An Alternative to Reduce Medium-Voltage

Transient Recovery Voltage Peaks

D. M. Nobre W. L. A. Neves B. A. de Souza

Departamento de Engenharia Elétrica - UFPB

Av. Aprígio Veloso, 882 – Bodocongó
58.109-970, Campina Grande – PB – Brazil
[email protected], [email protected], [email protected]

Abstract - Circuit breakers can fail to interrupt three- considering present and 2003 substation configurations.
phase fault currents when power systems have
Transient Recovery Voltage (TRV) characteristics,
which exceed their ratings. This paper is the outcome of II. METHODOLOGY
a study for the reevaluation of three 15 kV breakers at
the CHESF (Companhia Hidroelétrica do São The TRV to which a circuit breaker is subjected depends
Francisco) Angelim substation, located in the northeast on the type of fault, its location and the type of switching
of Brazil. Digital simulations are carried out with ATP® circuit. These studies must consider:
(Alternative Transients Program) to compute the TRV
waveforms. Alternative actions are taken to bring • three-phase ungrounded terminal fault interruption
circuits within the 15 kV circuit breaker ratings when since it is usually the most severe TRV condition, with
the parameters of the simulated TRV exceed the the highest peak voltage magnitudes across the first
breaker ratings. A simple and cheap alternative to opening-pole of the breaker. This type of fault is used
reduce TRV peaks is the placement of ZnO devices as the basis for the breaker rating;
across the breaker terminals. • the short-line fault since it usually has the highest rate
of rise of the recovery voltage.
Keywords: Circuit Breakers, TRV, ZnO Devices,
Electromagnetic Transients. The ATP is used to compute the TRV waveforms and to
calculate its parameters [2]. These parameters are compared
with the TRV circuit breaker ratings.
I. INTRODUCTION

Recently, a study to examine TRV characteristics for III. CIRCUIT BREAKER RATINGS
medium-voltage circuit breakers was conducted by Swindler
et al [1]. It was considered 13.8 kV breakers in an oil In this TRV study, the NBR-7118 Brazilian Standard,
refinery in which circuit parameters as originally designed which is based in IEC-56 International Standard, 1987, is
would have exceeded the TRV (Transient Recovery Voltage) used [3]. Its purpose is to provide an application guide on the
ratings. Digital simulations were carried out with the EMTP TRV ratings for ac circuit breakers rated above 1000V. For
(Electromagnetic Transients Program) to analyze TRV circuit breakers rated 72.5 kV and below, the TRV is defined
parameters. Alternatives to bring circuits to within the 13.8 by the Two Parameter Method in which two TRV envelopes
kV circuit breaker ratings were presented. are used to define the TRV maximum value (Uc) and the rate
The circuit transient recovery voltage can be modified by of rise of the recovery voltage (RRRV). Fig. 1 shows TRV
the circuit breaker's design or by the use of additional envelope for 15 kV circuit breakers.
components like opening resistors, surge capacitors, grading
resistors or capacitors. The most common adopted solutions
are:

• when the TRV peak value is above the breaker rating,

the existing breaker is replaced by one of higher
voltage class;
• when the rate of rise of the recovery voltage exceeds
the specified value, surge capacitors are used to reduce
it.

This paper is the outcome of a study for the

reevaluation of three 15 kV breakers at the CHESF
Angelim substation, located in Pernambuco State, Brazil. Fig. 1. Two parameter envelope
TRV parameters are analyzed for bus and line breakers, for circuit breakers rated ≤ 72.5 kV.
 The TRV envelope comprises a two-parameter reference
line (two straight line segments) and a delay line of specified
value, where:
U c is the TRV maximum value, in kV;
t 3 , time to reach U c , in µs;
t d , specified delay time, in µs; u ′ = 1 3 U c , in kV;
t ′ , time to reach u ′ , in µs.

The required 15 kV circuit breaker TRV capabilities at

three interrupting current levels for terminal faults are
indicated in Table 1.

Table 1. 15 kV breaker's capability.

Percent of U c (kV) RRRV (kV/µs)
interrupting rating
0 - 30% 27.6 1.97
30% - 60% 27.6 0.95
60% - 100% 25.7 0.38
Fig. 3. Angelim configuration
Fig. 2 shows how to obtain these TRV parameters in
Table 2. 230 kV bus equivalent.
practical situations. The RRRV is defined as the slope of the
OP line segment. For the interruption process to be Sequence 1998-2002 2003-2007
successful, the computed curve must be under the two- R (Ω) X (Ω ) R ( Ω) X (Ω )
segment curve leading to calculated TRV parameters within
15 kV breaker ratings. Positive 0.0025 0.0205 0.0023 0.0187
Zero 0.0049 0.0224 0.0029 0.0165

Table 3. Transformer ratings.

T1 T1
T2
(1998) (2003)
MVA rating 33.3 5 100
positive-sequence impedance
73.15 125.2 13.2
(Sbase = 100 MVA ) (%)

Table 4. 69 kV and 13.8 kV load.

Voltage Bus 1998-2002 2003-2007
(kV) MW MVAr MW MVAr
Fig. 2. TRV parameters.
69 104 27.4 123 32.3
13.8 3 0 4.2 0
IV. CASE STUDY: 13.8 kV CIRCUIT BREAKERS
Table 5. Connecting lead parameters.

Digital simulations using ATP were made for the R XL C
Angelim substation. Two configurations were considered: Ω/km/phase Ω/km/phase nF/km/phase
the configuration for the year 1998 up to the year 2002 and 0.147 0.349 12.35
the one for the year 2003 up to the year 2007. The Angelim
substation diagram is shown in Fig. 3 and its data are
summarized in tables 2 to 6. Table 6. Lead lengths.
The power system is represented by its 230 kV bus 230 kV Bus – 33.3 MVA Transformer 45m
Thévenin equivalent (when performing transient 69 kV Bus - 5 MVA Transformer 30m
simulations it is recommended that one should represent at 5 MVA Transformer - 11T4 Breaker 15m
least two buses, besides the one in which the breaker to be 11T4 Breaker –13.8 kV Bus 3m
evaluated is connected) [4]. The 230 kV and 13.8 kV 13.8 kV Bus - 21Y5 Breaker 5m
transformer windings are wye connected with grounded 13.8 kV Bus - 21Y3 Breaker 10m
neutral. The 69 kV windings are delta connected. 13.8 kV Bus - 21Y4 Breaker 2m
 30.0
The bushing and winding capacitances of the
transformers were considered to be nearly 3000 pF [5]. The
effective capacitances of the circuit breakers were considered
to be 600 pF.

TRV - Phase A (kV)

20.0
A. Digital Simulations

The circuit breakers to be evaluated are located at the

11T4, 21Y4 and 21Y3 substation positions, according to
Fig. 3. For each breaker, using a time step of 1µs, the 10.0
simulated TRV waveforms are obtained for three-phase
ungrounded terminal faults. For the 21Y3 breaker, digital
simulations are also performed for short-line fault clearing
for 1, 3 and 5 km from the breaker terminals.
The effect of the 13.8 kV load on the short-circuit current 0.0
and on the TRV is taken into account for the following 380.0 400.0 420.0 440.0 460.0
situations: no load, 0.5 MW and full load. The effects of Time (microseconds)
bushing add winding capacitance to ground values of Fig. 4. TRV waveform for the 11T4 breaker.
transformers T1 and T2 are considered in TRV simulations. (13.8 kV bus with no load )
AC short-circuit current peak values, Isc , and the maximum
allowed circuit breaker currents, Imax , are shown in Table 7. 40

35
Table 7. Short-circuit breaking current levels.
30
Breaker Year I (kA) I max (kA) I (%)
TRV peak, kV

sc sc 25
1998 7.39 13.07 20
11T4 2003 9.30 40 16.44
15
1998 7.39 32.66
21Y4 16 10
2003 9.29 41.07
1998 7.38 32.60 5
21Y3 16
2003 9.27 40.98 0
11T4-1998 11T4-2003 21Y4-1998 21Y4-2003 21Y3-1998 21Y3-2003
Circuit Breaker - Year
no load 0.5 MW full load reference line
Three-Phase Faults
Fig. 5. Three-phase fault TRV peaks.
Fig. 4 shows the TRV waveform for the 11T4 breaker 3,0
when clearing a three-phase ungrounded terminal fault in
present configuration. The TRV peaks and rate of rise 2,5
calculated for each breaker opening operation are shown in
2,0
RRRV , kV/ µ s

figs. 5 and 6.
It is necessary to identify the short-circuit breaking 1,5
capacity for all breakers to be evaluated. From Table 7 and
Table 1, one can see that the interrupting current levels for 1,0
the 11T4, 21Y4 and 21Y3 breakers are in the range from 0
to 30%, from 30% to 60% and from 30% to 60% of their 0,5

rated short-circuit current, respectively.

0,0
For the 11T4 breaker the RRRV values are above the 11T4-1998 11T4-2003 21Y4-1998 21Y4-2003 21Y3-1998 21Y3-2003
standard ratings for all situations and the TRV peak Circuit Breaker - Year

exceeds the standard ratings in the following situations: no load 0.5 MW full load reference line

Fig. 6. Three-phase fault TRV rate of rise.

• no load at 13.8 kV bus, years 1998 - 2002;
• up to a 0.5 MW load, year 2003 - 2007. Short-Line Fault

For the 21Y4 and 21Y3 breakers, The RRRV values Fig. 7 shows the TRV waveform for the 21Y3 breaker
exceed the standard ratings up to a 3 MW load for the when clearing a fault at 1 km from the breaker terminals for
years 1998-2002 and up to a 4.2 MW load for the years 2003 configuration. The calculated TRV parameters are
2003-2007. The TRV peaks exceeds the standard ratings for: shown in figs. 8 and 9. The parameters for short-line faults
are within standard values. The three-phase ungrounded
• no load at 13.8 kV bus, year 1998 to 2002; terminal fault at the 13.8 kV bus with no load presented the
• up to a 4.2 MW load, year 2003 to 2007. most severe TRV condition in all simulated cases.
 0.0

V. ALTERNATIVE ACTIONS TO REDUCE TRV

-4.0

The placement of ZnO devices (three or four ZnO disks

TRT - Phase A (kV)

- commercial layout - 62 mm diameter and 23 mm height)

-8.0
across the breaker terminals (Fig. 10) is an alternative to
reduce TRV peaks. Figs. 11 and 12 show simulation results
-12.0
for the 11T4 and 21Y4 circuit breakers when clearing a
three-phase ungrounded terminal fault (the TRV waveform
for the 21Y3 breaker is similar to the one for the 21Y4
-16.0
breaker).

ZnO device
-20.0

8.10 8.15 8.20 8.25 8.30

Time (ms)
Fig. 7. TRV waveform for short-line fault at 1km from 21Y3
breaker terminals (13.8 kV bus with no load ).

Circuit breaker
30
Fig. 10. ZnO device to reduce TRV peaks.
25

TRV - without ZnO devices

20
TRV peak, kV

TRV - with three ZnO disks

30.0
TRV - with four ZnO disks
15

10
TRV - Phase A (kV)

5 20.0

0
1km-1998 1km-2003 3 km-1998 3 km-2003 5 km-1998 5 km-2003
Distance from breaker terminals - Year
no load 0.5 MW full load reference line 10.0

Fig. 8. TRV peaks (short-line fault).

2,5
0.0

400.0 440.0 480.0

2,0 Time (microseconds)

Fig. 11. TRV – 11T4 breaker.

RRRV , kV/ µ s

1,5

1,0 TRV - without ZnO devices

30.0
TRV - with three ZnO disks
TRV - with four ZnO disks
0,5

0,0
TRV - Phase A (kV)

20.0
1 km-1998 1 km-2003 3 km-1998 3 km-2003 5 km-1998 5 km-2003
Distance from breaker terminals - Year
no load 0.5 MW full load reference line

Fig. 9. TRV rate of rise (short-line fault).

10.0

Note in figs. 6 and 9 the reference line values for 21Y3

breaker are different since the AC short-circuit current
peak values calculated for interrupting short-line faults are
smaller than for three-phase ungrounded terminal faults. 0.0
Thus, as shown in Fig. 9 and according to Table 1, the 400.0 440.0 480.0
21Y3 breaker TRV capability is at 0-30% level of Time (microseconds)

interrupting rating (1.97 kV/µs). Fig. 12. TRV – 21Y4 breaker.

 The calculated TRV parameters are shown in tables 8 are nearly 0.65mA and 0.18mA, respectively. The used
and 9. If ZnO devices are used the TRV peaks will be within arrester v-i curve was the one for a temperature of 40ºC
standard values (Table 1). [7]. The energy absorbed by the ZnO blocks in either case
The simulated absorbed energy curve for a three ZnO is very small.
disk device is shown in Fig. 13. After 2ms the energy Comparing tables 8 and 9 with Table 1, the TRV rate of
increase is very small. rise values are still above the standard ratings. Thus, it is
recommended the installation of surge capacitors at the
Table 8. TRV calculated parameters - three ZnO disks
source side breaker terminals.
Uc (kV) RRRV (kV/µs) For the 11T4 circuit breaker a 37 nF (four ZnO disks)
Breaker
Year

without with without with surge capacitor would reduce the RRRV to values within
ZnO ZnO ZnO ZnO the standard rating in the two configurations. With this
device device device device surge capacitor, a RRRV reduction of nearly 50% is
accomplished for the year 2003-2007 configuration.
11T4 28.07 15.44 2.51 2.14
Similarly, for the 21Y4 and 21Y3 circuit breakers, 159
1998

21Y4 27.92 15.50 1.90 1.72 nF (four ZnO disks) surge capacitors would bring the
21Y3 28.17 15.56 1.99 1.72 RRRV values to standard rating.
The use of four ZnO disks to reduce TRV peaks is more
11T4 33.73 16.86 2.92 2.43 appropriate and secure comparing to the use of three ZnO
2003

21Y4 33.72 15.60 2.54 1.93 disks. The maximum continuous operating voltage (MCOV)
of the arrester and its residual voltage are nearly 11.2 and 20
21Y3 33.94 15.67 2.60 1.98 kV, respectively.

Table 9. TRV calculated parameters - four ZnO disks

VI. CONCLUSIONS
Uc (kV) RRRV (kV/µs)
Breaker
Year

without with without with

A study for the reevaluation of 15 kV breakers was
ZnO ZnO ZnO ZnO
performed for the CHESF Angelim substation. The
device device device device
calculated TRV responses were compared to the
11T4 28.07 20.49 2.51 2.24 capabilities of the circuit breakers.
1998

21Y4 27.92 20.60 1.90 1.83 In some situations, the TRV exceeded the breaker
ratings. From simulation results, it was shown that the
21Y3 28.17 20.62 1.99 1.84 placement of ZnO devices across breaker terminals reduces
11T4 33.73 20.64 2.92 2.66 TRV peaks. Surge capacitors, at the source side, were used
to reduce the TRV rate of rise.
2003

21Y4 33.72 20.72 2.54 2.18 Although simulations have shown that ZnO devices
21Y3 33.94 20.81 2.60 2.14 placed across the breaker terminals are effective in
reducing TRV peaks, it would be interesting to carry out
40.0
laboratory or field tests for validation purposes.

VII. ACKNOWLEDGMENTS
30.0

The authors are grateful to the reviewers for their

valuable suggestions. The authors would like to thank the
Energy (J)

Electrical Engineering Department of UFPB for its support

20.0
and CHESF for the substation data used in this work. The
authors also would like to thank M. O. B. C. Melo, F. R.
Alves, J. S. Monteiro (CHESF Engineers), and prof. E. G.
10.0
Costa, D. Sc. (UFPB). Miss D. M. Nobre wishes to
acknowledge to Brazilian National Research Council
(CNPq) for financial support.

0.0

0.0 2.0 4.0 6.0 8.0 VIII. REFERENCES

Time (ms)
Fig. 13. Three ZnO disk devices absorbed energy.
[1] D.L. Swindler, P. Schwartz, P.S. Hamer, S.R. Lambert,
At steady-state the rms voltage across the ZnO device is "Transient recovery voltage considerations in the
nearly 8 kV. The maximum continuous operating voltage application of medium-voltage circuit breakers", IEEE
(MCOV) for the power system is 8.8 kV. The steady-state Transaction on Industry Applications, vol. 33, nº. 2,
current peak values to three and four block ZnO devices March/April 1997, pp. 383-388.
[2] Leuven EMTP Center, ATP - Alternative Transients [5] IEEE Application Guide for Transient Recovery Voltage
Program Rule Book, Belgium, July 1987. for AC High-Voltage Circuit Breakers Rated on a
Symmetrical Current Basis - IEEE C37.011-1994
[3] NBR-7118 High Voltage Circuit Breaker Brazilian Standard, New York, NY, June 1995.
Standards (In Portuguese), Rio de Janeiro, Brazil, 161 [6] D.M. Nobre, Transient Recovery Voltage Reevaluation
p., September 1994. Studies for 15 kV-class Circuit Breakers (In Portuguese),
M.Sc. Dissertation. UFPB. Campina Grande, Brazil, 85
[4] CCON ((North-Northeast Operation Coordinating p., November 1999.
Committee), Methodology and Criteria for the Analysis
of Circuit Breakers Adequacy to Transient Recovery [7] E. G., COSTA, ZnO Surge Arrester Performance
Voltages (In Portuguese), Recife, Brazil,214 p., June Analysis (In Portuguese), D.Sc. Thesis. UFPB Campina
1991. Grande, Brazil, 177 p., April 1999.