Power System Protection and Switchgear

On silver contacts silver oxide does not form readily. Even when formed, its
thickness does not exceed 10 ¢and hence can easily be moved aside by high pres.
sure or wiping action. Humid sulphurous and high temperature atmosphere causes
corrosion. In polluted atmospheres where coal fires are used, silver sulphide is formed
readily, especially in the presence of heat and humidity. It is not breakable like oxides
but it is soft and thus canbe squeezed aside by high pressure. Athin petrolatum coat
ing can reduce corrosion of contactswithout increasing their resistance. It is helpful
in polluted atmospheres.
provided at the
A dust-proof casing is usually used for modern relays. A filter is
casing is
back to trap any dust and to allowthe relay to breathe. A relay with such a
poor ventila
quite suitable for a dusty or otherwise dirty atmosphere. In arelay with
tion, particularly in a sealedrelay, high resistance polymers may appear on contacts.
insulation material. AIl insulating
This is due to organic emanations from the coil
degree. Phenolic resin
materials, except teflon give off organic vapour to a certainep0xide varnishes now
Polyester and
gives off organic vapours more than others.
satisfactory for coil insulation. Relay
available have good performance and are quite
minimise the collectionof high
casings with good ventilation and having dust filters case of reed
contacts as in the
resistance polyners on the contacts. Encapsulated
Alternatively, the relay coil can also
relays provide the best solution to the problem.
beencapsulated.
usually of a small rating.
An electromechanical relay used with comparators is protected
they are to be
When such relays controlauxiliary relays and timing units, connected across the
capacitor
with spark quenching circuits. A series resistor and
contacts is a simple spark quenching circuit.
2.3 STATIC RELAYS
electrical quantities is performed
In astaticrelay, the comparison or measurement of
the tripping of a circuit breaker.
by a static circuit which gives an output signal for
relay as a slave relay. The
Mostof the present day static relays include a de polarised
function of comparison or
slave relay is an outputdevice and does not perform the of its low cost. In a fully
measurement. It simply closes contacts. It is used because
slave relay. The elec
staticrelay, a thyristor is used in place of the electromagnetic output contacts at lovw
tromechanical relay used as a slave relay provides a number of
much simpler than an
cost. Electromagnetic multicontact tripping arrangements are
equivalent group of thyristor circuits.
A static relay (or solid state relay) employs
semiconductor diodes, transistors.
Now-a-days, integrated
zener diodes, thyristors, logic gates, etc. as its components. reliable and compact.
circuits are being used in place of transistors. They are more
directional relays.
Earlier, induction cup units were widely used for distance and
Later these were replaced by rectifier bridge type static relays which employed de
polarisedrelays as slave relays. Where overcurrent relays are needed, induction disc
ultimately static relays will
relays are in universal use throughout the world. But
relays and d
supersede all electromagnetic relays, except the attracted armature
polarisedrelays as these relays can control many circuit at low costs.
 Relay Construction and
2.3.1 Merits and Demerits of Static Relays Operating Principles 51
he advantages of static relays over
electromechanical relays are as follows
(i) Low burden on CTs and VTs. The static
relays
most of the cases they draw
power from the consume less power and in
auxiliary dc supply
(ii) Fast response
(iiü) Long life
(iv) High resistance to shock and vibration
(w) Less maintenance due to the
absence of moving parts and bearings
(vi) Frequent operations cause no deterioration
(vii) Quick resetting and absence of overshoot
(viii) Compact size
(ix) Greater sensitivity as amplification can be
provided easily
(x) Complex relaying characteristics can easily be
obtained
(xi) Logic circuits can be used for complex
protective schemes
The logiccircuit may take decisions to operate
under certain conditions and not to
operate under other conditions.
The demerits of staticrelays are as follows:
(i) Static relays are temperature sensitive. Their
characteristics may vary with the
variation of temperature. Temperature compensation can be made by using
thermistors and by using digital techniques for measurements, etc.
(ii) Static relays are sensitive to voltage transients. The semiconductor
Components
may get damaged due to voltage spikes. Filters and shielding can be used for
their protection against voltage spikes.
(ii1) Static relays need an auxiliary power supply. This can however be easily sup
plied by a battery or a stabilized power supply.
2.3.2 Comparators
When faults occur on asystem, the magnitude of voltage and current and phase
angle between voltage and current may change. These quantities during faulty con
ditions are different from those under healthy conditions. The static relay circuitry
Is designed to recognise the changes and to distinguish between healthy and faulty
conditions. Either magnitudes of voltage/current (or corresponding derived quanti
ies) are compared or phase angle between voltage and current (or corresponding
derivedquantities) are measured by the static relay circuitry and atrip signal is sent
t0 the circuit breaker when a fault occurs. The part of the circuitry which compares
the two actuating quantities either in amplitude or phase is known as the comparator.
Ihere are two types of comparators--amplitude comparator and phase comparator.
Amplitude Comparator
Aamplitude comparator compares the magnitudes of two input quantities, irrespec
Uve of the angle between them. One of the input quantities 1s an operating quantity
and the other a restraining quantity. When the amplitude of the operating quantity
Cxceeds the amplitude of the restraining quantity, the relay sends a tripping signal.
Tne actual circuits for comparators will be discussed in subsequent chapters.
 Relay Construction and Operating Principles 61
through a delay circuit. The delay circuit gives a delayed output. The output is delayed
2.30(b) and (c). The
by an angle &from the staring point of the block as shown in Fig.circuit) still coincide.
delay Sis kept 90°. If the block and pulse (output of the delay
2.30(b). If the block and
the second AND gate will give an output, as shown in Fig.
in
pulse do not coincide, the second AND gate does not give any output as shown
so
Fig. 2.30(c). Thismeans that the output of the first gate has to persist for a period 8
that the second gate may operate and send a tripping signal. This technique is more
suitable for multi-input comparators. However, it is subject to false tripping by a false
transient signal whereas phase comparators discussed earlier are not.

M
AND 8 AND -+Trip

Pulse
(a)

M N M

(b) (c)

Fig. 2.30 Time bias type phase comparator

2.4 NUMERICAL RELAYS

With the tremendous developments in VLSI and computer hardware technology,
microprocessors that appeared in the seventies have evolved and made remarkable
progress in recent years. Fast and sophisticated microprocessors, microcontrollers,
anddigital signal processors (DSPs) are available today at low prices. Their applica
tion to power system protection have resulted in the availability of compact, faster,
more accurate, flexible and reliable protective relays, as compared to the conven
tional ones.
Numerical relays which are based on numerical (digital) devices e.g. micropro
cessors, microcontrollers. DSPs etc. are the latest development in the area of power
system protection. In these relays, the analog current and voltage signals monitored
through primary transducers (CTs and VTs) are conditioned, sampled at specified
instants of time and converted to digital form for numerical manipulation, display
and recording. Thus, numerical relays, having monitored the current and voltage
signals through transducers, acquire the sequential samples of these ac quantities in
numeric (digital) data form through the data acquisition, system, and process the data
62 Power System Protection and Switchgear
numerically using an algorithm to calculate the fault discriminants and make trip
decisions.With the continuous reduction in digital circuitcosts and increases in their
functionality, considerable cost-benefit improvement ensues. At present micropro
cessor/microcontroller-based numerical relays are widely used. There is a growing
trend to develop and use numerical relays for the protection of various components
ofthe modern complex power system. Numerical relaying has become a viable alter.
native to the traditional relaying systems employing electromechanical and static
relays. Intelligent numerical relays using artificial intelligence techniques such as
artificial neural networks (ANNs) and Fuzzy Logic Systems are presently under
active research and development stage.
The main features of numerical relays are their economy, compactness, flexibility,
reliability, self-monitoring and self-checking capability, adaptive capability, muliple
functions, metering and communication facilities, low burden on transducers (instru
ment transformers) and improved performance over conventional relays.
The schematic diagram of a typical numerical relay is shown in Fig. 2.31

Signal S/H
CT Microprocessor
conditioner CKt
system
Analog ADC
Multiplexer
Signal S/H
VVT conditioner CKt |Microcomputer/
Microcontroller

Trip signal +

Fig. 2.31 The schematic diagram of a typical numerical relay

The levels of voltage and current signals of the power system are reduced by
voltage and current transformers (VT and CT). The outputs of the CT and VT (trans
ducers are applied to the signal conditioner which brings real-world signals into digi
tizer. The signal conditioner electrically isolates the relay from the power system,
reduces the level of the input voltage, converts current to equivalent voltage and
remove high frequency components from the signals using analog filters. The output
of the signal conditioner are applied to the analog interface, which includes sample
and hold (S/H) circuits, analogmultiplexer and analog-to-digital (A/D) converterS.
These components sample the reduced level signals and convert their analog levels
to equivalent numbers the are stored in memory for processing.
The signal conditioner, and the analog interface (i.e., S/H CKt, analog multiplexer
and A/D converter) constitute the data acquisition system (DAS).
The acquired signals in the form of discrete numbers are processed by a numeri
cal relaying algorithm to calculate the fault discriminants and make tripdecisions. If
there is a fault within the defined protective zone, a trip signal is issued to the circutt
breaker.
 Relay Construction and Operating Principles 63
2.5. COMPARISONBETWEEN ELECTROMECHANICAL
RELAYS AND NUMERICAL RELAYS
Sr. No. Feature Electronmechanical Relay Numerical Relay
1. Size
Bigger Compact
2. Characteristics Fixed Selectable
3. Flexibility No flexibility Flexibility due to
programmability.
4. Communication feature Not available Available
5. Blocking feature Not available Available
6 Self-supervision Not available Available
7. Adaptability Not adaptable Adaptable to changing
system condition.
8 Multiple-functions Not possible Possible
9 Accuracy ± 5% or more + 2%
10. Speed of operation Slow Fast
11. Burden on Transducers Very high Extremely low
(CTs and VTs)
12. Consistancy of Deteriorate No effect on calibration even
calibration with time after use of 20-25 years
13. Setting Through plug Software based settings.
setting in fix steps.
14. Memory feature No memory of any Several memory features
type is available. are available.
15. Maintenance Cumbersome and Maintenance free relays
frequent mainte
narce required.
16. Output relay programming Not available Available
17. Accessibility of relay Not possible Remote accessibility is
from remote place available
18. Status of service values Not available Available
19. Safety of personnel Not adequate due Adequate safety is provided.
to non-accessibility
at remote location.
20. Spares requirement There is need to Universal designs minimise
stock several items the spares requirement.
as spare.
21. Upward connectivity Not possible Possible
for a present or future
system such as SCADA,
EMS etc.
 Numerical Protection
11.1 Introduction
Dctoctive relays, which started out as meters with contacts, have undergone tremendous
evolution over the years. They were soon replaced by electromechanical relays which were
sensitive and accurate. When vacuum tubes were in vogue, protection engineers
implemented relays using vacuum tubes. Within a year of invention of the transistor, its
use in protective relays was reported. With the development of large-scale integrated
circuits,these were extensively used in the protective relays. The microprocessor that was
invented around 1971 revolutionized the electronics scene in its entirety and the
development of a microprocessor-based relay followed soon thereafter. However, a subtle
shift in the paradigm takes place when we move on to the microprocessor-based relay,
which works on numbers representing instantaneous values of the signals. Hence, the
name numerical relay. Other popular nomenclatures for such relays are digital relay,
computer-based relay or microprocessor-based relay.
In numerical relays, there is an additional entity, the software, which runs in the
background and which actually runs the relay. With the advent of numerical relays, the
emphasis has shifted from hardware to software. Hardware is more or less the same
Debween any two numerical relays. What distinguishes one numerical relay from the other
is the software.
perform
1ne conventional non-numerical relays are go-no-go devices. They
Ompar1son rather than straight numerical computation. In fact, the conventional relay
eriy bypasses the problem of computation by performing comparison. The numerical
relay does not have any such limitation because of its ability to performoldreal-time
relaying
Computation. Thus, the relay engineer need not merely implement the
Concepts but can devise entirely new computation-based concepts.balancing device to a
The modern numerical relay has thus evolved from a torque
programmable information processor. relaving technique.
concept using the numericalendless
"e can implement an existing almost and
However,
there the
is
possibilities of developing a new numerical relay are
of development of a new
numerical relay
is
shown very little standardization. The
in the flowchart of Figure 11.1
process
225
226 Fundamentals of Power System Protection

State the relaying problem

Mathematically model the relaying problem

Write algorithm

Run
Write program simulation
(High-level language)

Test with simulated data Data

Modify No
algorithm Working OK?
Yes
Generate machine code for the target
microprocessor/Digital signal processor
Relay test
bench
Download to numerical relay

Test with test inputs Analogue

voltages
and
Currents Real-life
inputs from
No
Working OK? CT and PT

Yes
Install in parallel with the
existing relay

Evaluate performance over a

period of 2-3 years

Ready for
Commercial launch

Figure 11.1 Development cycle of a new numerical relay.

11.2 Block Diagram of Numerical Relay

Figure 11.2 shows the block diagram of a numerical relay. The signals from the CTs and
PTs cannot be sampled directly and converted to the digital form. This is to ehan
that the signal does not contain frequency components having a frequency greater than
one half of the sampling frequency, This limit is enforced by the Shannon S sampling
theorem.
 Numerical Protection 227
Power system
Trip

Sample Select
command channel
Command Start conversion
Command
Analogue Anti
signals aliasing Sample Micro
from CT filters and hold Multiplexer procesSor
and PT J(Analogue) circuits Trip
Port
Analogue
to digital
Converter Data

End of Algorithm
conversion

Figure 11.2 Block diagram of numerical relay.

Therefore, the signals are first passed through a low-pass filter, which has to be an
analogue type of filter, because digital processing can only take place after the frequency
spectrum of the signal is properly shaped.
Next, the analogue signal is sampled and held constant during the time the value is
converted to digital form. The sample and hold circuit is an absolute must. T'he range of
Trequencies that can be handled by the analogue-to-digital converter (ADC) without the
Sample and hold (S/H) circuit is extremely low as shown in Table 11.1.

able 11.1l Comparison of maximum allowable frequency with and without S/H
Without S/H circuit With S/H circuit

1, du VAull scale du Vfull scale

1.
dtWhere
ax TADCeny
2rTADCOnv is the
dtmax 2 Ts/H aperture
where TsH aperture is the acquisition
time of the ADC. conversion time of the S/H circuit.

2. rmax 1 1
2. fmax
212TADConv
n= ADC word
length = 16 bits n
212"TsH aperture
=ADC word length = 16
bits
= 250ps (Typical)
TADCOny = 10 us
Vm= Aull scale (Typical)
TsH aperture
Vm= Vfull scale
Gives: Gives:
fnax = 9.7 kHz
tmax = 0.24 Hz
ADC can now handle
"Thus, without SH, the ADC can handle
Only extremely low
With S/H, the same
much higher frequencies.
frequencies.
 Protection
Fundamentals of Power System
246
any sampling instant is a function of mnumber of
and
Notekthat
number output at past flter is
the of past outputs. Thus this type of filter has feedback, hence the
inputs
known as recursive (which literally means running backwards).
also
transfer function of the filter in the z-domain is given by
The -1
Y(z)
X(2) 1-b,2-...- b, zk (11.44)
output of the filter is a
A recursive digital filter is stable if the
sequence, i.e. for n tending to infinity yn Should not exceed some non-increasing
positive numbe
irrespective of the choice of initial conditions.
Fiters
11.6.5 Comparison Between FIR and IIR
The following table lists several important issues concerning the FIR and IIR flters

FIR Filter IIR Filter

Output is a function of past m inputs. Output is a function of past m inputs
Therefore, non-recursive. as well as k past outputs. Therefore
recursive.
Finite impulse response. Infinite impulse response.
Always stable since there is no Because of feedback, possibility of
feedback. instability exists.
Has less number of coefficients. Has more number of coefficients.
Transfer function has only the numerator Transfer function has both the numerator
terms. and denominator terms.
Higher-order filter required for a given Lower-order filter required for a given
frequency response. frequency response.
Has linear phase response. Has nonlinear phase response.
Yn = aon t a1n-1 +"t Yn = ao In t ajxn-1 t* + amn-m
amn-m

Y(2) Y(z)
X(2) X(z) 1- b,z-...-b,z
Very simple to implement. Not as simple as the FIR filter.

11.7 Numerical Over-current Protection

Numerical over-current numerical relay.
protection
Here we describe a possible is a
straightforward application of the
Over-current relay.
method
The algorithm first reads all theofsettings
implementing a
such as numerical
the type of characteristics to be
time
implemented, the pick-up value Ipu the time multiplier of inverse
over-current relay or the time delay in case of setting in casethefundamental
The algorithm of a DTOC relay.
component of numerical
the fault over-current relay will first extract establish

value. Full cyclecurrent mss from the post-fault of current purpose

and asit
its rms samples
window Fourier transform may be used for this
effectively filters out the dc offset. It will then
Numerical Protection 247
withthe pick-up
setting and
compute the
plug compare this fundamental
whichthe relay
will be operating. setting multiplier, given component
by Ums/Ipu)Irms
Equipped with the PSM value, the relay at
will either compute or look up the
time delay depending upon the type of
implemented. The algorithm will then
cause the over-current
relay to gocharacteristic
required
that is
equalto the operating time. into a wait state for being
a time
At the end of this time delay, the relay will once again
fundamental tofind if the fault has evaluate the rms value of the
current still persists then already been cleared
the relay will issue a trip bysome other relay. If the fault
processed to make it compatible with the
trip coil of
output. This signal will be suitably
over-current relay is shown in Figure 11.16. the circuit breaker. The
of a typical algorithm
Start

Read tvpe of OC relay

characteristic,
extremely inverse
for example, DTOC, IDMT
or any custom inverse.,
characteristic
Read plug setting
Read time setting in case of DTOC

Read time multiplier setting in case of inverse time OC

relays

Read samples

Find the rms value of fundamental Ins using full cycle window
discrete Fourier transform

No
Is Ims > pu?
Yes
Compute or look-up time delay Top

Wait for time equal to Top

Read samples

Find the rms value of fundamental Ims

No
Is ms > pu?
Yes
Issue trip output

Stop
over-current relay algorithm.
Figure 11.16 Flowchart for a numerical
248 Fundamentals of Power System Protection

11.8 Numerical Transformer Differential Protection

Figure 11.17 shows the conceptual block diagram for numerical both protection of a
transformer. The idea is to estimate the phasor value of the current on sides of tha
transformer and find the phasor difference between the two. If the magnitude of this
be
difference is substantial, an internal fault is indicated and the trip signal shouldthat
issued. The above is a description of the simple differential scheme. All the reasons the
Therefore.
prompt us to go in for the percentage differential relay exist here as well.
numerical relay algorithm should be made to implement the percentage differential relay.
Transformer

L
Current to
voltage Vs a
Converter
Current to
lvoltage converter

Trip
B Slope
% Bias B Numerical Restrain
Min pick-up lpu relay
Time setting hardware
'pu
Tap setting
(, + I2)2

Trip Algorithm
Figure 11.17 Block diagram of numerical protection of transformer.

Algorithm for percentage differential relay will consist of the following steps:
Read percentage bias B and minimum pick-up Ipu
Read i, samples -’ Estimate phasor I, using any technique.
Read i, samples ’ Estimate phasor I, using any technique.
Compute spill current Ispill = I, - I,.
Compute circulating current Içireulating = ( + I)/2.
" If Igpill > (Bleireulating + pu) then trip, else restrain.

11.9 Numerical Distance Protection of Transmission Line

11.9.1 Mann and Morrison Method
Let voltage at the relay location be described by
(11.45)
V= Vm Sin(ot + 0,)
 Numerical Protection 249
currentbv
andthe i= Im sin(ot + 0;)
noted that voltage and current are (11.46)
maybe and i' represent the assumed to be pure sinusoids.
It
Let v'
canwrite
derivatives of voltage and current, respectively. Then we
u' = 0Vm cos(ot + 0,)
(11.47)
= V, cos(ot + 9,)

and
i= olm cos(ot t 0:) (11.48)
Or

= Im cos(ot + 0:)

Combining expressions forvand (v/0), we get

(11.49)

Similarly
(11.50)

Frther, the phase angles can be found as follows:

(11.51)
tan - 0t
,,
(11.52)
0; = tan

and relay current will be given by

The phase angle between the relay yoltage (11.53)
= , - 0;
signals. However, the same can be adapted
this analysis refers to continuous values instead of instantaneous values and
frm
relaying by substituting sample continuous derivatives. The numerical
Dliical instead of
computed derivatives written as
numeri cally
derivatives at the nth sampling
and i, can be
instant, u,
Un+1- Un-1
(11.54)

2At

Similarly in+1-n-1
(11.55)

2At
and m, We get
expressions for Vm
in the
Substituting Eqs. (11.54) and (11.55)
U +
(n+1- V,
(2Ato)
(11.56)

Vm
250 Fundamentals of Power System Protection

;2
Im = Ln t
(2Atw? (11.57)
v,2 At
, = - 0tn
Un+1- Un-1 (11.58)
Wi, 2At
0; = wtn (11.59)
Ln+1 - n-1
= ., - ; (11.60)
Thus, using a window of three samples, we can estimate the peak and hence the rms
values of voltage and current at the relay location as well as the phase angle between
them. In other words, the phasors for voltage and current can be estimated.
We can keep on sliding the window to make a running estimate of rms value of
current, rms value of voltage, phase angle between voltage and current and, hence
compute the apparent impedance seen from the relay location.
The algorithms appear to be very attractive because of their simplicity. However,
closer examination reveals that the Mann and Morrison method is based on the following
assumption:
Waveforms of both the voltage and current consist of undistorted sine waves of
fundamental frequency which is accurately known and is constant.
Obviously, the above assumption will be seldom, if ever, true in a power system
relaying scenario. Thus, the method will introduce significant errors if the assumption is
not true. How then can we use the method in practice?
We can use the method in practice, provided we operate as close to the assumption
as possible. This can be achieved by heavily filtering the voltage and the current
waveforms so as to remove all traces of de offsets, harmonics and noise. Thus, the price
for the simplicity of the method has to be paid in terms of complex preprocessing of the
voltage and current signals.
In order to make a robust relay based on the above algorithm, we will have to employ
additional methods tomake sure that stray noise signals do not cause false tripping. One
such method is to increment a counter every time the estimated value is such that
tripping is called for. The actual trip output is issued only if this counter reaches a
predetermined threshold. If the counter does not reach the threshold, it is reset to zero
after a predetermined time-out. The counter threshold depends upon the expected level
of noise at the particular location and has to be determined empirically.
11.9.2 Differential Equcation Method
In this method, the faulted line is modelled as a lumped series R-L circuit. The distributed
nature of the line parameters as well as the shunt capacitance are neglected. Thus, we
can relate the voltage and current at the relay location and the resistance and inductance
up to the fault location with the help of the differential equation:

(11.61)
dt
 Numerical Protection 251
where
the faulted phase
isthe current in
4,
is
thefunction all of the three-phase currents
4
are the functions of resistance and
Rand L location.
tothe fault
inductance of the transmission line up
order to keep things simple while
In
single-phase transmission line. developing the concept, we first consider a
Applicationto single-phase transmission line
o o single-phase
transmissi0n line fed iromn one end. Let us assume a groundfault
at acertain distance x from the sending end. The transmission line can be modelled as
aseries R-L circuit. This is shown in Figure 11.18.

CB CT R
PT

Relay
Fault

gure 11.18 Faulted transmission line modelled by lumped series R-L circuit.
For the loop
The voltage and current at the relay location are v and i, respectively.
formed by the equivalent circuit of the faulted line, using KVL, we can write
di (11.62)
v= Ri + b
known
values of v and i are
Though Eq.
and the (11.62) isvalue a differential equation, the
of dildt can be computed,
numerical
the equation, in fact, is a linear
one more equation
in
algebraic nuumerical
Vand i to equation
solve forinthe two unknowns. Writing
two unknowns, R and L. We,
therefore.,
Ea. (11.62) need
for two different sampling

hstantsn and + 1 as shown in

n Figure 11.19, we get (11.63)

U, Ri, + L din
dt
(11.64)

t L
din +1
Un +1 = Ri,.1 dt
252 Fundamentals of Power System Protection

Vn+1
Vn Vn+2

Vn-1

'n+1 n+2

Slide
Window
Time

Time
Sampling instants
Figure 11.19 Numbering of samples of voltage and current for differential equation algorithm.

From Figure 11.19, we can easily see that the numerical derivative of the current at
instants n and n + 1is
di, Ln+1- n-1 (11.65)
dt 2At

din +1 n+2- n (11.66)

dt 2At

where At is the sampling interval.

Expressing the simultaneous algebraic Egs. (11.63) and (11.64) in matrix notation
and solving for the unknowns R and L, we get
r+1-n-1
2At [R]
(11.67)
Un+1 L
n+1 24t

n+1- n-1
TRI 24t
(11.68)
|L| n+2-n
2At
 Numerical Protection 253
TR] 1
n+2- in -(n+1- in-1)
2At
D 2At
| L (11.69)
-in1
where
D= h2-i,
2At in+1-h-1 i1
24t
Arom which we get
R= (in+2- n)Un- (nt1- h1)n+1
i,(Gn+2- i)- in+1n+1- n-1) (11.70)

L =2At n Un +1 - n+1Vn
2A2- i) - n+E¡+1-n-i) (11.71)
Awindow of four samples is seen to be adequate for
computing one
values of R and L. In order to make the method more robust, the windowestimate of the
can be shifted
further and another estimate can be made. If the fault is a metallic fault, the estimated
values will converge on to a stable value.

Application to three-phase transmission line

The above discussion referred to a single-phase transmission line. However, in practice,
ne power transmission lines are always three-phase in nature. It will be pertinent,
eretore, to see how we can apply the differential equation approach to a three-phase
ansmission line. Consider the model of the three-phase transmission line shown in
ure l1.20. The elemental length dx of any phase is assumed to have a resistance, an
inductance and a mutual inductance parameter associated with it. The shunt capacitance
is neglected.

Self inductance
Ra La

Mutual
Lab inductance
Lh Lca
R

Va
Loc Le
Re
V
0000
Vo dx

line.
Figure 11.20 Model of three-phase
 Fundamentals of Power
System Protection
254
The voltage drop across the infinitesimal length da. of phase a can be written as

di, + (Ldx)
du, = (R,i, dx) + (L,dx) dig
dt +(Lgsda)dt dt
(11.72)
Or
d di,) di,
du,
dx
dt |+|Lae dt (11.73)
for phases b and ccan be written o
Similarly the voltage and current relationships
du, di, |+| R, +Ld
+ e didte (11.74)
dx dt
d
du, di, +R+ Le dt (11.75)
dx dt

transposed, we have
If the line is assumed to be ideally
(11.76)
R, = R, = R, = R,
(11.77)
L =Lp = L, = Ls
Lm (11.78)
Lab = Lne = Lba = Lpe = Lca = Lcb =
where
series resistance and self-inductance per unit length of each phase
Rs, L_ are the
between any two phases.
Ly is the mutual inductance per unit length
Eqs. (11.73) to (11.75), we obtain
By substituting Eqs. (11.76) to(11.78) into
di, + LM di, (11.79)
dx
R_ + Ls dt Lu dt dt

di, (11.80)
Lu digdt + R_ +Ls di
du, + L dt
dx
(11.81)
du.
dx
di,
dt + LM diy
dt
Rg + Ls dt
positive sequence
Now, the self- and mutual-parameters are related to the zero and
parameters as follows: (11.82)
Lo = L_ + 2LM (11.83)
RË = R_ (11.84)
and LË =L_ - LM
(11.85)
io =
3
 Numerical Protection 255
where
positive sequence resistance and
B, LË are
Lois
the zero sequence inductance
inductance, respectively
sequence current
i. isthe zero
By combining Eqs. (11.79) to (11.81) and Eqs. (11.82) to (11.85), we obtain
d
dua
dx R + hdt + (Lo- L)h
dt (11.86)
duy
dx =+ L i 4+ (, -L4) dio
dt (11.87)
du
dx
=R+ L4 dt , + (Lo - L) dio
dt (11.88)
Equations (11.72) to (11.75) can be used to compute the voltage drop between the relay
Jocation and the fault point for different types of faults as shown next.
Single-phase toground fault. Assume a metallic single-phase to ground fault on phase
aat adistance x from the relay location. The instantaneous value of voltage v, at the relay
location can be calculated using Eqs. (11.72) and (11.73) for untransposed lines and
Eqs. (11.79) to (11.81) for ideally transposed lines.
Using the instantaneous values of the voltages, currents and the rate of change of
currents, the voltage v, can be obtained by using Eqs. (11.72) to(11.73) as shown below:

dt
+ Lab i Lac ic (11.89)

di, (11.90)
where (11.91)
and i, =ia
(11.92)

expressed in the same way and în

Equations
this case
Egs.
elating to the
(11.79) to
transposed
(11.81) can be
lines
used
can
to
be
obtain the following relationship:
(11.93)
Va
diy
=xRi, + *Ls dt
where (11.94)
and
(11.95)

iy = i +
 System Protection
256 Fundamentals of Power

Phase-to-phase andthree-phase faults. When the fault involves two or

voltage between the faulted phases, say a and b, can be found as follows
three phases, the
d di, * Loa di,
dt + | R, +Ly;
dt
(11.96)
Equation (11.96) can be reaced to the more compact form as
di,
dt (11.97)
where

(11.98)
and
(L;- Lab)i (11.99)
L- Lab
When the line is assumed ideally transposed, Eq. (11.93) can be written in terms of
the difference between the currents on each faulted phase, giving

Va - Vb = xRi, + xLy dt
di, (11.100)
where in this case
(11.101)

The above analysis shows that the behaviour of the transmission line under fault
conditions is governed by a differential equation having the general form of

U= Ri, + L dty
(11.102)

The measured values of the currents and the voltages are taken in the form 0r
samples and the measurement is usually done simultaneously on all the three phases.
In Eq. (11.102), all the terms except R and L. are either the measured samples of ue
quantities easily computed from the samples. The values of R and L can be found Dy
solving the simultaneous equations involying R and L. One of the methods, Wnle
described in
involves computing the numerical derivatives has already been
Section 11.9.2.
integrating both
We can convert the differential equation into an integral equation by will have to be
sides, in which case the numerical integration rather than differentiation resultsin
performed. Integration, being an inherently low-pass operation, yields better current
the presence of higher frequency components present in the voltage and
waveforms.

11.10 Algorithms and Assumptions protection

We may easily get bogged down by the sheer number of algorithmsfor digital classified
which have been proposed in the be broadly
However, algorithms canexample,
into time-domain algorithmns and literature. -domain thesolution
frequency-¢ algorithms. For