Switching Transients Application Example

DIgSILENT PowerFactory *

Abstract Therefore, transients have to be expected and are ob-

served in the system during the change from the sit-
uation before to the situation after switching. Tran-
This example provides an introduction to switch- sients are abnormal patterns of current and voltage
ing transients. The application example contains that have a limited duration. They frequently exceed
various study cases for the investigation of elec- the currents and voltages met during steady-state op-
tromagnetic transient (EMT) phenomena such as eration. In order to ensure a safe and reliable oper-
transformer energistaion, overhead line energisa- ation of the power system, electrical equipment has
tion, fault and load switching as well as the analy- to be designed to withstand the stresses caused by
sis of transient recovery voltage in PowerFactory. transients.
The following topics and functionalities are dis-
cussed: Power System Modelling for Electromag- The network model in Figure 1 is used to demonstrate
netic Transients, Electromagnetic Transients Sim- the use of PowerFactory for the analysis of switching
ulation, Statistical Switching Analysis. transients. The network comprises parts of a meshed
400 kV transmission system and an underlying 110 kV
sub-transmission system.

1 General Description The transmission system consists of five substations,

a power plant and two large loads. The power plant is
connected to the substation North via a 400 kV cable
Switching operations in power systems are very com- system. The substations are interconnected by single
mon and must not jeopardize the system’s reliability and double circuit overhead lines according to Figure
and safety. Switching in power systems is necessary 1.
for the following reasons and duties:
The neighbouring network in the West is modelled by
• Taking into or out of service some sections of the a voltage source with an equivalent short circuit power
system, certain loads, or consumers. Typical ex- which acts as slack bus for the system. The transmis-
amples are: Energisation and de-energisation of sion system in the East is modelled by an equivalent
overhead lines, cables and transformers, switch- load. Shunt reactors and capacitor banks are used for
ing of shunt capacitor banks or shunt reactors. reactive power compensation.
• Transferring the flow of energy from one circuit
The 110 kV sub-transmission system comprises three
to another, e.g. in a substation from one busbar
substations which are connected to the Central sub-
to another.
station by overhead lines. Furthermore a wind farm
• Isolating certain network components because is connected to the substation through a high voltage
of maintenance or replacement. cable system.

• Isolating faulted sections of the network in order

to avoid damage and/or system instability. Ex-
amples are: terminal fault and short line faults. 2 Simulation Model
Switching in electrical power systems re-configures
the topology of an electrical network. It involves the The investigation of electromagnetic transients re-
making and breaking of circuits and causes a distur- quires an accurate representation of relevant power
bance of the steady state energy flow. system components. Depending on the transient be-

* DIgSILENT GmbH, Heinrich-Hertz-Str. 9, 72810 Gomaringen, Germany, www.digsilent.de

DIgSILENT PowerFactory, r4759 1

Switching Transients Example

ing analysed the network model should include: • Stray inductances and capacitances

• Valid representation of the model for a frequency • Electric arcing model

range which may vary from DC to several MHz
The following subsections describe the models of the
• Distributed parameter line models power system components of the transmission system
• Non-linear saturation characteristics of induc- shown in Figure 1.
tances

~
Power Plant SG Generator

C Bank
2
GT-1 GT-2
0 9

PP1 Load NE
PP2 Central North-East

Inactive
Out of Calculation
De-energised

Voltage Levels NPP-1 NPP-2

20 km 20 km
400, kV
110, kV
33, kV
27, kV
North1
Load N Shunt Reactor
CE-1.1
1
North2 25 km
West
Load E-1
WN NC-1 NC-2
100 km 100 km 100 km CE-1.2
25 km
0

East-1

CE-2.1
Grid TR_C1 30 km

Load E-2
TR_C2 CE-2.2
30 km
0

East-2
WC-1.1 WC-1.2 WC-1.3
100 km 100 km 100 km

Wind Farm
WFT-1
WC-2.1 WC-2.2 WC-2.3 WT
8
100 km 100 km 100 km CE-3
12 km 8

West 1 West 2

WF WFT-2

C4_1.1 C4_1.2 C1_1 C1_2

Figure 1: Single line diagram of the power system

DIgSILENT PowerFactory, r4759 2

Switching Transients Example

IS.R0.2

CCE-1-1
IS.R0.1 CB.R0

IS4.3
CNC-1_comp

IS.R6.3 IS4.2 CB4 IS4.1 CE-1.1

25 km
IS.R3.3

IS.R6.2 CB.R6 IS.R6.1

IS.R3.1 CB.R3 IS.R3.2

CNE-1 IS7.2 CB7 IS7.1
NC-2
100 km 100 km CE-1.2
25 km
IS.R5.2 CB.R5 IS.R5.1

CNE-2 IS7.3

IS.R5.3 100 km
CNC1

CCE-1-2
IS.R1.3

IS6.3

IS.R1.2 CB.R1 IS.R1.1

Shunt Reactor
IS.R4.1 CB.R4 IS.R4.2 IS6.1 CB6 IS6.2

CCE2-2
IS.R4.3 IS3.3
TR_C1

TR_C1-HV
IS3.2 CB3 IS3.1 CE-2.2

TR_C1-MV
CBS1 30 km
CBS2

CWC1_comp

IS.L2.3 IS2.2 CB2 IS2.1 CE-2.1

TR_C2-MV
30 km

TR_C2-HV
IS.L2.1 CB.L2 IS.L2.2 IS2.3
WC-1.3
100 km

TR_C2 IS5.1 CB5 IS5.2

CCE2-1
CWC1

IS.L3.2 CB.L3 IS.L3.1

IS5.3

IS.L3.3
CWC2

IS1.3

IS.L1.1 CB.L1 IS.L1.2

WC-2.3
100 km
IS1.2 CB1 IS1.1 CE-3
12 km
IS.L1.3

IS0.2

IS.L0.2
CWC-2_comp

CCE-3
CB0 IS0.1

IS.L0.1 CB.L0

C1_1 C1_2
C4_1.1 C4_1.2

Figure 2: Detailed substation layout diagram of substation Central

2.1 Transmission Lines

The transmission system contains the following over-

head lines (OHL) and cable systems:

• 400 kV double circuit OHL

• 400 kV single circuit OHL
• 400 kV double circuit cable system
• 110 kV double circuit OHL
• 110 kV single circuit cable system

Both cables and overhead lines are modelled based

on their geometry of the corresponding characteristics
of conductors and insulation layers.

Overhead Lines

Overhead lines can be modelled based on their ge- Figure 3: 400 kV double circuit overhead line structure
ometry and material characteristics of the conductors
and earth return path in PowerFactory.
The Y-values which are entered in PowerFactory cor-
The geometry is entered in PowerFactory using the respond to the average height of the conductor and
X-Y coordinates of the phase conductors and ground ground wire along the line. The average height in-
wires as shown in Figure 3. The graph shows the ge- cludes the sag and is calculated as follows:
ometry of the 400 kV and 110 kV double circuit over-
head line being used in the network model (Figure 1).

DIgSILENT PowerFactory, r4759 3

Switching Transients Example

The 110 kV overhead lines are untransposed in this

2 application example.
ℎaverage = ℎtower − · 𝑙sag
3
Phase A Phase B Phase C
Apart from the overhead line geometry the phase con-
ductors and earth wires have to be defined. As input upper
parameter the geometry of the corresponding conduc-
middle
tor, its DC-resistance and bundling configuration has
to be entered. The 400 kV overhead line is a bundled lower
conductor type configuration and comprises 4 subcon-
ductors per phase.
upper
The specific earth resistivity for the earth return path middle
is assumed with 100 Ωm for all overhead lines which
corresponds to a typical value [1]. Based on the lower
above input parameters, PowerFactory calculates the
impedance and admittance matrix for all phases of the
multiphase overhead line system. Figure 4: Transposition of the 210 km long 400 kV OHL
WC-1 and WC-2
The impedance and admittance matrix is then used
to calculate the reduced impedance matrix, sequence
impedances and corresponding distributed parameter High Voltage Cables
line models which will be required for the switching
transients studies. The high voltage cables are modelled in a similar way.
The geometry of the single core cable and its corre-
Depending on the frequency of the transient being in- sponding material characteristics are entered in Pow-
vestigated, one of the following models will be used for erFactory. The following layers are included in the
the EMT simulation: model of the single core cable:

• Lumped parameter model • Conductor

• Distributed parameter model with constant pa- • Sheath

rameter (frequently known as Bergeron model)
• Insulation
• Distributed parameter model with frequency de-
• Oversheath
pendent parameter (frequently known as J. Marti
model) • Semiconducting layers
Transposition The cross-section of the 400 kV single core cable is
displayed in Figure 5. The geometry is entered by
Due to the geometry of the overhead lines, the defining the thickness of each layer. Furthermore the
impedances of each phase differ from each other. In resistivity, relative permittivity and permeability is en-
order to mitigate unbalances in the system, overhead tered in PowerFactory.
lines are usually transposed.

In this example the transposition is modelled explic-

itly for the lines WC-1 and WC-2 between substation
West and Central. The phasing is entered in the line
coupling (ElmTow). Figure 4 shows the transposition
of OHL WC-1 and WC-2.

The other 400 kV overhead lines are transposed cir-

cuit wise in *.TypTow by selecting the option Trans-
position → Circuit Wise. By selecting this option the
positive and zero sequence offdiagonal elements of
the mutual sub-matrices between transposed circuits
are eliminated.
Figure 5: 400 kV single core cable

DIgSILENT PowerFactory, r4759 4

Switching Transients Example

The cables are layed directly in ground in flat formation tion of the wind farm with the 110 kV terminal of the
with a distance of 0.7 m apart. Two cable systems con- sub-transmission system.
nect the substation North with the Power Plant (see
Figure 1). They are laid in parallel. The layout of Saturation Model
the complete cable system is shown in Figure 6. The
earth resistivity is assumed to be 100 Ωm. For simulating non-linear, electromagnetic transients
such as the transformer energisation, the core satu-
ration needs to be included in the model of the trans-
former.

The saturation is defined in the transformer type. In

this simulation model the magnetising branch of the 3-
winding transformer is placed at the start point. The
magnetising branch of the 2 winding transformers is
located between the primary and secondary side.
Figure 6: 400 kV double circuit cable system laid in
ground The non-linear flux-current relationship of the 2-
winding transformers is modelled using a polynomial
approximation for the saturation curve. As an example
As for overhead lines, internal routines calculate the for a saturation curve Figure 7 shown the magnetising
impedance and admittance matrix of the cable system characteristic of the transformer WFT-1 in Figure 1.
and all required parameters for steady state, dynamic
and electromagnetic transient analysis. The saturation characteristic of the 3-winding trans-
former is modelled based on open circuit measure-
Line Compensation ment data. They are entered as RMS values (open cir-
cuit test). PowerFactory converted them internally to
Shunt reactors are used for the compensation of the current-flux peak values to model the saturation char-
charging current of the 400 kV cable system and the acteristic properly.
200 km single circuit and 210 km long double circuit
overhead line connecting substation West and Cen- The residual flux of the transformer is taken into con-
tral. sideration using a parameter event. The residual flux
is entered in PowerFactory in the dq-frame. The trans-
They are connected directly to the line at its sending formation of the a-b-c fluxes in the dq-frame is docu-
and receiving end and are designed to compensate mented in the Technical Reference of the transformer.
approximately 70% of the charging current. Furthermore, hysteresis can be included in the trans-
former core model.

2.2 Transformers

The network model comprises two-winding and three-

winding transformers [6], [7].

Transformer Types

3-winding transformers (TR_C1 and TR_C2) are in-

stalled in substation Central connecting the 110 kV
sub-transmission system to the 400 kV transmission
system. The transformers have a rated power of
275 MVA. The vector group is YN0yn0d11.

The substation PowerPlant in the North of the network

includes a YNd5 1600 MVA transformer which con-
nects the generator’s 27 kV terminal with the 400 kV Figure 7: Saturation characteristic of transformer
terminal of the transmission grid.

Furthermore two YNd5 50 MVA transformers are in-

stalled in the wind farm and connect the 33 kV substa-

DIgSILENT PowerFactory, r4759 5

Switching Transients Example

2.3 Generator
NPP-1 NPP-2 NNPS2_comp

The synchronous machine of the power plant has a NNPS1_comp

rated power of 1560 MVA and a nominal voltage of

27 kV. Switching transients will not be investigated in
close vicinity of the generator. Therefore the model is
not further described here. North 1

IS2.2 IS4.2

The wind farm has a rated power of 100 MVA and is

represented by an aggregated wind turbine as shown CB2 CB4

in Figure 1. The wind turbine is a full-scale converter

model. The converter is implemented in PowerFactory IS2.1 IS4.1

using a static generator element, which is configured

to operate as a current controlled voltage source. The ISa.2 ISb.2

converter currents are controlled using a classical d-

CBa CBb

q rotating reference system current controller in the

EMT simulation. NWN1 ISa.1 ISb.1 NNC2

IS1.1 IS3.1

2.4 Loads
CB1 CB3

The loads are modelled as constant impedance loads. IS1.2 IS3.2

North 2

2.5 Substations
NC
WN
The network model comprises 5 substations:

• Substation West Figure 8: Substation North with 1 1/2 breaker method

• Substation North
• Substation Central The substation layout includes two-breaker and 1 1/2-
breaker configuration. The two-breaker configuration
• Substation Power Plant is shown in Figure 2. In this case the circuit breaker,
branch disconnector and instrumental transformers
• Substation Wind Farm
are duplicated in each branch. Busbar interchange
All substations are modelled in detail as shown in Fig- and isolation of one busbar for maintenance is possi-
ure 2 for substation Central. The substation models ble. One branch breaker can be taken out for main-
include busbars, circuit breakers (CB), disconnector tained at any time without interrupting operation.
switches, surge arresters and shunt reactors for reac-
tive power compensation. The 1 1/2-breaker design is applied in Figure 8. Fewer
breakers are needed here for the same flexibility as
Generally busbars, circuit breakers and disconnec- above. Isolation without interruption is possible. All
tor switches are modelled as ideal elements without breakers are normally closed. Uninterrupted supply is
impedances and as ideal switches. If required, bus- thus maintained even if one busbar fails.
bars will be modelled as distributed parameters lines.
Relevant stray inductances and capacitances of sub-
station equipment (such as instrumental transformers) 2.5.1 Circuit Breakers
will also be considered if necessary.
Circuit breakers (CB) are modelled as ideal switches.
Arcing models and models for re-ignition are available
in PowerFactory but will not be considered in the study
cases which are discussed in this example.

For Transient Recovery Voltage (TRV) studies the

DIgSILENT PowerFactory, r4759 6

Switching Transients Example

breaker capability curve is modelled according to 3 Switching Transients

IEC 62271-100 with a 2-parameter or 4-parameter
curve depending on the voltage level and circuit
breaker rating. Switching in electrical power systems re-configures
the network topology for the following purposes:

2.5.2 Surge Arresters • Isolation and Earthing

• Busbar-Transfer Switching
Surge arresters are used to protect electrical equip-
• Load Switching
ment in substations, such as transformers, circuit
breakers and bushings, against the effects of over- • Fault-Clearing
voltages caused by incoming surges. Such overvolt-
ages can be caused by a direct or nearby lightning The following switching transients are analysed in this
strike and other electromagnetic transients such as application example:
e.g. switching operation in the power supply system. • Transformer Energisation
Surge arresters present a nonlinear resistor and are • Overhead Line Energisation
characterised by a highly non-linear U-I curve. During
normal operating voltages surge arresters have an ex- • Inductive Load Switching
tremely high resistance and a relatively low resistance • Fault-Current Interruption Switching
during transient overvoltages.
The transient phenomena following the above switch-
Metal oxide arresters (MOA) are usually used for ing actions will result in abnormal patterns of current
surge arresters and in most cases are connected from and voltage during the transient. The root causes of
phase to ground. The charactersitic U-I curve of an the abnormal currents and voltages are described in
surge arrester which is necessary for an EMT study the following section.
is usually provided in the vendor datasheet and is en-
tered in PowerFactory in tabular form.
3.1 Saturation

For simulating electromagnetic transients such as

transformer inrush currents or ferro-resonance, reac-
tor / transformer core saturation needs to be included
in the model of the transformer. The non-linear be-
haviour of a typical transformer iron core is shown in
Figure 10 [1].

The solid lines represent the steady state voltage and

flux and current relationship when no saturation oc-
curs. The dashed lines represent the behaviour dur-
Figure 9: Surge arrester 𝑈 − 𝐼 characteristic ing a transient, when the core is being saturated (e.g.
during transformer energisation).

2.5.3 Reactive Power Compensation The magnetic flux Ψ, representing the time-integral of
the voltage, lags in steady state the sinusoidal wave-
form of the voltage by 90∘ . Being still in the linear
For reactive power compensation a shunt reactor is in- part of the saturation characteristic, the current is si-
stalled in substation Central. The rated reactive power nusoidal as well. Thus, the current is proportional to
of the shunt reactor is 50 Mvar per step. The reactor the voltage.
is switchable and has 4 steps. Thus, the shunt reactor
can consume a maximum of 200 Mvar. When a completely demagnetised transformer is en-
ergised the flux Ψ in the iron core is zero at the instant
A capacitor bank is installed at the substation North- of switching. The instantaneous voltage in Figure 10
East with a rated reactive power of 50 Mvar per step. is zero at this instance. Due to the 90∘ phase shift
It is also switchable and has 10 steps. between the flux Ψ and the voltage, the flux ought to
be at its peak in steady state operation. However, due

DIgSILENT PowerFactory, r4759 7

Switching Transients Example

to the initial condition (Ψ equals zero) the flux starts at supply voltage 𝑢𝑆 .
zero. As a result the curve of the flux Ψ is shifted by
Ψ𝑚 upwards in the first cycle of the transient. Thus, At the receiving end of the line, the voltage 𝑢𝑅 is still
the flux starts at zero and then reaches a maximum of zero since the surge needs a certain time to travel
2Ψ𝑚 in the first cycle as indicated in Figure 10 [1]. from the sending to the receiving end of the line. The
surge propagates with the speed 𝑣 along the line and
Taking into a account a residual flux of Ψ𝑅 in the iron arrives at the sending end of the line with a delay equal
core, the curve is shifted upwards even further as in- with the propagation time 𝜏 .
dicated in Figure 10 and reaches a maximum peak
values of Ψ𝑅 + 2Ψ𝑚 . The magnetization curve is now The propagation speed 𝑣 is a function of the line in-
′ ′′
operating in the saturated region and the relationship ductance 𝐿 and capacitance 𝐶𝐿 . For a lossless line
between the flux Ψ𝑚 and the current 𝑖 are no longer the propagation speed 𝑣 is calculated by the following
governed by a linear relationship. This causes high equation [1]:
and distorted inrush currents as shown in the graph.
1
𝑣=√
ψ ψ, B 𝐿 · 𝐶′
′
ψR + 2 ψm
2 ∙ ψm
The propagation time constant 𝜏 is dependent on the
ψR propagation speed v and the line length
H, i
𝑙 √︀
𝜏= = 𝑙 · 𝐿′ · 𝐶 ′
u 𝑣
i
Figure 11 shows the voltage surge during the line en-
ergisation of an ideal single phase conductor for three
time steps of the transient:

• 1: Breaker is closed at the sending end

• 2: Surge propagates from sending to receiving

Figure 10: Saturation and inrush current end
• 3: Reflected surge travels back to sending end

3.2 Travelling Waves Transient voltage and current surges are thus a func-
tion of time and location along the line. Distributed pa-
rameter line models are therefore required to analyse
When investigating transient and high-frequency such phenomena. Depending on the transient being
steady-state phenomena, it is necessary to account investigated, line models with constant or frequency
for the distributed-parameter nature of conductors dependent parameters are used in simulation models.
such as overhead lines, cables and in the case of high
frequency transients, even of busbars. The constant distributed parameter model in Power-
Factory is based on Bergeron’s method, which calcu-
Typically, for low frequency steady-state analysis, lines lates the voltages and currents at one end of the line
and cables are modelled using the well-known lumped based on the voltage and current at the other end de-
parameter equivalent circuit, thus neglecting the dis- layed in time. For further information refer to the tech-
tributed nature without loosing too much accuracy in nical reference of the line models [4].
the results. In reality, a current in a conductor having
even a very short length, needs a certain time to travel With the exception of lossless and distortion-less
from its sending end to the remote end [1]. lines, the characteristic impedance 𝑍𝐶 and the propa-
gation constant 𝛾 are frequency dependent. To handle
When an overhead line or cable is energised from frequency-dependent parameters for overhead lines,
the grid as shown in Figure 11, a voltage and current PowerFactory uses the approach proposed by J.Marti
surge are injected into the line. Before the breaker is [4]. For cable systems the Universal Line Model is
closed, the voltage at both ends of the line is zero. available and offers a high accuracy and a phase-
Upon breaker closing at the sending end of the line, domain formulation [5].
the voltage changes instantaneously from zero to the

DIgSILENT PowerFactory, r4759 8

Switching Transients Example

𝑢𝑆 𝑢𝑅 𝑢2 2 · 𝑍2
=
𝑢1 𝑍1 + 𝑍2
𝑍 + 𝑍1
𝑢𝑆 𝑢𝑅 𝑢𝑟 = 2
𝑍1 + 𝑍2
𝑖1 2 · 𝑍1
=
𝑖2 𝑍1 + 𝑍2
𝑢𝑆 𝑢𝑅
𝑍 − 𝑍2
𝑖𝑟 = 1
𝑍1 + 𝑍2

Figure 11: Voltage surge line energisation Typical surge impedances for common power system
components are:

During the transient, the surge is travelling back and • Overhead lines: 200 Ω to 500 Ω
forward until damping and attenuation will eventually • Cable systems: 40 Ω to 70 Ω
result in a steady state condition. At each electri-
cal boundary the surge is reflected and refracted as • Transformers: 1 kΩ to 10 kΩ
indicated in Figure 11. Electrical boundaries arise
when the surge impedance of the network compo-
nents change, e.g. at an overhead line - cable tran- 3.3 Transient Recovery Voltage (TRV)
sition or at the open end of a line.
Transient recovery voltages (TRV) arise both at fault
interruption and load switching. In both cases a circuit
𝑍1 𝑍2 breaker is switching off. The circuit breaker opening
procedure is explained in Figure 13. The switching
𝑖1 𝑖2 command is usually initiated automatically by a relay
that detects a fault in the system or a switching com-
𝑢1 mand from the control center to trigger a change in the
𝑢2 operation scenario of the network.
𝑍2 > 𝑍1 : 𝑢𝑟
The tripping command activates the operating mech-
𝑢𝑟 anism and through its kinematic chain separates the
𝑢2
𝑢1 contacts in the circuit-breaker. After a certain opening
time, the circuit-breaker arcing contacts will open in all
three poles.

𝑖1 Breaker state

𝑖1 𝑖2
𝑖𝑟

Figure 12: Travelling voltages and currents at surge

impedance jump location
Switching Contacts start Arc extinction in last pole Contacts fully
command to separate  Current interruption separated

arcing time

The reflected and refracted voltage and current surge Total break time

are a function of the incoming voltage and current

waveform and the surge impedance on both sides of Figure 13: Circuit breaker operation switching off
the boundary. For an ideal surge with a very steep
rise / very short rise time, the reflected and refracted
surges are calculated according to the following equa- Upon contact separation, an arc is formed in the in-
tions: terruption chamber of each pole. The actual current

DIgSILENT PowerFactory, r4759 9

Switching Transients Example

interruption must wait for the zero crossing of the cur- side is defined by 𝐿𝐿 and 𝐶𝐿 and is a single frequency
rent. Depending on the type of circuit-breaker, the de- oscillation in case of linear inductive load.
vice may not be ready to interrupt at the first occurring
current zero after contact separation.
1 1
𝑓𝑅 = · √︀
For SF6 circuit breakers, it takes a certain minimum 2𝜋 𝐿𝐿 · 𝐶𝐿
arcing time before the electric arc extinguishes in the
circuit breaker because sufficient cooling pressure of The Transient Recovery Voltage (TRV) is the voltage
the extinction medium must be available and sufficient across the open circuit-breaker contacts which arises
contact distance must be reached. For vacuum circuit immediately after current interruption. It is the differ-
breakers sufficient contact distance has to be reached ence between the voltage-to-earth at the source side
in order to extinguish the electric arc [2]. 𝑢𝑆 and load side 𝑢𝐿 :

𝐿𝑆
𝑢TRV = 𝑢𝑆 − 𝑢𝐿

𝐶𝑆 𝑢𝑆 𝑢𝐿 𝐶𝐿 𝐿𝐿 The TRV always consists of at least two oscillatory

components, the source-side and load-side frequency
component. Depending on the switching action (e.g.
shunt reactor switching, terminal fault, short line fault
interruption) the TRV contains multiple oscillation fre-
quencies which superimpose.
Figure 14: Frequency spectrum of load side oscillation
The exemplary TRV for a single phase circuit is shown
in Figure 15 assuming the following values:
Fault currents and inductive load currents (e.g. discon-
nection of shunt reactors) lag the voltage by around
90∘ . Thus, the instantaneous voltage at both sides of 𝐿S = 17 mH; 𝐶S = 273 𝜇F
the breaker pole is at its peak when the current is in-
terrupted at the current zero-crossing. 𝐿L = 1000 mH; 𝐶L = 2 𝜇F

Immediately after current interruption a transient volt- Figure 15 shows that the TRV starts from zero at
age will oscillate on the load side and source side. current zero, makes an excursion to the momentary
Figure 14 shows the equivalent circuit of the load and power-frequency voltage, overshoots in a damped
source side after the switching action. oscillatory manner and continues to oscillate until a
steady-state condition is reached. The frequency of
The source side voltage is composed of two compo- the load side oscillation is 𝑓𝐿 = 3558.8 Hz and of
nents, the power frequency voltage (e.g. 50 Hz) and the source side 𝑓𝑆 = 73.9 Hz. Important parame-
the transient voltage oscillating between 𝐿𝑆 and 𝐶𝑆 . ters in a TRV study are the maximum transient recov-
The frequency of the transient part of the voltage is ery voltage 𝑈𝐶 and the rate of rise of recovery volt-
calculated by the following formula: age (RRRV). Both parameters are required for circuit
breaker design studies.
1 1
𝑓𝑆 = · √︀ Current chopping may need be considered as well in
2𝜋 𝐿𝑆 · 𝐶𝑆
the TRV study for some types of circuit breakers or
switching transients.
The oscillation frequency of the voltage on the load

DIgSILENT PowerFactory, r4759 10

Switching Transients Example

200,00

100,00

0,00

-100,00

-200,00

-300,00
0,00 2,00 4,00 6,00 8,00 [ms] 10,00
Source: Phase Voltage SP in kV

200,00

100,00

0,00

-100,00

-200,00

-300,00
0,00 2,00 4,00 6,00 8,00 [ms] 10,00
Load: Phase Voltage SP in kV

200,00 40,00

100,00 30,00

0,00 20,00

-100,00 10,00

-200,00 0,00

-300,00 -10,00
0,00 2,00 4,00 6,00 8,00 [ms] 10,00
Circuit Breaker: Voltage Difference in kV Circuit Breaker: Phase Current/Terminal i in A
Source: Power Frequency Voltage

Figure 15: Source side, load side and transient recovery voltage (TRV)

DIgSILENT PowerFactory, r4759 11

Switching Transients Example

4 Study Cases Large loads are connected to substation North and

North-East, each with a consumption of 1200 MW and
a power factor of 0.9 and 0.98 respectively. For reac-
The application example contains several study cases, tive power compensation the capacitor bank at sub-
each of them being discussed further below: station North-East is supplying 400 Mvar (tap 8). The
shunt reactor at substation Central is tapped to zero
• “Steady State Load Flow Analysis”: Intro- and thus not consuming any active power.
duction to the network model and components.
Load flow calculation and analysis (see section The load flow calculation is configured with automatic
4.1); tap adjustment of transformers at substation Central
• “Transformer Energisation”: Investigation of and at the wind farm. The load flow calculation will
instantaneous inrush currents caused by the en- determine the active and reactive power flows for all
ergisation of the wind farm transformer WF-1. branches, and the voltage magnitude and phase for
all nodes. Executing the load flow results in the phase
Peak currents and RMS voltage dip. Fourier
voltages displayed in Figure 16. The voltage at sub-
analysis of the inrush currents (see section 4.2);
station West is the reference voltage with an angle of
• “Overhead Line Energisation”: Investigation 0∘ and a symmetrical phase shift of 120∘ for each
of transient overvoltages occurring during the phase respectively.
energisation of the 400 kV overhead line WC-1.
Determination of peak and duration of transient The external grid serves as slack and is supplying
voltages and currents. Statistical switching tool 1900 MW for active power balancing and a reactive
(see section 4.3); power of 60 Mvar. The voltages of all busbars in
the transmission system and sub-transmission sys-
• “TRV Analysis of Shunt Reactor tem, are in the range of 0.99 to 1.03 p.u. in the steady
Switching off”: Analysis of the transient recov- state load flow calculation. The loading of lines and
ery voltage (TRV) during shunt reactor open- transformers varies between 14% and 50%.
ing. Circuit breaker design validation (see sec-
tion 4.4); The steady state load flow calculation results are the
• “TRV Analysis Terminal Fault”: Investigation basis for the calculation of the initial conditions and
of the TRV following a terminal fault at substa- thus the starting point of the EMT simulation.
tion Central. High voltage TRV modelling (see
section 4.5);
[deg]

6,00
4.1 Steady State Load Flow 5,00
Analysis 4,00

3,00

The load flow calculation is based on algorithms for 2,00

the unbalanced load flow taking into account unbal- 1,00

ances resulting from imperfect line transposition of -240, -210, -180, -150, -120, -90,0 -60,0 -30,0 30,0 60,0 90,0 120, 150, 180, [kV]
the multi-phase network and unsymmetrical loads and -1,00

generation. The active power control in the load flow -2,00

calculation is based on the dispatch of generators and -3,00

loads. The difference between load and generation is -4,00

balanced by the reference machine. -5,00

-6,00

The external grid connected to substation West rep- -7,00

resents the transmission system in the West is used -8,00

as reference machine (slack). It controls the voltage West 1: Line-Ground Voltage

North 1: Line-Ground Voltage
to 1 p.u. at substation West. The power plant con- C4_1.1: Line-Ground Voltage
North-East: Line-Ground Voltage

nected to substation North is set to dispatch 750 MW.

A station controller is used to regulate the voltage at Figure 16: Phase voltages of the transmission system
the HV-side of the Power Plant to 1.02 p.u. The wind
farm connected to the 110 kV sub-transmission sys-
tem is set to dispatch 30 MW at a power factor of 0.95
overexcited.

DIgSILENT PowerFactory, r4759 12

Switching Transients Example

4.2 Transformer Energisation phase C results in a phase current of 2.97 p.u. in the
saturation curve corresponding to transformer WF-1.
This study case investigates the energisation of the
A Fast Fourier Transform (FFT) is used to investigate
wind farm transformer WF-1. Before energisation, the
the inrush currents in order to quantify their harmonic
complete wind farm is disconnected. The 110 kV HV
content. The FFT is carried out with a 20 ms win-
cable which connects the wind farm to the point of
dow and 64 samples, resulting in a sampling rate of
common coupling (PCC) at substation Central is al-
3200 Hz. Figure 18 shows the FFT of the inrush cur-
ready energised and in steady state. The wind farm
rents in the same plot. As can be seen, phase C con-
transformer WF-1 is energised by closing the circuit
tains a significant DC-component. This is because the
breaker CB4. The residual flux in the transformer core
flux in phase C starts at zero when the transformer is
is assumed to be zero at the time when the trans-
energised, thus resulting in a fluctuating flux between
former is energised (𝜓𝐴 (𝑡0 ) = 𝜓𝐵 (𝑡0 ) = 𝜓𝐶 (𝑡0 ) = 0).
0 p.u. and 2 p.u. (fluctuating around 1 p.u.).
The EMT simulation is initialised at 𝑡 = −100 ms and
The energisation of the wind farm transformer causes
the integration step size is set to 𝑡 = 100 𝜇s . The cir-
a voltage drop at the grid connection point (Central
cuit breaker CB4 is closed at 𝑡 = 0 s. The simulation
substation). The RMS voltage during the energisation
results are documented in Figure 17 and Figure 18.
process is shown in Figure 18. Ahead of energisation
Following the switch event the flux in all three phases
the RMS voltage in all three phases is approximately
rises proportional to the source voltage at the HV side
1.015 p.u. The minimum RMS voltage during the en-
of the transformer with a 90∘ phase shift. The peak
ergisation occurs in phase C and is 0.99 p.u. which
flux arises in phase C and is -1.68 p.u. which corre-
corresponds to a voltage dip of approximately 2.5%.
sponds to a flux of 339.6 Vs.
Usually grid code requirements limit the maximum al-
During the energisation, the fluxes in all three phases lowable RMS voltage dip to 2.0%. Thus, countermea-
exceed 1 p.u. and drive the transformer into satura- sures such as point on wave switching (𝑃 𝑜𝑊 ) would
tion. The current-flux relationship is no longer linear have to be considered for the specific case.
and results in high inrush currents which contain har-
Energising both transformers instantaneously at
monic distortion as described in section 3.1. The non-
𝑡 = 0 s results in a minimum voltage of 0.97 p.u. This
linear flux-current relationship can also be plotted in
will most likely not be acceptable with regard to the
PowerFactory and is shown in Figure 17. The peak
grid code. Taking residual flux into consideration will
inrush current on the HV side of the transformer arises
increase the inrush currents and consequently the
in phase C and is 780 A. This is approximately 3 times
RMS voltage dip.
the rated current of 262.4 A. Thus, a flux of 1.68 p.u. in

DIgSILENT PowerFactory, r4759 13

Switching Transients Example

2 3 3

[p.u.] [p.u.] [p.u.]

Max. = 1,197 p.u.
1 2 Max. = 1,471 p.u. 2

0 1 1

-1 0 0

-2 -1 -1

Min. = -1,675 p.u.

-3 -2 -2
0 10 20 30 40 [ms] 50 0 10 20 30 40 [ms] 50 0 10 20 30 40 [ms] 50
WF: Phase Voltage A WF: Phase Voltage B WF: Phase Voltage C
WFT-1: Magnetising Flux A WFT-1: Magnetising Flux B WFT-1: Mag. Flux c C

4 800 3,00
[p.u.] [A] [p.u.]
2 400
2,00

0 0
1,00

-2 -400
0,00

-4 -800

-1,00
-6 -1200
0 10 20 30 40 [ms] 50 0 10 20 30 40 [ms] 50
WFT-1: Magnetising Current, Phase a WFT-1: Phase Current A/HV-Side -2,00
WFT-1: Magnetising Current, Phase b WFT-1: Phase Current B/HV-Side -6 -4 -2 0 2 [p.u.] 4
WFT-1: Magnetising Current, Phase c WFT-1: Phase Current C/HV-Side

Figure 17: Transformer HV side voltage, magnetic flux in the core, magnetizing current

1,03
[sec.V]
1,02

PCC\RMS Voltage: Phase A

1,01
PCC\RMS Voltage: Phase B
PCC\RMS Voltage: Phase C
1,00

0,99

0,98
0,0 100,0 200,0 300,0 400,0 [ms] 500,0

800,0
[A]
400,0

0,0 WFT-1: Phase Current A/HV-Side

WFT-1: Phase Current B/HV-Side
WFT-1: Phase Current C/HV-Side
-400,0

-800,0

-1200,0
0,0 100,0 200,0 300,0 400,0 [ms] 500,0

500,0
[A]
400,0

WFT-1: Phase A
300,0
WFT-1: Phase B
WFT-1: Phase C
200,0

100,0

0,0
0,000 50,00 100,0 150,0 200,0 [Hz]

Figure 18: RMS voltage dip at PCC, Inrush currents and their harmonic distortion

DIgSILENT PowerFactory, r4759 14

Switching Transients Example

4.3 Line Energisation

The energisation of high voltage overhead lines and

cable systems may cause considerable overvoltages.
With increasing operating voltages of transmission
systems, switching surge overvoltages determine the
insulation design rather than lightning overvoltages.

Transient overvoltages caused by internal line events

such as line switching are one of the main problem in
EHV and UHV systems with regard to insulation coor-
dination and are discussed in this study case. These
transients typically have durations ranging from a few
tens to several thousands of µs and belong to the
category of slow front transients. They usually have
complex waveforms with frequencies in the range of
100 Hz to several kHz superimposed on the power fre-
quency.

4.3.1 Overhead Line 400 kV

Figure 19: Frequency dependent parameters of the

This study case investigates the energisation of the
line model for Mode 1
overhead line WC-1. Before the energisation, the line
is disconnected from the grid and will be energised
from the substation Central while the other end of the
line remains disconnected. The initial condition before The EMT simulation is initialised at 𝑡 = −20 ms and
the energisation is based on the steady state load flow the integration step size is set to 𝑡 = 10 𝜇𝑠. A switch
calculation as described in section 4.1. The surge ar- event is defined to close all three phases of the high
resters connected to the overhead line WC-1 are in voltage circuit breaker CB.L2 simultaneously at 𝑡 = 0 s
service. The network topology is stored in the Oper- as shown in Figure 20. After the breaker contacts are
ation Scenario Energisation 400 kV OHL WC-1 which closed, transient voltage and current surges are in-
is activated with the Study Case Overhead Line Ener- jected and travel towards the receiving end at substa-
gisation. tion West with the propagation speed of approximately
300 m/µs.
For the purpose of this analysis a frequency depen-
dent, distributed parameter line model is used for
overhead line WC-1. The change of the line model
compared to the base case is stored in the Variation
Distributed Parameter WC. The modal transformation
matrix is calculated at a frequency of 500 Hz. The min-
imum and maximum frequency for the approximation
by rational functions of the propagation factor and the
characteristic impedance are 0.01 Hz and 1 MHz re-
spectively.

The frequency dependent parameters of the overhead

line model are displayed in Figure 19 for Mode 1.
Figure 20: Energisation of OHL WC-1 at substation
Central

DIgSILENT PowerFactory, r4759 15

Switching Transients Example

600 Max. =552,791 kV

300

-300

-600

-900
0 4 8 12 16 [ms] 20
L2: Phase Voltage A in kV
T1.6: Phase Voltage A in kV

500

250

-250

-500 Min. =-394,306 kV

-750
0 4 8 12 16 [ms] 20
L2: Phase Voltage B in kV
T1.6: Phase Voltage B in kV

500

250

-250

-500 Min. =-410,437 kV

-750
0 4 8 12 16 [ms] 20
L2: Phase Voltage C in kV
T1.6: Phase Voltage C in kV

Figure 21: Transient phase-to-ground voltages at the sending and receiving end during the energisation of
OHL WC-1

1200
[A]
800

WC-1.3: Phase Current A/Terminal j

400
WC-1.2: Phase Current A/Terminal j
WC-1.1: Phase Current A/Terminal j
0

-400

-800
0 300 600 900 1200 [us] 1500

1200
[A]
800

WC-1.3: Phase Current B/Terminal j

400
WC-1.2: Phase Current B/Terminal j
WC-1.1: Phase Current B/Terminal j
0

-400

-800
0 300 600 900 1200 [us] 1500

300
[A]
0

WC-1.3: Phase Current C/Terminal j

-300
WC-1.2: Phase Current C/Terminal j
WC-1.1: Phase Current C/Terminal j
-600

-900

-1200
0 300 600 900 1200 [us] 1500

Figure 22: Transient phase currents during the energisation of OHL WC-1

DIgSILENT PowerFactory, r4759 16

Switching Transients Example

Figure 21 shows the transient voltage at the sending the switch event) is varied randomly by ± 10 ms. Fur-
and receiving end of overhead line WC-1 during the thermore, the scatter time corresponding to the me-
energisation process. The voltage surges are injected chanical delay of the closing time of each individual
at the sending end of the overhead line and then du- phase of the breaker is varied randomly by 0...1 ms in
plicated and reflected at the receiving end (substation each simulation. Figure 25 shows the maximum over-
West). Then the surges travel back to substation Cen- voltages for the 𝑛 = 100 EMT simulations executed
tral. Due to attenuation, the magnitude of the wave automatically by the statistical tool in PowerFactory.
decreases and the waveform is distorted, as it prop-
agates along the line. The current and voltage wave Figure 25 documents the maximum line-to-earth over-
shapes become dissimilar, though they were the same voltages at substation West resulting from each sim-
initially. The following maximum overvoltages arise ulation executed during statistical analysis. The plot
during the energisation of overhead line WC-1: documents the overvoltages resulting from the energi-
sation of overhead line WC-1 from substation Central
• Phase A: 516 kV for all three phases.
• Phase B: 400 kV As demonstrated, the overvoltages significantly de-
pend on the breaker closing time and also the oper-
• Phase C: 420 kV
ation scenario, such as e.g. energisation from sub-
Figure 22 shows the current surge along the line. In station West or Central. For economic reasons the
phase A a current surge of approximately 1080 A is in- insulation design for self-restoring insulation is usually
jected into the line, in phase B the amplitude is 845 A not based on the maximum prospective overvoltage
and in phase C 267 A. The propagation time of the of all investigated cases. Instead the design is based
first line segment is approximately 234 µs. Thus the on a probabilistic approach to withstand only a certain
percentage, e.g. 98%, of the prospective overvoltages
injected current surge arrives at the first point of trans-
position only after 234 µs which can be observed in which are calculated in a statistical analysis.
Figure 22. The surge arrives at SS West (open end
of overhead line WC-1) approximately 712 µs after the According to the insulation coordination standard IEC
circuit breaker is closed. 60071-2 first the maximum overvoltages are deter-
mined using EMT simulations. The maximum over-
The overvoltages resulting from line energisation voltages are documented in Figure 25 and are based
highly depend on the magnitude of the injected surge on the statistical analysis. As a next step, the rep-
at the sending end of the line. The magnitude of the resentative overvoltage 𝑈𝑟𝑝 is derived from the max-
injected surge depends mainly upon the voltage in- imum overvoltages. Secondly, the coordination with-
stantaneous value (point on the waveform) at which stand voltage 𝑈𝑐𝑤 is calculated from the representa-
the circuit-breaker contacts close electrically. Since tive overvoltage using the coordination factor 𝐾𝑐 , as
the point on the waveform of the voltage depends on follows:
the circuit breaker closing instant, statistical studies
should usually be performed. As the circuit breaker
closing time is of random nature in reality, all potential 𝑈𝑐𝑤 = 𝑈𝑟𝑝 · 𝐾𝑐
breaker closing times and thus overvoltages should be
evaluated. The coordination factor 𝐾𝑐 makes allowance for lim-
itations in the modelling, shape and duration of an
The randomness of the circuit breaker closing time overvoltage. For the deterministic approach (non-self
is modelled in PowerFactory by means of statistical restoring insulation) in the insulation coordination 𝐾𝑐
analysis. The statistical analysis tool in PowerFactory is usually 1.0 . For a probabilistic approach it is based
runs a number of n simulations. In each simulation on the statistical analysis.
the breaker closing time is varied randomly within a
defined range. Additionally, breaker pole scattering For that purpose, the risk of failure 𝑅(𝑈 ) is calcu-
can be included in the statistical analysis in order to lated based both on the probability distribution of the
account for deviations in the closing time between the overvoltage 𝑓 (𝑈 ) and the probability of the insulation
three phases. strength 𝑃 (𝑈 ) and is illustrated in Figure 23.

An example of a statistical switching analysis is pro- ∫︁

vided in this section. The overhead line WC-1 from 𝑅(𝑈 ) = 𝑓 (𝑈 ) · 𝑃 (𝑈 ) 𝑑𝑈
the previous simulation is used for this purpose. For
the statistical analysis, 100 simulations are run. For
each simulation the breaker closing time (defined in Figure 23 shows the probability distribution of the

DIgSILENT PowerFactory, r4759 17

Switching Transients Example

overvoltages and the failure probability of the dielec- The mean 𝜇 of the distribution function and the stan-
tric strength of the insulation. According to the figure, dard deviation 𝜎 are determined using the statistical
a voltage of 1.5 p.u. is most likely. For that voltage tool in PowerFactory. The normal distribution function
the failure of the insulation is very small. Thus, the is shown in Figure 24 and shows the probability versus
risk of failure R(U) is also relatively small. For higher standard deviation 𝜎 in general.
voltages, e.g. 2 p.u., the probability for overvoltages is
smaller. However, the probability of insulation failure is The insulation for self-restoring insulation, such as
much higher. Thus, the overall risk of failure is higher. overhead lines, is usually designed for the 98%
For very high overvoltages the risk of insulation failure percent value of the prospective voltage stresses.
is significant. However, the probability of such high The statistical tool in PowerFactory determines the
overvoltages is almost zero and thus the risk of failure prospective overvoltage 𝑈 for a defined probability
is also nearly zero. 𝑓 (𝑈 ) (e.g. 98%). The analysis is based on the 𝑛 num-
ber of EMT simulations executed during the statistical
analysis. In this example (𝑛 = 100 EMT simulations)
1
the statistical results in Table 1 are obtained for en-
f(U) ergisation of overhead line WC-1 from the substation
0,8
P(U) Central.
Probability in p.u.

0,6 R(U)
Table 1: Statistical overvoltages for different bands
0,4
around the mean 𝜇
0,2
𝑎𝑟𝑒𝑎 𝑃 (𝑈 ) 𝜇 𝑢𝑚𝑎𝑥
0 (𝜇 − 𝜎, 𝜇 + 𝜎) 65.9% 471 kV 492 kV
-3 -2 -1 0 1 2 3 4
Stress, e.g. voltage in p.u. (𝜇 − 2𝜎, 𝜇 + 2𝜎) 95.5% 471 kV 548 kV
(𝜇 − 3𝜎, 𝜇 + 3𝜎) 99.73% 470 kV 593 kV
Figure 23: Risk of failure 𝑅(𝑈 )

The coordination withstand voltage 𝑈𝑐𝑤 is chosen

The distribution of the breaker closing time and thus
based on the 2% probability and is equal to 564 kV
the transient overvoltages 𝑓 (𝑈 ) is usually a normal
in this case. This means that only 2% of the simulated
distribution. Consequently, a normal distribution func-
overvoltages exceed this value.
tion can be assumed for the breaker closing time and
the resulting overvoltages: In the following insulation coordination study the re-
quired withstand voltage 𝑈𝑟𝑤 is now determined
1 1 𝑢−𝜇 2 based on the coordination withstand voltage 𝑈𝑐𝑤 and
𝑓 (𝑢) = √ · 𝑒− 2 ·( 𝜎 ) using the safety factor 𝐾𝑠 and altitude correction fac-
𝜎· 2·𝜋
tor 𝐾𝑎 :

0,4
Probability overvoltages in p.u.

𝑈𝑟𝑤 = 𝑈𝑐𝑤 · 𝐾𝑠 · 𝐾𝑎
0,3

0,2 Finally, the standard withstand voltage 𝑈𝑤 is selected

68.5 %
according to the tables in IEC 60071-1. In case the
required withstand voltage 𝑈𝑟𝑤 exceeds the desired
0,1 standard withstand voltage defined in IEC 60071-1,
95.5 %
the prospective overvoltages can be reduced by addi-
99.7 %
tional surge arresters, pre-insertion resistors or point
0
-4 -3 -2 -1 0 1 2 3 𝜎 4 on wave switching.

It should be noted that usually 1000 runs are recom-

Figure 24: Probability distribution of switching over- mended for a statistical switching study in order to en-
voltages sure the statistical significance of the study.

DIgSILENT PowerFactory, r4759 18

Switching Transients Example

600
overvoltages in kV
500
400
300
200
Phase A Phase B Phase C
100
0
0 20 40 60 80 100
number of simulation n
100
Phase A Phase B Phase C
80
probability in %

60
40
20
0
350 400 450 500 550 600
voltage in kV

Figure 25: Probability distribution of energisation overvoltages for OHL WC-1

DIgSILENT PowerFactory, r4759 19

Switching Transients Example

4.4 TRV Analysis of Shunt Reactor The network impedance at the shunt reactor is calcu-
Switching off lated using the Frequency Sweep Calculation Tool in
PowerFactory. The inductance is 68.51 mH at substa-
tion Central in case the shunt reactor is out of service.
For reactive power compensation a shunt reactor is
Based on the network capacitances and inductances
installed at substation Central as shown in Figure 2.
the approximate oscillation frequency can be calcu-
The shunt reactor is a three-phase five-leg core type
lated as described in section 3.3. The load side oscil-
with a rated reactive power of 200 Mvar. The reac-
lation frequency is approximately 1.58 kHz:
tor is switchable and has 4 steps. Thus, it is able
to consume 0 Mvar, 50 Mvar, 100 Mvar, 150 Mvar and
200 Mvar. The reactor is connected directly to the 1 1
𝑓𝑅 = · √︀ = 1.58 kHz
400 kV busbar via a 400 kV circuit breaker. 2𝜋 𝐿𝑅 · 𝐶 𝑅

In this study case the transient recovery voltage (TRV)

is analysed when the shunt reactor in substation Cen- Figure 28 shows the transient oscillation voltage on
tral is switched off. Before disconnecting, the shunt the shunt reactor side after disconnecting the device
reactor is operated at step 2, supplying the grid with together with the phase currents through the circuit
200 Mvar of reactive power. The reactor is then dis- breaker connecting the reactor with substation Cen-
connected by opening the CB. The objective of this tral.
study case is to analyse the transients following the
The first zero-crossing occurs in phase A after the
reactor drop out.
circuit breaker is triggered by a simulation event.
For the purpose of this analysis the variation Detailed At 𝑡 = 0.607 ms the current in phase A is zero and
CB model shunt reactor is activated which contains the breaker is opened. Then, 3.404 ms later at
a more detailed model of CB.R4 with a two param- 𝑡 = 4.011 ms the current in phase C goes to zero and
eter curve according to IEC 62271-100 to model the the ideal switch is opened for that phase. In phase B
dielectric strength of the breaker during the transient. the switch is opened at 𝑡 = 7.314 ms.
Furthermore, winding capacitances of the shunt reac-
Due to the 90∘ phase shift between the interrupted
tor and additional stray capacitances are included in
current and the supply voltage (inductive load), the
the model in order to do the TRV analysis.
voltage in each phase of the isolated shunt reactor is
Figure 26 shows a simplified single line diagram of the at its peak when the breaker contacts separate. Thus,
most important components of the switching bay of the the stray capacitances in each phase are charged with
shunt reactor. The shunt reactor is tapped at step 1. energy 𝐸𝐶 when the current is interrupted. The stored
The inductance of the reactor is 10.19 H. The capac- energy 𝐸𝐶 is:
itance on the source side of the CB corresponds to
the lumped capacitances of the instrumental voltage 1
and current transformers. They are assumed to be 𝐸𝐶 = · 𝐶 · 𝑢2
2
equal to 4 nF. The capacitances on the shunt reactor
side mainly represent the winding stray capacitances The energy is now oscillating between the inductance
of the shunt reactor and are assumed to be equal to and capacitance until the transient is damped out and
1 nF. the load side voltage is zero. Figure 28 shows the
transient oscillation on the load side. The peak value
68.51 𝑚𝐻 of the voltage is approximately 315 kV and the fre-
quency corresponds to the resonant frequency of the
circuit in Figure 26. Applying a Fast Fourier Trans-
10.19 𝐻

4 𝑛𝐹 1 𝑛𝐹 former (FFT) to the load side transient voltage results

in the plot in Figure 27. The dominant frequency is
between 1.8 kHz and 1.9 kHz as predicted by the sim-
plified hand calculation.

Following the current interruption in each phase, the

Figure 26: Equivalent circuit diagram showing the TRV, as described in section 3.3 arises across the
source inductance and stray capacitances of the shunt breaker poles. The TRV for each phase is depicted
reactor bay in Figure 29 together with the breaker capability curve
of the CB. According to IEC 62271-100 the TRV en-
velope is defined by a two-parameter curve. The input

DIgSILENT PowerFactory, r4759 20

Switching Transients Example

parameters for the TRV envelope are based on the

maximum peak voltages 𝑢𝑐 and rate of rise of recovery
voltage 𝑅𝑅𝑅𝑉 . In this example the following parame-
ters are assumed for the breaker capability curve:

• Peak TRV voltage 𝑢𝐶 : 787 kV

• Time to 𝑢𝐶 : 112 µs

In the investigated case the maximum peak voltages

and RRRV of the TRV are within the defined limits for
all three poles. The following results are obtained:

• Phase A: 𝑢𝑐 = 626 kV; 𝑅𝑅𝑅𝑉 = 2.57 kV/𝜇s

• Phase B: 𝑢𝑐 = 625 kV; 𝑅𝑅𝑅𝑉 = 2.47 kV/𝜇s
• Phase C: 𝑢𝑐 = 627 kV; 𝑅𝑅𝑅𝑉 = 2.53 kV/𝜇s

The interruption of small inductive currents frequently

results in current chopping and virtual current chop-
ping which might lead to re-ignition and even multiple
re-ignitions. Both effects can easily be included in the
breaker model used for this study.

Figure 27: Equivalent circuit of source and load side

DIgSILENT PowerFactory, r4759 21

Switching Transients Example

600

[kV]
400

200

-200

-400
0 4 8 12 16 [ms] 20
T1: Phase Voltage A
T1: Phase Voltage B
T1: Phase Voltage C

200

[A]
100

-100

-200

-300
0 4 8 12 16 [ms] 20
CB phase A: Phase Current/Terminal i
CB phase B: Phase Current/Terminal i
CB phase C: Phase Current/Terminal i

Figure 28: Load side oscillation voltagesand phase currents in the high voltage circuit breaker

1000

500

-500

-1000

-1500
0,0 2,5 5,0 7,5 10,0 [ms] 12,5
CB phase A: Voltage Difference in kV
TRV Limit Phase A T10: TRV_limit_neg
TRV Limit Phase A T10: TRV_limit_pos

1000

500

-500

-1000

-1500
0,0 2,5 5,0 7,5 10,0 [ms] 12,5
CB phase B: Voltage Difference in kV
TRV Limit Phase B T10: TRV_limit_neg
TRV Limit Phase B T10: TRV_limit_pos

1000

500

-500

-1000

-1500
0,0 2,5 5,0 7,5 10,0 [ms] 12,5
CB phase C: Voltage Difference in kV
TRV Limit Phase C T10: TRV_limit_neg
TRV Limit Phase C T10: TRV_limit_pos

Figure 29: Transient recovery voltage (TRV) and dielectric strength of circuit breaker

DIgSILENT PowerFactory, r4759 22

Switching Transients Example

4.5 TRV Analysis Terminal Fault

0,00
91,4
0,0
CBS1
CBS2
CWC1_comp
In this section the transient recovery voltage across
the breaker poles of the circuit breaker CB.L2 in sub- IS.L2.3
CB.L2
station Central is analysed, following a terminal fault
IS.L2.1 IS.L2.2
on the line side of the circuit breaker. For the purpose WC-1.3
of this analysis the CB.L2 is modelled by detailed cir-
cuit breaker model. The breakers include a model of

CWC1
IS.L3.2 CB.L3
the dielectric strength during the transient according
to IEC 62271-100. IS.L3.3

CWC2
The detailed breaker model is stored in the Variation
TRV Analysis Terminal Fault which is activated with WC-2.3 IS.L1.1 CB.L1 IS.L1.2

the study case TRV Analysis Terminal Fault. Fur-

IS.L1.3
thermore, all lines connected to the faulted terminal
are modelled as distributed parameter lines. The fre-
quency for travel-time estimation 𝑓𝜏 is chosen based CWC-2_comp IS.L0.2
on the line length 𝑙 and propagation time of each line
according to the following equation: IS.L0.1 CB.L0

0,00
91,4

0,00
91,4
0,0

0,0
𝑓𝜏 =
4·𝑙 C4_1.1 C4_1.2

Before the fault, the transmission system is in steady

state as described in section 4.1.The short circuit is Figure 30: Fault location of terminal fault
applied on the line side of the circuit breaker CB.L2
as depicted in Figure 30. To simulate the fault, a 3-
phase short circuit is triggered at 𝑡 = 0 s. After 50 ms The limits in Table 2 correspond to the short-circuit
switch events are triggered to isolate the faulted part duty tests T100, T60 and T30. The duty test T100 cor-
of the network. The following circuit breakers have to responds to a short circuit current equal to the short
be opened: circuit rating of the CB. In the duty test T60 the short
circuit current is only 60% of the short circuit rating of
• CB.L2.1 (substation Central)
the CB.
• CB.W.1 (substation West)
Table 2: Parameters of TRV envelope of 420 kV circuit
• CB.W.1 (substation West) breakers; terminal faults (T)

The current in each circuit breaker is interrupted at the Curve 𝑢1 /kV 𝑡1 /𝜇s 𝑈𝐶 /kV 𝑡2 /𝜇s
following current zero of each individual phase. The
line is re-connected after clearing of the fault. How- T100 334 167 624 668
ever, the re-closure action of the line is not simulated T60 334 11 669 666
in this study case since the investigation focuses on
T30 - - 687 137
the TRV capability of the HV circuit breaker during the
Terminal fault.

The breaker capability curve is modelled according to Figure 31 shows the short circuit current flowing from
the limits defined in IEC62271-100 for a voltage level substation Central to the fault location at the beginning
of 420 kV. The relevant parameters for the envelope of overhead line (OHL) WC-1.3. Initially the current
curve are listed in Table 2 and define a 4-parameter through the circuit breaker is in steady state. At time
envelope. The envelope curve defines the maximum 𝑡 = 0 s the fault is initiated with zero fault impedance
transient voltages which are allowed after the fault in- (𝑅𝑓 𝑎𝑢𝑙𝑡 = 0 Ω and 𝑋𝑓 𝑎𝑢𝑙𝑡 = 0 Ω).
terruption across the breaker poles and represents the
dielectric strength of the breaker. The steady state load current changes into a signif-
icantly higher short circuit current in an oscillatory
manner as shown in Figure 31. After a few ms the
short circuit current is nearly in steady state. At time

DIgSILENT PowerFactory, r4759 23

Switching Transients Example

𝑡 = 50 s a switch event triggers the CBs to open. The The maximum peak voltages 𝑢𝑐 and rate of rise of re-
short circuit current is then interrupted at the following covery voltage 𝑅𝑅𝑅𝑉 during the transient for the dif-
current zero (zero-crossing) in each phase. ferent phases are as follows :

During the short circuit the terminal voltage at substa- • Phase A:

tion Central drops from a peak line-to-earth voltage 𝑢𝑐 = 529 kV; 𝑅𝑅𝑅𝑉 = 0.32 kV/𝜇s
of 328.3 kV to 0 kV and afterwards returns approxi-
• Phase B:
mately to its original value. The peak short circuit cur-
𝑢𝑐 = 497 kV; 𝑅𝑅𝑅𝑉 = 0.24 kV/𝜇s
rent through the circuit breaker connected to substa-
tion Central is 16.7 kA. • Phase C:
𝑢𝑐 = 489 kV; 𝑅𝑅𝑅𝑉 = 0.14 kV/𝜇s
As described in section 3.3 a transient recovery volt-
age arises across the circuit breaker poles after the The maximum withstand voltage of the CB for the
current is interrupted. The TRV appearing across the T100 duty is 624 kV and the maximum RRRV is
circuit breaker poles is the instantaneous values sum 0.32 kV/µs. The transient recovery voltage (TRV)
of the voltage from the source and the line side phase- therefore does not exceed the limits defined in
to-earth voltage and is shown in Figure 32. The plot IEC62271-100. For this particular case a high volt-
shows the TRV arising in phase A, B and C together age circuit breaker with a fault current ratio of 100%
with the breaker capability curve of the circuit breaker of its rated short circuit capability is sufficient for the
CB.L2. application.

Short Circuit Switch Event

30

[kA]

20

10

-10

-20
0 20 40 60 80 [ms] 100
IS.L2.3: Phase Current A/Terminal i
IS.L2.3: Phase Current B/Terminal i
IS.L2.3: Phase Current C/Terminal i

Short Circuit Switch Event

600

[kV]

300

-300

-600

-900
0 20 40 60 80 [ms] 100
C4_1.1: Phase Voltage A
C4_1.1: Phase Voltage B
C4_1.1: Phase Voltage C

Figure 31: Short circuit current through the CB.L2 during the transient and voltage at substation Central

DIgSILENT PowerFactory, r4759 24

Switching Transients Example

1000

500

-500

-1000

-1500
50 54 58 62 66 [ms] 70
CB phase A(1): Voltage Difference in kV
TRV Limit Phase A: TRV_limit_neg
TRV Limit Phase A: TRV_limit_pos

1200

800

400

-400

-800
50 54 58 62 66 [ms] 70
CB phase B(1): Voltage Difference in kV
TRV Limit Phase B: TRV_limit_neg
TRV Limit Phase B: TRV_limit_pos

1000

500

-500

-1000

-1500
50 54 58 62 66 [ms] 70
CB phase C(1): Voltage Difference in kV
TRV Limit Phase C: TRV_limit_neg
TRV Limit Phase C: TRV_limit_pos

Figure 32: Transient recovery voltage (TRV) and dielectric strength of circuit breaker

References Line Models”, PowerFactory 2018, DIgSILENT

GmbH, Gomaringen, Germany, 2018

[1] Juan A. Martinez-Velasco: “Power System Tran- [5] DIgSILENT PowerFactory

sients: Parameter Determination”, CRC Press, Technical Reference Documentation “Cable Sys-
2009, ISBN 978-1420065299 tem”, PowerFactory 2018, DIgSILENT GmbH,
Gomaringen, Germany, 2018
[2] R. Smeets; L. Sluis; M. Kapetanoviae; D. Peelo;
A. Janssen: “Switching in Electrical Transmission [6] DIgSILENT PowerFactory
and Distribution Systems”, Wiley, 2014, Technical Reference Documentation “Two-
ISBN 978-1118381359 Winding Transformer”, PowerFactory 2018,
DIgSILENT GmbH, Gomaringen, Germany,
[3] Allan Greenwood: “Electrical Transients in Power 2018
Systems”, Wiley-Interscience, 1991,
ISBN 978-0471620587 [7] DIgSILENT PowerFactory
Technical Reference Documentation “Three-
[4] DIgSILENT PowerFactory Winding Transformer”, PowerFactory 2018,
Technical Reference Documentation “Overhead DIgSILENT GmbH, Gomaringen, Germany,
2018

DIgSILENT PowerFactory, r4759 25