proposed Evaluation of the Potential Risk of Voltage Collapse in the Cape

Network as a Result of GIC

Conference Paper · January 2014

CITATIONS READS

0 645

3 authors, including:

Abu-bakr Jakoet Akshay Kumar Saha

University of Cape Town University of KwaZulu-Natal
2 PUBLICATIONS 3 CITATIONS 155 PUBLICATIONS 940 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Akshay Kumar Saha on 09 September 2020.

The user has requested enhancement of the downloaded file.

It further provides a methodology to quantify the additional
reactive power absorbed by different transformers, in a
selected part of the South African transmission system,
under saturation because of GICs. Subsequently, a
methodology to conduct a voltage stability study of the
selected network is presented, considering different
contingency analysis scenarios for its comprehensive
evaluation.

2. EFFECTS OF GICs ON POWER TRANSFORMERS

Transformers form the backbone of an EPS as they step-up

voltages for power transmission across a region and step-
down voltages to different usable levels for the consumer.
In the past, power transmission over longer distances was Fig. 3: Relationship between flux and current with and without
not possible due to the high power losses experienced. DC-bias [4]
Resistive power losses in transmission lines are reduced
when stepping up voltages for power transmission across Due to the asymmetrical nature of the magnetization
long distances. Transformers can be used to vary the current during saturation, the exciting current of a half-
relative voltage of circuits, isolate them, or both. cycle saturated transformer is rich in both even and odd
harmonic components, which can lead to problems
Power transformers are designed to operate in the linear mentioned briefly, previously in this paper. The
region of their saturation characteristic (B-H curve) so as to characteristics and harmonic profile of the magnetization
minimise the amount of copper used in its construction, current during DC-bias, shown in Fig. 4, and the associated
thus minimising construction costs. Fig. 2 indicates the reactive power consumption associated therewith, has been
saturation characteristic of a specific ferromagnetic reported in [15], [16] and [17].
material used in transformer core construction.
The relationship between the applied DC, the excitation
The curves indicate the relationship between the magnetic current and VAR consumption are linearly related [15-17].
field strength H, and the magnetic flux density B. The A decrease in harmonic amplitude of the excitation current
characteristic varies with different core material types. The is observed as the harmonic frequency increases [4].
B-H curve on the left includes hysteresis and eddy current
losses, whereas the curve on the right ignores this A transformer’s susceptibility to GICs is dependent on its
phenomenon and can therefore be linearized as shown. The core construction, as this relates to how many complete DC
linear region of the curve is indicated by the red arrows in flux paths will be present in the core [18]. Single phase
the figure. transformers are considered to be the most susceptible to
GICs [19]. In a 3-phase three limb transformer, no
During a GMD, GICs of frequencies between 0.001 Hz and complete DC flux path exists in the core, making this type
1 Hz with periods of 1 second to 1000 seconds can flow in of design the least susceptible to GICs. The 3-phase five
or out of the neutrals of grounded Y transformers. These limb transformers, 3-phase seven limb and 3-phase shell
quasi-DC GICs cause half-cycle saturation of power form are next, in order of increasing susceptibility. It
transformers. Power transformer saturation is the root should be noted, that field tests of applied dc to 3-phase
cause of EPS problems associated with GICs [5,14]. three limb transformers indicated some signs that saturation
has occurred [15], so the statement that a 3-phase three
The application of DC to the neutral of power transformers limb transformer is not susceptible to GICs could not be
causes a one directional DC flux in the core, which entirely true and needs to be investigated in the future. A
depends on the reluctance path, the magnitude of the GIC possible explanation is that the transformer tank and
and the number of turns in the winding in which it flows structural members could perhaps supply a sufficient
[11]. This DC flux causes the AC flux in one half-cycle to return-path permeability to allow main-leg saturation. The
rise above the knee-point voltage (Vknee). The result is that effects of GICs on the transformer tanks have been studied
the core now operates in a very non-linear region of its B-H by [20] and include effects such as localised heating of the
curve causing short duration, high magnitude and tank, corrosion and paint damage.
asymmetrical magnetization current peaks. This is
illustrated in Fig. 3. An increase in leakage flux is observed
which can result in overheating depending on transformer
design and core construction.

Linear region of
B-H curve

X axis: H – magnetic field strength (T)

Y axis B – magnetic flux density (Wb) Fig. 4: Harmonic spectrum of excitation current under different
GIC conditions [4]
Fig. 2: A typical saturation characteristic of a ferromagnetic
material used in transformer core construction [10]
In order to accurately model the transformer for GIC transformers behaviour in the saturated region. Unlike the
studies, its low frequency and saturation behaviour under classical transformer model, the UMEC considers magnetic
DC-bias needs to be considered. coupling between windings of different phases, and
coupling between windings of the same phase. The UMEC
3. MODELLING OF GICS IN AN EPS transformer model is based primarily on core geometry and
uses the duality principle to solve magnetic circuits instead
A methodology for conducting load flow studies of an EPS of electrical circuits as in the classical transformer model.
considering the effects of GIC is described in [21]. It Normalized core parameters are used in the model so that
involves modelling the geo-electrical field produced by a requirements of physical data are minimized [25]. Fig. 5
GMD to determine the induced earth surface potential that illustrates a 3-limb transformer and 5-limb transformer and
drives GICs and modelling this as a DC voltage source in the required core geometry ratios for the UMEC model.
the transmission line. RTDS allows for the selection of a 3-phase, 3-limb or a 3-
phase, 5-limb transformer from the configuration menu.
GICs can be considered as DC when compared to the
50 Hz or 60 Hz operating frequency of power transformers, Selecting a 3-limb core requires the input of the ratio of
hence the term quasi-DC. The reactance of the transmission core yoke length (Ly) to the core winding-limb length
lines, transformers and other devices required in the EPS (Lw), as well as the ratio of core yoke area (Ay) to the core
(shunt capacitors and inductors) can thus be ignored when winding-limb area (Aw). Selecting a 5 limb core requires
determining the magnitude and direction of GIC flow. the ratio of core yoke length (Ly) to the core outer-limb
Delta winding transformers are also ignored and series length (Lo) and the ratio of core yoke area (Ay) to the core
capacitors block the flow of GICs along that path. outer-limb area (Ao) in addition to the two ratios required
Important parameters that influence the magnitude and for a 3-limb transformer. The name plate data of the
direction of GICs are the following [22]: specific transformer is also entered into the configuration
menu and the saturation curve can be entered discretely as
i. The magnitude of the induced earth surface points on a curve, shown in Table 1. For a more detailed
potential due to a GMD (modelled as a DC explanation of the UMEC model, see [25] and references
voltage source in the transmission line) therein.
ii. Resistance of the coils of grounded transformers
iii. Resistance of the series winding of auto-
transformers (including common winding if
grounded)
iv. Resistance of transmission lines
v. Substation grounding resistance

The GIC distribution in the network can then be calculated.

Hereafter, a method for determining the increased VAR
requirements under different levels of GICs should be
developed, subsequently, these VAR requirements should
be modelled in a load flow program, including realistic
loads, in order to provide suitable responses to voltage
stability problems.

In more recent work in the field of calculation of GIC,

Turnbull [23] improved the modelling of GIC in the United
Kingdom. It is recommended, that in order to improve GIC
modelling, specific substation transformer data is required
[23]. Marti et al [24] and references therein, shows that
GICs can be calculated using the excessive reactive power
absorption of transformers when saturated.

4. RTDS MODELLING OF TRANSFORMERS TO

DETERMINE MVAR CONSUMPTION
Fig. 5: Core structures of 3-limb and 5-limb transformers [26]
RTDS® Simulator (RTDS) is a power system simulator that
solves electromagnetic transient simulations in real time. The network to be studied is shown in Fig. 7. The black
The system is used for high speed simulations, closed-loop squares represent substations and the red stars represent
testing of protection and control equipment and hardware towns. High voltage transformers (132 kV and above) at
in the loop (HIL) applications. the different substations will be modelled using the UMEC
model in RTDS as shown in Fig. 8 to determine the
Transformers in a selected portion of the South African reactive power consumption of transformers when
transmission system will be modelled in RTDS to subjected to different magnitudes of GIC.
determine its excess reactive power consumption in the
saturation region of the B-H curve. The transformers will The UMEC model is computationally intensive, therefore
be injected with different magnitudes of GIC in its neutrals some banks of transformers would need to be lumped
and the corresponding reactive power will be recorded. A together to limit the computational power required. This
curve of MVAR vs. GIC will be derived for each different will be assessed on an as needed basis in the detailed
transformer in the network being studied. modelling, which is not presented in this paper.

The Unified Magnetic Equivalent Circuit (UMEC) model,

developed by Enright [25], will be used to model the
 V nom kV
Q lo s s V pu K I G IC
V n o m kV ,a ssu m ed (2)

Literature provides some guidance for the values of K for a

specific transformer core design, if they are not known
[3,11,16].

The proposed values for K are listed below:

- Single phase core: K = 1.18

- Shell form core: K = 0.33
- 3-limb core: K = 0.29
- 5-limb core: K = 0.66

K values are to be derived for different transformers, in the

Fig 7: Western Cape network to be studied indicating substations network to studied, using RTDS and the model shown in
and towns [28] Fig. 8. These will be compared to that proposed in
literature.
Table 1: Saturation curve input in RTDS for UMEC transformer
model
5. LOAD FLOW AND VOLTAGE STABILITY
Variable Name Description Value
STUDIES CONSIDERING THE EFFECTS OF GIC
X1E Point 1 Current as a % of rated current 0
Y1E Point 1 Voltage in p.u. 0
The South African EPS is currently very stressed and the
X2E Point 2 Current as a % of rated current 0.1774
system is run very tightly in order to meet demand while
Y2E Point 2 Voltage in p.u. 0.324129
performing maintenance on an ageing fleet. As previously
X3E Point 3 Current as a % of rated current 0.487637
mentioned, South Africa has long transmission lines
Y3E Point 3 Voltage in p.u. 0.61284
X4E Point 1 Current as a % of rated current 0.980856
spanning from the north of the country to the south.
Y4E Point 1 Voltage in p.u. 0.825118
Voltage stability problems are usually associated with very
X5E Point 1 Current as a % of rated current 2
stressed power systems and long lines [28].
Y5E Point 1 Voltage in p.u. 1
X6E Point 1 Current as a % of rated current 3.09543
The main factor causing voltage instability problems in an
Y6E Point 1 Voltage in p.u. 1.08024
EPS is the inability of the power system to meet the
X7E Point 1 Current as a % of rated current 6.52348
demand for reactive power. When performing voltage
Y7E Point 1 Voltage in p.u. 1.17334
stability studies generator reactive power and voltage
X8E Point 1 Current as a % of rated current 20.357 control limits, load characteristics, characteristics of
Y8E Point 1 Voltage in p.u. 1.26115 reactive power and voltage control devices need to be
X9E Point 1 Current as a % of rated current 60.215 considered. Voltage stability depends on how the variation
Y9E Point 1 Voltage in p.u. 1.36094 of reactive power as well as real power in the load area
X10E Point 1 Current as a % of rated current 124.388 affect the voltage at the load buses [28].
Y10E Point 1 Voltage in p.u. 1.49469
The additional reactive power loads drawn by transformers
Overbye et al [27] describes a linear relationship between will be quantified based on a one in a hundred year severe
the absorbed reactive power and the per-phase GIC flowing solar storm. The additional reactive power load of the
in the neutral of grounded Y transformers. An approximate different transformers at substations will be modelled as
equation to quantify the additional reactive power absorbed constant reactive current loads in a load flow program such
by the transformer is shown in Equation 1. as DIgSILENT Power Factory. The impact to voltage
stability will be assessed considering different “what if”
scenarios, e.g. loss of a critical SVC or loss of a critical
Q lo s s V kv kI G IC (1) line.

Where: An assessment of the impact of degraded voltage on

nearby industry and nuclear plants will be made based on
Qloss = additional reactive power absorbed by the the voltage stability results obtained by the study. Safe
transformer under GIC conditions operational voltage of nuclear plant equipment will be
reviewed during the study to ensure that the network
Vkv = the terminal voltage voltage remains within the safety limits during a GMD
event.
k = a transformer specific constant
6. CONCLUSION
IGIC = the per-phase GIC flowing in the neutral of the
grounded Y transformer The paper aims at providing a review of the existing
literature regarding the effect of GICs on power
In load flow calculations, bus voltages are usually transformers and load flow studies when GICs are present.
expressed in per-unit (p.u), we thus include the It reported a methodology to quantify the additional
transformers maximum nominal voltage in the transformer reactive power absorbed by different transformers, in a
specific constant, denoted as K, which has a unit of selected part of the South African transmission system,
Mvars/Amp. Equation 2 below indicates the form of the under saturation because of GICs and a methodology to
equation to be used for large system studies. Vpu is the per- conduct a voltage stability study of the selected network.
unit voltage at the bus [27]. Contingency analysis for a comprehensive evaluation of
voltage stability concerns is considered.
 Fig. 8: RTDS UMEC transformer test model for GIC vs. MVAR consumption determination

This literature review provides a basis for the evaluation of regions incorrectly considered to have low GIC-risk," IEEE
the potential risk of voltage collapse in the Western Cape PowerTech,Lausanne Switzerland, pp. 807-812, 2007.
network as a result of GIC.
[9] L. Bolduc, P. Langlois,; D. Boteler, and R. Pirjola, , "A
Power transformers in the network are simulated using the study of geoelectromagnetic disturbances in Quebec. I.
General results," IEEE Transactions on Power
UMEC model to determine their reactive power
Delivery,Vol. 13, No. 4, pp. 1251-1256, 1998.
consumption under different levels of DC applied to its
neutral. Predicted GIC levels, from literature for the area to [10] J. Berge, R.K. Varma, and L Marti,."Laboratory validation
of the relationship between Geomagnetically Induced
be studied, are to be used to determine the reactive power Current (GIC) and transformer absorbed reactive
consumption of transformers for a one in a hundred year power,"IEEE Electrical Power and Energy Conference
solar storm. These reactive loads are then modelled in a (EPEC), pp. 491-495, 2011.
load flow program so as to determine the impact on the [11] R.S. Girgis, and K. Vedante, ,"Effects of GIC on power
voltage stability of the network. The impact on associated transformers and power systems," IEEE PES Transmission
industry and nuclear power plant in the study area can then and Distribution Conference and Exposition (T D), pp. 1-8,
be assessed. 2012.
[12] M Lehtinen and R. Pirjola, "Currents produced in earthed
REFERENCES conductor networks by geomagnetically-induced electric
fields," Annales Geophysicae, Vol. 3, No.4, pp. 479-484,
1985
[1] Eskom, “How electricity is transmitted,” [Online].
Available: [13] L. Trichtchenko and D.H. Boteler, "Effects of Recent
http://www.eskom.co.za/c/article/209/how-electricity-is- Geomagnetic Storms on Power Systems," 7th International
transmitted Symposium on Electromagnetic Compatibility and
Electromagnetic Ecology, pp. 265-268, 2007
[2] Eskom, "Transmission Ten-Year Development Plan,"
[Online]. [14] T.S. Molinski, "Why utilities respect geomagnetically
Available: induced currents ," Journal of Atmospheric and Solar-
http://www.eskom.co.za/Whatweredoing/TransmissionDev Terrestrial Physics , Vol. 64, No. 16, pp. 1765-1778, 2002.
elopmentPlan/Documents/ TransDevPlanBrochure2013- [15] R. A. Walling and A.N. Khan,. "Characteristics of
2022.pdf transformer exciting-current during geomagnetic
[3] W.A. Radasky and J.G. Kappenman "Impacts of disturbances," , IEEE Transactions on Power Delivery,
geomagnetic storms on EHV and UHV power grids," Asia- Vol. 6, No. 4, pp. 1707-1714, 1991..
Pacific Symposium on Electromagnetic Compatibility [16] X. Dong,; Y. Liu, and J.G. Kappenman, "Comparative
(APEMC), pp. 695-698, 2010 analysis of exciting current harmonics and reactive power
[4] NERC, "2012 Special Reliability Assessment Interim consumption from GIC saturated transformers," IEEE
Report: Effects of Geomagnetic Disturbances on the Bulk Power Engineering Society Winter Meeting, Vol. 1, pp.
Power System," NERC, Tech. rep. 2012. 318-322, 2001.
[5] J.G. Kappenman and V.D. Albertson, "Bracing for the [17] J. Yao, M. Liu, C. Li, and Q. Li, , "Harmonics and Reactive
geomagnetic storms," Spectrum, IEEE, Vol. 27, No. 3, pp. Power of Power Transformers with DC Bias," Asia-Pacific
27-33, 1990. Power and Energy Engineering Conference (APPEEC),
pp. 1-4, 2010.
[6] R. J. Pijola, D. H. Boteler’ and H. Nevanlinna, "The Effects
Of Geomagnetic Disturbances On Electrical Systems At [18] N. T. Takasu,; T. Oshi,; F. Miyawaki, S. Saito, and Y.
The Earth’s Surface," Advance Space Research, Vol. 22, Fujiwara, "An experimental analysis of DC excitation of
pp. II-21, 1998. transformers by geomagnetically induced currents," IEEE
Transactions on Power Delivery, Vol. 9, No. 2, pp. 1173-
[7] W.A. Radasky, "Overview of the impact of intense
1182, 1994.
geomagnetic storms on the U.S. high voltage power grid,"
IEEE International Symposium on Electromagnetic [19] A.P. Meliopoulos, S. Elias, N. Glytsis, G.J. Cokkinides, and
Compatibility (EMC), pp. 300-305,2011. M. Rabinowitz, "Comparison of SS-GIC and MHD-EMP-
GIC effects on power systems," IEEE Transactions
[8] C.T Gaunt and G. Coetzee, "Transformer failures in
on,Power Delivery, Vol. 9, No. 1, pp. 194-207, 1994.
View publication stats

[20] B. Zhang,; L. Liu,; Y. Liu,; M. McVey, and R.M. Gardner, [26] RTDS Technologies, Real Time Digital Simulator Power
"Effect of geomagnetically induced current on the loss of System Users Manual. Canada, 2006.
transformer tank," IET Electric Power Applications, Vol. 4, [27] T.J. Overbye, T.R. Hutchins, , K. Shetye,; J. Weber,. and S.
No. 5, pp. 373-379, 2010. Dahman, "Integration of geomagnetic disturbance modeling
[21] V.D. Albertson, J.G. Kappenman; N. Mohan and G.A. into the power flow: A methodology for large-scale system
Skarbakka, "Load-Flow Studies in the Presence of studies," in North American Power Symposium (NAPS), pp.
Geomagnetically-Induced Currents," IEEE Transactions on 1-7, 2012.
Power Apparatus and Systems, Vol.100, No. 2, pp. 594- [28] Kundur,P, Power System Stability and Control, Mark G.
607, 1981. Lauby Neal J. Balu, Ed. New York, USA: McGraw-Hill,
[22] J. Berge and R.K. Varma, "A software simulator for 1994.
Geomagnetically Induced Currents in electrical power [29] Eskom, "Generation Connection Capacity Assessment of
systems," Canadian Conference on Electrical and the 2016 Transmission Network,"[Online]
Computer Engineering (CCECE), pp. 695-700. 2009 Available:
[23] K. Turnbull, "Modelling GIC in the UK," A&G, Vol. 51, http://www.eskom.co.za/Whatweredoing/GCCAReport/Do
pp. 5.25-5.26, 2010 cuments/GCCA_2016REV1.pdf
[24] L. Marti,; J. Berge, and. R.K. Varma,” Determination of
Geomagnetically Induced Current Flow in a Transformer
From Reactive Power Absorption”, IEEE Transactions on
Power Delivery , 2013
[25] W. Enright,; N. Watson, and O. Nayak, "Three-phase five-
limb unified magnetic equivalent circuit transformer
models for PSCAD V3.," International Conference on
Power System Transients, 1999.