Evaluation of Switching Capacitive Currents by

Disconnect Switches Using DIgSILENT Software Tool

Sharon Arigye-Mushabe
Department of Electrical Engineering
University of Cape Town
Cape Town, Western Province, South
Africa
arigyes@powerelec.ee.uct.ac.za

Komla A. Folly
Department of Electrical Engineering
University of Cape Town
Cape Town, Western Province, South
Africa
kfolly@ebe.uct.ac.za

ABSTRACT
This paper discusses the switching of capacitor currents
by disconnect switches. It presents a brief background
to the theory behind the switching of capacitive currents.
It seeks to demonstrate the type of effect this switching
operation has on a system by simulating a simple model
using the DIgSILENT software tool. It also attempts to
verify if the results from the DIgSILENT simulations are
comparable to the results obtained by other researchers.

In this paper, a simple model is used to simulate

switching operations using DIgSILENT. The model is
adapted from a model used by Baptista et al [6]. The
objective of the paper [6] was to study the effect that the
switching of a capacitor bank in a 60kV substation had
on a Portuguese electrical system, in particular the
effects experienced by the generating units in the
network.

1.

INTRODUCTION

Switching operations in substations or on transmission

lines have always caused problems with the transient
currents and overvoltages they generate. Even more so
now, as utilities and industries tend towards automation
and use of electronically controlled equipment, the
problems are aggravated. Increasing numbers of
capacitor banks are being located on 3-phase buses at
almost all voltage levels to improve the power factor or
in the case of transmission systems, to support voltage
and improve the power handling capability of
transmission lines [1, 2].
Switching of capacitive currents is a normal duty for
many medium and high voltage circuit breakers and
switches. As load conditions vary, these capacitive
currents are switched quite frequently. Typical cases are
the disconnecting of capacitor banks and the dropping of
unloaded overhead lines or cables [1, 2].
If the switching operation is unsuccessful, that is, if the
switch used for this purpose should restrike or reignite in
the course of the operation, severe transient overvoltages
may occur.
These switching transients produce
hazardous mechanical stresses and induce undesirable
transients in neighbouring circuits [2]. During capacitive
switching, the electrical stress experienced by the
switching device is normally relatively close to the rated
values of the device [2, 3, 4].
Disconnect switches are primarily intended for isolation
of equipment, bus sections, lines from buses, etc. They
are not specifically designed to interrupt current, but are
expected to be capable of interrupting the low currents
associated with normal no-load switching operations.
This is limited, however, by the space available for the
arc produced to get extinguished [5].

This paper, however, attempts to verify if the results

from the DIgSILENT simulations are comparable to the
results obtained in paper [6], also it includes in the
simulations with DIgSILENT, the effect de-energising a
transmission line with a disconnect switch has on the
network
This paper is structured as follows: section 2 gives a
background on the interrupting currents with disconnect
switches. Section 3 analyses the effect of the switching
of a capacitive current on a system, by simulating a
simple model using DIgSILENT. Section 4 presents the
model that was adapted from the paper by Baptista et al
[6]. Later on in section 4, the results obtained from the
switching simulations are shown and discussed. Further,
switching transient mitigating techniques are suggested
as methods to limit the transients to within acceptable
parameters. Finally, the conclusions are presented in
section 5.

2.

INTERRUPTING

CURRENTS
DISCONNECT SWITCHES

WITH

A fair amount of research has been performed

concerning the current interrupting capability of
disconnect switches [5, 7, 8]. This was done by
determining the maximum line charging current that
could safely be interrupted. In his paper, Peelo [5]
presents the maximum lengths of line and busbar for
various system voltages that can be switched by
disconnect switches. Some of them are: 61km of line for
a 15kV, 20km for 27.5KV or 3.6km for 72.5kV system
voltage. Andrews et al [7] found that for a 132kV line
charging current, a switch fitted with a horn-gap could
interrupt about 7A of charging current. In [8], Foti
demonstrated that on a 1.9mile section of line, charging
currents of 2.4A at 400kV and 3.0A at 505kV could be
interrupted by disconnect switches.
It should be mentioned that in all these studies, the
disconnect switches had to be fitted with a range of

fixtures to ensure their effectiveness and/or to reduce

wear on the fixed contacts. For example, the use of
arcing horns, high-speed whip type interrupting devices
[5], fixed resistors or non-linear resistors [8].
Different methods have been employed to try and lessen
the effects of capacitor-bank switching transients. These
include the use of circuit switchers with pre-insertion
impedances; (either an inductor or resistor in series),
whose objective is to control inrush currents. [8, 9]

3.

CAPACITIVE CURRENT SWITCHING

To analyse capacitive current switching, let us consider a

single-phase lumped element representation of a
capacitive circuit, as shown in Figure 1.
In Figure 1, V is the magnitude of the supply voltage, C
is the lumped capacitor that represents the capacitive
load, Ls, Rs and Cs represent the impedances at the
supply side of the switch (S) and L is the stray
inductance through which C is charged. It should be
noted that Ls L [3].
Switch S is opened and the current through L and C is
interrupted. After current interruption, the capacitor C is
fully charged to maximum voltage because of the
relative phase of current and voltage (current leads the
voltage by approximately 90). The capacitance, now
isolated from the source, retains its charge. As the supply
voltage reverses polarity, at half a cycle after the
interruption, the voltage across the switch (S) terminals
reaches twice the peak value of the supply voltage, as
can be seen in Figure 2.
The 2 pu voltage that develops across the switch is
potentially dangerous as there is an increased likelihood
of the switch re-igniting. When the re-ignition occurs,
the capacitance C discharges itself via the re-ignited arc
channel and the inductances Ls and L. This re-ignition
current is given by Equation 1 [2]:

i (t ) =

2V
Ls
C

sin 0 t ,

(1)

and it oscillates at a frequency, f0:

f0 =

1
2 Ls C

(2)

Figure 2: Current and voltage traces during the

interruption of a capacitive current. [3]

As previously mentioned, a single-phase representation

was used to analyse the interruption of a capacitive
current. For a three phase circuit, if the neutral of the
three-phase capacitor banks and the source neutral are
solidly interconnected, then the analysis used for a
single-phase circuit would apply to the individual phases
of the three-phase circuit without modification [2].

4.

CASE STUDY

The model investigated in this study is adapted from a

previous paper by Baptista et al [6]. The model was
chosen for its simplicity and it gives one the opportunity
to test various aspects concerning capacitive switching,
namely: the switching of capacitor banks, back-to-back
switching, de-energising of transmission lines and
monitoring the effect this type of switching has on the
generators and the transformers shown in the model.
Because not all information required was made available
to reproduce the network accurately, the parameters used
to model the network in DIgSILENT differed from those
provided in the paper [6]. .
A single line diagram of the model used is shown in
Figure 3. The network consists of two 60kV substations:
Valdigem and Telheira and a switching substation,
Soutelo; 3 generators G1, G2 and G3; 20MVar capacitor
bank at Valdigem and two 8MVar capacitor banks at
Telheira. The network frequency is 50Hz.
The analysis tool used to simulate the model network is
DIgSILENT PowerFactory Version 13.1.255.
4.1.

NETWORK ELEMENTS AND SIMULATION

4.1.1.

GENERATORS

The generators used in the simulation were salient pole

rotor types and rated voltage of 6kV. Table 1 shows the
data of the three generators
4.1.2.
Figure 1: Single-Phase lumped
representation of a capacitive circuit.

TRANSFORMERS

element

All the transformers were 20MVA, 60kV/6kV

transformers, with a phase orientation of YNyn0.

4.1.6.

NETWORK EQUIVALENT

The model shown in Figure 3 forms part of a bigger

network. This remaining network will therefore have
some influence in the way the model responds to various
operations. This external network, as shown in Fig. 3, is
modelled using the port element available in
DIgSILENT and is modelled as a slack bus with voltage
set point of 1.0pu and an angle of 0.
4.2.

RESULTS

The study was conducted by analysing the switching of

the capacitor banks for different topologies of the system
and to inspect the overvoltage effects de-energising and
energising of an unloaded transmission line has on the
system.
4.2.1.

Figure 3: Network Model [6].

Table 1: Generator characteristics

Generator:

G1

G2

G3

Power:

9MVA

10MVA

8MVA

Voltage set pt.:

0.98 pu

1.01 pu

1.03 pu

pf (cos )

0.9

0.9

0.9

4.1.3.

In the simulations, the capacitors are all initially

switched out.
During the simulation, the 8MVar
capacitors at Telheira substation are switched in at the
instant 20ms, and at 80ms, the 20MVar capacitor bank at
Valdigem is switched. The different system topologies
used in the DIgSILENT simulations are tabulated in
Table 3. For instance, the switching operation (Sw. op 3)
in Table 3 explains that before the capacitors are
switched as an initial condition, the generator G2 and G3
are removed from or switched out of the system while
G1 is generating: the G2 and G3 breakers are open,
while G1 is closed.
Table 3: Simulated topologies of the system
studied when cap banks were switched: 2X8MVar
at 20ms and 20MVar at 80ms

Switch situation

CAPACITOR BANKS

The capacitor banks were represented as pure capacities

connected in star, with a grounded neutral. These details
are given in Table 2:
Table 2: Capacitor bank details

Substation
Telheira
Valdigem

4.1.4.

SWITCHING THE CAPACITOR BANKS

Power
2 x 8 MVar
20 MVar

Voltage
60kV
60kV

Connection
Star (YN)
Star (YN)

Sw. op

G1

G2

G3

open

closed

open

open

open

closed

closed

open

open

open

closed

closed

closed

open

closed

closed

closed

open

closed

closed

closed

TRANSMISSION LINES

Transmission lines were modelled as lumped (pi)

parameters. They are all single circuit overhead lines,
with the exceptions of the 18.079km line between
Valdigem and Telheira substation and the first 7.283km
of line from Telheira substation to G1. These lines are
double-circuit lines.

The results of the simulations on the 7 topologies shown

in Table 3 are displayed in Table 4. The switches were
simulated to open immediately and all phases were
opened at the same time. Table 4 also compares these
results of DigSilent (Dig.Sim columns) with those
obtained from the paper [6].

4.1.5.

From Table 4, one notes that the results obtained from

the DIgSILENT simulations differ noticeably from those
documented in paper [6]. This can mainly be attributed
to the fact that the parameters, especially the
transmission line parameters, used to model the network
in DIgSILENT differ from those in the paper [6].

LOAD

There is only one load in the system. This load is

connected to the Telheira substation; with a rated power
of 40MW and a power factor of 0.8 lagging.

Table 4: Table comparing the results obtained

from Baptista paper [6] (Bap.) and results from
DIgSILENT 13.1 simulations (Dig.Sim)

Maximum overvoltages (p.u.)

Sw.
op

Alvadia (G2)

Covas (G3)

Pena suar (G1)

Bap.

Dig.Sim

Bap.

Dig.Sim

Bap.

Dig.Sim

2.0

2.01

2.25

1.52

1.94

1.78

1.88

1.52

2.19

2.03

2.12

1.47

1.67

1.79

1.58

1.83

1.43

1.95

1.62

1.93

1.89

1.61

1.73

1.88

2.11

1.95

1.49

1.59

1.70

1.53

1.86

1.91

1.64

Figure 4: Phase Voltages at generator G1, after

switching operation 3 (Sw.op 3) is simulated.

In their paper Baptista et al, state that the worst

transients experienced during the switching operations
were experienced at generator G1 during switching
operation three, where only generator G1 was in use. In
contrast, from Table 4, the worst transients on average
occurred at generator G3 (1.98 pu) and generator G2
(1.93 pu) during the simulations performed using
DIgSILENT. Figures 4 and 5 show the phase voltages
experienced at the generators during some of the
simulations. Figure 4 shows the phase voltages at
generator G1 during switching operation 3 (Sw. op3) as
per Table 3. The highest peak overvoltage is 1.47pu.
Figure 5 displays the phase voltages at generator 2, after
the switching operation 7 (Sw. op 7) is performed, here
the peak overvoltage is 1.70pu.
4.2.2.

SWITCHING

Figure 5: Phase voltages at generator G2, after

switching operation 7 (Sw.op 7) is simulated.

AN UNLOADED TRANSMISSION

LINE

An unloaded transmission line behaves the same as a

capacitor. During the switching operation 3, the 20.9km
line between Telheira substation and Soutelo was
generating approximately 0.63MVar of reactive power,
and carrying a capacitive current of 6.22A at 60kV. This
is likened to what Andrews et al [7], had determined to
be an interruptible charging current of 7A, however this
current was at a system voltage of 132kV.
After switching operation 3 was performed, simulation
to drop the 6.22A on the 20.9km line was performed in
DIgSILENT. A disconnect switch at Telheira substation
was switched open at 50ms to drop the line and at 80ms
the switch was closed. The resultant voltage waveforms
due to the line dropping at G2 are shown in Figure 6. No
overvoltages are measured during the de-energising
operation, but a maximum transient voltage of over 2pu
is registered when the line is re-energised.
Attempts were made to simulate some of the mitigation
techniques discussed earlier in the paper; but, using the
DIgSILENT software tool made it difficult to model
these methods. What is expected are significantly
reduced transients when using these methods, as
documented in various studies cited before [1,5,7,8].

Figure 6: Phase Voltages at generator G2 during

line dropping simulation at 50ms and energisation
of the same line at 80ms.

Unfortunately, DIgSILENT allows for only one method

to be simulated in order to test the effectiveness of the
mitigation techniques. This can be done by creating a
disconnect switch with an on-resistance and a
specified breaking time. The outcomes from these
simulation tests will follow in a future paper.

5.

CONCLUSIONS

A simple network with 3 generators, one load and 2

capacitor banks at different locations was modelled in
DIgSILENT. Several tests were performed by switching
the capacitor banks for varying system topologies of the
model. The results obtained are shown in Table 4.

Figures 4 and 5 show the overvoltages experienced at the

generators when the capacitor banks are switched.
When a line dropping simulation was performed, see
Figure 6, no transient overvoltages were experienced,
but when the line was energised, overvoltages in excess
of 2pu were recorded.
The results shown in this paper indicate that disconnect
switches can be used to switch small capacitive currents
however, transients induced into the network can be
extreme.
The future direction of this study is to ensure mitigation
techniques can be modelled simply and adequately with
DIgSILENT. Furthermore, to simulate all the switching
operations mentioned in this paper, using the same
network model and element parameters, but with another
power analysis software tool and compare the results
obtained to the results displayed in this paper.

[http://www.sandc.com/edocs_pdfs/edoc_02466
9.pdf]
[5]

D. F. Peelo, Current Interrupting Capability of

Air Break Disconnect Switches, IEEE
Transactions on Power Delivery, vol. PWRD-1,
no. 1, pp 212-216, January 1986.

[6]

J. M. R. Baptista, M. R. Cordeiro and A.

Machado e Moura, Overvoltages due to
Capacitors Bank Switching in a 60kV System,
International Conference on Power Systems
Transients. (IPST '01). Rio de Janeiro, Brazil,
June 24-28, 2001. Available Online:
[http://www.ipst.org/TechPapers/2001/IPST01P
aper038.pdf]

[7]

F. E. Andrews, L. R. Janes and M. A.

Andersson, Interrupting Ability of Horn-Gap
Switches, AIEE Transactions, vol.69, pp 10161027, 1950.

[8]

A. Foti ad J. M. Lakas, EHV Switch Tests and

Switching Surges, IEEE Transactions on
Power Apparatus and Systems, vol. PAS-83,
no. 4, pp 266-271, 1964.

[9]

S. Arigye-Mushabe and K. A. Folly, Load

Disconnect Switches as affected by switching
currents, Proceedings of the South African
University Power Engineers Conference
(SAUPEC
05),
University
of
the
Witwatersrand, Johannesburg, 27-28 January
2005.

[10]

DIgSILENT PowerFactory Version 13.1.255

APPENDIX

Line characteristics used: 60kV, rated current 500A.

R=0.0529/km;
X=0.529/km;
R0=0.1587/km;
X0=1.587/km; B=3.3081S/km; B0=0.1007S/km;
Table 5: Synchronous machine constants used to
model the generator in Figure 3

Xd
Xq
X'd

1.8
1.9
0.3

X'q

0.55

X''d

0.25

X"q
X2
X0
R(ac)
R2
T'd

0.25
0.2
0.1
0.0025
0
8s

T"d=T"q
H (inertia)

0.03
6.5s

REFERENCES
[1]

CIGRE WG 13.04, Capacitive Current

Switching-state of the art, Electra, vol.155,
Part 5-6, pp 32-63, August 1994.

[2]

A. Greenwood, Electrical Transients in Power

System. 2nd Ed. Wiley & Sons, New York,
1991.

[3]

L. van der Sluis, Transients in Power Systems.

Wiley & Sons, Chichester, 2002.

[4]

R. P. OLeary and R. H. Harner, Evaluation of

Methods for Controlling the Overvoltages
produced by the Energization of Shunt
Capacitor Bank, International Conference on
Large High Voltage Electric Systems (CIGRE),
Paris, 28 August 3 September 1988. Available
online: