460

connecting spindle

flange
(aluminium alloy)

elastic stirrup

exhaust pipe and rivet

overpressure device
in the upper and Zn O blocks
lower flanges
washer

fault indication
plate

spacer

thermal shield
exhaust pipe
porcelain enclosure

compression spring

flange
rubber seal

prestressed tightness
device

ring clamping overpressure device

device

Figure 5-48: example of the structure of a ZnO surge arrester in a porcelain enclosure
for 20 kV networks

peak kV
U

600
500
400 Zn O
300

200
linear
SiC
100

.001 .01 .1 1 10 100 1000 10000 I

SiC : non-linear resistor type gapped surge arrester made up of a silicon carbide resistor
ZnO : zinc oxide surge arrester
linear : U curve proportional to I

Figure 5-49: characteristics of two surge arresters having the same 550 kV/10 kA protection level
Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 439

Reactance earthing: (network reactance)

Voltage at circuit-breaker terminals Voltage at the terminals of the reactance

Resistance earthing: (network reactance)

Voltage at circuit-breaker terminals Voltage at the terminals of the resistor

: arc extinction voltage

Figure 5-35: transient voltage on circuit-breaker opening during a permanent phase-earth fault

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly orpartly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 482

n recapitulative table of touch voltages and overvoltages which occur for each earthing
system

TN − a IT − a TT − a IT − b TN − b TT − b IT − c

Touch voltage Y Y N N N N N

Overvoltage of LV
N N Y Y N N N
installation exposed
conductive parts

Overvoltage of
N N N N Y Y Y
substation exposed
conductive parts

Y : yes
N : no

Table 5-8: touch voltages and overvoltages which occur for

each earthing system

5.3.2.2. Rise in potential of the LV exposed conductive parts on occurrence of a

lightning impulse
When a lightning overvoltage from the distribution network flows to earth in an MV/LV
substation through a protection device (surge arrester or MV spark-gap), there follows a rise in
potential of the substation LV exposed conductive parts and/or those of the installation which
depends on the earthing system.

The level of overvoltages transferred in LV depends on the clipped value Ursd and the earth
electrode values.

To ensure protection of the LV switchgear against these overvoltages, LV surge arresters must
be installed and the resistance of the substation earth electrode limited so that the equipment
lightning impulse withstand voltage is not exceeded.

n limiting earth electrode impedances

As for the case of the MV earth fault, the limit values of the earth electrode impedances are
calculated for each earthing system.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 473

o power frequency withstand voltage

This corresponds to the equipment withstand to power frequency overvoltages likely to occur
on the network and the duration of which depends on the network operating and protection
mode.

The equipment withstand is tested by applying a sinusoidal voltage with a frequency of

between 48 Hz and 62 Hz for one minute. The test is valid for nominal network frequencies of
50 Hz and 60 Hz (see IEC 71-1).

o switching impulse withstand voltage

This characterises the equipment withstand to switching impulses (only for equipment with a
standard voltage above or equal to 300 kV).

The equipment test (see IEC 60-1) is performed by applying a wave with a front time of 250 µs
and a time to half-value of 2500 µs.

o lightning impulse withstand voltage

This characterises the equipment withstand to the 1.2 µs / 50 µs lightning voltage wave.
This withstand voltage concerns all voltage ranges, including low voltage.

o examples of equipment withstand (see table 5-5)

Highest voltage for the Standard short-duration Standard lightning impulse

equipment power frequency withstand withstand voltage (kV)
U m (kV) (1) voltage (kV) (peak value)
(r.m.s. value) (r.m.s.)
3.6 10 20
40
7.2 20 40
60
12 28 60
75
95
17.5 38 75
95
24 50 95
125
145
36 70 145
170
52 95 250
72.5 140 325

(1) Um is the highest rms value of the phase-to-phase voltage for which the equipment is specified.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 398

For a solid fault ( R f = 0 ) at the transformer terminals:

R(1) = RT = 0.056 Ω

R(0) = 3 Re + RT = 3 × 0.5 + 0.056 = 1.56 Ω

X(1) = XT = 1.22 Ω

X(0 ) = X(0 )T = 0.85 Ω

whence R(1) = 0.05 X(1) ≅ 0

R(0)
= 1.28
X(1)

X(0 )
= 0.70
X(1)

Figure 5-3 shows that k is between 1.4 and 1.5.

For a solid fault ( R f = 0 ) 5 km away from the transformer:

R(1) = RT + RC = 0.056 + 0.15 × 5 = 0.81 Ω

R(0) = 3 Re + RT + RC = 3 × 0.5 + 0.056 + 0.15 × 5 = 2.31 Ω

X(1) = XT + XC = 1.22 + 0.1 × 5 = 1.72 Ω

X(0 ) = X(0 )T + X(0 )C = 0.85 + 0.3 × 5 = 2.35 Ω

whence R(1) = 0.47 X(1)

R(0)
= 1.34
X(1)

X( 0)
= 1.37
X(1)

Figure 5-4 shows that k is between 1.2 and 1.3.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 475

5.3.2. Reduction in risks and overvoltage levels

The risks of overvoltages, and consequently the danger they represent for persons and
equipment, can be greatly reduced if certain measures of protection are taken:

- limiting substation earth electrode resistances in order to reduce power frequency

overvoltages

- reducing switching overvoltages by choosing suitable switchgear (interruption in SF6)

- making lightning impulses flow to earth by a first clipping operation (surge arrester or spark-
gap at the entrance to the substation) with limitation of the earth electrode resistances and
pylon impedances

- limiting the residual voltage from the first clipping operation by HV surge arrester which is
transferred to the downstream network by providing a second protection level on the
transformer secondary

- protection of sensitive equipment in LV (computers, telecommunications, automatic

systems, etc.) by connecting series filters and/or overvoltage limiters to it.

5.3.2.1. Rise in potential of LV exposed conductive parts following an MV fault in the

transformer substation
In this paragraph, we propose to study overvoltages in LV caused by an earth fault on the MV
side in an MV/LV substation, and the measures to be taken in order to protect equipment and
persons, in compliance with IEC 364-4-442.

The values of rises in potential of the substation or LV installation exposed conductive parts
depend on the values of the earth electrode resistances, the fault current values and the
earthing system.

n earthing in transformer substations

A single earth electrode must be installed in a transformer substation, to which must be

connected:

- the transformer tank

- the metallic coverings of high voltage cables

- the earth conductors of high voltage installations

- the exposed conductive parts of high voltage and low voltage equipment

- the extraneous conductive parts.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 474

Table 5-5: standard withstand voltages for 3.6 kV < U m < 72.5 kV

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 462

n choice of zinc oxide surge arresters in HV

The general method for choosing a zinc oxide surge arrester in HV consists in determining its
characteristic parameters using the network data, at the place where it will be installed.

The parameters characterising the surge arrester are:

- U C , steady-state voltage

- U r , rated voltage

- I nd , nominal discharge current

- discharge class and energy capacity

- mechanical characteristics.

The data relative to the network are:

- Um , highest phase-to-phase voltage applied to equipment

- TOV temporary overvoltages (appearing on occurrence of an earth fault or load shedding

on the public distribution network).

The choice of the surge arrester involves making a compromise between the equipment
protection levels and the energy capacity of the surge arrester.

The protection level must be as low as possible for the equipment withstand. This involves the
lowest voltage rating possible and thus greater difficulty withstanding temporary overvoltages.

o determining U C and U r

Ÿ simplified method using equipment characteristics

The voltages U C and U r may be directly determined using the highest voltage for the
equipment Um :

Um
UC ≥
3

Ur = 1.25 × UC

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 425

5.1.3.2. Load switching

n de-energizing loads

o inductive load

Ÿ single-phase circuit

Let us consider the equivalent single-phase diagram in figure 5-25 with an ideal circuit-breaker
CB which has a zero arc resistance the instant the contacts separate and which carries out
interruption when the current crosses zero. Before operation of the circuit-breaker, between
points A and B, there is a voltage drop due to the load current flowing through Ls .

At the instant of interruption, the voltage at B suddenly reaches the voltage at A and the
capacitance Cs is charged through Ls . The energy exchanges between Cs and Ls make
voltage oscillations at frequencies of 5 to 10 kHz occur.

The voltage at C suddenly decreases to zero and the capacitance C p is then discharged
through L . The energy exchanges between Cp and L create voltage oscillations at
frequencies going from 1 to 100 KHz.

A Is CB
B ID C IL
Ls

VA Cs Cp L

I0
Lp

Ls : network inductance upstream of the circuit-breaker

Cs : network capacitance upstream of the circuit-breaker
L : load inductance
L p : stray inductance
C p : network capacitance downstream of the circuit-breaker
CB : circuit-breaker
Figure 5-25: interruption in an inductive load network

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 418

In the case of VT with a single secondary, a charging resistor is installed on each phase
(see fig. 5-19).

100
A resistance value equal to 68 Ω is recommended for a secondary voltage of V .
3

VT VT VT

r
r
r

measurements

Figure 5-19: protection against risks of ferro-resonance using resistors with single secondary VT

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 448

n indirect lightning strikes (lightning protection rope or pylons)

When lightning strikes the line protection rope, part of the current flows through the pylon since
the protection rope is connected to it (see fig. 5-42).

This results in a potential rise at the top of the pylon the value of which depends on the self
inductance L of the pylon and the resistance R of the earth electrode:

 di ( t ) 
U (t ) = k  R i (t ) + L 
 dt 

k : ratio of the current shunted into the pylon by the incident current

lightning strike
i
k.i
U protection rope

k .i

 di 
U = k R × I + L 
 dt 

Figure 5-42: lightning strike on a protection rope

The voltage U may reach the impulse sparkover voltage of the insulators and cause a
breakdown. This is "back-flashover". Part of the current is then propagated along the affected
phase(s) towards the users. This current is in general greater than that of a direct lightning
strike.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction , either partly of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 437

We can see that reactance earthing of the neutral (case with restrike) clearly increases the
magnitude of the overvoltages. Resistance earthing is thus preferable. In the latter case, we
see that the overvoltages do not exceed 240 % when the ratio of the current in the earthing
resistor to the network capacitive current is equal to 2 (see fig. 5-34). In networks with
resistance earthing, the following relation should therefore always be respected if possible:

I rN > 2 I C

I rN : current in the neutral earthing resistor during the fault

I C : currents in the network phase-earth capacitances (see § 4.3 of Protection guide)
V3 CB
Ph 3
V2
Ph 2
V1
Ph 1

ZN C C C
or r If
N

ZN : neutral earthing impedance (or rN )

C : phase-earth capacitance
If : fault current

CB : circuit-breaker
V1 , V2 , V3 : single-phase voltages

Figure 5-32: phase-earth fault clearance

Xnet CB

~
ZN C
or r
N
If

IC
I rN

Xnet : network reactance

C : fault phase earth capacitance
Z N or rN : neutral earthing impedance (or resistance rN )
If : fault current

Figure 5-33: fault circuit on occurrence of a phase-earth fault

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 408

In high voltage, the substation earth electrode value ( Re ) is very low compared with the
transformer and line impedances. The fault current is thus independent of Re ; it is thus
considered to be a source of current with a value of I f .

The equivalent Thevenin’s diagram of the current source I f with an internal impedance of
Re is shown in figure 5-11.

Re

equivalent
If Re Ve Re I f

Figure 5-11: equivalent Thevenin’s diagram of the current source I f with an internal impedance of Re

The equivalent MV network diagram is thus that shown in figure 5-12.

Re Rc Lc

Z MV Z MV Z MV
Ve Re I f
C C C

Figure 5-12: equivalent MV network diagram on occurrence of an earth fault on the substation HV side

The transformer and cable impedances are negligible compared with the cable phase-earth
1
capacitance: Z MV << .
Cω

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 476

n symbols

In the following paragraphs, the symbols used have the following signification:

Im : part of the earth fault current in the high voltage installation which flows through the earth electrode of the
transformer substation exposed conductive parts
Re : transformer substation earth electrode resistance
V : low voltage installation phase-to-neutral voltage
U : low voltage installation phase-to-phase voltage
U f : fault voltage in the low voltage installation, between the exposed conductive parts and earth
U1 : stress-voltage in the transformer substation low voltage equipment
U 2 : stress-voltage in the installation low voltage equipment

n TN − a and IT − a earthing systems (see fig. 5-54)

In these two systems, the substation, neutral and installation earth electrodes are the same.
Inside the equipotential area, the ground and exposed conductive part potentials increase
simultaneously. The touch voltage U f is then zero.

On the other hand, outside this area, the ground potential remains equal to that of the remote
earth, while the potential of the exposed conductive parts increases to U f = Re I m .

Thus, when there are exposed conductive parts outside of the equipotential area and the
touch voltage U f = Re I m cannot be cleared in the time defined in tables 2-3-a and 2-3-b, the
TN − a and IT − a earthing systems are not acceptable in relation to the protection of
persons.

To overcome this drawback, the following provisions must be taken:

- TN − a earthing system: the neutral of the LV installation must be connected to a separate

earth electrode, which is the case in the TN − b earthing system (see fig. 5-55)

- IT − a earthing system: the exposed conductive parts of the LV installation must be

connected to a separate earth electrode from that of the substation, which is the case in the
IT − b earthing system (see fig. 5-56).

TN − b and IT − b earthing systems allow dangerous touch voltages to be cleared but make
overvoltages occur:

- in the installation LV equipment for the IT − b earthing system

- in the substation LV equipment for the TN − b earthing system.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 459

n characteristics

Variable resistor type surge arresters are characterised by:

- the rated voltage, which is the maximum specified value of the power frequency rms voltage
permitted between its terminals for which the surge arrester is designed to function
correctly. This voltage can be continuously applied to the surge arrester without this
modifying its operating characteristics.

- the sparkover voltages for the different wave forms (power frequency, switching impulse,
lightning impulse, etc.).

- the impulse current evacuation capacity.

5.2.3.2. Zinc oxide ( ZnO ) surge arresters

n operating principle

Figure 5-48 shows that, unlike the non-linear resistor type gapped surge arrester, the zinc
oxide surge arrester is only made up of a highly non-linear variable resistor.

The resistance goes from 1.5 MΩ at the duty voltage (which corresponds to a leakage current
below 10 mA) to 15 Ω during discharge.

Following the passage of the discharge current, the voltage at the terminals of the surge
arrester become equal to the network voltage. The current which flows through the surge
arrester is very weak and is stabilised around the value of the earth leakage current.

Because of the high non-linearity of the ZnO surge arrester a high current variation causes a
low voltage variation (see fig. 5-49).

For example, when the current is multiplied by 107, the voltage is only multiplied by 1.8.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 453

The value of v is close to 150 m/µs.

This gives us an idea of the way a lightning wave spreads along a conductor. Figure 5-45
shows how a lightning wave spreads along an overhead line in relation to time and space.

development
in time V front : 200 kV / µs

400 kV

2 µs

t (to x constant)

spread in
space V front : 0.66 kV / m

400 kV

600 m = 300 x 2 µs

x (to t constant)

Figure 5-45: diagram showing how a lightning wave spreads along an overhead line
in relation to time and space

Let us closely examine the phenomenon that is produced at a point M , where a change of
impedance exists, separating two circuits with characterstic impedances of Z1 and Z 2
(see fig. 5-46).

v1 v2
i1 i2

Z1 M Z2
v 1'
'
i1

Z1 , Z 2 : upstream and downstream characteristic impedances

v1 , i1 : incident wave upstream of M
v2 , i2 : wave transferred downstream of M
v1' , i1' : wave reflected upstream of M

Figure 5-46: propagation of a wave at a change of impedance point M

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction , either partly of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 406

The voltage U is common to the three elements.

We have the following relation:

1 1 
J = + + j Cω U
 R j Lω 

The resonance phenomenon occurs when I L = − IC :

U
= − jCωU
j Lω

L C ω2 =1

We thus have U = R J ; the inductance and capacitance behave like an open circuit.

For given values of L and C , the angular frequency ω r such that LC ω 2r = 1 is said to be
a resonant angular frequency.

An overvoltage factor is thus defined which is the ratio:

- between the voltage that is produced at the terminals of the parallel R , L, C circuit when
the resonance occurs

- and the voltage that would be produced on occurrence of the resonance if the inductance
(or capacitance) were the only circuit element

RJ
f =
Lω r J

R
f = = RCωr
Lω r

The most current example of parallel resonance is the case of a network having harmonic
currents (patterned by current sources) and reactive energy compensation capacitors
(see § 8.1.5).

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 397

o example

Let us consider a YNyn, 33 kV/11 kV transformer with a power rating of Sn = 24 MVA (see
IEC 909-2 table 3 A) supplying a network with 240 mm² aluminium cables the longest outgoing
feeder of which is 5 km. The neutral earth electrode resistance is 0.5 Ω.

- transformer characteristics:

Usc = 24.2 %

RT
= 0.046
XT

X(0 )T
= 0.7
XT

we can deduce
U2
XT = Usc × n = 0.242 ×
11 × 10 3
= 1.22 Ω
( )
Sn 24 × 10 6

RT = 0.056 Ω

X(0 )T = 0.85 Ω

Note: the value of Usc is extremely high in relation to the transformers feeding a network with a
limiting resistor earthed neutral. The transformer here is a United Kingdom transformer adapted
to the solidly earthed neutral system.

The short-circuit voltage has been chosen high on purpose so as to minimise the short-circuit

( )
R(0)
current. Indeed, if Usc is high, the value is minimised since X(1) = XT + XC , which
X(1)
decreases the overvoltage factor (see fig. 5-3 and 5-4).

- cable characteristics:

ρ L 0.036 × 1000
RC = = = 0.15 Ω / km
S 240

XC = 0.1 Ω / km

We assume that X(0 )C = 3 XC = 0.3 Ω / km .

Note: the value of X( 0 )C is highly variable (from 0.2 to 4 X(1) ) depending on what the cable is made
of and the return via the earth (remote earth, screen or earthing conductor).

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 424

R L

V C VA

I V VA

L , R : inductance and resistance equivalent to the network upstream of the circuit-breaker

C : upstream network capacitance

Figure 5-23: transient recovery voltage during circuit-breaker interruption

n fuse interruption

On occurrence of a short circuit, the value of the current flowing through the fuse is higher than
its nominal fusing value.

Interruption can thus occur at any instant and not necessarily the moment the current crosses
zero.

Figure 5-24 gives an example of a transient overvoltage which occurs on the network after a
wire fuse has fused.
Volts

1000

225

t
~ 1 ms

Figure 5-24: transient overvoltage on fusion of a wire fuse

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 400

For a transformer or a cable in low voltage, we can take the zero-sequence impedance to be
approximately equal to the positive-sequence impedance: Z(0)T = ZT and Z(0)C = ZC .

We thus have Z(0) = Z T + ZC + 3 Z PE

Z(1) = ZT + ZC

a 3 Z PE a Z PE
whence kM = 1− = 1− for a solid fault ( R f = 0 )
3 ( ZT + Z C + Z PE ) Z PE + ZT + Z C

j 2π
a=e 3 : rotation operator of 120°

The overvoltage will be maximum when Z T is negligible compared with Z PE + ZC , which is the
case for a long length cable.

a Z PE
Thus kM ≤ 1−
Z PE + ZC

k M will be maximum when the protective conductor cross-sectional area is as small as

possible, i.e. equal to half the phase conductor cross-sectional area; thus RPE = 2 RC .

For an aluminium cable cross-sectional area smaller than 120 mm², the reactance can be
neglected compared with the resistance, which thus gives us:

Z PE RPE 2
≅ = since RPE = 2 RC
Z PE + ZC RPE + RC 3

2
whence kM ≤ 1− a
3

2 1 3
kM ≤ 1− − + j 
3 2 2 

k M ≤ 1.45

We can show that for a cable with a large cross-sectional area (> 120 mm²), the overvoltage
will be lower than in the case of a small cross-sectional area.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 415

On saturation, L ( I ) becomes very weak and the voltage almost stops increasing when I
rises.

The network operating point is located at the intersection of curve (b) having the equation:

I
v = VL −
Cω

and ellipse (a) having the equation:

v = V 2 − R2 I 2

There are three possible operating points: M , N , P . M and P are stable, N is unstable.
A voltage disturbance can make the circuit move from point M to point P . This results in a
high current and high overvoltages at the inductance and capacitance terminals. Ferro-
resonance can be avoided if the resistance R is sufficiently high for ellipse (a) to stay within
zone OSA , even when there is a high overvoltage.

VL
resonance

VC

V ''' (a''')
X (b)

V
' S (a') T
V N (a)
M P
''
V Q A
(a'')
O IQ IM IS IN IA IP IT IX I

Figure 5-17: series ferro-resonance - graphic resolution

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 412

This equation can be graphically resolved by plotting, as a function of V, the curves

representing functions (see fig. 5-15):

V2
I = IT2 − (a)
R2

I = C ω V − IL (b)

For any value of IT , the intersection of curves (a) and (b) gives the V solutions of equation
(3); figure 5-15 shows the graphic resolution of this equation.

Curve (a) is an ellipse having the equation:

V2
2
+ I 2 = IT2
R

and having one half axis which is equal to IT and the other to R IT . An ellipse corresponds
to each total current value IT .

Curve I L (V ) presents a very steep slope when V increases owing to the saturation of the
V
coil's iron core: I L (V ) = .
L (V ) ω
On saturation, L (V ) becomes very weak and the current then highly increases (see fig. 5-15).

Curve I C = C ω V is a linear function of V (see fig. 5-15).

Curve (b) shows the development of I C − I L = (C ω V − I L ) as a function of the voltage.

The OSA portion of curve (b) corresponds to a lead current in relation to the voltage owing to
the preponderance of the capacitive current. On the other hand, the AB part corresponds to a
lag current, since the inductive current is preponderant. The intersection of ellipse (a) and
curve (b) can give:

- an operating point Q if ellipse (a) is inside ellipse (a") passing through point A

- three operating points M , N , P if ellipse (a) is between ellipses (a') and (a")

- two points S , T if ellipse (a) is equal to ellipse (a')

- a single point X if ellipse (a) is outside ellipse (a').

The ferro-resonant mechanism is described below.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 472

The clearance is directly related to the withstand of the equipment to different overvoltages.

distance
in air

creepage
distance
distance
in air

Figure 5-53: air clearance and creepage distance

n overvoltage withstand

The overvoltage withstand depends on the type of overvoltage applied (magnitude, wave form,
frequency and duration, etc.).

It is also influenced by external factors such as:

- ageing

- environmental conditions (humidity, pollution)

- variation in air or insulating gas pressure.

n withstand voltage

Electrical equipment is characterised by its withstand voltage to different types of overvoltages.

We can thus distinguish:

- the power frequency withstand voltage

- the switching impulse withstand voltage

- the lightning impulse withstand voltage.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 423

5.1.3. Switching overvoltages

When switching to energize or de-energize loads transient overvoltages occur on the network.
These overvoltages are all the more dangerous if the current interrupted is inductive or
capacitive. The magnitude, frequency and damping duration of these transient overvoltages
depend on the given network characteristics and the mechanical and dielectric characteristics
of the switching device.

5.1.3.1. Interrupting principle

Interrupting an electric current using an ideal device involves the resistance of the device going
from zero before interruption to an infinite value just after interruption. The interruption occurs
the instant the current crosses zero.

It is impossible to make such an ideal device, but with the interrupting techniques being based
on the behaviour of the electric arc in different dielectric media we can come close to it.

n circuit-breaker interruption

The instant the current is interrupted, an electric arc is created between the terminals of the
switching device. The conductive electric arc tends to be held by the ionizing phenomenon of
the dielectric caused by the energy dissipated.

Around current zero crossing, the dissipated energy decreases dropping below the thermal
energy supplied to the medium, the arc cools down and its resistance increases.

When the current crosses zero, the arc resistance becomes infinite and the arc is interrupted.

Between the start and end of interruption, the voltage between the poles of the switching
device goes from zero to the network voltage. This change gives rise to a high frequency
transient phenomenon called the transient recovery voltage (see fig. 5-23).

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 442

Similarly, the steepness of the wave front varies according to a log-normal distribution law. Let us
determine the probability of exceeding a given front steepness (see fig. 5-38). We can see that the
probability of exceeding a front steepness of 50 kA/µs of a negative stroke of lightning is 20 %.

Figure 5-38: probability of exceeding the front steepnesses of positive and negative
lightning currents according to IEEE (experimental statistic)

Ÿ standard wave form

The lightning impulse wave form given by IEC 71-1 is a 1.2/50 µs wave (see fig. 5-39):

- rise time to the maximum value of 1.2 µs

- time to half-value of 50 µs.

Figure 5-39: standard lightning impulse voltage wave form (IEC 71-1)

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly orpartly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 432

When the supply voltage reverses back, a half period later, the potential difference at the
device terminals then reaches 4 V$n . Such an overvoltage can obviously cause renewed
restrike between the device contacts, and the previously described oscillation mechanism is
renewed with increased magnitude, leading to a new rise in the phase-neutral voltage of the
capacitances 5 V$ . ( n)
The cumulative effect of multiple restrike is obviously highly dangerous for the network
components as for the device itself.

This rise in overvoltages can be avoided by choosing the appropriate equipment, i.e. which
does not allow restrike.

VC 5V$n

V 20 ms
VC V$n

V$n 2 V$n

t
4 V$n

interruption

VC 3V$n

Figure 5-30: voltage rise on separation of a capacitor bank from

the network by a slow operating device

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 428

Ÿ multiple restrike

If we consider the single-phase diagram in figure 5-25, we can see that in the case of restrike,
the voltage at point C almost instantaneously reaches the voltage at point B .

The capacitance C p is charged by a high frequency current (roughly 1 MHz) circulating in the
L p , Cs , CB and C p circuit.

This high frequency current very quickly crosses zero (1 µs).

If the circuit-breaker manages to interrupt the current at that moment, the restrike phenomenon
is repeated as the distance between the circuit-breaker contacts is still very small.

Furthermore, the peak-to-peak magnitude of the oscillation is then equal to 4 V$n .

The overvoltage increase makes the occurrence of a second breakdown highly probable.

Indeed, the increase in dielectric withstand through the increase in the distance between the
circuit-breaker contacts may be lower than the increase in overvoltage.

This is why a multiple restrike phenomenon occurs with overvoltages of increasing magnitude
(see fig. 5-28).

In theory, such a phenomenon may generate overvoltages having a peak value equal to the
dielectric withstand limit of the open device, without a definite interruption of the current being
obtained. In practice, this case remains exceptional as it is enough for one of the restrikes to
allow the power frequency current to be restored; a new current half wave then flows through
the circuit-breaker. The circuit-breaker interrupts this half-wave the moment it crosses zero
when the distance between the contacts is sufficient. Thus the types of circuit-breakers
undergoing multiple restrike usually manage to interrupt the current without causing
overvoltages of very high magnitude.

VC

Figure 5-28: voltage VC in case of interruption with multiple restrike

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 477

substation LV installation

U1 U2

MV ph 1 LV

ph 2

ph 3
PEN

Uf 0 Uf Re I m
equipotential zone outside
zone

U1 V U2 U1 V
Im Re

TN − a

U1 U2

MV LV

Uf 0
Uf Re I m
equipotential zone outsite zone

Im Re
U1 V 3*
U2 U1 V 3*
(*) a first LV fault is present

IT − a

Figure 5-54: rise in potential of TN-a and IT-a earthing systems

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 480

n TT − a and IT − b earthing systems

In these two cases the substation earth electrode and that of the neutral are common.
The LV installation earth electrode is separate.

The earth fault current flows through the common earth electrode (neutral/substation).

As shown in figure 5-56, we can see that there is a risk of breakdown for the LV equipment
whose earth electrode is separate from that of the substation.

The following conditions must be met:

UtM > Re Im + V for the TT − a earthing system

and UtM > Re I m + V 3 for the IT − b earthing system

whence  UtM − V
 Re < Im
for the TT − a earthing system
 UtM − V 3
 Re < for the IT − b earthing system
 Im

where:

U tM : power frequency withstand voltage of the installation LV equipment equal to 2V + 1000 for V = 220 to
250 V, i.e. 1500 V

Table 5-7 gives the values of Re for different values of Im .

TT − a IT − b

I m = 300 A 4Ω 3.5 Ω

I m = 1000 A 1.2 Ω 1Ω

I m = 5000 A 0.24 Ω 0.2 Ω

Table 5-7: maximum values of Re in TT − a and IT − b earthing systems

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 436

5.1.3.3. Circuit-breaker clearance of phase-earth faults

Let us consider the three-phase network shown in figure 5-32 in which phase 1 is affected by
an earth fault.

In this case, the network is equivalent to the diagram in figure 5-33 which corresponds to the
case examined in paragraph 5.1.2.1.

At the start of contact separation, the arc voltage is weak and remains constant.

On the other hand, just before interruption, this voltage, called the extinction voltage, increases
to a more or less high value which may exceed V$n . This voltage depends on the type of
circuit-breaker (air, oil, SF6 , vacuum) as well as the arc extinction technique (cooling,
lengthening, rotating arc).

When the current crosses zero, the arc is extinguished and the recovery voltage magnitude
will depend on the extinction voltage as follows:

- for the case of neutral earthing via resistance (the fault current is in phase in relation to the
voltage), the extinction voltage limits the magnitude of recovery voltage oscillations

π
- for the case of neutral earthing via reactance (the fault current is phase shifted by in
2
relation to the voltage), the extinction voltage increases the magnitude of oscillations.

After interruption, restrike may take place if re-generation of the dielectric medium is not fast
enough in relation to the rise in recovery voltage. In this case, the magnitude of oscillations
may reach double the size of the first recovery voltage.

If we neglect the transformer and line impedances, the voltage at the terminals of the neutral
earthing impedance (VN ) is equal to the difference between the supply voltage and the
voltage at the circuit-breaker terminals. The voltage V N is vectorially added to the voltage of
the healthy phases and may lead to the latter reaching higher overvoltages than the
overvoltages observed on the fault phase.

The curves in figure 5-34 give the overvoltage levels recorded on occurrence of an earth fault
in relation to the network characteristics and the earthing impedance.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 405

The resonance phenomenon occurs when U L = − U C :

1
j Lω I = −
j Cω

LC ω 2 = 1

We thus have U = R I ; the series inductance and capacitance behave like a short circuit.

For given values of L and C , the angular frequency ω r such that LC ω 2r = 1 is said to be
a resonant angular frequency.

An overvoltage factor f is thus defined which is the ratio of the voltage U L (or U C ) to the
supply voltage U :

U L Lω r I
f = =
U RI

Lω r 1
f = =
R RC ω r

parallel resonance

Figure 5-9 shows a parallel R , L, C circuit at the terminals of which a current source J is
applied.

IR IL IC

J C
L U
R

Figure 5-9: parallel R , L, C circuit fed by a current source J

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 446

n direct lightning strike (on phase conductors)

When lightning strikes the phase conductor of a line, the current i (t ) is shared out in equal
quantities on either side of the point of impact and is spread along the conductors. These have
a wave impedance Z the value of which is between 300 and 500 Ω. This impedance is that
seen by the wave front, is independent of the length of the line and of a different type from the
impedance at 50 Hz.

This results in a voltage wave of:

i (t )
U (t ) = Z .
2

which spreads along the line (see fig. 5-41).

U i
i
U =Z
2

Figure 5-41: lightning strike on a phase conductor

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction , either partly of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 481

substation LV installation
U1 U2

MV L1 LV

L2

L3
N

U1 V U2 R e Im V
Uf
Uf 0
Im Re RA

TT − a

U1 U2

MV LV

*
U1 V 3
Uf
U2 Re Im V 3*
Im Re Uf RA If UL If RA

(*) a first LV fault is present

IT − b

Figure 5-56: Rise in potential in TT − a and IT − b earthing systems

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 470

electrical switchboard

disconnecting
circuit-breaker
equipment to
be protected

surge arrester

PEN
Ph1
Ph2
Ph3
PEN

LV neutral main earth load earth

earth electrode (entrenched
terminal loop) electrode

TNC earthing system

electrical switchboard

disconnecting
circuit-breaker
equipment to
be protected

surge arrester
PE
PE

Ph1
Ph2
Ph3
N
PE

LV neutral
main earth (entrenched load earth
earth electrode
terminal loop) electrode

TNS earthing system

Figure 5-52-b: connection layout of an LV surge arrester

for TNC and TNS earthing systems

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 456

5.2.2. Spark-gaps
n operation

The spark-gap is a simple device made up of two electrodes, the first connected to the
conductor to be protected and the second connected to earth.

At the place where it is installed in the network, the spark-gap constitutes a weak point where
overvoltages can flow to earth and thus protects the equipment.

The sparkover voltage of the spark-gap is set by adjusting the distance in the air between the
electrodes so as to obtain a margin between the impulse withstand of the equipment to be
protected and the impulse sparkover voltage of the spark-gap (see fig. 5-47). For example,
B = 40 mm on French public EDF 20 kV networks.

bird proof rod

earth electrode phase electrode

45° 45°

electrode B
holder

rigid
anchoring chain

device for adjusting B

and locking the electrode

Figure 5-47: MV spark-gap with birdproof rod

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 401

n TT earthing system (see fig. 5-6)

ZT ZC
V3
ZT ZC
V2
ZT ZC
V1
If

load 1 load 2
If

VN Re VM1 RM1 R M2

Re : substation earth electrode resistance

R M1 : load 1 and fault load earth electrode resistance
RM 2 : load 2 earth electrode resistance
V M1 : load 1 and fault load phase-to-earth voltage

Figure 5-6: equivalent diagram of an earth fault in a TT earthing system

We want to know the overvoltage of the healthy phases in relation to the exposed conductive
part, which determines whether or not an insulation fault may occur on the other load:
V − VM V − VM
kM = 2 = 3
Vn Vn

In low voltage, the neutral and load earth electrode resistances are very high in relation to the
transformer and cable impedance ( Z T and Z C are roughly several tens of mΩ).

We can thus write that the fault current is:

V1
If = and ( ZT + ZC ) I f ≅0
Re + RM1

The exposed conductive part of load 1 is connected to phase 1 by the fault (zero impedance).
The voltage of one healthy phase of this load in relation to the frame is V2 − V1 or V3 − V1
(since ( Z T + ZC ) I f ≅ 0 ) , whence k M = 3 = 1.73 .

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 447

Depending on the magnitude of the lightning current, two cases may occur:

o full impulse propagation

 I 
If the maximum voltage U max = Z max  is below the sparkover voltage U a of the insulator
 2 
string, the entire (full) wave spreads along the line.

o chopped impulse propagation

In the case where U max ≥ U a , as a first approximation, insulator sparkover occurs at the value
of U a , and a phase-earth fault occurs at 50 Hz due to the arc being maintained. The lightning
that is propagated is thus broken at the maximum value corresponding to U a .

The lightning current causing this flashover, for a given line, is called the critical current IC
such that:

U
IC = 2 a
Z

For lines, the order of magnitude of IC is:

- 5.5 kA at 225 kV, which corresponds to a probability of exceeding the magnitude according
to the IEEE method of 95 % (see figure 5-37)

- 8.5 kA at 400 kV, which corresponds to a probability of exceeding the magnitude according
to the IEEE method of 92 % (see figure 5-37).

In medium voltage, flashover is systematic in the case of a stroke of lightning occurring due to
the small distances in the air of the insulator string. This flashover of the insulator gives rise to
a phase-earth fault current, called a follow current, which is held at the power frequency of 50
Hz until it is cleared by the protections.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction , either partly of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 441

Ÿ form and magnitude of the lightning wave

The physical phenomenon of lightning corresponds to a source of impulse current the actual
form of which is highly variable: it consists of a front rising up to the maximum magnitude of
several miscroseconds to 20 µs followed by a decreasing tail of several tens of µs (see
figure 5-36).

Figure 5-36: oscillogram of a lightning current

The magnitude of strokes of lightning varies according to a log-normal distribution law. We can
thus determine the probability of a given magnitude being exceeded (see figure 5-37). We can
see, for example, that for the average curve (IEEE), the probability of exceeding a magnitude
of 100 kA is 5 %. This means that 95 % of lightning strokes have a magnitude less than
100 kA.

Figure 5-37: probability of exceeding positive and negative lightning stroke magnitudes,
according to IEEE (experimental statistic)

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly orpartly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 449

In extra high voltage (> 220 kV), back-flashover is unlikely (the flashover level of the insulators
is high), which is why it is useful to install protection ropes thus limiting the number of service
interruptions. But below 90 kV back-flashover occurs even if the value of the earth electrode
resistance is low (< 15 Ω); the usefulness of protection ropes is thus limited (more frequent
service interruptions).

o induced impulse

A stroke of lightning that falls anywhere on the ground behaves like an electromagnetic field
radiation source.

The steeper the rising front of the lightning current the greater the radiation.

For front steepnesses of 50 to 100 kA/µs, the effects of this field will be felt several hundreds
of metres, if not kilometres, away.

The magnetic field H at a point located at a distance of r from a circuit through which a
current I flows, is given in the relation:

I
H=
2π r

This field creates induced voltages in the neighbouring circuits which are able to reach
dangerous values both for equipment and persons.

Ÿ case of a loop

Let us consider the loop formed by the supply cable and the telecommunication link in figure
5-43, with a surface S and located 100 m from the lightning impact which has a current rising
front steepness of 80 kA/µs.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction , either partly of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 410

Numerical application:

Let us take a 63/5.5 kV substation where:

5.5
Vn = = 3175
. kV
3

I f (63 kV ) = 3 kA

Re = 0.5 Ω

L ω
Q= c =4
Rc

The rise in potential is: Ve = Re × I f = 1 500 V .

The phase-earth voltage in the cables is: VC = Q × Ve .

VC = 1500 × 4 = 6 000 V

VC = 1.89 Vn

The overvoltage in the cables is roughly twice the nominal phase-earth voltage.

It can be dangerous if the substation earth electrode is of poor quality. Indeed, for Re = 3 Ω
we will have VC = 11.3 Vn .

It is thus essential to limit the value of Re .

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 466

Numerical application:

Let us consider the example illustrated in figure 5-50:

Ul = 125 kV , case of an MV/LV transformer complying with IEC 76.3

Ursd = 75 kV , residual voltage of the surge arrester

r = 300 kV / µs , voltage wave rise front steepness

v = 300 m / µs , for an overhead line (speed of light)

125 − 75
we then have D ≤ × 300
2 × 300

D ≤ 25 m

The surge arrester must therefore be installed less than 25 m away from the transformer for
the overvoltage not to exceed the lightning impulse withstand value.

lightning impulse

transformer
D
A B
overhead line ZC ZC

surge arrester

Figure 5-50: protection distance of a surge arrester protecting a transformer fed

by an overhead line

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 444

By applying an electro-geometrical model to a vertical rod with a height (see fig. 5-40-a),
we can show that there are two distinct zones:

- zone 1 : this is located between the ground and the parabola which is the locus of the
equidistant points of and the ground; the instant the flash occurs, any leader
located in this zone will touch the ground since it is nearer to this than to

- zone 2 : this is located above the parabola; the instant the flash occurs, any leader located
in this zone will be picked up by point on the vertical rod as soon as the
distance between and the leader is less than the striking distance .

Figure 5-40-a: diagram of different protection zones offered by a vertical rod

For a lightning current with a value of , and thus a given striking distance, the distance x between
the point of impact on the ground and the point where the rod is fixed to the ground (called the rod
pick-up radius) is:

if

if

The rod pick-up radius is thus all the greater the more intensive the lightning stroke.

For very weak currents, the pick-up radius becomes less than the height of the rod which is then
able to pick up the current along its length. This has been experimentally proved.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly orpartly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 411

n ferro-resonance

o parallel ferro-resonance (see fig. 5-14)

Let us take a circuit made up of a parallel-connected capacitance, a coil with a saturable iron
core and a resistance. Let R be the resistance, C the capacitance and L the inductance
which varies with the current flowing through the coil and the voltage at the circuit terminals.

IT

IR IL IC

R L C V

Figure 5-14: parallel ferro-resonance

The total current IT flowing through the circuit is then given by the relation (1):

V
IT = + j (C ω V − I L ) (1)
R

We cannot express I L as a function of V , owing to the saturation.

The rms values are given by the relation (2):

V2
+ (C ω V − I L )
2
IT2 = 2
(2)
R

We can thus write relation (3) as follows:

V2 (3)
I T2 − 2
= C ω V − IL
R

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 409

The simplified MV network diagram is thus that shown in figure 5-13.

Re Rc Lc

VL

Ve Re I f 3C VC

Figure 5-13: simplified diagram

Let VL be the voltage at the inductance terminals.

Lc ω
We have VL =
( )
Ve
Re + Rc + j Lc ω − 3 C1ω

In the case of a Petersen coil earthed neutral, (resonance) tuning between the inductance and
1
the MV cable capacitance is aimed at as far as possible. We thus have : Lc ω ≈ and
3Cω
VC ≈ VL whence VL = Lc ω Ve .
Re + Rc

To minimise the rise in substation earth potential (Ve ), the resistance earth electrode must be
as weak as possible (of the order of 0.5 Ω).

We can thus neglect Re compared with Rc , which thus gives us:

L ω
VL = VC = c Ve = Q Ve
Rc

VC = Q Re I f

Q : coil quality factor

VC : is equal to the MV cable phase-earth overvoltage in this case

The coil quality factor must not therefore be too high in order to avoid the risk of a very high
overvoltage.

This is why, in some cases, a resistor must be connected in parallel with the coil, in order to
reduce the quality factor.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 421

5.1.2.3. Neutral conductor breakdown

Let us consider the diagram in figure 5-22 , where Z1 , Z 2 and Z 3 represent the equivalent
impedances per phase of all the loads downstream of the neutral breakdown point.

If the phases are perfectly balanced, the voltage system is not disturbed.

In the event of load unbalance, the neutral point is displaced and the phase-neutral voltages
move close to the phase-to-phase voltage for the least loaded phases, while for the loaded
phases (weak impedance), they drop below the single-phase voltage.

Z3
V3
Z2
V2
Z1
V1 N

neutral breakdown

Figure 5-22: equivalent diagram of an LV network during neutral breakdown

Using the superposition theorem, we can show that:

 Z1 // Z2   Z1 // Z 3   Z2 // Z3 
VN =   V3 +   V2 +   V1 (1)
 Z3 + Z1 // Z2   Z2 + Z1 // Z 3   Z1 + Z 2 // Z3 

The voltage applied to the terminals of a single-phase load on phase 3, for example, will be:

V3 N = V3 − V N

 1 3
If we know that V2 = a 2 V1 and V3 = a V1 ,  a = − + j 
 2 2 

then we can calculate V3 N , for example, for the following impedances:

Z1 = R
Z2 = 2 R
Z 3 = 10 R

(We have taken resistive loads to simplify the calculations.)

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 463

Ÿ more accurate method using temporary overvoltages

The simplified method has a drawback as it does not take into account the real requirements
Um
of the network which are generally lower than .
3

The temporary overvoltages likely to occur in a network are of two types:

- overvoltages due to a phase-earth fault the clearance time of which depends on the
protection system (see table 5.1 - the earth overvoltage factor is equal to 1.73 for unearthed
or impedance earthed networks)

- overvoltage due to load-shedding on the public distribution network, of the order of 15 %

but able to reach 35 % in some networks.

The temporary overvoltage value to be taken into account is the product of the earth fault
overvoltage and load shedding factors.

- specific case

If one of the temporary overvoltages lasts over 2 hours, it is considered to be a steady-state

condition for the surge arrester and thus U C is chosen to be equal to this overvoltage and
Ur = 1.25 × UC

- general case

A surge arrester's capacity to withstand temporary overvoltages is given in relation to an

equivalent voltage lasting 10 seconds (U10s ) expressed in the following equation:
η
T
U10s = TOV   where η ≅ 0.02
 10 

T : overvoltage duration
TOV : overvoltage value

This formula allows the 10 second overvoltage which would cause the same stress on the
surge arrester to be calculated for each temporary overvoltage.

The duration of the temporary overvoltage must be between several seconds and two to three
hours ( U10s = 0.97 × TOV for T = 2 s and U10 s = 114
. × TOV for T = 2 hours ).

The rated voltage of the surge arrester will be chosen to be above or equal to the maximum
value of the equivalent 10 second voltages: U r ≥ max (U10s ) .

Um
We will take UC ≥
3

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 427

Ÿ three-phase circuit

When the three-phase circuit in figure 5-27 is interrupted, the first phase which sees the
current crossing zero interrupts this current. There follows a transient current circulating in the
two uninterrupted phases. Thus, if phase 1 interrupts the current first a transient voltage is
obtained between points C1 , C2 and earth which is capable of reaching a value of 2 V$n for
an ideal circuit-breaker. For an actual circuit-breaker, the overvoltage coefficient is higher than
or equal to 2.

V$n : peak value of the phase-neutral nominal voltage

Note: the current crosses zero on the following phase after 1/3 of a period (7 ms at 50 Hz), while the
period of oscillations is roughly 1 ms.

Ls B1 C1 L1
A1 V1

Cs Cp
Lp

Ls B2 C2 L2
A2 V2 N

Cs Cp
Lp

Ls B3 C3 L3
A3 V3

Cs Cp
Lp

Figure 5-27: equivalent diagram of a three-phase circuit during interruption

Ÿ restrike phenomenon

The instant a circuit is interrupted, the voltage at the terminals of the circuit-breaker quickly
increases (roughly from 0.1 to 0.5 kV/µs). If the circuit-breaker poles separate shortly before
the current reaches zero (for an inductive circuit, this corresponds to the maximum voltage),
regeneration of the dielectric medium may not be sufficient to withstand the stress-voltage.
Indeed, in this case, the voltage is maximum and the poles are closer together.

Renewed breakdown then occurs accompanied by overvoltages with a peak to peak

magnitude of 2 V$n . This phenomenon is called restrike.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 431

Ÿ virtual chopping current - simultaneous interruption of the three phases

The transients generated by the first phase that creates overvoltages may cause, owing to the
capacitive coupling between the phases, oscillating currents inside circuits L p , C p, Cs of the
other phases.

It is thus possible to obtain zero current in these phases, immediately (several hundreds of a
microsecond) after interruption of the first phase.

If the circuit-breaker interrupts such currents, a chopping current phenomenon is then created
with very high chopping current and overvoltage values.

Ÿ chopping current and multiple restrike

Current chopping and multiple restrike are frequently linked.

Overvoltages caused by current chopping can themselves lead to restrike. They are almost
systematic in the case of the virtual chopping current.

o capacitive loads (see fig. 5-30)

Interruption of capacitive circuits, such as a capacitor bank or off-load cable, raises less
difficulties than the interruption of inductive circuits.

Indeed, the capacitances remain charged at the peak value of the 50 Hz wave after extinction
of the arc when the current reaches zero and the recurrence of voltage at the switchgear
terminals is accompanied by a 50 Hz wave.

Nevertheless, one half period after interruption, the device is subjected to a voltage equal to
twice the 50 Hz peak voltage 2 V$n . ( )
If the speed and dielectric withstand of the device are not sufficient to withstand this stress,
restrike may occur. It is followed by a voltage reversal at the terminals of the capacitances,
raising them to a phase-neutral voltage equal to 3 V$n maximum (if damping is neglected).

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 458

5.2.3. Surge arresters

To overcome the drawbacks of spark-gaps, different models of surge arresters have been
designed with the aim of ensuring better protection of installations and good continuity of
service.

Non-linear resistor type gapped surge arresters are especially found in HV and MV
installations which have been in operation for several years. The current tendency is to use
zinc oxide surge arresters which provide better performance.

n definitions

Surge arrester discharge current

The surge or impulse current which flows through the arrester after a sparkover of the series
gaps.

Surge arrester follow current

The current from the connected power source which flows through an arrester following the
passage of discharge current.

Surge arrester residual voltage

The voltage that appears between the terminals of an arrester during the passage of discharge
current.

5.2.3.1. Non-linear resistor type gapped surge arresters (see IEC 99-1)
n operating principle

In this type of surge arrester, a variable resistor (varistor), which limits the current after the
passage of the impulse wave, is associated with a spark gap.

After evacuation of the impulse wave to earth, the surge arrester is only subjected to the
network voltage and the follow current is limited by the varistor.

The arc is systematically extinguished after the 50 Hz wave of the single-phase-to-earth fault
current has reached zero.

Owing to the variation of the resistance, the residual voltage is maintained close to the
sparkover level. Indeed, this resistance decreases with the increase in current.

Various techniques have been used to make non-linear resistor type gapped arresters. The
most conventional method uses a silicon carbide (SiC) resistor.
Some surge arresters also have voltage grading systems (resistive or capacitive dividers) and
arc blowing systems (magnets or blow-out coils).

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 403

n consequences on equipment selection

The overvoltage factor and fault duration influence the choice of equipment insulation voltage
level.

o solidly or limiting impedance earthed neutral in MV, or TT and TN earthing system

in LV

Rapid clearance of the fault, and thus a short overvoltage time, means that the switchgear
phase-earth insulation level does not have to be higher than the nominal single-phase voltage.

o unearthed neutral in MV or IT earthing system in LV

Since the power supply does not have to be interrupted on occurrence of a first fault, the
overvoltage is likely to occur for a long period of time (several hours). It is therefore advisable
to choose switchgear with a phase-earth insulation level that is suitable for the nominal phase-
to-phase voltage.

Note: some manufacturers give a phase-earth insulation withstand equal to the single-phase voltage,
but stipulate that their switchgear can be implemented in an unearthed neutral network. There
are also switchgear standards that specify an insulation level compatible with use in an
unearthed neutral network.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 483

The overvoltage at a point on the network where the impedance changes is given in the
relation:

Z2
v2 = 2 v1 (see § 5.1.4)
Z1 + Z2

v1 = U rsd : corresponds in this case to the clipped overvoltage

v2 : overvoltage of the substation exposed conductive parts
Z1 = Zc : characteristic impedance of the medium voltage line
Z2 = Ze : substation earth electrode impedance

We thus have:
Ze
v2 = . 2 Ursd
Zc + Ze

The equipment lightning impulse voltage U tc must be above the overvoltage v2 , whence:
Ze
Utc ≥ . 2 Ursd
Zc + Z e

Zc
Ze ≤
( 2 Ursd
Utc −1 )
For U rsd = 120 kV
and Z c = 330 Ω , the impulse impedance Z e is equal to 1.5 times the
Z
resistance Re measured in low frequency: Re = e .
1.5

The condition on the value of the substation earth electrode impedance is thus:

Zc
Re ≤
( U
1.5 × Ursd − 1
tc
)
The maximum values of Re for the different earthing systems are given in table 5-9.

Earthing system TN − b , TT − b , IT − c TT − a , IT − b
U tc (kV) 4 8 20 3

Re 3.8 7.7 20.2 2.7

Table 5-9: maximum values of the MV/LV substation earth electrode resistances

recommended for limiting MV atmospheric overvoltages transferred in LV

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 396

we can determine:
R(1) = RT + RC
X(1) = XT + XC
R(0) = 3 Re + RT + RC
X(0 ) = X(0 )T + X(0)C

Note: A factor 3 appears before Re . The reason for this is explained in figure 4-11 of the Industrial
network protection guide.

R(0) 8
X (1)
7
k = 1.7
6

5 k = 1.6

3
k = 1.5
2

k = 1.4
1
k = 1.3
k = 1.2
X(0)
1 2 3 4 5 6 7 8
X (1)

X(0 ) R(0)
Figure 5-3: earth fault factor in relation to ratios and
X(1) X(1)
for R(1) = 0 and R f = 0

R(0) 8
X (1)
7 k = 1.7
k = 1.6
6 k = 1.5

4
k = 1.4
3

2 k = 1.3 k = 1.5
k = 1.2
1

X(0)
1 2 3 4 5 6 7 8
X (1)

X(0 ) R(0)
Figure 5-4: earth fault factor in relation to ratios and
X(1) X(1)
for R(1) = 0.5 X(1) and R f = 0
Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 390

5. OVERVOLTAGES AND INSULATION CO-ORDINATION

Different types of overvoltage may occur in industrial networks. Devices must therefore be
installed to reduce their magnitude and the insulation level of equipment must be chosen so
that fault risks are reduced to an acceptable level.

5.1. Overvoltages
An overvoltage is any voltage between one phase conductor and earth, or between phase
conductors having a peak value exceeding the corresponding peak of the highest voltage for
equipment, defined in standard IEC 71-1.

An overvoltage is said to be of differential mode if it occurs between phase conductors or

between different circuits. It is said to be of common mode if it occurs between one phase
conductor and the frame or earth.

n origin of overvoltages

Overvoltages can be of internal or external origin.

o internal origin

These overvoltages are caused by a given network element and only depend on the
characteristics and structure of the network itself.

For example, the overvoltage that occurs when a transformer's magnetizing current is
interrupted.

o external origin

These overvoltages are caused or transmitted by elements outside the network, for example:

- overvoltage caused by lightning

- spread of HV overvoltage through a transformer to the internal network of a factory.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 404

5.1.2.2. Resonance and ferro-resonance

resonance

The presence of inductive L , capacitive C and resistive R elements, connected, either in

series or in parallel, causes spreading of current and voltage having values which may be
dangerous for equipment.

series resonance

Figure 5-7 shows a series R , L, C circuit at the terminals of which a voltage U is applied.

I R L C

Figure 5-7: series R , L, C circuit fed by a voltage U

The voltage U is the vectorial sum of the voltages at the terminals of each element:

U = U R + U L + UC
1
= R I + j Lω I +
jCω

The vectorial diagram in figure 5-8 shows that for certain values of L and C , the voltages at
the terminals of the inductance and capacitance may be higher than the network voltage U :

jL I
1
I
jC

RI

Figure 5-8: vectorial diagram of a series R , L, C circuit fed by a voltage U

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 450

The induced voltage is given in the relation:

dφ dB dH
e=− = −S = − µ0 S
dt dt dt

µ 0 = 4 π × 10 −7 : magnetic permeability of the vacuum

dH 1 dI 1 103
now = = × 80 × −6 = 127 × 106 A / m / s
dt 2 π r dt 2 π × 100 10

whence e = 4 π 10 −7 × 120 × 127 × 106 = 19 kV

A phase-earth overvoltage of 19 kV thus occurs on the loop. This has a very short duration
( ≈ 1µs ) but can cause insulation breakdown.

To avoid this risk, the surfaces of the loops must be reduced.

lightning impulse
front steepness = 80 kA/µs telecommunication link computer

circuit
loop

surface = 120 m²
supply cable
magnetic field
100 m
printer

phase-earth insulation
subjected to 19 kV ( 1 µs)

earth

Figure 5-43: circuit loop

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction , either partly of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 392

- disturbance of control, monitoring and communication circuits by conduction or

electromagnetic radiation

- electrodynamic stress (destruction or deformation of equipment) and thermal stress

(elements melting, fire, explosion) essentially caused by lightning impulses

- hazard to man and animals following rises in potential and occurrence of step and touch
voltages.

5.1.2. Power frequency overvoltages

Power frequency overvoltages are generally caused by:

- an earth fault

- resonance or ferro-resonance

- neutral conductor breakdown

- a generator voltage regulator or transformer on-load tap changer fault

- overcompensation of reactive energy following a varmeter regulator fault

- load shedding, notably when the supply source is a generator

5.1.2.1. Overvoltage caused by an earth fault

Overvoltages caused by the occurrence of an earth fault greatly depend on the neutral
earthing system of the given network.

n unearthed (MV or LV) or impedance earthed (MV) neutral

Figure 5-1 shows that on occurrence of a solid earth fault, the voltage between the neutral
point and earth becomes equal to the single-phase voltage:

VNeutral = Vn

Vn : nominal single-phase voltage

For a fault on phase 1, VNeutral = − V1 .

The phase-earth voltage of healthy phases thus becomes equal to the phase-to-phase
voltage:

V2 E = VNeutral + V2 = V2 − V1
V3E = VNeutral + V3 = V3 − V1

whence V2 E = V3E = 3 Vn
Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 471

5.3. Insulation co-ordination in an industrial electrical network

5.3.1. General
Co-ordinating the insulation of an installation consists in determining the insulation
characteristics necessary for the various network elements, in view to obtaining a withstand
level that matches the normal voltages, as well as the different overvoltages.

Its ultimate purpose is to provide dependable and optimised energy distribution.

Optimal insulation co-ordination gives the best cost-effective ratio between the different
parameters depending on it:

- cost of equipment insulation

- cost of overvoltage protections

- cost of failures (loss of operation and destruction of equipment), taking into account their
probability of occurrence.

With the cost of overinsulating equipment being very high, the insulation cannot be rated to
withstand the stress of all the overvoltages studied in paragraph 5.1.

Overcoming the damaging effects of overvoltages supposes an initial approach which consists
in dealing with the phenomena that generate them, which is not always very easy. Indeed, if
using the appropriate arc interruption techniques the switchgear switching overvoltages can be
limited, it is impossible to prevent lightning strikes.

n clearance (see fig. 5-53)

This term covers two notions:

- gas clearance (air, SF6, etc.), which is the shortest path between two conductive parts.

- creepage distance: this is also the shortest path between two conductors, but following the
outer surface of a solid insulating material (e.g. insulator).

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 420

o example of series ferro-resonance (see fig. 5-21)

Figure 5-21 shows a solidly earthed network feeding a three-phase transformer having a delta-
connected primary. This can also apply to a star-connected transformer with an unearthed
neutral. If, when the switch is closed, one of the poles remains accidentally open or closes
late, for example the pole of phase 1, series ferro-resonance may occur in the circuit including:

- the magnetizing inductance of transformer windings AC or BC

- the capacitance of phase 1 in relation to earth.

Very high overvoltages can occur at the transformer terminals and between phase 1 and the
earth.

This type of ferro-resonance has frequently been encountered on HV networks with solidly
earthed neutral. It may also occur when a switch is opened. The means of protecting against
this type of ferro-resonance consists in inserting a resistor in the supply transformer neutral
point earthing. This solution does not however provide total protection since ferro-resonance
can, for example, occur in the circuit including the transformer AC winding and the
capacitances of phases 1 and 3 in relation to earth.

V3 switch
A

V2 Ph 3
L
L
V1 Ph 2 B L

Ph 1 If C

C C C

Figure 5-21: series ferro-resonance

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 419

In the case of VT with two secondaries, a resistor is installed in the open delta of one of the
two (see fig. 5-20).

It is recommended that a power above 50 W be dissipated in the resistor on occurrence of a

phase-earth fault.

100
For a secondary voltage of V , on occurrence of a solid earth fault, the voltage at the
3
resistor terminals is equal to 100 V; the resistance value is then determined:

R≤
(100) 2
50

R ≤ 200 Ω

VT VT VT

measurements r

Figure 5-20: protection against the risks of ferro-resonance via a resistor with two-secondary VT

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 445

Ÿ application to equipment protection using a lightning conductor

The lightning conductor diverts lightning to itself in order to protect equipment. Its principle is
based on the striking distance; tapered rods are placed at the top of equipment to be
protected, they are connected to the earth by the most direct path (the lightning conductors
surrounding the structure to be protected and interconnected with the earthing network).

The electrogeometric model allows the zone to be protected to be determined using the fictive
sphere method.

The point of impact of the lightning is determined by the object on the ground the closest to the
leader starting distance d . Everything happens as if the leader was surrounded by a fictive
sphere with a radius d moving with it. For good protection, the fictive sphere rolling on the
ground reaches the lightning conductor without touching the objects to be protected
(see fig. 5-40-b).

Protection against direct lightning strikes is approximately good in a cone the top of which is
the top of the lightning conductor and the half-angle at the top is 45 °.

leader

d = critical striking distance

protected
zone fictive
(cone) sphere

lightning
conductor
45°

Figure 5-40-b: determining the zone protected by a lightning conductor

using the "fictive" sphere method
Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction , either partly of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 443

Ÿ lightning density

On a world-wide scale, 63 billion discharges are recorded on average each year which
corresponds to 100 discharges per second. In France, this figure varies from 1.5 to 2 million
lightning strokes per year.

We then define the lightning density as being the number of days per year on which
thunder has been heard in a place.

In France, the average value of is 20 with a variation range going from 10 in channel
coastal regions up to over 30 in mountainous regions.

The value of may be much higher and reach 180 in tropical Africa or Indonesia.

Ÿ lightning strike density

The lightning strike density represents the number of lightning strikes per km2 per year,
whatever their current value levels.

In France, varies between 2 and 6 lightning strikes/km2/year depending on the region.

o lightning impact mechanism

The lightning impact mechanism begins with a leader from a cloud which approaches the
ground at a low speed. When the electric field is sufficient, sudden conduction is established
giving rise to the lightning discharge.

An experimental practical approach has enabled the relation linking the current of the
lightning strike to the distance between the starting point (leader position) and discharge point
(point of impact connected to the earth) to be found:

or according to the authors.

: striking distance, in m
: lightning current, in kA

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly orpartly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 465

5.2.3.3. Installation of HV and MV surge arresters

In HV and MV electrical networks, surge arresters are installed at the entrance to the
substation to ensure protection of the substation transformer and equipment. This protection
only works if the protection distance and the installation rules are respected.

n protection distance

The wave propagation phenomenon studied in § 5.1.5. shows that at the point of reflection
(e.g. MV/LV transformer), the overvoltage reaches double the value of the incident wave.

The surge arrester peaks at a sparkover voltage U rsd (equal to the residual voltage for ZnO
surge arresters).

If it is located a considerable distance away, the maximum voltage at the location of the
equipment to be protected will thus be 2 U rsd . Now, the equipment impulse withstand is
generally lower than 2 U rsd .

To overcome this drawback, the surge arrester is installed at a shorter distance away than the
"protection" distance D . The surge arrester then undergoes the sum of the incident wave and
the reflected wave. It is thus sparked for an incident wave below U rsd .

Assuming that at the equipment termination point, the wave is totally reflected, we can show
D
that the overvoltage in relation to the equipment is limited to U = U rsd + 2 r
v

dV
r= : rise front steepness of the voltage wave, kV/µs
dt
v : wave propagation speed, m/µs

For a lightning impulse withstand voltage Ul , the surge arrester must therefore be located at
a distance D such that:

D
Ursd + 2 r ≤ Ul
v

U − U rsd
whence D≤ l ⋅v
2r

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 440

5.1.4. Atmospheric overvoltages

n general

The earth and the electrosphere, the conductive area in the atmosphere (about 50 to 100 km
thick), constitute a natural spherical capacitor which is charged by ionization, thus producing
an electric field directed towards the ground of roughly several hundred volts/metre

Since air is not very conductive, there is thus an associated permanent conduction current, of
roughly 1 500 A for the entire earth's globe. Electrical balance is ensured during discharges by
rain and strokes of lightning.

The formation of storm clouds, masses of water in the form of aerosols, is accompanied by
charge separation electrostatic phenomena: the positively charged light particles are driven by
the rising air currents, and the negatively charged heavy particles fall because of their weight.
At the base of the cloud, there may also be islets of positive charges where heavy rains are
located.

On an overall macroscopic scale, a dipole is created.

When the breakdown withstand limit gradient is reached, a discharge is produced inside the
cloud or between clouds or between the cloud and the ground. In the latter case, it is referred
to as lightning.

The cloud-ground electric field can reach 15 to 20 kV/metre on flat ground. But the presence of
obstacles deforms and locally increases this field by a factor of 10 to 100 or even 1 000
depending on the form of the obstacles (also called the "peak effect"). The atmospheric air
ionizing threshold is thus reached, i.e. roughly 30 kV/cm, and corona effect discharges are
produced. When these discharges are located on fairly high objects (tower, chimney, pylon)
they may divert lightning to this objects.

o classification and characteristics of strokes of lightning

Strokes of lightning are classed according to the origin of the discharge (or leader) and their
polarity.

Depending on the leader origin, the stroke of lightning may be:

- either descending from the clouds to the ground in the case of fairly flat land
- or ascending from the ground to the clouds in the case of mountainous land.

Depending on the polarity the following distinctions between lightning strokes are made:

- negative when the negative part of the cloud is discharged, which represents 80 % of cases
in temperate countries
- positive when the positive part of the cloud is discharged.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly orpartly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 435

The loads affected by these phenomena are off-load transformers, neutral point coils (neutral
reactance earthing) and motors during the starting period for inductive circuits as well as
capacitor banks for capacitive circuits.

Transformers undergo impulse wave dielectric tests; because of this, they are better built than
motors to be able to withstand the transients caused by restrike (see IEC 76-3).

The case of motors is different. At each start, they must withstand the transients caused by
prestrike. Moreover, even if interruption during the starting period does not occur very often, it
is nevertheless a possibility and they are then subjected to multiple restrike.

Motors are thus especially sensitive to multiple prestrike, because of its high rate of
occurrence, as well as to multiple restrike, due to the magnitude of the overvoltages produced.
These overvoltages cause deterioration of the insulation of the first turns.

In order to limit overvoltages, Zn0 type surge arresters can be connected in parallel with the
load.

But the best method consists in using switching devices suitable for the type of application.

Table 5-3 gives the behaviour of medium voltage switchgear with respect to the phenomena
relating to the switching overvoltages studied.

Switchgear Multiple Current Multiple Overall behaviour

prestrike on chopping restrike
closing
Puffer-type SF6 circuit- no weak no No problem. Below 300
breaker kW, use a rotating arc
SF6 circuit-breaker.
Rotating arc SF6 circuit- no no no No problem.
breaker and contactor
Vacuum circuit-breaker yes yes yes Use surge arresters

Vacuum contactor yes weak yes Use surge arresters

Magnetic blast circuit- no no no No problem.

breaker and contactor
Minimum oil circuit-breaker no yes yes Use surge arresters

Table 5-3: behaviour of medium voltage switchgear

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 414

o series ferro-resonance (see fig. 5-16)

Let us take a series circuit made up of a resistance, a coil with a saturable iron core and a
capacitance. We have:

 I 
V = R I + j  VL −  (1)
 Cω 

We cannot express VL as a function of I , owing to the saturation.

If we move to rms values, we can write:

2
 I 
V 2 = R 2 I 2 + VL −  (2)
 Cω 

2
 I 
or: V 2 − R2 I 2 = VL −  (3)
 Cω 

I
V 2 − R2 I 2 = Lω I − (4)
Cω

R L C
I

VR VL VC

Figure 5-16: series ferro-resonance

As for the parallel circuit, this equation can be graphically resolved as a function of I ,
by plotting curves (see fig. 5-17):

v = V 2 − R2 I 2

I
and v = VL −
Cω

Curve VL ( I ) presents a very small slope when I increases owing to the saturation of the
coil's iron core VL ( I ) = L ( I ) ω V .

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 467

5.2.4. protection of LV installations

n general

LV installations are protected against overvoltages by installing devices in parallel; 3 types of

devices are used:

- overvoltage limiters located on the secondary of MV/LV transformers (only in an IT

earthing system); they only provide protection against power frequency overvoltages

- low voltage surge arresters installed in LV switchboards or incorporated in loads

- surge diverters designed to protect telephone networks, LV terminal boxes and loads.

The main technologies used are:

- zener diodes

- gas discharge tubes

- zinc oxide varistors.

Zener diodes have the drawback of only ensuring the protection of a precise point in the
network. The gas discharge tube requires the addition of a varistor to prevent follow current.
Variable resistor-type surge arresters are currently the most cost-effective solution owing to
their simplicity and reliability

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 416

o example of parallel ferro-resonance - unearthed neutral three-phase network

(see fig. 5-18)

Let us consider a three-phase network with unearthed neutral having a capacitance C

between each phase and the earth. Furthermore, a voltage transformer, with a similar
magnetizing inductance to a saturable core reactor, is connected between each phase and
earth. A parallel inductance-capacitance circuit thus appears between each phase and earth.
Parallel ferro-resonance can then be sparked between the capacitance and voltage
transformer of the same phase.

This ferro-resonance may occur following a transient overcurrent or overvoltage caused by a

switching operation and notably when the network is energized. Owing to the existing phase
displacements between the voltages of the three network conductors, the overcurrents and
switching overvoltages do not have the same magnitude in the three phases. Ferro-resonance
can thus very easily occur on only two phases, phases 2 and 3 for example. The voltages of
these two phases in relation to earth correspond to points located on portion AB of curve (b)
(see fig. 5-15). The voltage of phase 1 corresponds to a point located on the OS part of this
curve.

For phases 2 and 3, the capacitance-inductance assembly behaves like an inductance, and for
phase 1, like a capacitance. If we plot the voltage vector diagram, we can see:

- that the phase 1 voltage in relation to earth is weak

- that the voltages in relation to earth of the other two phases are very high

- that there is a very high potential difference between the neutral point and earth
(see fig. 5-18).

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 451

n impulse wave transference in a transformer (see IEC 71-2 - appendix A)

In lightning impulse conditions, the transformer behaves like a capacitive divider with a ratio of
s ≤ 0.4 . It is equivalent to a capacitance Ct (see figure 5-44-a).

equivalent Ct
lightning wave

U0 sU1
U1 U0 sU1

U1 : impulse voltage on the high voltage terminal

U 0 : no-load voltage transferred

Figure 5-44-a: impulse wave transference in a transformer

U 0 represents the no-load overvoltage, i.e. when the secondary outgoing terminals are not
connected to any cables or lines. This overvoltage is generally not acceptable by the
transformer.

In reality, the transformer is connected to a network with a capacitiance Cs . This plays the
role of a voltage divider with the transformer capacitance Ct (see fig. 5-44-b).

Ct

U 0 = sU1 Cs U2

U 2 : voltage transferred to the secondary with a network

Figure 5-44-b: transformer with its equivalent network

The voltage transferred to the secondary is thus:

Ct
U2 = s U1
Ct + Cs

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction , either partly of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 417

These overvoltages will cause a breakdown in equipment insulation if provisions to limit them
are not taken.

V3
Ph 3
V2
Ph 2
V1
Ph 1
N

v1T C C C
L L L

V1
vN
v3T
v2T
N
V3 V2

Ferro-resonance occuring between two phases

Figure 5-18: parallel ferro-resonance in an unearthed neutral network

Ÿ protection against the risks of parallel ferro-resonance

A voltage transformer (VT ) charged by a resistor r behaves like a saturable (magnetizing)

inductor in parallel with this resistor.

Thus, in an unearthed network, if a charging resistor is connected to the secondary of the

voltage transformers, the L-C parallel circuits, made up of these transformers and network
cable capacitances, are transformed into R-L-C parallel circuits, such that if the resistors are
correctly sized, the risk of ferro-resonance outlined previously can be avoided (ellipse (a)
remains inside the zone 0SA - see fig. 5-15):

- the resistors must be sufficiently weak to be efficient

- they must not be too weak, so that the VT are not overcharged and their accuracy is
maintained.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 413

With the circuit being initially unused, the total current IT is zero, as well as the voltage V , and
ellipse (a) is reduced to point O . If the current increases, the length of the axes of ellipse (a)
increases and the voltage rises, the operating point M moves along branch OS of curve (b).
When the total current exceeds the value I T' for which ellipse (a') cuts curve (b) at S , the
operating point suddenly jumps from point M to point T located on branch AB of curve (b), it
then moves along this branch. The voltage thus suddenly increases, going from VS to VT , and
then it continues to increase if the current IT increases.

If the total current now decreases, the operating point moves along branch AB and stays
there, even if the current drops below the value I T' corresponding to ellipse (a'). When the
current reaches the value IT , the operating point is P instead of M . It only returns to
branch OS if the current drops below the value I T'' corresponding to ellipse (a") passing
through point A . When this occurs, the operating point suddenly jumps from A to Q , and
the voltage from V A to VQ .

We can thus see that two stable operating conditions, for which the voltage at the circuit
terminals takes very different values, for example V M and VP , can correspond to the same
rms current value IT .

Finally, if the initial operating conditions correspond to a weak voltage (branch OS ), with a
resulting capacitive current, it is possible that, following a sudden change in operating
conditions leading to a transient phenomenon (overcurrent or overvoltage), the resulting
current becomes inductive and the voltage maintains a high value, even once the disturbance
has disappeared.

Ferro-resonance can be avoided if the resistance R is sufficiently weak for ellipse (a) to
remain within zone OSA , even when there is a high overcurrent.

I IL inductive operating conditions

IC C V
(b)
resonance C V IL

capacitive operating B
conditions
X (a''')
IT'''

' S T (a')
IT
IT P (a)
M N
IT''
Q (a'') A
O VQ VM VS V N VA V P VT V

Figure 5-15: parallel ferro-resonance - graphic resolution

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 434

If the bank is energized shortly afterwards, a breakdown due to the application of a voltage of
opposite polarity may give rise to an overvoltage of 3 V$n .

L CB

U C

Figure 5-31: closing operation of a capacitive circuit

To ensure the safety of persons, the capacitor banks are fitted with a discharging resistor
having a time constant allowing 75 V to be reached after 3 minutes in LV and 10 minutes in
HV.

n Means of protecting loads

The phenomena created by de-energizing (or energizing) loads, which we have studied, lead
to transient overvoltages which may be dangerous for both loads and other network elements.

Table 5-2 gives the level of overvoltages and their characteristics for each phenomenon
studied.

Occurrence of Number of Overvoltage dU/dt order of Remark

phenomenon overvoltage value magnitude
peaks

Chopping at every 1 2 to 4 V$n 0.1 kV/µs favours restrike

current interruption
Multiple interruption with 0 to 20 2 to 7 V$n 10 kV/µs
restrike separation
close to zero
current
Prestrike at every closing 1 to 50 2.5 V$n 10 kV/µs

V$n : phase-neutral voltage peak value

Table 5-2: different types of overvoltage

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 452

The values of Ct are generally between 1 and 10 nF. The capacitance of a cable is close to
0.4 nF/m. Thus, several tens of metres of cable will greatly reduce the overvoltage transferred
to the secondary.

In general, the network is sufficiently widespread for the overvoltage transferred not to raise
any difficulties.

However, in the case of a short cable, e.g. between a specific transformer and a load (arc
furnace, etc.), the overvoltage transferred may be unacceptable for the equipment on the low
voltage side.

To reduce the magnitude of the impulse transferred, it is possible to:

- use a surge arrester with a lower sparkover voltage on the high voltage side

- install a surge arrester on the low voltage side between each phase and earth

- increase the capacitance between each phase and earth on the low voltage side.

5.1.5. Propagation of overvoltages

Overhead lines and cables constitute a propagation media for any overvoltage wave likely to
occur on a network.

For high frequencies (case of switching and lightning overvoltages), the line is characterised by
its so-called "characteristic" or "wave" impedance:

L
Zc ≈
C

L : line inductance
C : line capacitance

We can see that this impedance is independent of the length of the line.

The speed of the wave propagation on an overhead line is close to the speed of light:

c = 3 × 108 m/s (300 m/µs)

c
for cables, it is equal to v =
εr

εr : relative permittivity of the cable insulating material

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction , either partly of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 407

example: resonance in a Petersen coil earthed HV/MV substation

Figure 5-10 shows the diagram of a Petersen coil earthed HV/MV substation when an HV
earth fault flows through the common earth electrode.

HV HV / MV transformer Z MV

If C
Lc , Rc
Re1

Ve Re

If : HV earth fault current

Lc , Rc : Petersen coil inductance and resistance
Re , Re1 : earth electrode resistances
C : MV cable phase-earth capacitance
Ve : rise in substation earth potential
Z MV : sum of MV cable and transformer impedances

Figure 5-10: HV earth fault in an HV/MV substation with a Petersen coil earthed neutral

The symmetrical component method gives us the fault current value as (see § 4.2.2 of the
Network protection guide):

3 Vn
If =
Z(1) + Z( 2 ) + Z( 0 )

where Z(1) = ZT + Z l

Z( 2 ) = Z T + Z l

(0 Z( )T + Z 0)l + e Re1

ZT , Z( 0)T : HV transformer positive-sequence (or negative-sequence) and zero-sequence impedances

Z l , Z(0 )l : HV line positive-sequence (or negative-sequence) and zero-sequence impedances

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 464

o nominal discharge current I nd

In practice, for the voltage range 1 kV ≤ U m ≤ 52 kV , two values of I nd are available: 5 kA and
10 kA.

The value I nd = 10 kA is chosen for areas with a high lightning density.

o discharge class and energy capacity

These are determined by testing or comparison with identical projects.

o mechanical characteristics

The IEC 99-4 and 99-5 standards fix the allowable pressure limit (expressed in "kA") which
must be met for the three-phase short circuit at the surge arrester terminals.

The surge arrester characteristics will also be checked in relation to:

- the ambient temperature

- the altitude

- the level of pollution

- the mechanical resistance to the wind, seismic stress, frost.

o surge arrester protection level

The protection level of the surge arrester at the installation point corresponds to the residual
voltage (U rsd ) at its terminals when its nominal discharge current flows through it.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 468

n LV surge arrester installation rules

The equipment is only protected properly if certain installation rules are followed:

- rule 1
The length of the connection between the surge arrester and its disconnecting circuit-breaker
must be below 0.5 m.

disconnecting
circuit-breaker

L < 50 cm load to be
protected

Figure 5-51: diagram of connections

- rule 2
The outgoing feeders of the protected conductors must be connected to the terminals of the
surge arrester and its disconnecting circuit-breaker.

- rule 3
The loop surfaces must be reduced by tightly grouping together the incoming, phase, neutral
and PE wires.

- rule 4
the incoming wires of the surge arrester (polluted) must be moved away from the protected
outgoing wires (healthy) in order to avoid any possible electromagnetic coupling.

- rule 5
The cables must be flattened against the metal structures of the box in order to reduce frame
loops and thus benefit from a reduction in disturbances.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 469

n connection layout according to the earthing system

In figures 5-52-a and 5-52-b the connection layouts of the LV surge arrester are shown for
different earthing systems.

electrical switchboard

disconnecting
RCD circuit-breaker
equipment
to be protected

surge arrester
PE
PE

Ph1
Ph2
Ph3
N

LV neutral
main earth load earth
earth electrode (entrenched
terminal electrode
loop)

TT earthing system

electrical switchboard

disconnecting
circuit-breaker
equipment
to be protected

surge
arrester
PE
PE

Ph1
Ph2
Ph3
N

PIM overvoltage
limiter
main earth (entrenched load
LV neutral terminal earth
earth electrode loop)
electrode

IT earthing system

Figure 5-52-a: connection layout of an LV surge arrester for TT and IT earthing systems

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 455

5.2. Overvoltage protection devices

5.2.1. Principle of protection

The protection of installations and persons against overvoltages is greatly improved when
disturbances flow to earth, and this is done as close as possible to the sources of disturbance.
This requires low impedance earth electrodes to be implemented.

Thus, three overvoltage protection levels can be distinguished:

n 1st protection level

The objective is to avoid a direct impact on structures by catching the lightning and directing it
towards designated flow points, via:

- lightning conductors, whose principle is based on the striking distance; a rod placed at the
top of a structure to be protected captures the lightning and evacuates it through the
earthing network (see fig. 5-40-b)

- meshed or Faraday cages

- lightning protection ropes (see fig. 5-42).

n 2nd protection level

Its aim is to ensure that the basic impulse level (BIL) of the substation components has not
been exceeded.

In HV, this type of protection is established using elements ensuring that the lightning wave
flows to earth, such as:

- spark-gaps
- HV surge arresters.

n 3rd protection level

Used in LV as an extra protection for sensitive equipment (computers, telecommunication

devices, etc.).

It uses:

- series filters
- overvoltage limiters
- LV surge arresters.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 479

U1 U2

MV ph 1 LV

ph 2

ph 3
PEN

U1 Re Im V
Uf
Im Re RB U2 V
Uf 0

TN − b

U1 U2

MV ph 1 LV

ph 2

ph 3
N

U1 Re Im V
Im
Uf
Re RB RA
U2 V
Uf 0

TT − b

U1 U2

MV LV

U1 Re I m V 3 *
U2 V 3
Im Re Z If RA U f
Uf RA If UL

(*) a first LV fault is present

IT − c

Figure 5-55: rise in potential in TN − b , TT − b and IT − c earthing systems

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 429

Ÿ chopping current (weak inductive currents)

When weak currents, notably lower than the nominal current of the circuit-breaker, are
interrupted, the arc which occurs occupies a small volume. It consequently undergoes
considerable cooling linked to the circuit-breaker's capacity to interrupt much higher currents.

Owing to this fact, the arc becomes unstable and its voltage may present relatively large
variations, while its absolute value remains lower than the network voltage (case of SF6 or
vacuum). These voltage variations may generate high frequency oscillating currents, with a
magnitude that may reach 10% of the current at 50 Hz, in the nearby capacitances ( Cs , L p , C p
circuit in figure 5-25). Superposing these high frequency currents on the current at 50 Hz
results in multiple crossings of the current through zero around zero of the fundamental wave
(see fig. 5-29).

The circuit-breaker interrupts the current the first time it crosses zero while the load current
(only the current at 50Hz) is not zero. The value of this current represents what we call the
chopping current Ichop . ( )
current in the
circuit-breaker

I chop
"chopping"
current

extinction
possible

50 Hz wave

Figure 5-29: superposition of a high frequency oscillating current

on a power frequency current

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 399

n TN earthing system

The current of an earth fault circulates in the protective conductor. The neutral earth electrode
resistance is thus not used to determine the zero-sequence impedance
(see fig. 5-5).

ZT ZC
V3
ZT ZC
V2
V3 VM
ZT ZC
V1
V2 VM

Z PE VM
Re

V1 , V2 , V3 : single-phase voltages
ZT : transformer impedance
ZC : cable impedance
Z PE : protective conductor impedance
VM : potential of exposed conductive parts (masses) in relation to earth
Re : neutral earth electrode resistance

Figure 5-5: equivalent diagram of an earth fault in a TN earthing system

We are interested in the overvoltage of the healthy phases in relation to the exposed
conductive part, which determines whether or not an insulation fault may occur on the other
V − VM V − VM
load: k M = 2 = 3 .
Vn Vn

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 391

n classification of overvoltages

Standard IEC 71-1 gives the classification of overvoltages according to their duration and form.

According to the duration, a distinction is made between temporary overvoltages and transient
overvoltages:

- temporary overvoltage: power frequency overvoltages of relatively long duration (from

several periods to several seconds).

- transient overvoltage: short-duration overvoltage lasting only several milliseconds, which

may be oscillatory and is generally highly damped.

Transient overvoltages are divided into:

. slow-front overvoltage
. fast-front overvoltage
. very-fast-front overvoltage.

n standard voltage forms

Standard IEC 71-1 gives the standardised wave forms used to carry out tests on equipment:

- short-duration power frequency voltage: this is a sinusoidal voltage with a frequency

between 48 Hz and 62 Hz and a duration equal to 60 s.

- switching impulse: this is an impulse voltage having a time to peak of 250 µs and a time to
half-value of 2500 µs.

- lightning impulse: this is an impulse voltage having a front time of 1.2 µs and a time to
half-value of 50 µs.

n consequences of overvoltages

Overvoltages in electrical networks cause equipment degradation, a drop in service continuity

and are a hazard to the safety of persons.

The consequences can be very varied depending on the type of overvoltages, their magnitude
and their duration. They are summed up as follows:

- breakdown in the insulating dielectric of equipment in the case where the overvoltage
exceeds the specified withstand

- degradation of equipment through ageing, caused by non-destructive but repetitive

overvoltages

- loss of power supply caused by the destruction of network elements

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 394

n solidly earthed neutral (HV or MV)

On occurrence of an earth fault on one network phase, a high current is generated which
circulates in the circuit formed by the fault phase, earth and neutral earth electrode
(see fig. 5-2).

At the fault point, the three-phase voltage system is disturbed. The fault phase voltage in
relation to earth is almost zero if we neglect the fault resistance. The voltages of the other two
phases in relation to earth are higher than the single-phase voltage, while remaining lower
than the phase-to-phase voltage.

V3
ZT ZC
V2
ZT ZC
V1
ZT ZC
fault
V1E V2 E V3E
Re Rf

V1 , V2 , V3 : single-phase voltages
ZT : transformer impedance
ZC : cable impedance
Re : neutral earth electrode resistance
Rf : fault resistance

Figure 5-2: equivalent diagram of a phase-earth fault when the neutral is solidly earthed

Thus, we can define an earth fault factor k characterising the phase-earth overvoltage
occurring on the healthy phases:

V2 E = V3E = k Vn

Vn : nominal single-phase voltage

The symmetrical component calculation method (see § 4.2.2. of the Protection guide) can be
used to determine the value of k in relation to the positive, negative and zero-sequence
impedances:

Z(1) + a 2 Z( 2 ) + a Z( 0 )
k = 1−
Z(1) + Z( 2) + Z( 0 ) + 3 R f
Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 457

n advantages

The main advantages of spark-gaps:

- their low price

- their simplicity
- the possibility of setting the sparkover voltage.

n drawbacks

- The sparkover characteristics of the spark-gap are highly variable (up to 40 %) depending
on the atmospheric conditions (temperature, humidity, pressure) which modify the ionization
of the dielectric medium (air) between the electrodes.

- the sparkover level depends on the overvoltage.

- spark-gap sparkover causes a power frequency phase-to-earth short circuit owing to the arc
being maintained. The short circuit lasts until it is cleared by the switching devices (this
short circuit is called a follow current). This means that it is necessary to install shunt circuit-
breakers or rapid reclosing system on the circuit-breaker located upstream. Because of this,
the spark-gaps are unsuitable for the protection of an installation against switching
overvoltages.

- the sparkover caused by a steep front overvoltage is not instantaneous. Due to this delay,
the voltage actually reached in the network is higher than the chosen protection level. To
take this phenomenon into account, it is necessary to study the voltage-time curves of the
spark-gap.

- sparkover causes the appearance of a steep front broken wave which could damage the
windings of the transformers or motors located nearby.

Although still used in certain public networks, spark-gaps are currently being replaced by surge
arresters.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 478

n TN − b , TT − b and IT − c earthing systems (see fig. 5-55)

In these three systems, we can see a rise in potential of the exposed conductive parts of the
substation U1 such that:

U1 = Re I m + V for TN − b and TT − b earthing systems

U1 = Re Im + V . 3 for IT − c earthing systems with the presence

of a first fault on the LV side

Depending on the maximum current value Im , the values of Re must be limited so that U1
remains below the power frequency withstand voltage U tp of the substation equipment.

U1 ≤ U tp

Table 5-6 gives the maximum values of Re for different values of I m and U tp .

Values at Re not to be exceeded

Fault current Im U tp = 2 000 V U tp = 4 000 V U tp = 10 000 V

(A) Class I Class II Special class

TN − b ; TT − b IT − c TN − b ; TT − b ; IT − c TN − b ; TT − b ; IT − c

300 A 5.9 Ω 5.3 Ω 12 Ω 30 Ω

1 000 A 1.8 Ω 1.6 Ω 3.6 Ω 10 Ω

5 000 A 0.35 Ω 0.32 Ω 0.72 Ω 2Ω

Table 5-6: maximum values of Re in TN − b , TT − b and IT − c earthing systems

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 393

V3
3
V2
2

V1
1

earth
VNeutral ZNeutral fault V1E V2E V3E

V2 V3
Neutral

V2E V3E
V1

V1E 0

V1 , V2 , V3 : phase-neutral voltages
V1E , V2 E , V3E : phase-earth voltages
Z Neutral : earthing impedance ( Z Neutral = ∞ for an unearthed neutral)

Figure 5-1: overvoltage on an unearthed or impedance earthed network

on occurrence of a phase-to-earth fault

Note 1 : for an impedance earthed neutral, the value of Z Neutral is much greater than the value of the
transformer and cable impedances and the fault resistance, which is why VNeutral = − V1 .

Note 2 : in overhead public distribution networks, there are highly resistive faults (several kΩ), having a
value close to or higher than the earthing impedance. In this case, a highly resistive fault will
cause an overvoltage lower than 3 Vn .

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 422

By applying formula (1), we find:

 4 a2 + 9 
VN =   V1
 16 

we then have V3 N = V3 − VN = a V1 − VN

 − 15 + j 10 3 
V3 N =   V1
 16 

whence V3 N = 1.43 Vn

Similarly, we can determine: V2 N = 114

. Vn and V1N = 0.6 Vn

Vn : nominal single-phase voltage

We can see that once the most sensitive single-phase loads have broken down, there are
successive breakdowns, following the development of the phenomenon which worsens the
unbalance ( Z 3 increases after the breakdowns and consequently V3 N increases); this is an
avalanche phenomenon.

This risk thus underlines that it is preferable to well balance the loads on the three phases.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 461

n characteristics

ZnO surge arresters are characterised by:

- the steady-state voltage which is the permitted specified value of the power frequency rms
voltage that can be continuously applied between the terminals of the surge arrester

- the rated voltage which is the maximum power frequency rms voltage permitted between its
terminals for which the surge arrester is designed to operate correctly in the temporary
overvoltage conditions defined in the operating tests (a power frequency overvoltage of 10
seconds is applied to the surge arrester - see IEC 99-4)

- the protection level defined at random as being the residual voltage of the surge arrester
when it is subjected to a given current impulse (5,10 or 20 kA according to the class), with a
wave form of 8/20 µs

- steep front current impulse (1 µs), lightning impulse (8/20 µs), long duration impulse, and
switching impulse withstand

- nominal discharge current.

Table 5-4 gives an example of the characteristics of a phase-to-earth ZnO surge arrester for
a 20 kV public distribution network (with tripping on occurrence of the first fault).

Maximum steady-state voltage (phase-earth) 12.7 kV

Rated voltage 24 kV

Residual voltage for nominal discharge current < 75 kV

Nominal discharge current (8/20 µs wave) 5 kA

Impulse current withstand (4/10 µs wave) 65 kA

Table 5-4: example of the characteristics of a ZnO surge arrester for a 20 kV network

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 433

n energizing a load

o inductive circuit

When a device closes, on an inductive circuit (no-load transformer, motor on starting), there is
a moment when the dielectric withstand between its contacts drops below the applied voltage.

A breakdown occurs causing sudden zero voltage at the device terminals.

This is accompanied by oscillations with stray capacitances which cause high frequency
currents to circulate in the circuit-breaker.

Depending on the speed of the device, prestrikes may or may not occur up to complete closing
of the poles.

Multiple prestrike is accompanied by successive overvoltages which decrease until the device
is completely closed.

The phenomenon is highly complex and involves several parameters:

- the characteristics of the switching device

- the characteristic impedance of the connections
- the natural frequencies of the load circuit

which means that a mathematical simulation model is required to pre-determine the

overvoltage values.

o capacitive circuit (capacitor bank)

When a capacitor bank is energized via a slow operating device, prestrike occurs between the
contacts close to the wave peak of 50 Hz.

A damped oscillation in the system in figure 5-31 then occurs at a frequency above
LC
50 Hz concentrated around the peak. In this case the maximum overvoltage is 2 V$n . It
corresponds to the maximum overvoltage admissible by the capacitors (see IEC 831-1 for LV
and 871-1 for MV or HV).

With a faster device, prestrike does not necessarily occur around the 50 Hz peak and
consequently the overvoltage is smaller.

When put out of service, the bank remains charged at a voltage going from 0 to the peak
voltage of the network.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 430

The current is then interrupted as in the case in figure 5-25 except for the peak-to-peak
1 
magnitude of the oscillations, due to the presence of energy stored in L  L I a2  which is
2 
1 
added to that in the capacitance C p  C p V$n2  .
2 
If V$ is half the peak-to-peak maximum value of the oscillation at point C, we can write:
c max

1 1 1
C p V$c2max = C p V$n2 + L I a2
2 2 2

L 2
V$c max = V$n2 + I a in single-phase.
Cp
V$n : phase-neutral nominal voltage peak value

For a three-phase circuit V$n must be added in order to take into account the transient
operating conditions linked to the non-simultaneous interruption of the phases, whence:

L 2
V$c max = V$n + V$n2 + Ia
Cp

This phenomenon is notably problematic in the case of an arc furnace transformer power
supply.

Indeed, the transformer is generally connected not very far away from the busbar. Thus the
value of C p is very weak and therefore the value of V$c max high.

We can determine V$c max by taking:

L : transformer leakage inductance

C p : capacitance of the cable linking the circuit-breaker to the transformer
I a : transformer magnetizing current

Schneider carried out an analysis for a single-phase arc furnace transformer where:

15000 V
Vn = ; L = 8.26 H ; C p = 14.75 nF ; Ia = 4.36 A
3

we find V$c max = 8.5 V$n

Installing an R , C circuit in parallel with the circuit-breaker allowed the overvoltage to be

reduced to 2 V$ . n

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 426

The phenomena observed are illustrated by curves in figure 5-26.

VA

VB

VC

Is

ID

IL

VD = VB − VC

t
t0 t1

t0 : separation of contacts
t1 : zero current

Figure 5-26: interruption cycle of an ideal device

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 402

The exposed conductive part of load 2 is at the same potential as the remote earth.

The voltage of one healthy phase of this load in relation to the exposed conductive part is
therefore V2 − VNeutral or V3 − VNeutral :

Re V1 Re  a Re 
V2 − VNeutral = V2 − Re I f = V2 − = V2 − a V2 = V2 1 − 
Re + RM Re + RM  Re + RM 

a Re
whence kM = 1 −
Re + RM

for RM = Re , k M = 1.32

for RM > Re , k M < 1.32

The earth electrode resistance of a group of loads is in general higher than the substation
earth electrode resistance. The overvoltage coefficient will thus be lower than 1.32 on load 2.

The overvoltage factor is maximum in the TT earthing system for a load having an exposed
conductive part connected to the same earth electrode as the fault load, we thus have
kM = 3

n recapitulative table of maximum earth fault overvoltages in relation to the neutral

earthing system

Medium and high voltage (1) Low voltage (2)

solidly earthed neutral unearthed or TN system TT system IT system

(HV or MV) impedance
earthed neutral
(MV)

< 1.73 * 1.73 1.45 1.73 1.73

(generally 1.2 to 1.4)

(1) : phase-earth overvoltage

(2) : phase-exposed-conductive-part overvoltage
(*) : a network with a solidly earthed neutral is generally made up so as to limit overvoltages to values close to 1.2
to 1.4.

Table 5-1: maximum overvoltage factor in relation to neutral earthing system

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 454

Upstream of M , we have:

v1 = Z1 i1 and v1' = − Z1 i1' (1)

immediately downstream of M :
v 2 = Z 2 i2 (2)

at point M :

v2 = v1 + v1' and i2 = i1 + i1' (3)

We can thus deduce:

v2 = v1 + v1' = v1 − Z1 i1' = v1 − Z1 ( i2 − i1 )

whence Z2
v2 = × 2 v1
Z 2 + Z1

In particular:

- for a line short-circuited to earth, Z 2 = 0 ; we can deduce from this that v2 = 0 and v1' = − v1

- for a conductor without a change of impedance, Z 2 = Z1 ; we can deduce from this that
v2 = v1 and v1' = 0

- for an open line, Z 2 = ∞ ; we can deduce from this that v2 = 2 v1 and v1' = v1 .

To conclude, at the point of change of impedance, the maximum voltage value may reach
double the incident wave. This is the case of a line feeding a transformer as its impedance in
relation to the lightning wave is very high in relation to the characteristic impedance of the line.

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction , either partly of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 438

High resistance earthing with restrike in the Limiting resistor earthing with restrike in the
fault or circuit-breaker, case of industrial fault or circuit-breaker, case of public
networks for which IrN < 20 to 30 A distribution networks for which I rN is equal
(see Protection guide - § 10.1.1.). to several hundred to 1 000 A. The overvoltage
I rn rN
The overvoltage depends on the ratio depends on the ratio
IC Xd
transient voltage as transient voltage as
% of the nominal % of the nominal
single-phase voltage single-phase voltage
peak value peak value
%
%
250
healthy
460
phases
200

300
260 healthy 150 neutral
240 phases
200
100 fault
neutral phase
100 fault 50
phase
I rN rN
0.5 1 1.5 2 2.5 IC 3 4 5 6 8 10 20 30 50 70 90 X (1)
I rN = Vn / rN : current in the neutral resistor during rN : neutral point earthing resistance
the fault X(1) : network positive-sequence reactance
I C = 3 Cω Vn : vectorial sum of current in the phase-
earth capacitances
If I rN ≥ 2 I C , the overvoltage does not exceed 240 %

Reactance earthing, case of public distribution networks for which I XN is equal to 1 000 to
several thousand amps
Case without restrike in the circuit-breaker Case with restrike in the circuit-breaker
transient voltage as transient voltage as
% of the nominal % of the nominal
single-phase voltage single-phase voltage
peak value peak value
%
% 500

400 400
B
C
300 theoretical limits 300 theoretical limits
B
without without
damping A damping
200 A 200
C
N
100 N 100

XN XN
0 2 4 6 8 10 12 14 16 X (1) 0 2 4 6 8 10 12 14 16 X (1)
X(1) : network positive-sequence reactance B, C : healthy phases
X N : neutral point earthing reactance N : voltage at reactance terminals
A : earth fault phase
Figure 5-34: transient overvoltages depending on the type of neutral earthing
during a phase-earth fault
Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE

 395

In most networks, generators are sufficiently far away to take the approximation Z(1) = Z( 2 ) ; we
thus have:

k = 1+
(
a Z(1) − Z( 0) )
2 Z(1) + Z(0 ) + 3 R f

Nomographs can be used to determine factor k for a zero fault resistance ( R f = 0 ) in relation
R(0) X(0 )
to the ratios and for R(1) = 0 and R(1) = 0.5 X(1) (see fig. 5.3. et 5.4.).
X(1) X(1)

where:

R(1) : positive-sequence resistance seen from the fault point

X(1) : positive-sequence reactance seen from the fault point
R(0 ) : zero-sequence resistance seen from the fault point
X( 0) : zero-sequence reactance seen from the fault point

When the fault resistance is not zero, we can see in the formula expressing k that the
overvoltage is weaker. The calculation of the overvoltage with a zero fault resistance thus
provides an excess value.

If we again use the diagram in figure 5-2, we can determine these impedances for a practical
case:

by taking:
ZT = RT + j XT 
 positive - sequence impedances
ZC = RC + j XC 

Z( 0 )T = RT + j X( 0 )T 

 zero - sequence impedances
Z( 0 )C = RC + j X( 0 )C 

Publication, traduction et reproduction totales ou partielles de ce document sont rigoureusement interdites sauf autorisation écrite de nos services.
The publication, translation and reproduction, either wholly or partly, of this document are not allowed without our written consent.

Industrial electrical network design guide T&D 6 883 427/AE