See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/3273408

Digital time-domain investigation of transient

behavior of coupling capacitor voltage
transformer

Article in IEEE Transactions on Power Delivery May 1998

DOI: 10.1109/61.660947 Source: IEEE Xplore

CITATIONS READS

39 188

5 authors, including:

M.R. Iravani
University of Toronto
102 PUBLICATIONS 5,175 CITATIONS

SEE PROFILE

All content following this page was uploaded by M.R. Iravani on 19 July 2013.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
622 IEEE Transactions on Power Delivery, Vol. 13, No. 2, April 1998

I~E-DO~~IN IENT
BE COUPLING C SFORMER
M.R. Iravani, X. Wang I. Polishchuk, J. Ribeiro A. Sarshar
Department of Electrical and Haefely-Trench Haefely-Trench
Computer Engineering 390 Midwest Road 71 Maybrook Drive
University of Toronto Scarborough, Ontario Scarborough, Ontario
Toronto, Ontario Canada M1P 3B5 Canada M1V 4B6
Canada M5S 364

Abstract-This paper reports a set of digital time-domain systems. The model is developed based on the use of the
simulation studies conducted on TEHMPl61A Coupling EMTP. The objectives of this joint project are:
Capacitor Voltage Transformer (CCVT) of Haefely-Trench. The
Electro-Magnetic Transients Program (EMTP) is used to develop To evaluate, compare and quantify impacts of CCVT
the CCVT model and conduct the transient studies. The accuracy component parameters and protectivelsuppressive
of the CCVT model is verified through comparison of the EMTP devices on its transient response, e.g. the
simulation results with those obtained from test results. The phenomenon of ferroresonance.
investigations demonstrate that the developed model can
accurately predict CCVT transient response, e.g. the phenomenon
of ferroresonance. The model is developed (1) to determine
0 To predict and quantify impact of CCVT transient
impact of transients on CCVT response, (2) to design, optimize behaviour on protection systems.
and compare protective and ferroresonance suppressor devices of
CCVT, and (3) to predict CCVT transient response on power e To investigate impact of power system transients,
system monitoring and protection schemes. e.g. faults and plannedunplanned switching
incidents, on CCVT transient behaviour.
Keywords: CCVT, Ferroresonance, Simulation, EMTP,
Protection, Relaying. Salient feature of the developed model as compared with
the reported models [4,5,6,7,8] is that it represents details
1. INTRODUCTION of (1) CCVT step-down transformer including its
saturation characteristic and tap positions, (2) CCVT series
CCVT is a well known apparatus to transform high- reactor including its tap positions, (3) CCVT protective
voltage (input) to low-voltage levels (output) at which devices, (4) ferroresonance suppressor circuitry, and ( 5 )
monitoring devices and protection relays operate. various burden models. Due to space limitation, this paper
Theoretically, the output waveform should be an exact only reports some of the studies conducted on
replica of the input waveform under all operating TEHMP161A CCVT model of Haefely-Trench and
conditions. Under steady-state conditions, this highlights the conclusions.
requirement can be satisfied based upon proper design and
tuning of the CCVT. However, under transient conditions, The rest of this paper is organized as follows. Section 2
e.g. faults and switching incidents, the CCVT output introduces TEHMP161A CCVT system used for the
waveform may deviate from the input waveform due to the reported studies. Section 3 reports the results of the
impacts of capacitive, inductive and nonlinear components frequency-domain studies carried out on the CCVT system.
of the CCVT [1,2]. Therefore, fidelity of CCVT during Section 4 compares the EMTP simulation results with the
transients must be well known and quantified [3]. The test results and verifies the accuracy of the developed
other concern is thermal overstress and consequently model. Section 5 reports the CCVT response to various
deterioration of CCVT components due to its internal simulated transient scenarios. Conclusions are
transient phenomena, e.g. the phenomenon of summarized in Section 6.
ferroresonance.
2. TEHMP161A CCVT CIRCUITRY
To address the above issues, the Instrument Transformer
Division (ITD)of Haefely-Trench and the University of Figure 1 shows schematic diagram of the CCVT
Toronto have embarked upon development of a circuitry. Switches S1, S2, S3 and S4 are not part of the
comprehensive digital time-domain model of CCVT CCVT circuitry, and included in the EMTP model to
simulate various transient scenarios imposed on the CCVT.
Major CCVT components of Fig. 1 are: voltage divider C1
PE-479-PWRD-0-01-1997 A paper recommended and approved and C2, drain coil Ld, step-down transformer (SDT), series
by the IEEE Transmisslon and Distribution Committee of the IEEE
Power Engineering Society for pubhcatlon in the IEEE Transactions reactor, harmonic suppression filter, protective device, and
on Power Delivery. Manuscript submitted July 31, 1996; made burden. Capacitors Cm, Ct and Cc are lumped
available for printing January 8, 1997. representations of stray capacitances of STD and series
reactor. STD and series reactor have multiple tap positions
which are not identified on Fig. 1, but have been included
in the EMTP model. Various protective devices examined
for the CCVT system of Fig. 1 are: MOV, triac and spark-
gap. The EMTP provides basic functions and component
models to construct required models of the above protective

0885-8977/98/$10.00 0 1997 IEEE

 623
I,,, ........................ f 1........................ -55 0
i A -60
sTDr
I II -65

E -70
-
c
-75

t
1:Lc=38 H
2:Lc=40 H
-I
-5
3:Lc=42H 3i
-80
-90
400 450 500 550 600 650 700
(1) Rotective Device (2) Harmonic Suppression Filter (3) Burden Frequency (Hz)

Fig. 1 Schematic diagram of a TEHMP161A Haefely- Fig. 3 Effect of net series reactance 6,) on CCVT
Trench CCVT. frequency response.
r
devices. Typical data for the CCVT of Fig. 1 is given in
Appendix A.

3. FREQUENCY-DOMAIN SENSITIVITY
ANALYSIS
A set of frequency-domain studies [5, 71 are conducted
z-
U
= .
Y

on the CCVT linear model to identify output sensitivity to - -70 1: Ct=O.O pF

various parameters of the CCVT system. Such information 6 2:Ct=140 PF I

3:Ct=700 pF
is helpful (1) to determine appropriate CCVT component -a0 - 4: ct=i400 PF
characteristics for EMTP studies, and (2) to verify the
EM simulation results, e.g. to correlate simulation and
measurement results. Figures 2 to 7 show a set of
frequency domain study results.
-90
I j
10 1o2 1@ 1o4
Figure 2 shows that the CCVT frequency response (20 Frequency (Hz)
log (uz/u1)) is significantly affected by Ld at frequencies Fig. 4 Effect of Ct on CCVT frequency response.
higher than 600-Hz. Figure 3 shows that the frequency -55 I
response is not very sensitive to changes of the inductance
(Lc) of the series reactor. The studies also show that the -60
frequency response is not sensitive to variations of STD
-65
leakage inductance.

Figure 4 indicates that at frequencies above lOO-Hz, the

-g
v
-70
effect of primary side stray capacitance (C,) of the STD on .-C
the CCVT frequency response cannot be ignored.
g -75
7 -80

-85

/
400 500 600 700 800
a
o_ Frequency (Hz)
-
C
tu Fig. 5 Effect of Cc on CCVT frequency response.
Sensitivity studies show that variation of mutual stray
2: With Drain Coil capacitance Cm, from 0.0 pF to 220 pF, does not influence
-80 1 the CCVT frequency response. Figure 5 shows that the
notch frequency illustrated in the frequency response is
1 I noticeably affected by the stray capacitance (C,) of the
10 1oz 10 10
series reactor. Figure 6 indicates that VA of burden does
Frequency (Hz)
not have any noticeable influence on the frequency
Fig. 2 Effect of Ld on CCVT frequency response. response. To the contrary, Fig. 7 shows that power-factor
624
-50 1 n s4 C STD SI
pf = 0.8 lagging
n
.c
m
(j- / U '
I \I
1:burden=l200 VA
2:burden=800 VA
3:burden=400 VA
-80 - 4:burden=200 VA

I -.I
IO' 1oz 1o3 10'
Frequency (Hz) Fig. 8 CCVT test set-up.
Figure 10 compares the EMTP simulation and test
Fig. 6 Effect of burden VA on CCVT frequency results corresponding to ferroresonance test (secondary
response.
short-circuit test). Initially S1 (Fig. 8) is open. The STD
n secondary side is subjected to a short-circuit by closing S1
and considering a burden with zero resistance (for the
burden=400 VA

-soj /a: actual test a burden of I-VA is used). The short-circuit is

later removed by opening SI. Figures 1O(a) and lO(b)
show close agreement between the recorded and simulated
results. Figure 11 also compares the simulation and
recorded results corresponding to the ferroresonance test
1:pf=l .o when S 1 is opened at a peak instant.
2:pf=0.8 lagging
-70 Close agreement between the simulation and the test
3:pf=0.6 lagging
results indicates that the EMTP model can be used:
4:pf=0.4lagging
5:pf=O.O * To determine impact of CCVT parameters (e.g. stray
-a0 1 J capacitances) and component characteristics on the
10' 1o2 10' I o4
Frequency (Hz)

Fig. 7 Effect of burden power-factor on CCVT frequency

response.
of burden can noticeably affect the frequency response at
frequencies above 3 0 0 - H ~as well as low frequencies
below 60 Hz. Similar frequency-domain studies also were
conducted to determine the effect of various CCVT
parameters on phase-angle deviation of the output voltage
with respect to that of input voltage.

4. VERIFICATION OF EMTP SIMULATION

MODEL

The existing library of component models of the EMTP

is used to construct the overall model of TEHMP161A
CCVT. Accuracy of the developed EMTP based model is
verified by comparing the EMTE' simulation results with
the corresponding test results. Figure 8 shows the
I'I Voltage

simplified equivalent circuit of the set-up for the test cases.

\ I t I

Figure 9(a) shows the recorded CCVT output voltage as \

a result of the transient response test. The CCVT burden is \
set at 400-VA (switch S1 of Fig. 8 is initially closed) and 1
switch S4 is opened at a peak instant of input voltage \
-. I /
(intemption angle of 90O). Figure 9(b) shows the EMTP
simulation of the transient response test for the same initial
\
.I ,
0

--0
-/-- / '
/

TIME(S)
1'
operating condition. Note that for the sake of clarity, the
simulated response has been enlarged. Figure 9 shows that 0 0.006 ,0.003
the pattern of oscillatory transients and frequency of Fig. 9 CCVT recorded (a) and simulated (b) output
oscillations depicted in the simulation and test results voItage corresponding to transient response test.
closely agree.
 625

I
1 A

TIMUS)

I I I 1 I 1 -300 ; I I I I
I
0 0.1 02 03 0.4 05 0 a1 0.2 03 0.4 0.5

Fig. 10 C O T recorded (a) and simulated (b) output Fig. 11 CCVT recorded (a) and simulated (b) output
voltage corresponding to fenoresonance test (S1 voltage corresponding to fenoresonance test (S 1
opens at voltage zero-crossing). opens at peak voltage).
CCVT transient response. voltage component at the CCVT output. These oscillations
are due to energy exchange between capacitive voltage
To determine impact of power system transients divider and Ld. The frequency of this oscillatory mode is
(e.g., faults and capacitor energization) on the
fidelity of CCVT response. defined by the natural frequency of the loop formed by C1,
C2, and Ld (13.956-kHz). In practice, the net resistance
To determine impact of burden characteristics on the associated with C1, C2 and particularly Ld limits the peak
CCVT transient behaviour. value and dampens this oscillatory mode. This indicates
that accurate Q-factor of the drain coil must be represented
To investigate impacts of CCVT transient response in the EMTP model when the impact of system transient
on digital relaying/protection systems. are investigated.
To modifykhange CCVT design to meet particular Figure 12(c) shows the CCVT transient response to the
operating requirements and achieve desirable same switching scenario when the drain coil is not
transient response. included in the CCVT system. The high-frequency mode
in Fig. 12(c) is due to the presence of stray capacitance
5. EMTP STUDY RESULTS Cm. Comparison between Figs. 12(b) and 12(c) illustrates
Due to space limitation, only the simulation results sensitivity of the CCVT transient response to the
corresponding to system fault and secondary-winding inductance of the drain coil.
short-circuit study cases conducted on the TEHMP161A
CCVT of Figure 1 are reported in this paper. We intend to Figure 12(d) shows the CCVT transient response when
report a comprehensive set of results in a subsequent paper. the protective block of the CCVT (Fig. 1) is represented by
a spark gap system. As compared with Figs. 12(b) and
5.1 System Fault 12(c), Fig. 12(d) shows that the maximum encountered
voltage is limited by the spark gap protective level.
T h e CCYT rmponse to a temporary, dose-in, line-to-
ground fault is simulated by a close-open operation of Figure 12(e) shows the CCVT transient response to the
switch S3, Fig. 1 (initially S4 and S1 are closed and S2 is same switching scenario when the drain coil (Ld) and the
open). Figure 12(a) shows the CCVT input voltage. mutual stray capacitance (C,) are not included in the
model. Comparison of Fig. 12(b) and 12(c) with Fig. 12(e)
Figure 12(b) shows the CCVT response (output voltage) reveals that the high-frequency oscillations are due to the
to the fault. Closure of S 3 resuIts i n a high-frequency presence of Ld and c,. Therefore, (1) proper design to
626
x IOS x io5
2

0
,. ....... ... ........ ... . . . . .,
-0
.I\.
, a) ,a)
-3 -2 L 1
0.05 0.1 0.15 0 0.1 0.2 0.3 0.4 0.5
5000 500

0 0

-5000 -500 I .,b)

0 0.05 0.1 0.15 0 0.1 0.2 0.3 0.4 0.5
500
1
0 0
I 1
cl ,c)
0
,
0.05 0.1 0.15
-500 I
0 0.1 0.2 0.3 0.4 0.5
500
I IQ (VI 1
0

-500
0
,
0.05
d)
0.1 0.15
-500
0
1
0 0.1
ir Yd)
0.2 0.3 0.4 0.5

, e)
"
0 0.05 0.1 0.15 0 0.1 0.2 0.3 0.4 0.5
Time (S) Time (S)
Fig. 12 CCVT transient response to closure and opening Fig. 13 CCVT transient response to closure and opening
of Switch S3 of Fig. 1. of Switch S2 of Fig. 1.
(a) input voltage (a) input voltage
(b) output voltage (with Ld, with c,, without (b) output voltage (no burden (VA=O))
protective device) (c) output voltage (with protective device,
(c) output voltage (without Ld, with c,, without burden: 400-VA, pf=O.S)
(d) output voltage (with kIOV protective device,
protective device) burden: 400-VA, pf=O.8)
(d) output voltage (without Ld, with C, with (e) MOV absorbed energy corresponding to case
protective device) (d) above.
(b) output voltage (without Ld, without c,, effectively damped out.
without protective device)
minimize C, and ( 2 ) proper selection of Q-factor of Ld, Figure 13(d) shows the CCVT response to the same
can practically eliminate the high-frequency component transient when the protective device is represented by an
shown in Fig. 12. MOV element in the CCVT system. Figure 13(d) shows
that presence of MOV can effectively mitigate the
5.2 Secondary W&zd&g XhoTt-C&euk phenomenon of ferroresonance Fig 13(e) shows the
energy absorbed by the MOV element as a result of the
Figure 13 shows the CCVT ferroresonance response to a short-circuit.
temporary short-circuit at the secondary (burden) side of
SDT. The short-circuit is imposed by closing S2 at peak 6. CONCLUSIONS
voltage and then opening S2 after 10 cycles. Figures 13(a)
and 13(b) show the CCVT input voltage and output This paper presents the results of digital time-domain
voltage. Figure 13(b) shows that the switching process and frequency-domain studies conducted on TEHMPl61A
results in subharmonic ferroresonance. CCVT of Haefely-Trench. The Electro-Magnetic
Transients Program (EMTP) is used for digital time-
Figure 13(c) shows the CCVT response to the short- domain simulation of the CCVT transient response.
circuit when the protective device is represented by a spark Comparison of the EMTP results with those of test results
gap and the CCVT burden is 400-VA. Figure 13(c) verifies the accuracy of the EMTP model of the CCVT.
illustrates that the ferroresonance phenomenon is The investigations conclude that:
 627
Time-domain and frequency-domain analyses of
CCVT provide complementary information to
accurately predict steady-state and transient
behaviour of CCVT system, and to properly
desigdtune its protective and suppressor devices.

EMTP model of CCVT can accurately demonstrate

ferroresonance of a CCVT circuitry.

EMTP model of CCVT can identify characteristics

1 6.5 109998
7.1188458
105.0369035517
106.9125625437

and rating of protective and ferroresonance

suppression devices. The model also predicts 15.6395688 114.7303092223
thermal stresses imposed on such devices. 21.3303888 116.2983601396
26.2709430 117.86641 10569
EMTP model of CCVT also can be used to
investigate impacts of system transients and burden M.R. Iravani (M85) received his B.Sc. de-= in electrical engineering
characteristics on the CCVT transient response. from Tehran Polytechnique University and s d his career as a consulting
Such information is of significant importance for engineer. He received his M.A.Sc. and Ph.D. degrees also in electrical
engineering from the University of Manitoba Canada in 1981 and 1985
proper operation of monitoring devices, relays and respectively. Presently he is a professor af the Cniversity of Toronto. His
protection systems. research interests include power electronics and power system dynamics and
control.
7. REFERENCES
Xiaolin Wang was bom in P.R. China in 1951. He received his B.S.
degree and M.A. degree from High Voltage Divhon, Electrical Engineering
IEEE System Relay Committee, Transient Response of Coupling Department, Xian Jiaotong University in 1982 and 1988. Since 1982, he
Capacitor Voltage Transformers, IEEE Trans., Vol. PAS-100, No. has been working in Electrical Engineering lkpamnent, Xian Jiaotong
12, pp. 4811-4814, December 1981. University. Currently, he is a visiting scholar at the University of Toronto.
A. Sweetana, Transient Response Characteristics of Capacitive His research areas are electromagnetic transients and msulation coordination
Potential Devices, IEEE Trans., Vol. PAS-90, No. 6, pp. 1989-2001, in power systems.
Sept./Oct. 1971.
G.A. Gertsch, F. Antolic, F. Gygax, Capacitive Voltage Transformers I. Polishchuk received his B.S.E.E. from St Petesburg Electrotechnical
and Protective Relays, CIGRE Report 31-14, 1968. Institute, Russia. He joined Westinghouse Canada and later ABB where he
J.R. Lucas, P.G. McLaren, W.W.L. Keeahipala, R.P. Jayasinghe, worked as a power transformer design en-&eer. In 1989 he joined Haefely
Improved Simulation Model for Current and Voltage Transformers, Trench as a Product Manager and presently is responsible for transmission
IEEETrans., Vol. PWRD-7, No. 1, pp. 152-159,January 1992. products engineering at the Instrument Transformer Division. Mr.
M. Kezunovic, Lj. Kojovic, V. Skendzic, C.W. Fromen, D.R. Sevick, Polishchuk is a registered professional en-&eer in the province of Ontario
S.L. Nilssen, Digital Models of Coupling Capacitor Voltage and a member of the IEC working group.
Transformers for Protective Relay Transient Studies, IEEE Trans.,
Vol. PWRD-7, No. 4, pp. 1927-1935, October 1992. Joe Ribeiro is the CVT Technical Leader, Haefely Trench Canada. He
A.K.S. Chadhary, K.S. Tam, A.G. Phadke, Rotection System obtained B.A.Sc. from the University of Toronto in 1988 and in 1990
Representation in the Electromagnetic Transients Program, IEEE became a registered professional engineer in Ontario. He joined Trench
Trans., Vol. PWRD-9, No. 2, pp. 700-71 1, Apnl 1994. Electric Instrument Transformer Division in May 1988 as a R & D Engineer.
Lj. Kojovic, M. Kezunovic, V. Skenddc, C.W. Fromen, D.R. Sevick, From 1988 to 1991 he has been involved in development and testing of both
A New Method for the CCVT Performance Analysis Using Field capacitor voltage transformers and current uansformers during which time he
Measurements, Signal Processing and E W P Modelling, IEEE gained considerable experience in high voltage design and testing techniques.
Trans., Vol. PWRD-9, No. 4, pp. 1907-1915,October 1994. In 1991 Jorge assumed the responsibilities of CVT Technical Leader,
H.J. Vermeulen, L.R. Dann, J. van Roojien, Equivalent Circuit manager of CVT Engineering at the Instrument Transformer Division in
Modelling of a Capacitive Voltage Transformer for Power System Scarborough, Canada. From 1991 to 1993, J o s e was also responsible for
Harmonic Frequencies, IEEE Trans., Vol. PWRD-10, NO. 4, pp. the high voltage test laboratory and testing depatments. As an ANSI C93
committee member Jorge actively participates in the revision of standards on
1743-1749, October 1995. Power Line Canier (PLC) and Coupling Capacitor Voltage Transfomers.

APPENDIX A Arash S a n h a r (M1991) received his B.Sc. and M.A.Sc. degrees in

electrical engineering from Purdue University and the University of Toronto
CCVT Parameters in 1987 and 1991 respectively. From July 1990 to December 1991 he was
with Federal Pioneer Ltd. as a design en,+eer. He h s been with the R & D
Department of Trench Electric since January 1992 as a system engineer. His
Voltage Divider: main area of specialization is HVDC harmonics and HVDC-AC switching
C,= 14611- p F , C2 = 118400 -pF, Ld = 10- mH transients.
Series Reactor:
L, = 4 2 - H , C c = 2 4 2 - p F
Harmonic Suppression Filter:
Non-linear Branch: 4 . 5 3 1.3542
Linear Branch: 75-R
STD Parameters:
R1=284.5-Q, L1=7.37-H
R2=0.019-Q, L2=0.063-mH
R3=0.020-Q, L3=0.072-mH
Rm=4.96-ML2
 required to create the plot are summarized as follows:
David A. N. Jacobson (Manitoba Hydro, Winnipeg, MB, d
Burden: 400 VA, .8 pF lagging
Canada):
Manitoba Hydro is in the process of replacing several 230 kV
b
C,=1500 pF, Ct=l 0*9 pF, C,=100 pF
STD tap ratio (high side to Xl-X2 winding): 206.7:l
wound potential transformers with CCVTs. The replacement
was determined as the best long term solution to eliminate an The mode at 14 kHz is clearly seen. To make the figure more
identified ferroresonanceproblem [ 11. We are interestedin the readable, the input and output voltages were per-unitized. A
transient model of a CCVT in order to evaluate whether the gain if 0 db represents no attenuation (instead of -63 dB in the
replacement will have any adverse affects on the performance authors paper).
of our HVdc converter. The only problem which has been
The mode at 14 kHz (mentioned in Section 5.1) is clearly
identified is the impact on valve timing. During transient con-
seen. The first notch could only be duplicated if the series
ditions, the valve timing signal output from a CCVT deviates
reactor stray capacitance was increased to 1500 pF. Could the
significantly from the actual commutating bus voltage. There-
authors comment on this discrepancy?
fore, the existing valve timing PTs will not be replaced. Are
the authors aware of any instances where the frequency
response of the CCVT has impacted the control loop of a
FACTS device or HVdc converter?
The authors have presented an EMTP model of a CCVT. The
model was verified by comparing the output voltage transients
resulting from a standard CSA ferroresonance test (CAN3-
C13.1-M79, Sectionl3.4) with digital simulation. The devel-
oped model is similar to previously reported models except
the step down transformer and ferroresonance suppression
circuitry (FSC) are represented in more detail. Previous stud-
ies have used frequency response methods to determine the
network parameters. The authors did not provide enough -30
I Y
information to duplicate the transient response waveforms -40 , ,.....I. I I ,,,...t , , .,,... I , , . , , ,
101 101 1o5 1o4
(i.e. Fig. 13). A very important parameter is the saturation Frequency (HI)
characteristic of the FSC. It should saturate before the inter-
mediate transformer in order to insert the 1.35 ohm resistor.
Could the authors provide details of the saturation characteris- 4.Can the authors comment as to how the parameters typi-
tic? cally vary as the rating of a CCVT increases from 115 kV to
I have some additional modeling questions: 500 kV? It would be convenient to have a general CCVT
model which could be used at each voltage level requiring
1. Using the saturation data of the intermediate transformer only minor modifications.
provided in Appendix A, results in a magnetizing reactance of
16 H. Based on published results of a 138 kV CCVT, this 5. Are the authors planning to perform field tests on an actual
number is normally between 1600 and 10000 H. Should the CCVT to validate some of the transients identified in their
current quoted be in millihperes rather than Amps? EMTP study? It is not clear, for example, if the transients
shown in Fig. 12b are realistic or not. At frequencies above 10
2. The authors dont state what parameters were used as base kHz other parasitic capacitances not represented in the model
values in their frequency response plots (Figs. 2-7). Did the may play a role in the measured transients.
authors perform frequency response measurements to confirm
the parameter values? REFEWNCES
3 What configuration of components was used for the tran- [l] Jacobson, D A N ,Swatek, D R I and Mazur, R.W.,
sient tests? It would be valuable to show the frequency Mitigating Potential Transformer Ferroresonance in
response corresponding to the parameters used in the tests. a 230 kV Converter Station, presented at the 1996
The excited frequencies should correspond to particular linear IEEE T&D Conference, Los Angeles, Sept. 15,
resonances (e.g. Fig. 9b and Fig. 12b). 1996.
As an example, I used the authors data (Appendix A) to
derive the following frequency response plot. Additional data Manuscript received February 19, 1997.
 629

M.R. Iravani (University of Toronto): 5. Configuration of Fig. 1 of the paper was used
The authors would like to thank the discusser for his for transient tests reported in the paper. The
interest in the paper. The following comments are in frequency response provided by the discusser
response to the questions and comments of the indicates that the first notch appears slightly above
discusser. 600 Hz corresponding to CC = 1500 pF. This is
1. We are not aware of any reported case where consistent with our results of Fig. 5. At CC= 1500 pF,
CCVT frequency response has impacted control loops the first notch based on our studies appears at 640 Hz.
of power electronic based apparatus. However, it is Our studies indicate that the first notch exists
required that the CCVT ferroresonance to be damped regardless of the magnitude of CC, Figs. 2 to 6. The
out rapidly (e.g. 2 cycles) when it is used for voltage only case where we can eliminated the first notch is
measurement at SVC terminals. when the burden is adjusted at the rated value (400-
2. The discusser is correct in stating that the unit for VA) and pf = 0.6. There is no particular pattern for
current in the STD saturation data is mA and not A. variation of CCVT parameters as the rated voltage
3. Each frequency-response plot in the paper is increases to 500-kV.
simply the ratio of output voltage to the input voltage 7. Currently we are conducting some studies to
in (db). identify sensitivity of CCVT transient response to
4. The following table provides saturation various parameters. We intend to verify our
characteristic of the suppression filter. conclusions and simulation results based on
comparison with test results.
Current (A) Flux(V S)
0.0494 13 0.27000
Manuscript received October 1, 1997.
0.358222 0.3 1886
0.980298 0.33762
2.538552 0.35637
10.000000 0.40000

View publication stats