ASSIGNMENT NO.

2
Of

HIGH VOLTAGE ENGINEERING(EED-421)

On

Insulation Coordination and Over Voltages in Power systems

Submitted to: Prof. Yog Raj Sood

Submitted By: Pallvi Dogra (16257)

ELECTRICAL ENGINEERING DEPARTMENT

NATIONAL INSTITUTE OF TECHNOLOGY
HAMIRPUR 177005 (INDIA)

April, 2020
 Insulation Coordination and Over Voltages in Power systems

It is essential for electrical power engineers to reduce the number of outages and preserve the
continuity of service and electric supply. Therefore, it is necessary to direct special attenuation
towards the protection of transmission lines and power apparatus from the chief causes of over
voltages in electric systems, namely lightning over voltages and switching over voltages.
Lightning overvoltage is a natural phenomenon, while switching over voltages originate in the
system itself by the connection and disconnection of circuit breaker contacts or due to initiation
or interruption of faults. Switching over voltages are highly damped short duration over voltages.
They are ‘temporary over voltages’ of power frequency or its harmonic frequency, either
sustained or weakly damped, and originate in switching and fault-clearing processes in power
systems. There is a system operating voltage at which the emphasis changes from lightning to
switching surge design, this being important above 500 kV. In the range of 300 kV to 765 kV,
both switching over voltages and lightning over voltages have to be considered, while for ultra-
high voltages (>700 kV), perhaps switching surges may be the chief condition for design
considerations.

1.1 NATURAL CAUSES FOR OVERVOLTAGES-LIGHTNING PHENOMENON

Lightning phenomenon is a peak discharge in which charge accumulated in the cloud discharges
into a neighboring cloud or to the ground. The electrode separation, i.e., cloud-to-cloud or cloud-
to ground is very large, perhaps 10 km or more.

1.1.1 Charge Formation in the Clouds

The factors that contribute to the formation or accumulation of charge in the clouds are too many
and uncertain. But during thunderstorms, positive and negative charges become separated by the
heavy air currents with ice crystals in the upper part and rain in the lower parts of the cloud. This
charge separation depends on the height of the clouds, which range from 200 to 10,000 m, with
their charge centers probably at a distance of about 300 to 2000 m. The volume of the clouds that
participate in lightning flashover are uncertain, but the charge inside the cloud may be as high as
1 to 100 C. Clouds may have a potential as high as 107 to 108 V with field gradients ranging from
100 V/cm within the cloud to as high as 10 kV/cm at the initial discharge point. The energies
associated with the cloud discharges can be as high as 250 kWh. It is believed that the upper
regions of the cloud are usually positively charged, whereas the lower region and the base are
predominantly negative except the local region, near the base and the head, which is positive. The
maximum gradient reached at the ground level due to a charged cloud may be as high as 300
V/cm, while the fair-weather gradients are about 1 V/cm. A probable charge distribution model
is given in Fig. 1.1 with the corresponding field gradient near the ground.
Fig. 1.1 Probable field gradient near the ground corresponding to the probable charge
distribution in a cloud

Cloud Model according to simpson’s theory:

Fig. 1.2 Cloud model according to Simpson’s theory

According to the Simpson’s theory (Fig. 1.2) there are three essential regions in the cloud to
be considered for charge formation. Below region A, air currents travel above 800 cm/s, and no
raindrops fall through. In region A, air velocity is high enough to break the falling raindrops
causing a positive charge spray in the cloud and negative charge in the air. The spray is blown
upwards, but as the velocity of air decreases, the positively charged water drops recombine with
the larger drops and fall again. Thus, region A, eventually becomes predominantly positively
charged, while region B above it, becomes negatively charged by air currents. In the upper regions
in the cloud, the temperature is low (below freezing point) and only ice crystals exist. The impact
of air on these crystals makes them negatively charged, thus the distribution of the charge within
the cloud becomes as shown in Fig. 1.2.
However, the above theory is obsolete and the explanation presented is not satisfactory.
Recently, Reynolds and Mason proposed modification, according to which the thunder clouds
are developed at heights 1 to 2 km above the ground level and may extend up to 12 to 14 km
above the ground. For thunder clouds and charge formation air currents, moisture and specific
temperature range are required.

Rate of Charging of Thunder Clouds:

If λ is a factor which depends on the conductivity of the medium, there will be a resistive leakage
of charge from the electric field built up, and this should be taken into account for cloud charging.
 Let E be the electric field intensity, v be the velocity of separation of charges, and ρ the charge
density in the cloud. Then, the electric field intensity E is given by

(1)

(2)

This equation assumes initially E = 0 at t = 0, the start of charge separation, i.e. there is no
separation initially.
Let Q, be the separated charge and Qg be the generated charge, then

(3)

where ε0 is the permittivity of the medium, A is the cloud area and h is the height of the charged
region.
From Eq. (2), on substitution

(4)

where M= Qs · h
The velocity of separation of charges, v = 10 to 20 m/s.
Substituting these values, we get
1.1.2 Mechanism of Lightning Stroke

1.1.3 Parameters and Characteristics of the Lightning Strokes

The parameters and characteristics of lightning include the amplitude of the currents, the rate of
rise, the probability distribution of the above, and the waveshapes of the lightning voltages and
currents.
Typical oscillograms of the lightning current and voltage waveshapes on a transmission line
are shown in Figs 1.3 and 1.4. The lightning current oscillograms indicate and initial high current
portion which is characterized by short front times up to 10 µs. The high current peak may last
for some tens of microseconds followed by a long-duration low current portion lasting for several
milliseconds. This last portion is normally responsible for damages (thermal damage). Other
important characteristics are time to peak value and its rate of rise. From the field data, it was
indicated that 50% of lightning stroke currents have a rate of rise greater than 7.5 kA/µs, and for
10% strokes it exceeded 25 kA/µs. The duration of the stroke currents above half the value is
more than 30 µs.
 Fig. 1.3 Typical lighting current oscillograms

Fig. 1.4 Typical lightning stroke voltage on a transmission line without ground wire

Lightning strokes on transmission lines are classified into two groups—the direct strokes and the
inducted strokes. When a thunder cloud directly discharges onto a transmission line tower or line
wires it is called a direct stroke. This is the most severe form of the stroke. However, for bulk of
the transmission systems the direct strokes are rare and only the induced strokes occur.
Sometimes, when a direct lightning stroke occurs on a tower, the tower has to carry huge
impulse currents. If the tower footing resistance is considerable, the potential of the tower rises
to a large value, steeply with respect to the line and consequently a flashover may take place along
the insulator strings. This is known as back flashover.
1.1.4 Mathematical Model for Lightning
When a streamer discharge occurs to ground by first a leader stroke, followed by main strokes
with considerable currents flowing, the lightning stroke may be thought to be a current source of
value I0 with a source impedance Z0 discharging to earth. If the stroke strikes an object of
impedance Z, the voltage built across it may be taken as
 (5)

The source impedance of the lightning channels are not known exactly, but it is estimated to
be about 1000 to 3000 Ω. The objects of interest to electrical engineers, namely, transmission
line, etc. have surge impedances less than 500 Ω (overhead lines 300 to 500 Ω, ground wires 100
to 150 Ω, towers 10 to 50 Ω, etc.). Therefore, the value Z/Z0 will usually be less than 0.1 and
hence can be neglected. Hence, the voltage rises of lines, etc., may be taken to be approximately
V=I0Z, where I0 is the lightning stroke current and Z the line surge impedance.
The incidence of lightning strokes on transmission lines and sub-stations is related to the
degree of thunderstorm activity. It is based on the level of Thunderstorm Days (TD) known as
Isokeraunic Level defined as the number of days in a year when thunder is heard or recorded in a
particular location. But this indication does not often distinguish between the ground strokes and
the cloud-to-cloud strokes. If a measure of ground flashover density (Ng) is obtained, then the
number of ground flashovers can be computed from the TD level. From the past records and the
past experience, it is found that

Ng = (0.1 to 0.2) TD/strokes/km2-year.

The ground flashover density (Ng) is given by

Ng = K1(TD)b where
K1 = 0.04 and b = 1.25 (Anderson) or
K1 = 0.054 and b = 1.1 (Mac Gorman)

From the experimental and field data collected, the number of lightning strokes or flashes on
to an overhead line in open grounds with no nearby ground or tall objects can be estimated as
N = Ng (2.8 h0·6 + b/10)
(a) Over Voltages due to Indirect Strokes

where Vi = induced voltage

ϕ = electric potential caused by static charges (electrostatic)
 A = magnetic vector potential due to lightning currents
(electromagnetic)
h = height of conductor above the ground

(b) Model for Lightning Stroke Effect and Computation of Induced Voltages

In lightning studies, the strokes to ground are of importance as they induce over-voltages on
transmission lines and in nearby objects. Hence, a model based on lightning return stroke is
presented here which assumes that the lightning channel is straight, vertical and normal to the
ground plane. The return stroke current I (Z, t) is assumed as function of vertical co-ordinate z
and time t and the initial current I (o, t) is of importance in engineering applications.

i (z, t) = i (t − z/v) z < vt

i (z, t) = 0 z > vt

1.1.5 Travelling Waves on Transmission Lines

Any disturbance on a transmission line or system such as sudden opening or closing of line, a
short circuit or a fault results in the development of over voltages or over currents at that point.
This disturbance propagates as a travelling wave to the ends of the line or to a termination, such
as, a substation. They may be reflected, transmitted, attenuated or distorted during propagation
until the energy is absorbed. Long transmission lines are to be considered as electrical networks
with distributed electrical elements. In Fig. 1.5, a typical two-wire transmission line is shown
along with the distributed electrical elements, R, L, C and G.

Fig. 1.5 Distributed characteristic of a long transmission line

The propagation of any travelling wave, say, a voltage wave can be analyzed by considering
an elemental length of the line dx. The voltage drop in the positive x-direction in the elemental
length dx due to the inductance and resistance is

(6)
Here, δΨ is the change of flux linkages and is equal to i.L.dx, where i the current through the
line.

(7)

The shunt current through the leakage conductance (G) and capacitance (C) is

(8)
Here, ∇Φ is the change in electrostatic field flux and is equal to V C·dx, where V is the potential
at the point x.

Hence, the above equations can be written as

Taking Laplace transform with respect to the time variable t, the equations can be put in the
operation form as

Eliminating i and V from above equations and differentiating w.r.t. x, we get

(9)
where the product YZ = γ2 = RG + (RC + LG) s + LCs2.
The above two equations are called wave equations or telegraphic equations. The solutions
for the above equations can be written in the form
 (10)
where f1(t) and f2(t) are any arbitrary functions that satisfy the boundary conditions. The operator
γ is simplified as

(11)
Y(s) is called the surge admittance, the reciprocal of which
(12)
Z(s) is called the surge impedance of the transmission line.

(a) Classification of Transmission Lines

Transmission lines are usually classified as
(i) lines with no loss or ideal lines,
(ii) lines without distortion or distortion less lines,
(iii) lines with small losses, and
(iv) lines with infinite and finite length defined by all the four parameters.

(i) Ideal Lines:

A line is said to be an ideal line if R = 0 and G = 0. In this case the surge impedance of the line
is and the surge admittance of the line is The solution for the voltage and
current waves for this type of line is obtained as

(13)

The propagation velocity for either wave is given as

(ii) Distortion less Lines:

If, for any line R/L = G/C = α then, and the surge impedance is
(14)
For these conditions, the solution for the wave Eqs (8.10) and (8.10a) will be modified as

(15)

Under these conditions,

(16)
(iii) Line with Small Losses:
In this type of lines, the time constants of the line are large, i.e., R/L and G/C are small. Then γ
can be approximated to be equal to (s + α/v), and Y(s) to be equal to
Under these conditions, the solutions for the voltage and the current waves become

(17)

(iv) Exact Solution for Lines of Finite or Infinite Length Defined by all the Four
Parameters:
The exact solution of the wave equation of this type of lines is quite complex and is normally of
little practical importance. However, some of the inferences that can be drawn are

• the current and the voltage wave are dissimilar

• the attenuation and distortion due to normal line resistance and leakage conductance are of
little consequence
• the surge impedance Z(s) = e(s)/i(s) is a complex function and is not uniquely defined.
(b) Attenuation and Distortion of Travelling Waves
As a travelling wave moves along a line, it suffers both attenuation and distortion. The decrease
in the magnitude of the wave as it propagates along the line is called attenuation. The elongation
or change of wave shape that occurs is called distortion. For distortionless lines, the attenuation
is approximated as a loss function φ (V), considering that the attenuation is due to the energy lost
per unit length of the line in the resistance as the wave travels. It can be shown that

((18)
For different line conditions, φ(V) and attenuation are as follows:

(i) For Lines Having all the Parameters R, L, G and C

(19)
Where α is called the attenuation factor.
(20)
(ii) The Skilling Formula
If φ(V) is assumed to be equal to β (V − Vc), where Vc is critical corona voltage; then

and if the initial voltage at t = 0 is taken as V0, then

(21)
(iii) The Quadratic Formula
If φ (V) is assumed to vary as

Integrating the above equation, we get

(22)
(iv) The Foust and Manger formula
Here, φ (V) is assumed to be equal to λV3, so that

It follows from the above equation that

(23)
(v) Attenuation due to Corona
The effect of corona is to reduce the crest of the voltage wave under propagation, limiting the
peak value to the critical corona voltage.
Skilling gives the formula for corona power loss as P = K(V − Vc)2, where Vc is the critical
corona voltage and K is a constant.

(c) Reflection and Transmission of Waves at Transition Points

Fig. 1.6 Transition point (T) and the propagation of the wave at the transition point

(24)

Successive Reflections and Lattice diagrams:

The principles observed in the lattice diagrams are as follows:

(i) All waves travel downhill, i.e. into the positive time.
(ii) The position of the wave at any instant is given by means of the time scale at the left of
the lattice
diagram.
 (iii) The total potential at any instant of time is the superposition of all the waves which arrive
at that point until that instant of time, displaced in position from each other by time
intervals equal to the time differences of their arrival.
(iv) Attenuation is included so that the amount by which a wave is reduced is taken care of.
(v) The previous history of the wave, if desired can be easily traced. If the computation is to
be carried out at a point where the operations cannot be directly placed on the lattice
diagram, the arms can be numbered and the quantity can be tabulated and computed.

Fig. 1.7 Reflection lattice of a travelling wave

In the arrangement shown in the figure, there are two junctions 1 and 2. The travel times for
the waves are different through Z1, Z2, and Z3. The lines with surge impedances Z1, Z2, and Z3 are
connected on either side of the junctions. Let α and β be the attenuation coefficients for the two
sections Z2 and Z3.

1.1.6 Behavior of Rectangular Travelling Wave [Unit Step Function U/(t)] at Transition
Points—Typical Cases
Case (i) Open-ended Transmission Line of Surge Impedance Z
Let the voltage of travelling wave incident on the line be
e = EU(t)
then, Z1 = Z and Z2 = ∞
Case (ii) Short-circuited Line
Voltage of the wave, e = EU(t)
Surge impedances Z1 = Z and Z2 = 0

.·. voltage of the reflected wave, e' = Γ·e

= − EU(t)
The voltage of the transmitted wave, e" = (1 + Γ) e = 0
Case (iii) Line Terminated with a Resistance Equal to the Surge Impedance of the Line

In this case, Z1 = Z and Z2 = R = Z

.·. coefficient of reflection,

.·. voltage of the reflected wave, e' is Γ e = 0

The voltage of the transmitted wave is (1 + Г) е = e
Case (iv) Line Terminated with a Capacitor
In this case, e = E U(t),

Z1 = Z, and Z2 =

The voltage of the reflected wave, e' = Γ e

After solving we will get

Case (v) Line Terminated by an Inductance L

In this case, e = EU(t)
Z1 = Z, and Z2 = Ls
 The coefficient of reflection
Voltage of the reflected wave, e' = Γe
After solving we will get

The voltage across the inductor initially rises to double the value of the incident wave and
decays exponentially. This is of importance when long lines are terminated with inductors or
transformers on open circuit.

Case (vi) Line having a Series Inductor

Let the surge impedance of the line be Z before and after the series inductor L. Considering the
junction point after the inductor L, e = E U(t), Z1 = (Z + Ls), and Z2 = Z.

The coefficient of reflection

Voltage of the transformed reflected wave, e'(s) = Γe(s)

Taking the inverse transform,

Case (vii) Line Terminated with a Transformer (Taken as an L-C Parallel Combination)

1.2 OVERVOLTAGE DUE TO SWITCHING SURGES, SYSTEM FAULTS AND OTHER

ABNORMAL CONDITIONS
1.2.1 Characteristics of Switching Surges
The waveshapes of switching surges are quite different and may have origin from any of the
following sources.

(i) De-energizing of transmission lines, cables, shunt capacitor, capacitor banks, banks,
etc. (ii) Disconnection of unloaded transformers, reactors, etc.
(iii) Energization or reclosing of lines and reactive loads
(iv) Sudden switching off of loads
(v) Short circuits and fault clearances
(vi) Resonance phenomenon like ferro-resonance, arcing grounds, etc.
Typical waveshapes of the switching surges are given in Figs 1.8a to e.

Fig. 1.8 Typical waveshapes of switching surge voltages

From the figures of the switching surges it is clear that the over voltages are irregular (oscillatory
or unipolar) and can be of high frequency or power frequency with its harmonics. The relative
magnitudes of the over voltages may be about 2.4 p.u. in the case of transformer energizing and
1.4 to 2.0 p.u. in switching transmission lines.

(a) Switching Over Voltages in EHV and UHV Systems

The different situations under which this happens are summarized as

(i) interruption of low inductive currents (current chopping) by high speed circuit breakers.
This occurs when the transformers or reactors are switched off
(ii) interruption of small capacitive currents, such as switching off of unloaded lines, etc.
(iii) ferro-resonance condition
This may occur when poles of a circuit breaker do not close simultaneously
(iv) energization of long EHV or UHV lines
The other situations of switching that give rise to switching over voltages of short duration (0.5
to 5 ms) and lower magnitudes (2.0 to 2.5 p.u.) are

(a) single pole closing of circuit breaker

(b) interruption of fault current when the L-G or L-L fault is cleared
(c) resistance switching used in circuit breakers
(d) switching lines terminated by transformers
(e) series capacitor compensated lines
(f) sparking of the surge diverter located at the receiving end of the line to limit the lightning
over voltages
The over voltages due to the above conditions are studied or calculated from

(a) mathematical modelling of a system using digital computers

(b) scale modelling using transient network analyzers
(c) by conducting field tests to determine the expected maximum amplitude of the Over
Voltages and their duration at different points on the line. The main factors that are
investigated in the above studies are

(i) the effect of line parameters, series capacitors and shunt reactors on the magnitude
and duration of the transients.
(ii) the damping factors needed to reduce the magnitude of over-voltages
(iii) the effect of single pole closing, restriking and switching with series resistors in
circuit breakers on the over voltages, and
 (iv) the lightning arrester spark over characteristics.
It is necessary in EHV and UHV systems to control the switching surges to a safe value of less
than 2.5 p.u. or preferably to 2.0 p.u. or even less. The measures taken to control or reduce the
over voltages are

(i) one step or multi-step energization of lines by preinserting of resistors,

(ii) phase controlled closing or circuit breakers with proper sensors,
(iii) drainage of trapped charges on long lines before the reclosing of the lines
limiting the over voltages by using surge diverters.

In Table 1.1, a summary of the extent of over voltages that can be developed under various
conditions of switching is given.

Table 1.1 Over voltages due to switching operations under different conditions Maximum value
of the system line-to-ground voltage = 1.0 p.u.
S.no. Type of operation Over
voltage
(p.u.)
1 Switching an open-ended line with:
(a) infinite bus as source with trapped charges on line 4.1
(b) infinite bus as source without trapped charges 2.6
(c) de-energizing an un-faulted line with a restrike in the circuit breaker 2.7
(d) de-energizing an un-faulted line with a line to ground fault (about 270 km in 1.3
2 length)
(a) Switching a 500-kV line through an autotransformer, 220 kV/500 from the
L.V. side 2.0
2.2
(b) switching a transformer terminated line
2.2
(c) series capacitor compensated line with 50% compensation 2.6
3 (d) series capacitor compensated line with shunt reactor compensation 3.6
High speed reclosing of line after fault clearance

1.2.2 Power Frequency Over Voltages in Power Systems

The power frequency over voltages occur in large power systems and they are of much concern
in EHV systems, i.e., systems of 400 kV and above. The main causes for power frequency and
its harmonic over voltages are

(a) sudden loss of loads,

(b) disconnection of inductive loads or connection of capacitive loads,
(c) Ferranti effect, unsymmetrical faults, and
(d) saturation in transformers, etc.

1.2.3 Control of Over Voltages due to Switching

The over voltages due to switching and power frequency may be controlled by

(a) energization of transmission lines in one or more steps by inserting resistances and
withdrawing them afterwards,
(b) phase controlled closing of circuit breakers,
(c) drainage of trapped charges before reclosing,
(d) use of shunt reactors, and
(e) limiting switching surges by suitable surge diverters.

1.2.4 Protection of Transmission Lines against Over Voltages

Protection against Lightning Over Voltages and Switching Surges of Short Duration
Over Voltages due to lightning strokes can be avoided or minimized in practice by

(a) shielding the overhead lines by using ground wires above the phase wires,
(b) using ground rods and counter-poise wires, and
(c) including protective devices like expulsion gaps, protector tubes on the lines, and surge
diverters at the line terminations and substations.

(a) Lightning Protection Using Shielded Wires or Ground Wires

Fig. 1.9 Shielding arrangement of overhead lines by ground wires

(b) Protection Using Ground Rods and Counter-Poise Wires
When a line is shielded, the lightning strikes either the tower or the ground wire. The path for
drainage of the charge and lightning current is (a) through the tower frame to ground, (b) through
the ground line in opposite directions from the point of striking. Thus, the ground wire reduces
the instantaneous potential to which the tower top rises considerably, as the current path is in
three directions. The instantaneous potential to which tower top can rise is

(25)

where, Zr = surge impedance of the tower, and

Zs = surge impedance of the ground wire.

(c) Protective Devices

In regions where lightning strikes are intense or heavy, the overhead lines within these zones are
fitted with shunt protected devices. On the line itself two devices known as expulsion gaps and
protector tubes are used. Line terminations, junctions of lines, and sub-stations are usually fitted
with surge arresters.

(i) Expulsion Gaps

(ii) Protector Tubes

(iii) Rod Gaps

(iv) Surge Arresters or Lightning Arresters

Surge diverters or lightning arresters are devices used at sub-stations and at line terminations to
discharge the lightning over voltages and short duration switching surges. These are usually
mounted at the line end at the nearest point to the sub-station. They have a flashover voltage
lower than that of any other insulation or apparatus at the sub-station. These are capable of
discharging 10 to 20 kA of long duration surges (8/20 μs) and 100 to 250 kA of the short duration
surge currents (1/5 μs).

1.3 PRINCIPLES OF INSULATION COORDINATION ON HIGH-VOLTAGE AND

EXTRA HIGH-VOLTAGE POWER SYSTEMS
Electric power supply should ensure reliability and continuity to the utility concerns. Hence, the
power lines and sub-stations are to be operated and protected against over voltages such that the
number of failures are as few as possible. At the same time, the cost involved in the design,
installation, and operation of the protective devices should not be too high. Hence, a gradation of
system insulation and protective device operation is to be followed, keeping in view the
importance of the various equipment involved.
Generally, sub-stations contain transformers, switchgear, and other valuable equipment with
no self-restoring insulation, which have to be protected against failures and internal destruction.
For other apparatus, which contain self-restoring insulation, like string insulators, they may be
allowed to flashover in air. But the flashovers should be kept to a minimum so that the system
disturbances are the least. Hence, lightning and switching surge protection requires establishment
of protective voltage levels called shunt protection levels, by means of protective devices like
lightning arresters.
The lightning impulse withstand level known as the Basic Impulse Level (BIL) is established
for each system nominal voltage for different apparatus. Various equipment and their component
parts should have their BIL above the system protective level, by a suitable margin. This margin
is usually determined with respect to air insulation by statistical methods. For non-self-restoring
insulation like the transformer insulation, the margin limit is fixed using conventional methods.
For system voltages below 400 kV, the switching surges are not of importance. If the BIL is
chosen correctly, relative to the prevailing protective level, the equipment will have an adequate
switching surge level. For higher system voltages, since the switching surges are of higher
magnitude compared to the lightning over voltages, switching surge magnitudes are of
importance and the following criterion is to be adopted.

(i) The flashover voltage of a protective derive is chosen such that it will not operate for
switching over voltages and other power frequency and its harmonic over voltages. But
other long duration over voltages, namely, sustained overvoltage due to faults and even the
above-mentioned over voltages may sometimes cause thermal overloading due to leakage
currents. Therefore, the BIL has to be higher.
(ii) For EHV systems, it may be economical to use a protective device for limiting the over
voltages due to lightning as well as switching surges to a particular level. At present, there
are surge arresters which operate for both types of over voltages mentioned above. In such
cases, it is preferable to assign to each protected equipment a Switching Impulse Level
(SIL), so that there is a small margin above the controlled switching surge level, so that the
surge diverters operate on switching surges, only when the controlling devices fail.
Normally, only rod gaps and lightning arresters are used as protective devices for
protection, and their characteristics are considered here.
The ideal requirements of a protective device connected in parallel or in shunt are the following:

(a) It should not usually flashover for power frequency over voltages.
(b) The volt-time characteristics of the device must lie below the withstand voltage of the
protected apparatus or insulation. The marginal difference between the above two should be
adequate to
allow for the effects of distance, polarity, atmospheric conditions, changes in the
characteristics of the devices due to ageing, etc.
(c) It should be capable of discharging high energies contained in surges and recover insulation
strength quickly.
(d) It should not allow power frequency follow-on current to flow.
The behavior of shunt connected protective devices like rod gaps and surge diverters along
with transformer insulation is given in Fig. 1.10.
In Fig. 1.10a, the transformer insulation strength is given as a volt-time characteristic. Figure
8.10b gives the relative insulation strengths of the transformer (curve A), rod gaps (curves B and
C), and that of a lightning arrester (curve D).
 Fig. 1.10 Volt-time characteristics of transformer rod gaps and surge diverters

A lightning arrester protects the transformer insulation in the entire time region. The rod gap
protects the transformer insulation, only if the rate of rise of surge is less than the critical slope
(curve
X). Thus, if the surge voltage rise is as shown by curve
1, rod gap flashes and protects the transformer, if the surge voltage rise follows
curve 2, only the surge diverter can protect the transformer insulation.
Rod gaps are simple and cheap devices but do not meet all the requirements of a protective
device. Moreover, their flashover characteristics depend on the atmospheric conditions, polarity
of the wave, and wave shape. Also, it may give rise to very steep impulse waves on the
transformer windings as chopped waves, because no current limiting resistance is used. Chopped
impulse waves may lead to the destruction of the transformer turn-to-turn insulation. But still, rod
gaps provide reasonable protection where lightning surge levels are low, and steep fronted surges
are controlled by overhead ground wires.
1.3.1 Surge Arresters
These are non-linear resistors in series with spark gaps which act as fast switches. A typical surge
arrester or lightning arrester is shown in Fig. 1.11b and its characteristics are given in Fig. 1.12.
A number of non-linear resistor elements made of silicon carbide are stacked one over the other
into two or three sections. They are usually separated by spark gaps (see Fig. 1.11b). The entire
assembly is housed in a porcelain water-tight housing. The volt-ampere characteristic of a
resistance element is of the form
(26)

where, Id = discharge current,

 V = applied voltage across the element, and k and a are constants depending on the
material and dimensions of the element.
The dynamic characteristics are shown in Fig. 1.11 a.
When a surge voltage (Vi of Fig. 8.11b) is applied to the surge arrester, it breaks down giving
the discharge current id and maintains a voltage Vd across it. Thus, it provides a protection to the
apparatus to be protected above the protective level Vp (see Fig. 1.11b).

Fig. 1.11a Metal Oxide (ZnO) surge arrester in polymer housing

Fig. 1.11b Non-linear element surge arrester

 Fig. 1.12 Characteristics of a surge arrester

(a) Surge Arresters for EHV Systems The selection of surge arrester voltage rating for EHV and
UHV systems depends on

(i) the rate of rise of voltage,

(ii) the type of system to be handled, i.e., whether effectively grounded or grounded through
an impedance etc., and
(iii) operating characteristic of the arrester.
The usual type of surge arresters used for the above purposes are

(i) silicon carbide arresters with spark gaps,

(ii) silicon carbide arresters with current limiting gaps, and (iii) the
gapless metal oxide (zinc oxide) arresters.
The first two types of arresters have a V − I characteristic of the nature of V = KIn, where n
varies between 0.5 and 0.6 for the elements.
The V − I characteristics of silicon carbide and zinc oxide arresters are shown in Fig.8.13a for
comparison.
 Fig. 1.13a Typical V-I characteristics of silicon carbide (SiC) and zinc oxide (ZnO) surge
diverters

Fig. 1.13b Current/Voltage characteristic of non-linear metal oxide arrester

Fig. 1.13c Typical voltage and current oscillograms obtained for positive negative impulse surge
test on an arrester

The advantages of zinc oxide arresters for EHV systems are

(i) they are simple in construction,
(ii) they have flat V-I characteristic over a wide current range, and
(iii) the absence of a spark gap that produces steep voltage gradients when sparking occurs.
The main disadvantage of zinc oxide arresters is the continuous flow of power frequency
current and the consequent power loss. Voltage grading system is not needed for each of the units
of the zinc oxide arresters used in EHV systems. A typical 400 kV line arrester may be rated at
15 kA and may have a resistance of 100 ohms at the peak current rating.
(d) Protection of Lines with Surge Arresters Since surge arresters are devices that provide
low resistance paths for overvoltages through an alternate ground path, their operating
characteristics and application is of importance. The spark gap inside the arrester acts as a fast
acting switch while non-linear arrester elements provide the low impedance ground path. The
arrester voltage at its terminal when connected to a line of surge impedance Z to ground, is given
as

(27)

where, Z is the line surge impedance,

The Thevenin equivalent circuit for the arrester is shown in Fig. 1.14.

Fig. 1.14 Equivalent circuit of a surge arrester

1.3.2 Equipment Insulation Level and Insulation Coordination of Sub-stations

Table 1.2 Insulation levels (BIL) for various system voltages

 Table 1.3 Standard insulation levels for equipment (> 300 kV)

The illustrate the principles of insulation co-ordination, an example of a 132 kV sub-station is

given below.
Nominal system voltage : 132 kV
Highest system voltage : 145 kV

Highest system voltage to ground

Expected switching surge overvoltage
(Table 8.1) 3.0p.u. : 3 X 119 = 357 kV (peak)
(a) Surge arrester
Rating : 123 kV
Front of wave sparkover
voltage: 510kV(peak)
Vd (discharge voltage at
10 kA, 8/20 µs
impulse current wave) : 443 kV (peak)
(b) Transformer
Impulse withstand voltage : 550 kV (peak)
Induced voltage (withstand) level : 230 kV (rms)

Impulse protective margin :

(c) Switchgear
Impulse withstand voltage : 650 kV (peak)
Bus insulation impulse withstand voltage: 650 kV (peak)

1.3.3 Insulation Levels at Sub-stations with Protective Zones

(a) Magnitude and Shape of the Incoming Voltage Surges

The stations are completely shielded from direct strokes. The shielding is sometimes affected up
to about 2 km on either side of the station. This protected zone gives the surges to originate only
from outside regions. Usually, the voltage wave at the station entrance is estimated by assuming
a voltage magnitude at the beginning of the protected zone as equal to 1.3 times the negative
critical flashover voltage of the line insulation. It is also important that the sloping off of the point
of the surge is helpful for the effective performance of the surge arrester.

(b) Equipment Insulation Level

For steep-fronted travelling waves, the voltages at different points in the sub-station can exceed
the protective level by amounts that depend on the distance from the arrester location, the
steepness of the wave front and other electrical parameters. Hence, it is necessary to decide the
number of locations at which surge arresters are to be located and their ratings. It necessary to
keep this number to a minimum. Also, care must be taken regarding switching over voltages
generated due to current chopping which may destroy the transformer or the equipment near the
circuit breakers. The Basic Impulse Level (BIL) is often determined as simply 1.25 to 1.30 times
the protective level offered by the surge arrester. Usually, the next higher BIL value from the
standard values is chosen. This is quite sufficient for smaller stations and station ratings up to 220
kV.

(c) Distance Effect Usually the circuit breaker, the transformer and other equipment are placed
at finite distances from the surge arrester and connected through a short distance overhead line or
cable. When a surge arises, it suffers multiple reflections between each of the equipment which
may give rise to over voltages of considerable magnitude (the travel time is usually less than a
µs). It can be shown that when a surge arrester, a breaker and a transformer are in line, the voltage
that can build up at a distance D from the surge arrester point is given as V(D) = Vp + 2ST, where
Vp is the sparkover voltage/protective level, S is the steepness of the wave front, and T is the travel
time = D/v. Here, v is the velocity of the wave travel, assuming that the line extends to a large
distance such that no reflections come from the line end. The maximum value of V(D) is attained
when 2T = T0, the sparkover time of the arrester. The above simple expression shows that the
transfer surge impedance is very high. The ratio of the transformer terminal voltage, VT to that of
the protective level Vp is a function of (T/T0). For steep fronted waves, sometimes VT may exceed
even 2Vp. It has been shown that in a 330 kV sub-station, a 1.2/50 ms, 1500 kV incoming wave
can give rise to a surge of peak value 1250 kV having rise time of 2 ms at the bus terminals
(neglecting the transfer capacitance), whereas the protective level offered by the surge arrester at
the transformer terminal is only 750 kV.

The transformer terminals may get a surge voltage of 930 kV peak. This has been verified by
computer calculations for a 1/50 µs wave, on a line connected to a 330-kV station.

1.3.4 Insulation Co-ordination in EHV and UHV Systems

The insulation design of EHV and UHV stations is based on the following principles

(i) stations have transformers and other valuable equipment that have non-self-restoring
insulation, and
(ii) the protective levels for lightning surges and switching surges are almost equal and even
overlap. If the basic impulse level for the equipment or the system is chosen, then this level
cannot give protection against the switching impulses. Hence, a separate switching impulse
level (SIL) has to be chosen. It is, therefore, desirable to use protective devices for limiting
both lightning and switching over voltages. As such, the switching impulse insulation level
above the controlled switching surge level has to be adopted so that the surge arresters
operate only rarely on switching over voltages when the controls of the control devices for
 switching voltages fail. A general guideline that can be adopted for different EHV and
UHV system for maximum switching surge levels are given in Table 1.4.

Table 1.4 Maximum switching surge level at different line voltages

Let P0 (Vi) dVi be the probability of a surge voltage occurring as an overvoltage between Vi
and (Vi + dVi). Let the probability for flashover of the insulation be P0(Vi). Then probability of
both the above events occurring simultaneously in Vi and (V0 + Vi) will be given by P0(Vi) Pd(Vi).
The risk of failure over the entire voltage range then becomes

Fig. 1.15 Risk of failure as a function of the probability of the occurrence of a surge voltage
[Po (Vi)] and the probability of the insulation flashover [Pd(V)]

If a number of insulation structures such as string of insulators are subjected to a switching

surge, then the failure is adjusted for appropriate probability, considering all the individual
probabilities simultaneously.
The risk of failure is graphically represented in Fig. 1.15.
 Fig. 1.16 Relation between the risk of failure (R) and the statistical safety factor (τ)

In normal practice, the insulation level and the protective safety margin are arrived at by

(i) selecting the risk of failure R,

(ii) the statistical safety factor, γ, and
(iii) then fixing the withstand voltage and designing the insulation level of any equipment or
apparatus corresponding to 90% or 95% of the withstand voltage thus fixed.

Fig. 1.17 The statistical safety factor (τ) and its relation to the risk of failure (R)
For proper insulation co-ordination, a certain margin of safety has to be provided by properly
choosing the “protective level” the protective devices, such as spark gaps and surge arresters, and
proper insulation level for the equipment and the apparatus. The correlation between the two is
illustrated in Fig. 1.18.

In the figure, Rg gives the risk factor for the protective gap with Pg (Vi) as its probability density
function for failure and the overvoltage probability density function Po (Vi) occurring. The
probability density function for insulation (to be protected) is given by Pi (Vi) and Ri is its risk
factor. The safety margin which is the difference between Pg (Vi) and Pi (Vi) is also shown. In
reality, this computation is not simple and deviation in the occurrence of over voltages greatly
influences the risk factors and the safe margin usually gets reduced.

Fig. 1.18 Protective margin (margin of safety) and risk of failure provided by a protective device

The insulation co-ordination and safety margins fixed for a typical 750 kV station can be as
follows:
surge-arrester voltage rating : 590 kV
maximum sealed-off value of the
power frequency harmonics : 950 kV
lightning surge protective level : 1450
kV
switching surge protective level : 1200
kV
transformer or reactor BIL : 1800
kV
transformer or reactor SIL : 1500
kV
If the 765-kV line uses a V type insulator string with 35 standard type disc insulators, then the
CFO for lightning surges will be about 1350 kV. These values overlap with the SIL of the circuit
breaker and other apparatus. Then, a suitable design for a conductor to tower clearance and for
the conductor to ground clearance have to be given.
Surge diverter or lightning arrester is the main device that is used in power systems and
substations to limit the overvoltages. Hence, proper selection and approach to determine the
rating of arrester without damage to itself is of importance. In EHV and UHV systems the
arresters used are gapless ZnO type. The arrester must have proper

(i) MCOV: The maximum power frequency continuous operating voltage

(ii) TOV: Temporary power frequency overvoltage that can appear on it
(iii) Energy: The energy-dissipating capacity, i.e., energy to be absorbed by it when limiting the
surge voltages.
To conclude insulation coordination in power systems at sub-stations is to choose a protective
device (surge arrester) such that its characteristics give sufficient but not very high protective
margin