You are on page 1of 11

Chalmers University of Technology, Department of Electric Power Engineering 2002-02-06

EEK145 High Voltage Technology

CE Sölver
Electric arcs and arc interruption

1. Introduction

Electric arcs are encountered in a variety of situations in the field of electrical engineering. In
some cases the arcs are unwelcome results of lightning strokes or insulation failures, but in
many cases the arcs are deliberately used, and their advantages exploited.

Overstress of the insulation system in an electric power network may lead to electrical
flashover and short-circuit in the system. At the location of the fault, the short-circuit current
will then normally flow through an electric arc. Such a high-current arc may severely damage
nearby equipment, and in addition the short-circuit current will affect the whole network. It is
a major task of the protection system to quickly deenergise the faulty part of the network.

Electric arcs are created every time the contacts of a switch or circuit-breaker are opened in
order to interrupt current – either a moderate load current or a much higher short-circuit
current. In high-voltage switches or circuit-breakers the arc is used as a conducting element
between the contacts which allows the current to continue to flow until a natural current zero
of the AC current is reached, where the arc is extinguished. The electric arc is – with the
exception of power semiconductors – the only known element that is able to change from a
conducting to a nonconducting state in a sufficiently short time.

Electric arcs are used for heating purposes in arc furnaces and arc welding equipment.

2. Creation of electric arcs


An electric arc may be initiated either by an electric flashover between two electrodes, or by
separation of two contacts from each other.

An electric flashover may start either as a Townsend avalanche or as a streamer. Once a


filamentary conductive channel has been established between the electrodes, the current will
continue to grow, provided that the supply network is strong enough. The voltage drops, and
the spark turns into an arc. The transition from spark to arc will occur at a current value of the
order of 0.1 A. See Figure 2.1.

Figure 2.1 The U-I characteristics of a breakdown process

When the contacts of a circuit-breaker separate, the current through the contacts will continue
to flow, driven by the magnetic energy stored in the inductances of the power system. At the

CD Sölver Page 1
Chalmers University of Technology, Department of Electric Power Engineering 2002-02-06
EEK145 High Voltage Technology

last moment, just before the contacts separate, they touch each other only at a very small
surface area. The resulting high current density leads to strong heating, and the contact
material will melt and evaporate. This leads to a gas discharge – arc – between the parting
contacts in the surrounding medium that may be air, oil or SF6, depending on the type of
circuit breaker.

3. Properties of electric arcs


The arc channel between two electrodes (anode and cathode) can be divided into an arc
column, a cathode region and an anode region.

Cathode Anode
region region
Arc column
Anode
Cathode
Ie (Electrons)
I+ (Ions)

Space charge Contraction Contraction Space charge


zone zone zone zone

Figure 3.1 The regions of an arc channel

3.1 The arc column


The temperature in the arc column is very high, from around 5000 K to 20000 K. At such
high temperature the gas molecules are largely dissociated into free atoms. The travel
velocities of electrons and atoms are so high that ionization takes place when they collide. At
the same time, there is also a recombination process where electrons and positively charged
ions form neutral atoms. At thermal equilibrium, the rate of ionization is in balance with the
rate of recombination. The gas is in plasma state and there is a high amount of free electrons
and positive ions.

The fraction f of atoms that is ionized may be calculated by means of Saha’s equation:

f2
⋅ P = 3,16 ⋅ 10 − 7 ⋅ T 5/ 2 ⋅ exp( − eVi / kT )
1− f 2

with e = 1,6.10-19 As, the charge of an electron


Vi = ionisation potential of the gaseous medium
k = 1,38.10-23, Boltzmann’s constant
P = the gas pressure in bar
T = the temperature in K

Saha’s equation is shown graphically in Figure 3.2 for oxygen and nitrogen.

CD Sölver Page 2
Chalmers University of Technology, Department of Electric Power Engineering 2002-02-06
EEK145 High Voltage Technology
Degree of ionization

0.99667 1

0.8
Degree of ionization

0.6
F( 12.2 )

F( 15.5 )
0.4

0.2

2.31393e-019
0
4 4 4 4 4
0 5000 1 10 1.5 10 2 10 2.5 10 3 10
2000 T 26000
Temperature, K
Temperature, K
oxygen
nitrogen

Figure 3.2 Degree of thermal ionization for oxygen and nitrogen, at atmospheric pressure

Figure 3.2 indicates that the thermal ionization can be used to switch between a conducting
state (f close to 1) and a nonconducting state (f close to 0). Because of the relatively steep
slope of the function between the temperature and the degree of ionization, reduction of the
arc temperature by means of cooling with cold gas is an effective way to bring the arc column
from a conducting state to a non-conducting state. This principle is used in most circuit
breakers.

Although the arc column is strongly ionized, there is no space charge present. There is a
balance between the electron charges and the positive ion charges. The electrons have much
higher (three orders of magnitude) mobility than the positive ions. Therefore almost the entire
current flow is due to the electrons.

Due to the strong ionization, a high current arc is a relatively good electrical conductor. A
typical conductivity of the arc plasma is 10-100 S/cm, which is comparable to the
conductivity of e.g. carbon.

The total arc voltage, and also the voltage gradient along the arc, depends on the current
magnitude, the type of gas, and the pressure. When it is in thermal balance, the arc column
adjusts itself in such a way that the power supplied to the column (the ohmic heating) attains a
minimum value. If there is a disturbance from this situation, which tends to increase the
resistance of the arc, then the ohmic heating would increase, the temperature and diameter of
the arc would increase, and automatically counteract the disturbance. If, on the other hand,
there is a disturbance which tends to increase the temperature or diameter of the arc, then the
power losses would increase, and tend to bring the arc back to the original situation.

Figure 3.3 shows a typical potential distribution along an arc channel.

CD Sölver Page 3
Chalmers University of Technology, Department of Electric Power Engineering 2002-02-06
EEK145 High Voltage Technology

Figure 3.3 Typical potential distribution along an arc channel

3.2 Pinch effect


The current that flows in the arc channel is associated with magnetic forces, which lead to an
internal overpressure in the arc channel. The phenomenon is often referred to as “pinch
effect”, and may be understood from figure 3.4. The total current in the column may be
imagined as consisting of a number of parallel current filaments. Each separate filament will
be attracted to each of the others by a magnetic force. The resulting, overall force on each
filament will be directed towards the center of the column.

H(r)

Force R
r
r+dr

Simplified model Basis for determination


of overpressure

Figure 3.4 Magnetic forces in an arc column

The resulting overpressure in the center of the arc column may be calculated, under the
assumption that the current density j is constant over the whole cross section of the arc. With
radius R the total current is
I = π ⋅ R2 ⋅ j
At a radius r, the magnetic field H(r) can be calculated by Ampere’s law
2 ⋅ π ⋅ r ⋅ H (r ) = π ⋅ r 2 ⋅ j
r
H (r ) = ⋅ j
2

CD Sölver Page 4
Chalmers University of Technology, Department of Electric Power Engineering 2002-02-06
EEK145 High Voltage Technology

Consider now a ring segment at radius r, and with thickness dr. The current that flows in this
ring, in interaction with the magnetic flux, will result in a mechanical force, that is in each
point of the ring directed towards the center.
The total current in the ring is
I ring = 2πr ⋅ dr ⋅ j
and the total force (per unit of length of the ring):
Fring = µ o ⋅ H (r ) ⋅ Iring = µ o ⋅ π ⋅ r 2 ⋅ dr ⋅ j 2

The force per unit of area is


Fring r
F (r ) = = µ o ⋅ ⋅ dr ⋅ j 2
2πr 2
and the corresponding pressure gradient
dP F (r ) r
= = µo ⋅ ⋅ j2
dr dr 2

Finally the resulting overpressure P in the center of the arc column is obtained by integration:
R 2
r R

P = dP = ∫ µ o ⋅ ⋅ j 2 ⋅ dr = µ o ⋅ j 2
2 4
0
µ ⋅ I2
P= o
4π 2 ⋅ R 2

Close to the electrodes, the arc diameter is often smaller than further away from the
electrodes. This means that there will be a gradient in current density, and therefore also a
gradient in the internal pressure. This pressure gradient will contribute to a transport of
plasma and metal vapour from the electrodes into the arc column.

3.3 The contact regions


The role of the cathode, and the surrounding cathode region, is to emit the current carrying
electrons into the arc column. Some electrode materials have such a high boiling temperature
that significant thermionic emission of electrons starts already well below the evaporation
temperature. Such materials are e.g. carbon, tungsten and molybdenum. Other electrode
materials have lower evaporation temperature. In this case the major emission mechanism
will be field emission, where electrons will be emitted due to the high electric field strength
close to the surface. Copper is a typical example of such a material.

Close to the cathode there will be an accumulation of positive ions, arriving from the arc
column. Due to this space charge, there will be high electric field strength close to the cathode
surface (the cathode drop). This high field strength is essential for efficient (field) emission of
electrons into the arc.
The anode mainly serves as a collector of electrons, arriving from the cathode. The electrons
will arrive at high speed, and deliver all their energy to the anode. The anode surface will
therefore be kept at a high temperature. Close to the anode there will be a lack of positive
ions, since they tend to drift away. The surplus of electrons leads to high electric field
strength close to the anode surface (the anode drop).

CD Sölver Page 5
Chalmers University of Technology, Department of Electric Power Engineering 2002-02-06
EEK145 High Voltage Technology

4. Static and dynamic arc characteristics


For low to moderate values of arc current, there is an inverse relationship between a static arc
current and the corresponding arc voltage: a higher current value corresponds to a lower arc
voltage. See Figure 4.1. The arc channel has a certain thermal inertia, and a sudden rise in
current magnitude from an otherwise stationary arc current will have to be accompanied by an
initial rise in the arc voltage. This is necessary in order to build up the additional temperature
and stored energy that corresponds to the new current magnitude. After some time (the
thermal time constant of the arc), the arc would settle down to a new steady state,
corresponding to the new current magnitude. The opposite would happen if there would be a
sudden decrease in current magnitude.

4.1 DC arc with AC modulation


If the current consists of an AC modulation on an otherwise stationary DC value, the
corresponding voltage variation will be such that the operating point will follow a loop, as
shown in Figure 4.1. The loop shape depends on the frequency of the AC modulation. At very
low frequency the variation will (almost) follow the static characteristic. On the other hand, at
extremely high modulation frequency, the arc behaves like a linear resistance. Its thermal
inertia will prevent it from adjusting to such rapid changes in the current.

Im: amplitude of AC modulation

a: low frequency
b: medium frequency
c: high frequency

Fig 4.1 Current-voltage relation of a DC arc with superimposed AC modulation

4.2 AC arcs; conditions at current zero


In the case of AC arc current, the variation of the current leads to associated, dynamic
changes in the arc properties. Due to the thermal inertia of the arc plasma, the maximum
temperature and stored energy during a current loop will occur somewhat later than the peak
value of the current. The arc temperature will still be high when the current reaches zero. At
current zero the power supply to the arc is zero, and there is a possibility that the plasma will
cool down and the arc disappear.

At current zero there is also the complexity of the polarity change, forcing the anode and
cathode to change places.

When the current continues to flow after current zero, it will be due to either of two reasons:
- Thermal reignition
If the arc plasma is still sufficiently ionized at current zero, and if the new cathode emits
a sufficient amount of electrons right from the beginning, then the current may pass more
or less continuously through zero. The arc channel will pass through a minimum in
temperature and stored energy, and then start to heat up again.

CD Sölver Page 6
Chalmers University of Technology, Department of Electric Power Engineering 2002-02-06
EEK145 High Voltage Technology

- Dielectric reignition
If the arc plasma at the moment of current zero is so cold that it is only weakly ionized,
then it may continue to cool down, and the current will be interrupted. The arc channel
will still be comparatively hot and may also contain space charges, and therefore the
dielectric withstand capability between the two contacts will be quite low. Due to the
voltage that appears between the electrodes there may be a flashover, a new arc will be
established, and the current starts to flow again.

5. Arcs in open air


In electric power systems, AC arcs in open air will occur mainly in two situations:
- Arcs with relatively low current will appear between opening (and closing) disconnector
contacts
- Arcs with high current may occur in connection with short-circuits in the system

5.1 Low current arcs


The relationships between arc length, arc current and voltage are complicated. Figure 5.1
shows experimental results for DC arcs with length up to 200 mm, and copper electrodes. At
low current values, the curves have an “inverse” characteristic: when the current increases,
then the voltage decreases. At high current values the curves tend to be flat, with a relatively
constant arc voltage, independent of the current. When the arcs are short, the voltage is
determined mainly by the anode and cathode drops (around 20 V). When the arcs are long,
and the current not too low, the arc voltage tends to be of the order of 10 V/cm.

Figure 5.1 Current-voltage characteristics for DC arcs in air, with copper electrodes

5.2 High current arcs


Arcs that occur in connection with short circuits in power systems may conduct currents of
many kA. The arc voltage may vary considerably, but the order of magnitude is 10-20 V/cm,
for arc currents up to the order of 50 kA, and arc lengths up to the order of 2 m.

In case of a three phase short circuit, e.g. in a busbar system, there will often be only two arcs,
as illustrated in Figure 5.2. Each of these two arcs will conduct phase short circuit current.

CD Sölver Page 7
Chalmers University of Technology, Department of Electric Power Engineering 2002-02-06
EEK145 High Voltage Technology

Busbar
Pole A
arc
Pole B
arc
Pole C

Figure 5.2 Arcs during a three-phase short circuit

Due to the high power – or rather energy – released by high current arcs, they can lead to
severe damage to adjacent equipment. A special problem is the pressure rise that will occur in
substation cubicles, GIS enclosures, or buildings as a result of internal arcs.

Short-circuit arcs e.g. in a busbar system will attempt to move, as a result of the force created
by the magnetic field of the connecting bars. The travel direction will be away from the
current source. In order to limit damage from traveling arcs, the busbars are often insulated
(coated), or split up into sections with intermediate walls with bushings.

6. Switching arcs in circuit-breakers


In high voltage circuit breakers the electric arc is used as the switching element that rapidly
changes from a good conductor to a good insulator at current zero. The principal function is
shown in Fig 6.1. As shown in the figure, it is quite normal that there is an interval with
arcing between the contacts that lasts more than half a cycle of the AC current. The reason is
that the circuit breaker often has still not reached conditions for efficient extinguishing of the
arc at the first current zero. At the final current zero the arc is extinguished. In order for the
interruption to be successful, two conditions have to be met:
- The arc has to be cooled down to nonconducting state at current zero
- After current zero the contact system must be able to withstand the recovery voltage that
will appear between the contacts

open
contact
contact position
separation
arcing time

closed time

current

Figure 6.1 Contact separation and arc interruption in a circuit breaker

In most circuit-breaker designs the arc between the separating contacts will be subjected to a
strong gas blast in a nozzle. The flow of cold gas will keep the arc concentrated, and help to
cool it at current zero.

CD Sölver Page 8
Chalmers University of Technology, Department of Electric Power Engineering 2002-02-06
EEK145 High Voltage Technology

Figure 6.2 Circuit-breaker arc in axial blast nozzle

6.1 Conditions at current zero, power balance of the arc


At current zero the hot arc channel between the contacts has to be cooled down to such a low
temperature that it is no longer electrically conducting. Due to the stored thermal energy, the
arc has a certain inertia, and when the current approaches zero there will still be a certain
amount of electrical conductivity left in the arc path. This gives rise to what is called a “post
arc current”. See Fig 6.3. Whether the interruption is going to be successful or not is
determined by a race between the energy removed from the arc by cooling and the energy
input into the arc path by the post arc current, driven by the recovery voltage. The time scale
of this process is very short, only a few microseconds.

The input of power to the arc, and therefore the difficulty to interrupt the arc, is related to the
rate of decrease of the current towards zero, di/dt, and to the rate of rise of the recovery
voltage after current zero, du/dt.

Figure 6.3 Thermal interruption regime with post arc current. Times are in the microseconds
range

6.2 Conditions after arc interruption


As soon as the arc between the interrupter contacts has been extinguished, a recovery voltage
will build up between the contacts. The shape and magnitude of the voltage will be
determined by the surrounding electrical network. Typical magnitudes for a high voltage
circuit breaker are rates of rise of some kV/µs, and peak values of several hundred kV. The

CD Sölver Page 9
Chalmers University of Technology, Department of Electric Power Engineering 2002-02-06
EEK145 High Voltage Technology

contact system now has to withstand the high voltage stress without electrical breakdown,
even though the contacts and the gas are still hot, as a result of the arc that has just been
extinguished.

6.3 Interrupter design


Most modern high voltage circuit breakers use SF6 gas for electrical insulation and arc
interruption. In “puffer” designs, the blast of gas that is necessary for cooling of the arc is
achieved by means of mechanical compression during the opening operation of the breaker.
In “autopuffer” or “selfblast” designs some of the energy used for compression of the gas is
also taken from the electric arc itself.

Fig 6.4 shows the working principle of an autopuffer design. When interrupting small currents
(up to some kA) the circuit breaker operates as a pure puffer circuit breaker: gas is
compressed in the auxiliary puffer cylinder V2 and flows through volume V1 and the nozzle.
In the case of a short-circuit current, the pressure rise necessary for the extinguishing of the
arc is built up in the self-blast volume V1, through heating by the arc. A check valve between
the volumes V1 and V2 prevents the high pressure from escaping to the auxiliary puffer
cylinder. The pressure in V2 is relatively independent of the current. It is limited to a
moderate level by means of a spring-loaded valve (overpressure valve), which means that the
compression energy required from the operating mechanism is limited.

Figure 6.4 Autopuffer SF6 circuit breaker

6.4 Interruption of DC current


In the case of DC current there are no natural current zeroes where a circuit-breaker arc may
be extinguished. In this situation there are two possibilities to interrupt the current. One
method is to use an interrupter design where the arc resistance is increased so much that the
corresponding voltage will approach the supply voltage of the network, and thereby force the
arc current towards zero. This is, however, only realistic for moderate voltages. Another
method is to create an artificial current zero by means of a current oscillation, superimposed
on the DC current. The current may then be interrupted by means of a normal AC circuit
breaker.

CD Sölver Page 10
Chalmers University of Technology, Department of Electric Power Engineering 2002-02-06
EEK145 High Voltage Technology

Figure 6.5 Creation of an artificial current zero in an HVDC circuit

The principle for creation of an oscillating current is shown in Fig 6.5. I1 is the normal DC
current that is to be interrupted. The capacitor C is precharged, and an oscillating current I2
starts to flow when the auxiliary switch S is closed. In practical designs, additional circuit
components will be required in order to limit the rate of rise of the recovery voltage, and to
absorb the magnetic energy stored in the circuit inductances.

HVDC circuit breakers based on the oscillating current principle have been built, but are not
used in actual operation.

CD Sölver Page 11

You might also like