IET Generation, Transmission & Distribution

Research Article

Parameters affecting the arcing time of HVDC ISSN 1751-8687

Received on 16th February 2018

circuit breakers using black box arc model

Revised 28th July 2018
Accepted on 18th September 2018
E-First on 18th January 2019
doi: 10.1049/iet-gtd.2018.6264
www.ietdl.org

Osama E. Gouda1, Doaa Khalil Ibrahim1 , Adel Soliman2

1Faculty of Engineering, Cairo University, Egypt
2ElectricityCompany, South Cairo Distribution, Egypt
E-mail: [email protected]

Abstract: Arc interruption of high voltage direct current (HVDC) circuit breakers (CBs) is one of the main challenging factors for
using HVDC grids. To evaluate the arc interrupting capability in HVDC CBs, black box arc models are used to represent the
nonlinear arc conductance depending on Cassie and Mayr dynamic arc equations. Extensive simulation studies are carried out
to investigate the effect of controlled and uncontrolled parameters on the CB arcing time. A real line represents a part of 500 kV
electrical connection systems between Egypt and the Kingdom of Saudi Arabia is simulated to be a faulty load. It is found that
the arcing time of the HVDC CB can be reduced by increasing the value of cooling power coefficient (p) and decreasing the
value of arc time constant (τ). It is also deduced that the arcing time is reduced by the increase of the commutation capacitance
value (C) and decreasing the commutation inductance (L) value and vice versa. Moreover, it is concluded that the arcing time is
greatly affected by the fault location and the fault arc resistance (Rf ) according to fault conditions.

1 Introduction quickly to zero [20]. Besides, a circuit element is used to control

the rate of rise of recovery voltage (RRRV), and an absorber
High voltage direct current (HVDC) circuit breakers (CBs) play a element to absorb the energy stored in the system inductance after
vital role in the growth of HVDC power systems, which can widely arc interruption. These two circuit elements are used with the arc
help in replacing fossil fuels with renewable energy sources [1–8]. simulation by both Mayr and Cassie models [18, 32].
The main usage of the HVDC system is to connect two AC This study examines the influences of different parameters that
networks with different frequencies and to transmit large amounts affect the arc interruption time and determines the best values of
of power through long distances [9–15]. cooling power coefficient (p), arc time constant (τ), the
The key challenge that faces the expansion of HVDC CBs is the commutation capacitance (C) and the commutation inductance (L).
absence of natural current zero crossing. As a matter of fact, in AC It investigates also the influence of fault arc resistance (Rf ) with its
interrupting processes the current decrease to zero naturally. On the
maximal and minimal values at different circumstances. Moreover,
other hand, in DC interrupting processes the current needs to be
it investigates the influence of changing the fault location on the
forced to zero. Therefore to form an HVDC CB, it is necessary to
CB arcing interruption time. DC fault test bed modelling is carried
install additional components on conventional AC CB to form
out by Matlab/Simulink software to evaluate the capability to
artificial current zero crossing [16–18]. A lot of works have been
protect the HVDC overhead transmission line, connecting Badr
directed to create artificial zero crossing in the DC current by using
substation in Egypt and Elnabaq switching station. Such a line
active and passive commutation types [18–20].
represents a part of 500 kV electrical connection systems between
Arc is considered the main aspect of the interruption process.
Egypt and the Kingdom of Saudi Arabia.
Many electric arc models are developed for describing arc
behaviour [20–23]. Conventional arc models are classified into
physical arc models and black box arc models. The physical arc 2 Modelling of HVDC CB
model describes the entire arc behaviour during the interruption. Fig. 1 shows the puffer type of the SF6 gas CB structure, including
Therefore, it can be used to investigate the arc behaviour in detail,
the simulated arc, rated voltage of 525 kV, lightning impulse
but it is very complicated to be applied. Otherwise, the black box
withstand voltage of 1175 kV, rated normal current of 2000 A,
arc model can be considered the proper method to describe the arc
rated short time current (1 s) of 40 kA, rated peak withstand current
behaviour [20, 24–27]. Black box models represent the nonlinear
of 100 kA, and rated short-circuit making current of 100 kA.
arc conductance variation with time [20–24]. The choice of black
As it is known, the HVDC arc can be considered as a nonlinear
box model equations to determine their parameters is essential;
phenomenon, which occurs due to two factors. The first one is the
hence it requires making certain assumptions about the arc
high short circuit current that generates heat leading to the CB
behaviour [16–19].
contacts and quenching medium temperatures increase,
As the main purpose of the black box arc model is to describe
consequently, a sufficient quantity of electrons are emitted. In
the interaction between the arc and the electrical circuit during the
addition, at the arc initiation when the value of the voltage between
current interruption process, Cassie and Mayr dynamic arc
the two contacts exceeds the ionisation voltage of the inter-
equations can be considered the most representative arc models
electrode gas, the gas may be sufficiently ionised [30].
[16–18].
SF6 gas has electronegativity characteristics, which means that
For applying Mayr's model to interrupt DC current, a resonance
the SF6 gas takes the free electrons far away from the field, as a
inductor–capacitor (LC) circuit is coupled in parallel with the CB
result of this action; the SF6 gas becomes electrically unstable and
to generate self-excited oscillatory current superimposed on the DC
leaves the electrons quickly. Consequently, the electrons move in a
current forming zero current crossing [16, 21–34]. However, when
random motion, where the only way to reignite the arc is to achieve
applying Cassie's model, there is no need to use a parallel
an oriented motion of the free electrons.
resonance circuit with the breaker, it is only required to make the
steady state arc voltage over supply voltage to decrease the current

IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 4, pp. 461-467 461
© The Institution of Engineering and Technology 2018
 17518695, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-gtd.2018.6264 by Cochrane Romania, Wiley Online Library on [06/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
electrical arc. The description of these circuit elements is illustrated
in the following sections.

2.1 Commutation circuit

The commutation circuit element contains two parts. The first one
is an auxiliary switch to insert the commutation element in parallel
with the HVDC CB during the fault clearing process. After the arc
voltage reaches a certain value near to 50 V, this auxiliary switch
closes. The second part is the commutation inductance and
capacitance which are used to generate a high-frequency
oscillatory current superimposed on the DC current in the nominal
path, in order to create artificial current zero crossing. At this
instant the HVDC SF6 gas CB has the ability to interrupt the
electrical arc [16, 18, 32]. Actually, this element is only used with
Mayr model, while there is no need for using it in Cassie model.
Fig. 1 Puffer type of SF6 CB structure
2.2 Controlling RRRV circuit element
Black box arc models are applied using the models of Mayr and
This circuit element is connected in parallel with the interrupter. It
Cassie. Mayr's equation can be deduced from the energy balance
is used to control the RRRV. After the HVDC CB has extinguished
equation as given in (1) [24]
the electrical arc, the current is deviated to flow through the
commutation path. As a result of this action, the commutation
dQ
= PH −PQ, (1) capacitor keeps charging up. Consequently, a very high restriking
dt
voltage is generated across the breaker contacts. Therefore, it is
important to use R–C circuit element or use a capacitor only (C p),
where P H is the heat generated with the assumption that the
to control the RRRV. The value of the capacitor (C p) should be less
cooling power P Q and the arc conductance g Q are arbitrary
than the commutation capacitor value (less than one-tenth) [18].
functions of Q.
This circuit element is used with both Mayr and Cassie models as
The general form of the dynamic arc equation of current i is
shown in Figs. 2a and b.
represented as follows [24]:

dg 1 i2 2.3 Absorbing element

= −p . (2)
dt dQ/dg g This element is connected in parallel with the interrupter. It is not
only used to absorb the energy stored in the system inductance
By further assumption of Q = Q0. ln (g /G0), where Q0 and G0 are after current interruption, but also used to limit the recovery
constants describing the arc, the thermal time constant is defined voltage, by choosing the proper value of clipping voltage. The
by τ = Q0 / p. Thus, Mayr dynamic arc equation can be presented as energy absorber element may be a nonlinear element or a linear
given in (3) element in addition to an auxiliary switch. This auxiliary is
switched on after the value of the voltage across the breaker
Pτ dg i2 terminals reaches the clipping voltage value. After the energy
⋅ = −p . (3) stored in the system, inductance has dissipated through the energy
g dt g
absorber element, the current in the main path falls to almost zero.
It is worth mentioning that by eliminating the hypothesis that the At this moment the interruption process is assumed to be
time constant τ and cooling power p are constants, it leads to a completed. It is important to select the clipping voltage value and
generalised form of Mayr's equation as given below in (4), where u the energy rating in a proper way, in order to clear the fault quickly
is the arc voltage [24]: [16, 18, 32]. This branch is used with Mayr and Cassie models as
shown in Fig. 2b.
1 dg 1 gu2
= −1 . (4) 3 Evaluating proposed DC interrupter
g dt τ p
3.1 Tested transmission line and applied models
Meanwhile, Cassie defines the arc behaviour with the following
hypotheses and assumptions: (i) the arc column has a cylindrical The proposed HVDC interrupter is evaluated in protecting a 450
shape filled with highly ionised gas and free electrons, (ii) arc km overhead DC transmission line connecting Badr substation in
cylindrical column has uniform temperature and current density, Egypt and Elnabaq switching station. Such a line represents a part
but its diameter is altered in time and accommodate the change in of 3000 MVA, 500 kV HVDC electrical interconnection systems
current, (iii) arc voltage is considered constant during the arc between Egypt and the Kingdom of Saudi Arabia. The conductor
process, and finally (iv) the power dissipation is considered has 4 × 862 mm2 Aluminum-Clad Steel Reinforced/American Wire
proportional to the column cross-sectional area [20]. With these (ACSR/AW) Lapwing per pole and 4 × 775 mm2 ACSR/AW
assumptions, a linear relation between arc conductance and energy Bobolink of neutral conductor split into separate two-bundle
storage capacity of the arc is presented by the following (5), which configuration on either side of the towers. The DC fault test bed is
represents the last form of Cassie's equation [20]: implemented using Matlab/Simulink to carry out transient response
simulations of the switching arc. Fig. 2a illustrates the applied
1 dg 1 u2 Mayr's black box arc model test circuit while Fig. 2b shows the test
= −1 , (5) circuit for the simulated Cassie's black box arc model.
g dt τ Uc2
Table 1 illustrates the parameters used for simulation both
where τ is the arc time constant; p is the arc power loss coefficient; Mayr's and Cassie's models. Test bed includes 500 kV HVDC
g is the arc conductance (i/u); u is the arc voltage and Uc is the source supplying a load of 650 Ω (R load) through 450 km DC
steady state arc voltage. overhead transmission line with resistance R1 of 16.3 Ω, inductance
After simulating the nonlinear arc conductance using black box of L1 of 45 mH Ω, energy absorber element with clipping voltage
arc models, it should be noted that there is a need to use additional of V ref = 650 kV, and switching arc block represents a single
circuit elements that are connected with HVDC CB to interrupt the break SF6 CB puffer type.

462 IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 4, pp. 461-467
© The Institution of Engineering and Technology 2018
 17518695, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-gtd.2018.6264 by Cochrane Romania, Wiley Online Library on [06/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Fig. 2 Applied black box arc models
(a) Test circuit of Mayr's model, (b) Test circuit of Cassie's model

Table 1 Parameters for simulating Mayr's and Cassie's

models
Model Parameter Default value
Mayr model p, MW 100
τ, µs 10
Lc, mH 0.30
Cc, µF 20
Rp, Ω 75
C p, µF 0.1 Fig. 3 Response of DC CB for a line to line fault using Mayr's model
(a) Current, (b) Voltage, (c) Current of resonance circuit, (d) Current of absorber
Cassie model Uc, kV 750
element
τ, µs 100
Rp, Ω 75 the electrical arc is generated as it is observed in Figs. 3a and b. At
C p, µF 0.001 a value of almost 50 V, the series LC resonance circuit is connected
in parallel with the DC CB to generate a high-frequency oscillatory
current by self-excitation, this oscillatory current is superimposed
3.2 Results and discussions on the dc current of the main path and its amplitude grew to form a
current zero-crossing.
Fig. 3 illustrates the results of applying the Mayr model test bed. It is also noticed that from Fig. 3a at the point of 0.0167 s, the
The test is done for the CB rated current 750 A and line to line HVDC CB completely interrupts the arc current and extinguishes
fault current 20.4 kA occurs at 0.005 s as shown in Fig. 3a. It is the arc. As the clipping voltage value in this system is 650 kV as
noticed that when the DC current reaches ∼20 kA, the fault is shown in Fig. 3b, it is noticed from Figs. 3c and d that the fault is
detected and the DC CB successfully starts to open at 0.01 s and cleared after the energy absorber element has absorbed the energy

IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 4, pp. 461-467 463
© The Institution of Engineering and Technology 2018
 17518695, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-gtd.2018.6264 by Cochrane Romania, Wiley Online Library on [06/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 2 Mayr's model parameters for different simulated
tests at fixed values of commutation parameters
Parameter Default value Tested value
P, MW 100 65, 85, 115
τ, µs 10 5, 15, 20
Lc, mH 0.25 0.25
Cc, µF 25 25
Rp, Ω 75 75
C p, µF 0.1 0.1

the second current zero shortly. This principle of having second

current zero shortly after the first allows proper recovery during
several microseconds and enhances the chance to clear the fault.
Meanwhile, it must be pointed out that the Cassie model concept
will not be considered in the rest of the paper as it does not ensure
the interrupt capability of the breaker.

4 Parameters affecting the arcing time

Many researchers have investigated the parameters that affect the
arc interruption time and the DC arc current interruption [21–24].
The best values of the arc parameters that achieve the best arcing
time are also investigated in [16]. Yet, investigating the effect of
arc fault resistance on arcing time has not been fully addressed in
the literature. In the following sections, extensive tests are carried
out to investigate the effect of arc parameters, commutation
parameters, arc fault resistance and fault location on arcing time.
These parameters are divided into two groups: controlled
parameters that can be changed by the manufacturer of the
protection system, and uncontrolled parameters that cannot be
controlled by the manufacturer of the CBs.

4.1 Controlled parameters

4.1.1 Arc parameters: To well test the effect of arc parameters on
arcing time for Mayr's model, one of the two parameters either
cooling power coefficient (p) or arc time constant (τ) is kept
Fig. 4 Response of DC CB for a line to line fault using Cassie's model constant while varying the value of the other one as given in
(a) Current, (b) Voltage, (c) Current of the absorber element Table 2. In fact, for SF6 gas CBs, the cooling power coefficient
depends on the gas pressure by the puffer cylinder. When the gas
stored in the system inductance, and the current ceases to almost pressure is high, the cooling power coefficient will take a high
zero. value and the arc length is increased, leading the arc voltage to rise
Fig. 4 shows the results of applying the Cassie model test bed and extinguish the hot arc quickly. So it appears clear that a higher
where the rated current and line to line fault circumstances are cooling power coefficient does not only increase the arc voltage but
similar to those applied in Mayr's model. It is also observed that also helps to extinguish the hot arc, reduce arcing time, and vice
when the DC current reaches 20 kA, the fault is detected and the versa. In fact, cooling power coefficient can be changed by
DC CB contacts started to open at 0.01 s as shown in Fig. 4a. changing the diameter of buffer cylinder and the gas pressure [21,
The test shows that for Cassie model there is no need for using 23].
resonance LC circuit connected in parallel with the HVDC CB, The effect of varying the value of cooling power coefficient (p)
while the value of the steady state arc voltage should be over than on the arcing time during the interrupting process of the arc in
the supply voltage to decrease the value of fault current to reach HVDC CB, with a fixed value of arc time constant is clearly
zero quickly, otherwise the breaker cannot effectively interrupt the illustrated in Fig. 5a. As shown, the arc interruption time has a
fault current. It is noticed that during the interval time near current large value (19.8 ms) for the smallest value of p at 65 MW, and
zero crossing, not only the arc current is decreased, but also the arc while p is increased to 85 MW, the arcing time is reduced to 16.39
voltage is decreased as shown in Fig. 4b. The current waveform of ms and further decreased to 14.31 ms when p is further increased to
the absorbing path current is also illustrated in Fig. 4c. Actually, as 115 MW. It is also observed that for the low value of cooling power
the principle of Cassie arc model depends on the convection loss in coefficient (p), the HVDC CB cannot interrupt the current at the
the region of high current, so that the Cassie arc model is not fit first current zero crossing.
near current zero crossing. As it is known, the voltage ignition time depends on the arc
By comparing the two applied models, it is deduced that it is time constant, when the arc time constant grows, the starting point
essential to connect LC resonance circuit in parallel with the of voltage ignition is delayed so that the fault duration time is
HVDC CB to generate self-excited oscillatory current increased and the current waveform is getting higher and wider.
superimposed on the dc current forming current zero-crossing as in Moreover, the arc time constant is responsible for making the
Mayr's model. However, for the Cassie model, there is no need to electrical arc exhibits negative resistance characteristics, by
do that. It is only required to make the steady state arc voltage over choosing the proper value of arc time constant. In this case, the arc
supply voltage and decrease the value of arc time constant. conductance can follow the changes in the arc current. Meanwhile,
Consequently, Mayr's model is considered more flexible to modify the HVDC CB is able to interrupt the electrical arc. On the other
the waveform and it is found that Mayr's model creates the most hand, if the value of the arc time constant exceeds a critical value,
realised diagram. For Mayr's model, it appears that the breaker the electrical arc will exhibit positive arc characteristics. In this
cannot interrupt the current for many cases at first current zero case, the arc conductance cannot be able to follow the changes in
crossing, but soon later, the breaker is anticipated to interrupt it at the arc current at all. As a result of this action, the HVDC CB

464 IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 4, pp. 461-467
© The Institution of Engineering and Technology 2018
 17518695, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-gtd.2018.6264 by Cochrane Romania, Wiley Online Library on [06/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
cannot be able to interrupt the electrical arc [30]. Therefore, the
effect of varying the value of arc time constant (τ) on arc
interruption time during the arc interrupting process of HVDC CB
at a constant value of cooling power coefficient (p) is fully
investigated in Fig. 5b. It is found that for τ = 5 µs, the arcing time
is 14.7 ms, and by increasing τ to 17.5 µs, the arcing time is
increased to 15.98 ms and finally by increasing τ to 5 times (τ = 25
µs), the arcing time is increased to 16.6 ms. It is deduced that when
the arc time constant grows, the model predicts the failure of
interruption at the first current zero, but soon thereafter, the breaker
is predicted to interrupt at the second current zero.
This principle of having a second current zero crossing shortly
after the first not only exposes the ability of the SF6 gas CB to
recover through few microseconds but also enhances the capability
to clear the fault. Actually, arc time constant can be changed by
changing the value of the speed of opening contacts [23].
Fig. 5c summarises the effect of varying the value of arc time
constant (τ) at a constant value of cooling power coefficient (p),
and the effect of varying the value of cooling power coefficient (p)
at a constant value of arc time constant on arcing time. The
relationship between the arcing time and cooling power coefficient
in one side and the arc time constant on the other side given in
Fig. 5c could be exponentially fitted as given in (6) and (7) as
follows:

T s = 0.1215 × p−0.4725, (6)

T s = 0.0116 × τ0.0744 . (7)

According to Fig. 5c, the best values of the arcing time constant (τ)
and cooling power coefficient (p) to get less arcing time (T) can be
estimated, by determining the cross point of the line which
represent the relationship between the cooling power coefficient
and the arc interruption time, with the line which represent the
relationship between the arc time constant and the arc interruption
time, such values are: p = 97 MW and τ = 17.5 µs. These findings
are in agreement with that reported in [16].

4.1.2 Commutation parameters: To check the effect of

commutation parameters on the arcing time of HVDC CB by using
Mayr's model, one of the two parameters: commutation inductance
(L) or commutation capacitance (C) is maintained constant at the
default value while the value of the other parameter is changed
using different test values as given in Table 3.
Consequently, the effect of varying the commutation
capacitance (C) on the HVDC CB arcing time at the default value
of commutation inductance (L) is examined as given in Fig. 6a.
The results show that for the same value of inductance (L), the
arcing time is reduced with the increase of the capacitance (C), and
the discharging current is also increased, and thus it is easier to
produce a higher commutation current. The arcing time is therefore
reduced because the artificial zero current is generated more
quickly. As noticed from Fig. 6a, the arcing time is increased when
the capacitance (C) is reduced, although the frequency of the Fig. 5 Investigating Mayr's model responses for arc parameters change
oscillatory current is lower, its amplitude is higher, and this results (a) Transient responses at a constant value of time constant (τ) and different values of
in a shorter arcing time. It is observed that for the smallest value of cooling power (p), (b) Transient responses at a constant value of cooling power
C (15 µF), the arc interruption time has the largest value of 14.95 coefficient (p) and different values of time constant (τ), (c) Effects of P and τ values on
ms which is significantly reduced to 14.44 and 13.6 ms when the arcing time
value of C is increased to 20 and 35 µF, respectively.
On the other hand, the effect of varying the commutation Table 3 Mayr's model parameters for different simulated
inductance (L) on the arcing time at the same value of commutation tests at fixed values of arc parameters
capacitance (C) is studied in Fig. 6b. The achieved results show
that the arcing time is reduced with reducing the inductance (L) for Parameter Default value Tested value
the same value of capacitance (C), and the arcing time is increased P, MW 110 110
with the increase of the inductance (L). It is observed also that for τ, µs 15 15
the largest value of L (0.30 mH), the arc interruption time has the Cc, µF 30 15, 20, 35
largest value of (15.82 ms) which is significantly reduced to Lc, mH 0.20 0.10, 0.15, 0.30
(12.99 ms) and (12.22 ms) when the value of (L) is decreased to
Rp, Ω 75 75
(0.15 mH) and (0.10 mH), respectively.
Fig. 6c summarises the effect of varying the commutation C p, µF 0.1 0.1
capacitance (C) at a constant value of commutation inductance (L)
on the arcing time and vice versa. According to Fig. 6c, the best
values of commutation capacitance (C) and commutation
IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 4, pp. 461-467 465
© The Institution of Engineering and Technology 2018
 17518695, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-gtd.2018.6264 by Cochrane Romania, Wiley Online Library on [06/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Fig. 7 Investigating Mayr's model responses for fault arc resistance
change
(a) Transient responses at different values of fault arc resistance, (b) Effects of fault
arc resistance value on arcing time

(from 0.1 to 10 Ω) on the arcing time of the HVDC CB as shown in

Fig. 7a. It is observed from such figure that the increase of the fault
resistance leads to fast arc interruption and fewer oscillations of arc
current with lower magnitudes. This may be due to the high value
of fault resistance helping in limiting the fault current to a lower
value, while the arc voltage increases quickly forcing the current to
decrease quickly to zero. However, for the low value of fault
resistance, the fault current increases rapidly and the arc voltage
Fig. 6 Investigating Mayr's model responses for commutation parameters decreases, and the interrupter needs a long time to interrupt the
change fault current.
(a) Transient responses at constant value of inductance (L) and different values of The effect of resistance value change on arc interruption time is
capacitance (C), (b) Transient responses at constant value of capacitance (C) and summarised in Fig. 7b. It ensures that the arc interruption time is
different values of inductance (L), (c) Effects of commutation inductance (L) and reduced from 20.6 ms at 0.1 Ω (the minimal value of fault
capacitance, (C) values on arcing time resistance) to 13.76 and 11.31 ms when the resistance value is
increased to 4 and 10 Ω, respectively.
inductance (L) that meet the best arcing time can be estimated, by
determining the point of the intersection of the curve represents the 4.2.2 Fault location: The fault impedance of the HVDC
relationship between the commutation capacitance and the arcing transmission line is computed from the point of fault location to the
interruption time, with that represents the relationship between the supply of the system. So that if the point of fault location is far
commutation inductance and arcing interruption time, which are: L away from the supply point, the fault impedance of HVDC
= 0.21 mH and C = 32 µF. Furthermore, it is found that the two transmission line increases, on the other hand, if the point of fault
relationships are not straight, because in many cases the HVDC CB location is near to the supply point, the fault impedance of the
cannot interrupt the electrical arc at the first current zero crossing. HVDC transmission line decreases. Table 4 summarises the effect
of changing the fault location on arc interruption time. It ensures
4.2 Uncontrolled parameters that the arc interruption time is reduced from 70.86 ms for a fault
located at 10% of the line to only 13.76 ms if the fault occurs at
4.2.1 Fault arc resistance: According to [28], the minimal value 100% of the line. It is deduced that when the fault location is far
of fault resistance for line-to-line faults is 0.1 Ω and tends to be away from the supply, the value of the fault impedance which is
<0.1 Ω for fault current >10 kA, while the maximal value of fault considered as a part of fault resistance is increased, consequently,
resistance is 4 Ω for instantaneous protections, and 10 Ω for the fault current decreases and accordingly the interruption time
delayed protections. Meanwhile, extensive simulated tests are decreases, also and vice versa.
carried out to investigate the effect of arc fault resistance change
466 IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 4, pp. 461-467
© The Institution of Engineering and Technology 2018
 17518695, 2019, 4, Downloaded from https://ietresearch.onlinelibrary.wiley.com/doi/10.1049/iet-gtd.2018.6264 by Cochrane Romania, Wiley Online Library on [06/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table 4 Mayr's model parameters for simulated tests at [13] Zhang, X., Bai, J., Cao, G., et al.: ‘Optimizing HVDC control parameters in
multi-infeed HVDC system based on electromagnetic transient analysis’, Int.
different fault locations J. Electr. Power Energy Syst., 2013, 49, pp. 449–454
Fault location Line fault impedance Interruption time, ms [14] Chang, B., Cwikowski, O., Barnes, M., et al.: ‘Multi-terminal VSC-HVDC
10% line 1.63 + j 0.0045 70.86 pole-to-pole fault analysis and fault recovery study’. Proc. IEEE 11th IET Int.
Conf. on AC DC Power Transmission, Birmingham, UK, 10–12 February
20% line 3.26 + j 0.009 47.44 2015, pp. 1–8
30% line 4.89 + j 0.0135 34.75 [15] Mohammadi, M., Avendano-Mora, M., Barnes, M., et al.: ‘A study on fault
ride-through of VSC-connected offshore wind farms’. Proc. IEEE Power
40% line 6.52 + j 0.018 27.44 Energy Soc. General Meeting, Vancouver, BC, Canada, 21–25 July 2013, pp.
50% line 8.15 + j 0.0225 23.02 1–5, doi:10.1109/PESMG.2013.6672514
[16] Xiang, B., Zhang, L., Yang, K., et al.: ‘Arcing time of a DC circuit breaker
60% line 9.78 + j 0.027 19.62 based on a superconducting current-limiting technology’, IEEE Trans. Appl.
70% line 11.41 + j 0.0315 17.67 Supercond., 2016, 26, (7), pp. 1–5
80% line 13.04 + j 0.036 15.75 [17] May, T.W., Yeap, Y.M., Ukil, A.: ‘Comparative evaluation of power loss in
HVAC and HVDC transmission systems’. IEEE Region 10 Conf. (TENCON),
90% line 14.67 + j 0.0405 14.73 Singapore, November 2016, doi:10.1109/TENCON.2016.7848080
100% line 16.3 + j 0.045 13.76 [18] Darwish, H.A., Izzularab, M.A., Elkalashy, N.I.: ‘Enhanced commutation
circuit design of HVDC circuit breaker using EMTP’. IEEE/PES
Transmission and Distribution Conf. and Exhibition, Dallas, Texas, USA,
May 2006, doi:10.1109/TDC.2006.1668635
5 Conclusions [19] Qin, T., Dong, E., Liu, G., et al.: ‘Recovery of dielectric strength after DC
interruption in vacuum’, IEEE Trans. Dielectr. Electr. Insul., 2016, 23, (1), pp.
In this study, the black box arc model is used to represent the 29–34
[20] Lim, S.-W., Amir Khan, U., Lee, J.-G., et al.: ‘Simulation analysis of DC arc
nonlinear arc conductance depending on Mayr and Cassie models. in circuit breaker applying with conventional black box arc model’. 2015,3rd
Mayr's model has proved more flexibility to study the effect of Int. Conf. on Electric Power Equipment – Switching Technology (ICEPE-ST),
different controlled and uncontrolled parameters on the arcing time Busan, Korea, 25–28 October 2015
of HVDC CBs. It is found that the arcing time is reduced by [21] Nakao, H., Nakagoshi, Y., Hatano, M., et al.: ‘DC current interruption in
HVDC SF6 gas MRTB by means of self-excited oscillation superimposition’,
increasing the value of cooling power coefficient, decreasing the IEEE Trans. Power Deliv., 2001, 16, (4), pp. 687–693
arcing time constant, increasing the value of commutation [22] Smeets, R.P.P., Kertész, V.: ‘Application of a validated AC black-box arc
capacitance and decreasing the value of commutation inductance. model to DC current interruption’. 2nd Int. Conf. on Electric Power
The most suitable values of the arc controlled and uncontrolled Equipment - Switching Technology (ICEPE-ST), Matsue, Japan, 2013, pp. 1–
4, doi:10.1109/ICEPE-ST.2013.6804332
parameters that can achieve the best arcing time are determined. [23] Zhu, K., Li, X., Jia, S., et al.: ‘Study of the switching arc characteristics of a
It is also concluded that the fault arc resistance and fault site on 500 kV HVDC self-excited oscillatory metallic return transfer breaker’, IEEE
the transmission line play important factors in influencing the Trans. Dielectr. Electr. Insul., 2015, 22, (1), pp. 128–134
HVDC CBs arcing time. [24] Walter, M., Franck, C.: ‘Improved method for direct black-box arc parameter
determination and model validation’, IEEE Trans. Power Deliv., 2014, 29, (2),
pp. 580–588
6 References [25] Pauli, B., Mauthe, G., Ruoss, E., et al.: ‘Development of a high current
HVDC circuit breaker with fast fault clearing capability’, IEEE Trans. Power
[1] Franck, C.M.: ‘HVDC circuit breakers: a review identifying future research Deliv., 1988, 3, (4), pp. 2072–2080
needs’, IEEE Trans. Power Deliv., 2011, 26, (2), pp. 998–1007 [26] Greenwood, A., Kanngiessner, K., Lesclae, V., et al.: ‘Circuit breakers for
[2] Meah, K., Ula, S.: ‘Comparative evaluation of HVDC and HVAC meshed multi-terminal HVDC systems part I: Introduction DC side substation
transmission systems’. IEEE Power Engineering Society General Meeting, switching under normal and fault conditions’. Electra No. 163, 1995, pp. 98–
Tampa, Florida, USA June 2007, pp. 1–5, doi:10.1109/PES.2007.385993 122
[3] Mobarrez, M., Kashani, M.G., Bhattacharya, S., et al.: ‘Comparative study of [27] Garzon, R.D.: ‘High voltage circuit breakers: design and applications’
DC circuit breakers using real time simulations’. 40th Annual Conf. of the (Marcel Dekker INC, New York, USA, 2002, 2nd edn.)
IEEE Industrial Electronics Society, (IECON), Dallas, Texas, USA, 2014, pp. [28] Andrade, V.D., Sorrentino, E.: ‘Typical expected values of the fault resistance
3736–3742, doi:10.1109/IECON.2014.7049056 in power systems’. IEEE/PES Transmission and Distribution Conf. and
[4] Lu, W., Ooi, B.: ‘Optimal acquisition and aggregation of offshore wind power Exposition: Latin America (T&D-LA), Sao Paulo, Brazil, 2010, pp. 602–609,
by multi-terminal voltage-source HVDC’, IEEE Trans. Power Deliv., 2003, doi:10.1109/TDC-LA.2010.5762944
18, (1), pp. 201–206 [29] Ohtaka, T., Kertész, V., Smeets, R.P.P.N.: ‘Black-box arc model validated by
[5] Meyer, C., Hoeing, M., Peterson, A., et al.: ‘Control and design of DC grids high-voltage circuit breaker testing’, IEEE Trans. Power Deliv., 2018, 33, (4),
for offshore wind farms’, IEEE Trans. Ind. Appl., 2007, 43, (6), pp. 1475– pp. 1835–1844
1482 [30] Esenwein, F.D., Colclaser, R.G.: ‘Simulation of arc-circuit instability in a low
[6] DLR, German Aerospace Center, Institute of Technical Thermodynamics: current DC automotive switching application’, IEEE Trans. Veh. Technol.,
‘Section Systems Analysis and Technology Assessment’. Tech. Rep., Trans- 1995, 44, (2), pp. 348–355
Mediterranean Interconnection for Concentrating Solar Power, 2006 [31] Pugliese, H., Michael, V.K.: ‘Direct current circuit breaker primer’. IEEE
[7] Billon, V., Taisne, J., Arcidiacono, V., et al.: ‘The Corsican tapping: from Industry Applications Society 57th Annual Petroleum and Chemical Industry
design to commissioning tests of the third terminal of the Sardinia-Corsica- Conf. (PCIC), San Antonio, Texas, USA, 2010, doi:10.1109/
Italy HVDC’, IEEE Trans. Power Deliv., 1989, 4, (1), pp. 794–799 PCIC.2010.5666866
[8] Kirby, N., Xu, L., Luckett, M., et al.: ‘HVDC transmission for large offshore [32] Lee, J.G., Khan, U.A., Lee, H.Y., et al.: ‘Impact of SFCL on the four types of
wind farms’, Power Eng. J., 2002, 16, pp. 135–141 HVDC circuit breakers by simulation’, IEEE Trans. Appl. Supercond., 2016,
[9] McCallum, D., Moreau, G., Primeau, J., et al.: ‘Multi-terminal integration of 26, (4), pp. 1–6
the Nicolet converter station into the Quebec-New England phase II HVDC [33] Bizzarri, F., Brambilla, A., Gruosso, G., et al.: ‘A new black-box model of SF
transmission system’. Proc. CIGRE 35th Int. Conf. on Large High Voltage 6 breaker for medium voltage applications’. 43rd IEEE Annual Conf. of
Electric Systems, Paris, France, August/September 1994, vol. 1, pp. Industrial Electronics Society (IECON), Beijing, China, 2017, pp. 32–37,
14-103/1–14-103/9 doi:10.1109/IECON.2017.8216010
[10] Greenwood, A., Kanngiessner, K., Lesclae, V., et al.: ‘Circuit breakers for [34] Park, K.H., Lee, H.Y., Asif, M., et al.: ‘Assessment of various kinds of AC
meshed multi-terminal HVDC systems. Part II: switching of transmission black-box arc models for DC circuit breaker’. IEEE 4th Int. Conf. on Electric
lines in meshed MTDC systems’. Electra No. 164, February 1996, pp. 62–82 Power Equipment-Switching Technology (ICEPE-ST), Xi'an, China, 2017,
[11] Li, Y., Zhang, Z., Rehtanz, C., et al.: ‘Study on steady and transient-state pp. 465–469, doi:10.1109/ICEPE-ST.2017.8188878
characteristics of a new HVDC transmission system based on an inductive
filtering method’, IEEE Trans. Power Electron., 2011, 26, (7), pp. 1976–1986
[12] Guo, C.Y., Liu, W.J., Zhao, C.Y.: ‘Research on the control method for
voltage-current source hybrid-HVDC system’, Sci. Chin. Technol. Sci., 2013,
56, (11), pp. 2771–2777

IET Gener. Transm. Distrib., 2019, Vol. 13 Iss. 4, pp. 461-467 467
© The Institution of Engineering and Technology 2018