Renewable and Sustainable Energy Reviews 52 (2015) 280288

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: [Link]/locate/rser

Progress on power swing blocking schemes and the impact

of renewable energy on power swing characteristics: A review
M.K. Gunasegaran a,n, ChiaKwang Tan a, A.H.A. Bakar a, H. Mokhlis b, H.A. Illias b
a
UM Power Energy Dedicated Advanced Center (UMPEDAC), Level 4, Wisma R&D, University of Malaya, Jalan Pantai Baharu, 59990 Kuala Lumpur, Malaysia
b
Department of Electrical Engineering, Faculty of Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia

art ic l e i nf o a b s t r a c t

Article history: The occurrence of power swing had been reported to threaten the seamless operation of power system.
Received 18 November 2014 Several cascaded outages and major blackouts in the world have been traced to mal-operation of
Received in revised form distance relays due to power swing occurrences. As such, this paper will present a consolidated review
15 May 2015
on the power swing phenomenon, covering the causes of power swing, its impact to consumers and
Accepted 11 July 2015
generators, as well as its detection and relay blocking schemes to prevent the mal-operation of distance
relay. Given the wide arrays of power swing detection techniques and blocking schemes available up to
Keywords: date, particular attention is concentrated at the various power swing detection techniques on high
Power swing voltage power system. In addition, the effect of modern renewable energy generations to the power
Power swing blocking schemes
swing characteristics is also presented in this paper.
Distance relay
& 2015 Elsevier Ltd. All rights reserved.
Cascaded blackout

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
2. Background studies: power swing phenomenon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
2.1. Causes of power swing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
2.2. Impacts of power swing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
3. Power swing blocking scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
3.1. Conventional methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
3.1.1. Concentric characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
3.1.2. Blinder scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
3.1.3. Decreased resistance method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
3.1.4. Swing center voltage (Vcos) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
3.1.5. Superimposed current method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
3.1.6. Three phase power variation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
3.2. Signal analysis techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
3.2.1. Fast fourier transform (FFT) analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
3.2.2. Wavelet transform (WT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
3.2.3. S-transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
3.2.4. Prony method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
3.3. Intelligence based diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
3.3.1. Adaptive neuro-fuzzy inference system (ANFIS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
3.3.2. Support vector machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285

n
Corresponding author.
E-mail address: mu2kumaran@[Link] (M.K. Gunasegaran).

[Link]
1364-0321/& 2015 Elsevier Ltd. All rights reserved.
 M.K. Gunasegaran et al. / Renewable and Sustainable Energy Reviews 52 (2015) 280288 281

4. Impact of renewable generation on power swing characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286

5. Further research potentials . . . . . . . . . . . . . . . . . . ............. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287

1. Introduction
paper will be presented in Section 3. In the section, the widely used
schemes, its working principle, advantages, and disadvantages will
Power system stability plays a vital role in ensuring the systems
be presented. In addition, the section will also discuss the latest
reliability. In steady state, a balance power ow is maintained
methods and ndings in detecting power swing by means of signal
between the generation and load. However, unpredicted events or
analysis. Section 4 will present the potential impact to power swing
large disturbances such as fault, line switching, generator discon-
characteristics due to the integration of renewable energy genera-
nection, and the loss or application of large blocks of loads may
tions. New and signicant information generated from this review
interrupt this balance and cause oscillation of power known as
that require further analysis and understanding are presented in
power swing [1,2] in the system. In severe circumstances, the
Section 5. Finally, the conclusion is drawn in Section 6.
power swing may lead to loss of synchronism between the
generator and load, resulting in generator tripping and subse-
quently risking cascaded power outages and major power black-
2. Background studies: power swing phenomenon
outs. However, not all power swings are unstable for some of the
power swings are stable swing, where the generators can recover
Power swing is dened as large oscillation of power between
from the temporary transient instability. During these stable swing
two areas of a power system [8]. It can be categorized into two
events, the distance relays are very likely to detect a temporary
types, stable and unstable swing. If the swing is stable, the
decrement in impedance, which can be wrongly interpreted as a
uctuation settles in a short period of time. On the contrary,
fault by the relay. As oppose to a fault, it is undesirable to trip the
unstable swing leads to advancement of rotor angle causing
line under these stable swing conditions given that the network
instability to the system.
has the ability to return to stability. Consequently, the undesirable
operation of distance relay under a stable swing scenario is termed
as mal-operation of distance relay or nuisance tripping. 2.1. Causes of power swing
The North American Electric Reliability Corporation (NERC)
reported an alarming statistics where 75% of the major distur- Power system faults, line switching, generator disconnection,
bances leading to cascaded outages and major blackouts involved and the loss or application of large blocks of load causes uctua-
the mal-operation of distance relays [3]. Table 1 lists the countries tions in electrical power although the mechanical power remains
that experienced major blackout due to nuisance tripping of relatively constant in a generator [1,913]. When one of these
distance relay [1,47]. disturbances occurs, it will cause changes in the electrical power
Acknowledging the problem of power swing causes nuisance parameters as in following equation.
tripping of distance relay, various power swing blocking schemes P g Eg E1 =Xsin 1
have been developed. Given the availability of various blocking
schemes over the years, this paper aims to present a consolidated Where:
review of the various power swing blocking schemes up to date.
On the other hand, the increasing demand for energy and the Pg Electrical power
pressure on environmental issues over the years gave rise to the Eg Internal voltage
penetration of renewable energy into the power system. However, E1 Load voltage
the impact of large penetration of inverter interfaced renewable XReactance between the generator and the load.
energy generation has yet to be clearly established. Besides Angle that the internal voltage leads the load voltage
beneting the environment and generating renewable energies,
there are looming concerns that the renewable generations may Table 2
Condition and impacts due to power swing occurrence [25].
also signicantly impact the power swing characteristics and
affect the distance relay operation. Consequently, the impact of Condition Impact
modern renewable energy generations to the power swing char-
acteristics will also presented in this paper. Loss of load Loss of synchronism between voltages
The organization of this paper is as follows. Section 2 presents Generator disconnection Loss of synchronism between phase angles
Addition of load Loss of synchronism between phase sequence
the background of power swing phenomenon. The various power Line switching Loss of synchronism between frequencies
swing blocking schemes which forms the main contribution of this

Table 1
Record of blackouts due to zone 3 nuisance tripping [1,47].

Year Affected countries

September 2003 Malaysia Blackout for ve (5) hours including capital Kuala Lumpur
September 2003 Italy major power failure
January 2005 Malaysia Major blackout in northern region of the country
November 2009 Blackout in eight states of Brazil
July 2012 India Blackout in the western grid.
282 M.K. Gunasegaran et al. / Renewable and Sustainable Energy Reviews 52 (2015) 280288

X X Blinder B
outer zone
outer zone
Blinder A
Line
X
inner zone Z
inner zone

Impedance trajectory of
R R
a power swing situation

Load
Fig. 1. Concentric impedance characteristics [35].
R

For example, the occurrence of fault will cause voltage drop (E1
will change), while the opening of the circuit breaker (isolating the
line) will change the reactance, X. Another possible scenario occurs Fig. 2. Blinder scheme characteristics [27].
when a generation unit trips, the required electrical power from the
remaining generators increases. Therefore, when the load on a 3.1. Conventional methods
machine is suddenly increased, the energy furnished by the rotor
results in a decrease in the rotor angular velocity causing severe This section will discuss the types of conventional methods
power swings [1,2,8,1422]. used to block the distance relay from tripping during power swing.
In addition, their advantages and disadvantages are also presented.

2.2. Impacts of power swing

3.1.1. Concentric characteristics
The concentric characteristic is developed such that the inner
At the utility side, when a generator is exposed to power swing,
zone and outer zone for power swing detection in Mho relay is
instability will occur causing the generator to lose synchronism
shown in Fig. 1 [2730].
and experience pole slip [23]. Pole slip condition causes large
During fault, the rate of change of impedance is seen larger
current ow from the external system into the machine in an
than the rate of change of impedance during power swing.
attempt to re-establish the magnetic lock. This can cause electrical
Therefore, the working principle of concentric impedance is based
stress to the system and damage the pole (coil) [23,24]. Table 2
on observing the difference in the rate of change of impedance
shows the impact of power swing phenomenon due to the
(Z) to differentiate between power swing and fault.
different types of disturbances [25].
Two concentric impedances are set together with threshold (K)
The occurrence of power swing will cause uctuations in the
timer that record the duration of impedance travel between the
power, voltage and current. This will cause a drop of impedance
outer and inner concentric circles. If the travel of impedance crosses
measured by the distance relay at transmission line. When the
the outer and inner concentric circles before the set time (K), the
distance relay detects a drop of impedance traveling into its
scheme sees it as a fault. If the impedance crosses the concentric
operating zones, it may mal-operate and cause the circuit breakers
circles for period longer than set time (K), then the scheme sees it as
to trip. This may lead to complicated cascaded outages and
power swing and the relay will be blocked from tripping [10,28,30].
blackouts to the customers.
During slow swing ( o1 Hz), the travel of impedance is very
Since NERC had pointed that the mal-operation of distance
slow and it spends longer or sufcient amount of time in each
relay is one of the causes of major blackouts, it is of utmost
zone [6]. Therefore the scheme can detect the swing and will block
importance for utility engineers to block the distance relays from
the relay from tripping. Zadeh in [30] reported that this scheme
tripping during stable swing. To overcome this issue, researchers
can also detect high resistance ground faults.
have developed a wide array of power swing blocking schemes
However, it was found that this scheme is unable to detect fast swing
over the years which will be presented in the next section.
(45 Hz) due to the fact that the impedance travels very fast across the
different zones [10,27,28,3032]. Zadeh [30] simulated a power swing
event and proved that this scheme is unable to detect fast swing. From
the experiment, Nayak [Link] in [10], concluded that the scheme was also
3. Power swing blocking scheme
unable to detect fast swing during single line to ground fault [10,28].
Several authors have also considered a more complicated
The operating principle of relay is that it should trip only when
scenario where a fault occurs during power swing. Zadeh in [30]
fault or unstable swing occurs. For transmission line protection,
simulated a power swing event and intentionally initiated a fault
the commonly preferred relays are distance relay due to its speedy
during the power swing. Ideally, the fault has to be isolated by
reaction [26]. However, the power swing phenomenon may cause
operating the distance relay. However, it was found that, the
mal-operation of distance relay. Unintended tripping occurs when
scheme does not trip the fault during power swing. In this scenario,
the impedance trajectory enters the impedance locus of the
the relay is unable to reset and subsequently continue to block the
distance relay (usually zone 3). The subsequent operation of
relay from tripping. Adding on to its disadvantage, Zadeh [30] stated
distance protection in response to the power swing will lead to
that this scheme may mal-operate for high impedance fault during
unnecessary outages to consumers. In order to mitigate this
a power swing event by blocking the relay from tripping.
problem, researchers have developed power swing blocking
(PSB) schemes in distance relays to prevent unwanted tripping
during stable swing. The PSB schemes can be generally categorized 3.1.2. Blinder scheme
into conventional methods, signal analysis techniques and intelli- The blinder scheme or commonly known as double blinder was
gence based diagnosis. This section compiles the ndings of all the developed to meet the disadvantages of the concentric characteristic
developed power swing blocking schemes up to date. [33]. The blinders are usually positioned in parallel to the line
 M.K. Gunasegaran et al. / Renewable and Sustainable Energy Reviews 52 (2015) 280288 283

impedance as shown in Fig. 2. Reason being, when out of step distinguish between power swing and fault resulting to mal-
condition occurs during power swing, the impedance vectors will operation of relay. On the other hand, when a three phase fault
enter the protection zones approximately perpendicular to the line occurs at power angle close to 1801, the value of resistance element
impedance. This was discovered to be the most effective way for is already low. Hence, it is difcult to recognize fault during slow
measuring Z to detect the out of step impedance trajectories and fast swing at power angle close to 1801. Moreover, a time delay
[16,27,3336]. Fig. 2 shows the blinder scheme elements. is needed for the detection scheme even when the unblocking
The working principle for blinder is the same as concentric condition is satised due to the initial transient of fault [10].
characteristics. A timer starts when impedance vector crosses
blinder A (outer element) and stops when blinder B (inner
3.1.4. Swing center voltage (Vcos)
element) is crossed. If the measured time is more than the
All the above mentioned schemes are dependent on the line
threshold (K), power swing is detected. Vice versa, a fault is
and bus impedance unlike, the swing center voltage (SCV). SCV is
detected when the measured time is less than the threshold (K).
best represented by using Vcos , where V is the magnitude of
The advantages of blinder scheme are that it can be applied
locally measured voltage and is the angle between locally
independently in each zone, load impedance can lie inside the
measured voltage (V) and current (I) [34].
blinder impedances, and slow swing will not cross blinder A.
During power swing SCV changes continuously, but when fault
Therefore the relay will not trip unnecessarily [16,34,35].
occurs SCV remains constant. This criterion is used to distinguish
Solving the disadvantages of concentric scheme, the blinder
between a power swing and a fault. Relay can be controlled by
scheme comes with its own disadvantages. In order to implement
setting a threshold (k) using the criteria of dVcos/dt. When
the blinder scheme, sophisticated grid analysis is required to identify
dVcos/dt4 k, stable swing is identied and if dVcos/dt less than
the partition between the inner and outer elements. In addition, the
k, then it is an unstable swing or fault [10].
relay may mal-operate for stable swing unless very precise threshold
The advantage of SCV is that it is independent of source
estimation is made to differentiate between a fault and a power
impedances and line impedances. Therefore sophisticated grid ana-
swing event as experimented by Ambekar and Dambhare in [16].
lysis is not required. Besides that, the previously mentioned schemes
require the user to have some understanding on the overall grid and
distance relay congurations. In contrast, SCV is bounded with a
3.1.3. Decreased resistance method
lower limit of zero and an upper limit of one per unit. This allows the
Since the blinder scheme needs thorough grid studies and may
user to successfully apply power swing detection with limited
mal-operate for stable swing, the decreased resistance method
knowledge on the dynamic response of the power system.
was introduced as an alternative solution. Through experiment, it
However, several literatures presented the disadvantages of this
was found that the resistance element of measured impedance
scheme. Nayak [Link] in [10] carried out an experiment and found
seems to show continuous changes during power swing [10,30]. As
that when a three phase fault is initiated, there is some delay in the
for fault, the resistance element changes drastically only during
detection due to the inception of fault. Similar ndings were
the inception of fault and remains the same during the fault
reported by Ambekar and Dambhare in [16]. On the other hand,
period. These characteristics were used to differentiate between a
when power swing occurs at power angle close to 1801, the
power swing and fault for distance relay blocking.
threshold (k) must be at a very low value to detect the power
It must be noted that this blocking scheme can detect fault
swing [16]. Furthermore, when single line to ground fault with high
during power swing as experimented by Zadeh [30]. When a fault
ground resistance was simulated, Nayak [Link] found that the SCV did
occurs during power swing, the scheme will reset and allows the
not remain constant [10]. Consequently, the relay will not trip as the
relay to trip. Zadeh [30], experimented different scenarios and
SCV characteristic is similar to that of a power swing. In addition,
proved that this method works for all the simulated situation. In
Zadeh simulated a three phase fault during a slow swing at a
addition, this detection method does not require as thorough
maximum angle of 1801 and found that the scheme continues to
understanding of the system as in the blinder scheme [37].
block the relay from tripping even after the fault had occurred [30].
However, this detection scheme comes with its unique disad-
vantages. It must be noted that when fault occurs, the resistance
drops signicantly. Nayak [Link] in [10] reported that during swing 3.1.5. Superimposed current method
with low slip frequency at the power angle of 1801, the change in The occurrence of power swing produces continuous superim-
resistance over time (R) drops signicantly. Therefore, a swing posed current when the present value of currents is compared with a
with low slip frequency will be treated as a fault by this detection buffer that is taken two cycles earlier. Fig. 3 shows the working
scheme. The change in resistance during different slip frequencies principle of superimposed current detection technique [38].
was also presented in [30]. Additionally, when a three phase fault A change in the rate of current (I) is detected if the difference in
occurs during slow swing, it will also be difcult for this scheme to current is greater than 5% of the nominal current. Continuous I

Fig. 3. Superimposed current detection technique [38].

284 M.K. Gunasegaran et al. / Renewable and Sustainable Energy Reviews 52 (2015) 280288

measurement for three cycles indicates a power swing condition and frequency will be equivalent to the nominal frequency.
the relay will be blocked from tripping. The advantages of this However, the above literatures did not offer any insights for a
method are that, it will detect fast and slow swing and ensure the value of threshold (K) to be selected. As such, a novel approach to
correct blocking of zones. Besides that, this method is also reported determine the threshold (K) was presented by Mahamedi and Jian
to work efciently for faults occurring during a power swing [29,39]. Gao in [20]. All of these techniques require a proper selection of
However, Nayak [Link] [10] reported that the scheme prevents the threshold value for the fault to be detected [20,32,41,42].
the relay from tripping when three phase fault occurs at power
angle close to 1801 and single phase high resistance ground fault
3.2.2. Wavelet transform (WT)
occurs during the power swing. Reason being the small I in both
Wavelet transform has been used in many power system
scenarios. Similar ndings were reported by Zadeh in [30].
applications [9,43,44]. The WT was introduced to overcome the
disadvantage of FFT where the WT is able to retain the time
3.1.6. Three phase power variation information [45]. However despite this advantage, WT fails in
This scheme, also known as power derivation method is frequency location. Most of the recent research using WT are
proposed by Xiangning [Link] in [40]. During power swing, it is concentrated in addressing this disadvantage.
found that the three-phase active and reactive power changes. The authors in [9,22,43,4547] stated that during power swing,
Both the rate of change of active power (dP/dt) and reactive power the frequency of a system varies over a range around the nominal
(dQ/dt) are proportional to the angular speed ( d/dt) [18]. frequency. Due to the fact that fault causes transient, WT is
During power swing the normalized value of |dP/dt| and |dQ/dt| is efcient for power system transient analysis and feature extrac-
greater than 0.7, while during fault both values drop and stays tion. Bramha in [45] studied the different sampling frequency (fs)
close to zero. The normalized value of |dP/dt | and |dQ/dt| are used used for the WT to detect power swing and symmetrical fault
for the swing detection. during power swing. Bramha claimed that the sampling frequency
It is reported that this method has improved sensitivity in of 40.96 kHz, is ideal for detecting power swing and symmetrical
detecting power swing. According to Xiangning [Link] [40], the fault at any location to capture the rise and fall of the frequency
scheme is not affected by factors like fault inception time and variation. Even though it is better to use higher fs, but most
system parameters. numerical relays cannot accommodate high sampling rate. Thus,
However, several studies presented few limitations in this Bramha developed a relay logic as in [48] to utilize the information
scheme. First, it might operate incorrectly during a stable power captured for numerical relay using fs 40.96 kHz.
swing, because the speed might reach zero when the power angle On the other hand, Chengzong and Kezunovic in [9] developed
nears its maximum limit. Besides that, it only works for symme- a high speed WT algorithm to detect symmetrical fault during
trical faults. In addition, the latest discovery by Esmaeilian [Link] in power swing with fs 10 kHz. This sampling frequency will t
[18] found that, it has a longer delay in detection as compared to most of the presently available numerical relays without the need
SCV due to the transient from the inception of fault. To improve the of further modications.
disadvantages of this scheme, Mahamedi in [32] proposed an In [46], Mahamedi [Link] performed WT on the impedance
enhanced scheme based on this method to detect power swing variation to detect power swing and faults. The authors presented
phenomenon. Esmaelian [Link] in [18] did the latest comparison that a maximum delay of 30 ms was observed for fault detection
between SCV and this three phase power method but the compar- using this method. Using fs as low as 1 kHz, this method is claimed
ison did not include the scheme proposed by Mahamedi in [32]. to be suitable for high voltage (HV) and extra high voltage (EHV)
transmission system [46].
3.2. Signal analysis techniques In contrast, Dubey [Link] [21] used WT to analyze the current
variation for detecting faults of various resistances during power
The parameters that vary during power swing can also be swing. Here, the authors selected fs 20 kHz for the WT. The output
observed through waveforms. When digital signal processing of the WT is then reduced to 10 kHz before being fed into the
(DSP) eld expands, numerous researchers have proposed various numerical relay. Correspondingly, Dubey [Link] proposed an algorithm
methods using DSP to detect power swing. This section will compile to detect fault in the presence of power swing. The latest studies
the methods developed for detecting power swing using DSP. using WT was presented by the authors in [21] through the
introduction of entropy as in [22]. It was found that this method
detects both stable and unstable swing and offers reliable decision to
3.2.1. Fast fourier transform (FFT) analysis
the relay besides its capability to detect fault during power swing.
Various FFT techniques have been proposed by different
authors over the years for power swing detection. Each of the
proposed technique carries its own strengths and weaknesses. 3.2.3. S-transform
Karegar and Mohamedi in [41] used FFT approach based on DC S-transform can be derived from two methods which are
component of the current to detect fault during power swing Gaussian function on short time Fourier transform (STFT) and
occurrence for both symmetrical and asymmetrical fault. In this phase correction on WT [49]. However, most research has only
method, each phase current is measured and scrutinized by using derived it using Gaussian function, where the derivation is as
FFT analysis. The DC component obtained from the FFT analysis stated in [4955]. Most importantly, it has shown to have absolute
will be compared to settable threshold (K). If the threshold is referenced phase information and frequency invariant amplitude
higher than K, then fault is detected. response. Another key feature is its accurate time-frequency
Another approach based on FFT analysis is presented in [32], (amplitude-phase) information [8,52,55].
where only symmetrical faults were discussed. The method proposed Abidin [Link] in [8] introduced an analysis by using S-transform
by the authors in [32] is done by detecting the frequency component to extract information and characterize fault, stable swing, and
of three phase active power using FFT analysis with fundamental of unstable swing. This information is used as input signals to the
50 Hz. The author also demonstrated that during power swing, the probabilistic neural network (PNN) based arrangement scheme as
three phase real power frequency is equivalent to its slip frequency presented in [8] to differentiate between stable and unstable
(fslip). In contrast, when fault occurs, the three phase real power's power swing conditions and to detect fault during power swing.
 M.K. Gunasegaran et al. / Renewable and Sustainable Energy Reviews 52 (2015) 280288 285

The authors also compared the performance of multi-layer per- 3.3.1. Adaptive neuro-fuzzy inference system (ANFIS)
ceptron neural network (MLPNN) with PNN and reported that PNN A fuzzy logic system (FLS) is an intelligent system developed to
performs better than MLPNN in detecting and classifying stable generate its own decision making. It is pre-trained and registered
swing and unstable swing, fault, and fault during power swing. to ensure the correct output. The advantage of FLS is that it can
Mohamad [Link] in [51] introduced the use of S-transform in have many inputs. Very commonly, the FLS composed of ve layers
detecting fault during power swing. This was done using IEEE 9 bus which are, input, fuzziers, rules, inference, and de-fuzziers.
system. This scheme analyses the input signals to distance relay One of the methods to develop a fuzzy network is Adaptive
including the voltage, current, three phase active and reactive power. neuro-fuzzy inference network (ANFIS). ANFIS is an enhanced
It was demonstrated that when fault occurs, S-transform of the active version of FLS with extraordinary features such as, generalization
and reactive power will deliver value larger than 0, and during power capability, noise immunity, robustness, and fault tolerance [6163].
swing the value equals to 0. Zadeh and Li in [63] used the rates of positive sequence
Mohamad [Link] also presented another method to detect symme- impedance, positive and negative sequence current, and the SCV
trical fault during power swing in [54]. Here the input signals (voltage as input signals to ANFIS. In the output layer, the ANFIS has only
and current) to distance relay are calculated and processed using one node, which can have value from 0 to 1. For blocking scheme
S-transform tools to generate the S-matrix. The magnitude of S-matrix classication purposes, the authors pre-specied a threshold (k) of
are as presented by the authors in [54]. From the S-matrix, the 0.5 where in case of power swing, the output will be more than or
minimum value of voltage and current magnitude are plotted. It was equal to 0.5 and the blocking scheme will be activated. When the
reported that during power swing, less samples was observed and output is less than 0.5, it will be identied as fault and the relay
when fault occurs during power swing more samples was observed. will be unblocked to allow it to trip. The ANFIS approach uses
However, the schemes in [51,54] may face detection problem for high Gaussian membership function for fuzzy sets, linear function for
resistance fault. To resolve this disadvantage, the authors proposed a rule ouputs and Sugeno's inference mechanism [63]. The authors
new scheme as in [49] using IEEE 14 bus system as the test case. A presented test results at various fault location, fault inception time,
three phase fault with up to 200 resistances were simulated. The various load, single pole open condition, different slip frequencies
advantage of this developed scheme is that, fault during power swing and the test results are available in [63]. They concluded that the
can be detected at any inception time. scheme was able to operate for high resistance fault, clear the
The latest ndings in this particular area was presented by blocking if fault occur during swing, detect slow and fast swing,
Mohamad [Link] in [55], which applies S-transform analysis on active detect fault during power swing even for fault that occurs at
power at the distance relay. The authors calculated the magnitude of power swing center at 1801, operates during single pole open
S-matrix using equation as stated in [55], a standard deviation , is condition, and reset the relay in case of fault during power swing.
evaluated. The detection criterion is dened as such that, if 41 the On the other hand, Esmaeilian [Link] in [61] presented another
power swing is of unstable type, whereas if r1 the power swing is algorithm using ANFIS. Here, the authors used three inputs which
of stable type [55]. The authors simulated the scenario using IEEE 39 are the rate of change of currents (positive and negative sequence),
bus system. normalized active power, and normalized reactive power. The rst
order Sugeno's fuzzy model were implemented in ANFIS, linear
functions for the rules output and bell functions for membership
functions. The authors concluded that the method was able to
3.2.4. Prony method
detect the feature as in [63]. In addition, this method [61] is not be
Very recently prony method has been introduced in power system
affected by factors such as system parameters, fault inception time,
protection eld [2,56] and it was reported to have better result than
fault position, and pre fault load ow condition.
FFT [15]. In order to estimate the analyzed signal parameters, Prony
converts the complex damped sinusoidal signal into samples in time.
This is done as per the mathematical equations as presented by Lobos
3.3.2. Support vector machine
[Link] in [57].
Support vector machine (SVM) is a new solution for separating
Lotfard [Link] in [56] had proposed an algorithm by using the
functions in classication tasks such as pattern recognition [11]. As
current waveform to detect a three phase fault during power swing.
mentioned in [11,64,65], SVM has been known as a very potential tool
The method is based on the exponential drop in the DC current which
in solving the classication problems. Being introduced by Vapnik [65]
signies fault. The authors simulated a three phase fault during power
from the Statistical Learning Theory (SLT), it is found that SVM
swing and reported that this algorithm can unblock the relay if fault
performs better than Articial Neural Networks (ANN) because, SVMs
occurs during swing [56], thus allowing the relay to trip and isolate
essence follows on Structural Risk Minimization (SRM) while ANN is
the fault.
based on the Empirical Risk Minimization (ERM) [11,12,19,64,66]. The
However, it was found that the algorithm does not unblock the
typical steps for SVM comprised of extracting the input features, and
relay for other types of fault. To overcome this disadvantage, Thakal-
training the SVM classier. The SVM classier can be categorized into
lapeli [Link] in [2] proposed an algorithm based on DC component of
few sections depending on the necessity. For power swing detection,
current waveform. Unlike [56] where low impedance faults were
most researchers divide it into power swing classier, fault classier
discussed, high impedance faults were simulated in this paper and
and ground fault classier for discriminating purposes.
Prony method was used to differentiate between power swing and
Kampeerawat [Link] [64] proposed a new power swing blocking
high impedance faults.
method using a least square support vector machine (LS-SVM)
approach. This approach is claimed to be another signal processing
method to be used with numerical distance relay. The authors
3.3. Intelligence based diagnosis used the current and voltage signal at the relaying point as an
input for LS-SVM. In the paper, the event classications are based
Articial intelligence (AI) is widely used for pattern recognition on the rate of change of different measured quantities such as rate
and classication which is essential for developing protection of change in positive sequence impedance (Zp), rate of change in
techniques. Several AI based methods have been implemented in positive current (Ip), rate of change in negative sequence curr-
the blocking scheme of protection system and have been proven to ent (In) and rate of change in power swing center voltage
not be affected by parameters variation in the system [5860]. (Vscv). Three (3) LS-SVM classication such as LS-SVM1,
286 M.K. Gunasegaran et al. / Renewable and Sustainable Energy Reviews 52 (2015) 280288

LS-SVM2, and LS-SVM3 were developed to identify fault, power Besides that, the application of large scale power converters
swing and ground fault detection respectively. During fault, the produces interharmonic and subharmonic frequencies component
schemes uses fault current signal as input for fault detection. which can contaminate the input voltage and current signal to the
When power swing occurs, the scheme uses Zp and VSCV as distance relay. The current lters (Cosine and Fourier digital lters)
input for detection. For ground fault, the scheme uses zero which are used in distance relay are unable to reject the interharmo-
sequence component of fault current signal to detect ground nics and subharmonics frequency components. As a consequence, an
faults. Therefore, the method uses three-layer LS-SVM classier error estimation of impedance will cause the distance relay to under-
to distinguish power swing and fault in order to ensure the correct reach and overreach. Consequently, Trujillo and Leonowicz designed a
relay operation. The authors have simulated all types of fault, and Butterworth analog lter using the Prony method to mitigate the
concluded that this scheme is able to differentiate between fault, problem [15]. The authors reported that the proposed method when
power swing, and fault during power swing. compared to cosine lter managed to reduce the reach error from
Kampeerawat [Link] proposed another approach in [66], where 210% to 75%, and the operation time of the relay was also reduced
Wavelet Packet Transform (WPT) and LS-SVM were used. The from 15.05% to 0.86%. This nding has opened up new possibilities in
authors generated a double circuit transmission line model and designing new families of digital lters for power system protection.
simulated a power swing by inducing a three-phase fault. Here, the
extracted features were decomposed from the current and voltage
signal at relaying point by using WPT Daubechies based on multi- 5. Further research potentials
resolution analysis. The authors selected fs10 kHz. The training of
LS-SVM results were obtained using wavelet entropy (WPEn) and This section will present the new and signicant information
the rate of change of wavelet packet total energy (WPEtot ). The generated from this review that require further analysis and under-
authors concluded that the proposed method is able to distinguish standing. From the above comprehensive review of all the available
between power swing, fault and fault during power swing. PSB schemes (Section 3), it is observed that the intelligent based
Seethalekshmi [Link] [11] proposed another scheme using two diagnosis seems to be very promising as it can ensure correct relay
separate logic blocks for the summation of AND gate in relay in operation for almost all scenarios. However, it has to be noted that
order to distinguish between power swing and fault using SVM the intelligent based diagnosis only can be used for digital relays.
based classier. The input features are positive sequence magni- Since distance relays of Mho type are still preferred due to its speedy
tude and angle of voltage, positive sequence magnitude of the reaction, continuous research in PSB scheme to enhance the detec-
current, real power, and reactive power. These parameters are tion of fault during swing for Mho type distance relays is important.
selected based on the Principal Component Analysis (PCA) by On the other hand, the integration of inverter driven machines
utilizing the voltage and current signal at the relaying point. The has been reported to produce interharmonic and subharmonic
authors also compared the performance of SVM scheme with frequencies which could contaminate the voltage and current
ANFIS scheme and concluded that the SVM performs better that signals that is being fed into the distance relays, leading to under-
ANFIS due to the selection of the input features using PCA and a reach and overreach problems. This can be overcome by installing a
better generalization capability of SVM. lter at the input of distance relays for both signals. Few recent
On the other hand, Chotani [Link] [19] has proposed another works have reported the performance of Butterworth lter to be
power swing and fault discrimination using SVM based technique. better than other available lters. However, the performance of
The proposed algorithm uses fs 1 kHz at 50 Hz nominal fre- Bessel Low-Pass-Filter compared to other lters is unknown. As
quency and the inputs are the half cycle samples of three phase such, the future research can also be concentrated on the digital
current signals at the relaying point. A half cycle waveform is used lter designs to prevent mal-operation of distance relay.
to distinguish a fault and a power swing [19]. Subsequently, the Next, the integration of inertia-less nature of solar PV generator
necessary steps are taken to block or unblock the distance relay could also present a higher risk of distance relay mal-operation as
from tripping. The accuracy of the algorithm is reported to be compared to a grid consisting entirely of synchronous machines.
98.71% for a half cycle waveform and 98.93% for a full cycle When large scale solar PV generators are connected to the
waveform. However, the authors used a half cycle waveform to transmission system, the swing characteristics during transient
reduce the computing burden and decision making time. Simula- events may be different from the swing in a network consisting
tions were carried out for all types of fault at different distances on entirely of synchronous machines. Given the unknown swing
the line and fault during power swing. The authors also used a characteristics, the integration of large scale solar PV generator
maximum of 50 for ground fault at different lines considering could jeopardize the presently established PSB schemes which will
various factors such as variation in mechanical torque, line isola- lead to mal-operation of distance relay. Thus, it is important to
tion and line switching. It was reported that the algorithm analyze the severity of the swing in a solar PV integrated grid as
delivered promising results under these varied scenarios. well as their impact to the established PSB schemes.

6. Conclusion
4. Impact of renewable generation on power swing
characteristics Acknowledging the severe consequences of power swing occur-
rences, constant progress were observed on the development of
The escalating fuel price has promoted the interest in large new power swing detection techniques to prevent mal-operation of
scale renewable energy generations. However, the penetration of distance relays. As technology grows, new ndings, opportunities
these renewable generations into the grid might signicantly and solutions were developed to mitigate the impact of power
impact the power swing characteristics. As such, few literatures swing to the grid. This paper has reviewed the relevant aspects
have conducted studies in this area. Yusoff [Link] initiated power related to power swing phenomenon. The study evolves from the
swing by simulating a three phase bolted fault using IEEE 9-bus sources of power swing occurrence, its impact to consumers and
system with and without Photovoltaic (PV) generators. The generators, as well as its detection methods to prevent the mal-
authors found that the integration of solar PV generator will lead operation of distance relay. Special attention is paid to the various
to decrease in active power after installing the PV generators [25]. power swing detection techniques on high voltage power systems
 M.K. Gunasegaran et al. / Renewable and Sustainable Energy Reviews 52 (2015) 280288 287

which is used to ensure effective decision making for the distance [17] Hassan LH, Moghavvemi M, Almurib HAF, Steinmayer O. Current state of
relay operation. The various methodology presented by earlier neural networks applications in power system monitoring and control. Int J
Electr Power Energy Syst 2013;51:13444.
researchers revealed that, sophisticated grid analysis and pre- [18] Esmaeilian, A, Ghaderi, A, Tasdighi, M, Rouhani, A. Evaluation and perfor-
trained algorithm is needed for Mho-type and Numerical-type mance comparison of power swing detection algorithms in presence of series
distance relays respectively. Besides that, given the latest develop- compensation on transmission lines. In: proceedings of the 10th international
conference on environment and electrical engineering (EEEIC); 2011. p. 14.
ment in renewable energy generations, the paper has also pre- [19] Chothani NG, Bhalja BR, Parikh UB. New support vector machine-based digital
sented the impact from the integration of renewable energy relaying scheme for discrimination between power swing and fault. IET Gener
generation to the power swing characteristics. It can be concluded Transm Distrib 2014;8:1725.
[20] Mahamedi B, Jian Guo Z. A novel approach to detect symmetrical faults
that the uncertainties involved in the static generator integration
occurring during power swings by using frequency components of instanta-
into a synchronous machine environment have yet to be estab- neous three-phase active power. IEEE Trans Power Deliv 2012;27:136876.
lished. Hence, further research is required to enhance the under- [21] Dubey, R, Samantaray, SR, Tripathy, A, Babu, BC, Ehtesham, M. Wavelet based
standing of the impact of large scale inverter driven generation to energy function for symmetrical fault detection during power swing. In:
Proceedings of the students conference on engineering and systems (SCES);
the power swing characteristics and distance relay behavior. 2012. p. 16.
[22] Dubey R, Samantaray SR. Wavelet singular entropy-based symmetrical fault-
detection and out-of-step protection during power swing. IET Gener Transm
Distrib 2013;7:112334.
Acknowledgment [23] Mozina, CJ. Impact of power system instability on generator protection.
Transmission and distribution conference and exposition (T&D), IEEE
PES2012; 2012. p. 18.
The authors would like to express their appreciation to the [24] Ambekar, VA, Dambhare, SS. Out of step detection using synchronized swing
Centre of Research Grant Management, University of Malaya (UMRG impedance and resistance measurement. In: Proceedings of the 11th interna-
fund: RG299-14AFR), Fundamental Research Grant Scheme (FRGS tional conference on developments in power systems protection, (DPSP);
2012. p. 14.
fund: FP043-2013B) and High Impact Research (HIR grant: [25] Yusoff, N, Abidin, AF. The effect of solar photovoltaic (PV) system to the
H-16001-D00048) for providing nancial support in the form of characteristic of power swing. In: Proceedings of the IEEE 3rd international
research grant. The corresponding author would like to express his conference on system engineering and technology (ICSET); 2013. p. 99102.
[26] Abniki, H, Esfahani, AS, Razzaghi, R, Razavi, MTN. Precise Analysis of power
gratitude to the Ministry of Higher Education, Malaysia for nan-
swing effects on distance relay zones. In: Proceedings of IPEC, Conference;
cially sponsoring him to pursue his Masters of Philosophy through 2010. p. 3737.
MyBrain15 scholarship. [27] Martuscello, L, Krizauskas, E, Holbach, J, Yuchen, L. Tests of distance relay
performance on stable and unstable power swings reported using simulated
data of the August 14 2003 system disturbance. In: Proceedings of the 62th
annual conference for protective relay engineers; 2009. p. 23655.
References
[28] Xiangning L, Zhengtian L, Shuohao K, Yan G. Theoretical fundamentals and
implementation of novel self-adaptive distance protection resistant to power
[1] Bakar AHA, Yatim FM, Yusof S, Othman MR. Analysis of overload conditions in swings. IEEE Trans Power Deliv 2010;25:137283.
distance relay under severe system contingencies. Int J Electr Power Energy [29] Apostolov, A, Vandiver, B. Requirements for testing of power swing blocking
Syst 2009;32:34550. functions in protection IEDs. In: Proceedings of the 64th annual conference for
[2] Thakallapelli, A, Mehra, R, Mangalvedekar, HA. Differentiation of faults from protective relay engineers; 2011. p. 1259.
power swings and detection of high impedance faults by distance relays. In: [30] Khoradshadi-Zadeh, H. Evaluation and performance comparison of power
Proceedings of the IEEE 1st international conference on condition assessment swing detection algorithms. In: Proceedings of the IEEE power engineering
techniques in electrical systems (CATCON); 2013. p. 3747. society general meeting, Vol. 2; 2005. p. 18428.
[3] Chen-Ching, L, Juan, L. Patterns of cascaded events in blackouts. In: Proceed- [31] Abidin Fb. Adaptive distance protection to prevent false relay tripping during
ings of the IEEE power and energy society general meeting conversion and voltage collapse and power swing. Malaysia: Universiti Kebangsaan Malaysia;
delivery of electrical energy in the 21st century; 2008. p. 12. 2010.
[4] Chunyan, L, Yuanzhang, S, Xiangyi, C Analysis of the blackout in Europe on [32] Mahamedi, B. A very fast unblocking scheme for distance protection to detect
November 4, 2006. In: Proceedings of the International power engineering symmetrical faults during power swings. In: Proceedings of the international
conference, IPEC; 2007 p. 93944. power electronics conference (IPEC); 2010. p. 37883.
[5] Conti JP. The day the samba stopped. Eng Technol 2010;5:467. [33] Zadeh, HK, Li, Z. Articial neural network based load blinder for distance
[6] Corsi, S, Sabelli, C. General blackout in Italy sunday September 28, 2003, h. protection. In: Proceedings of the IEEE power and energy society general
[Link]. In: Proceedings of the IEEE power engineering society general meeting conversion and delivery of electrical energy in the 21st Century;
meeting; 2004. p. 1691702. 2008. p. 16.
[7] Pentayya, P, Gartia, A, Saha, SK, Anumasula, R, Kumar, C. Synchrophasor based [34] Fischer, N, Benmouyal, G, Daqing, H, Tziouvaras, D, Byrne-Finley, J, Smyth, B.
application development in Western India. In: proceedings of the 2013 IEEE Do system impedances really affect power swings; applying power swing
Innovative smart grid technologies Asia (ISGT Asia); 2013. P. 16. protection elements without complex system studies. In: Proceedings of the
[8] Abidin AF, Mohamed A, Shareef H. Intelligent detection of unstable power 65th annual conference for protective relay engineers; 2012. p. 10819.
swing for correct distance relay operation using S-transform and neural [35] Holbach, J. New out of step blocking algorithm for detecting fast power swing
networks. Expert Syst Appl 2011;38:1496975. frequencies. power systems conference: advanced metering, protection, con-
[9] Chengzong P, Kezunovic M. Fast distance relay scheme for detecting symme- trol, communication, and distributed resources, PS' 062006; 2006. p. 18299.
trical fault during power swing. IEEE Trans Power Deliv 2010;25:220512. [36] Tziouvaras, DA, Daqing, H. Out-of-step protection fundamentals and advance-
[10] Nayak, PK, Rao, JG, Kundur, P, Pradhan, AK, Bajpai, P. A comparative assessment ments. In: Proceedings of the 57th annual conference for protective relay
of power swing detection techniques. In: Proceedings of the power india joint engineers; 2004. p. 282307.
international conference on power electronics, drives and energy systems [37] Kang, H, Subramanian, S, Hassan, F, Yao, L. Distance protection under extreme
(PEDES); 2010. p. 14. conditions. In: Proceedings of the 10th IET international conference on
[11] Seethalekshmi, K, Singh, SN, Srivastava, SC. SVM based power swing identi- managing the change, developments in power system protection (DPSP);
cation scheme for distance relays. In: Proceedings of the IEEE power and 2010. p. 15.
energy society general meeting; 2010. p. 18. [38] Verzosa, Q. Realistic testing of power swing blocking and out-of-step tripping
[12] Seethalekshmi K, Singh SN, Srivastava SC. A classication approach using functions. In: Proceedings of the 66th annual conference for protective relay
support vector machines to prevent distance relay maloperation under power engineers; 2013. p. 42049.
swing and voltage instability. IEEE Trans Power Deliv 2012;27:112433. [39] Apostolov, AP, Tholomier, D, Richards, SH. Superimposed components based
[13] IEEE. Power swing and out-of-step considerations on transmission lines, IEEE sub-cycle protection of transmission lines. In: Proceedings of the IEEE power
PSRC WG D6; 2005. systems conference and exposition, vol.1; 2004. p. 5927.
[14] Mooney, J, Fischer, N. Application guidelines for power swing detection on [40] Xiangning L, Yan G, Pei L. A novel scheme to identify symmetrical faults
transmission systems. In: Proceedings of the 59th annual conference for occurring during power swings. IEEE Trans Power Deliv 2008;23:738.
protective relay engineers; 2006. p. 10. [41] Karegar, HK, Mohamedi, B. A new method for fault detection during power
[15] Trujillo, GLA, Conde, EA, Leonowicz, Z. Application of the prony method for swing in distance protection. In: Proceedings of the 6th international con-
compensation of errors in distance relays. In: Proceedings of the 12th ference on electrical engineering/electronics, computer, telecommunications
international conference on environment and electrical engineering (EEEIC) and information technology (ECTI-CON); 2009. p. 2303.
l; 2013. p. 56872. [42] DApuzzo M, DArco M. A time-domain approach for the analysis of nonsta-
[16] Ambekar, VA, Dambhare, SS. Comparative evaluation of out of step detection tionary signals in power systems. IEEE Trans Instrum Meas 2008;57:196977.
schemes for distance relays. In: Proceedings of the IEEE fth power india [43] Fu, L, He, ZY, Bo, ZQ. Wavelet transform and approximate entropy based
conference; 2012. p. 16. identication of faults in power swings. In: Proceedings of the IET 9th
288 M.K. Gunasegaran et al. / Renewable and Sustainable Energy Reviews 52 (2015) 280288

international conference on developments in power system protection Proceedings of the IEEE student conference on research and development
(DPSP); 2008. p. 5904. (SCOReD); 2012. p. 710.
[44] Parameswariah C, Cox M. Frequency characteristics of wavelets. IEEE Trans [55] Mohamad, NZ, Abidin, AF, Musirin, I. The use of S-Transform for distance relay's
Power Deliv 2002;17:8004. power swing detection. In: Proceedings of the IEEE jordan conference on applied
[45] Brahma SM. Use of wavelets for out of step blocking function of distance electrical engineering and computing technologies (AEECT); 2013. p. 14.
relays. In: Proceedings of the IEEE Power Engineering Society General Meet- [56] Lotfard S, Faiz J, Kezunovic M. Detection of symmetrical faults by distance
ing; 2006. p. 5. relays during power swings. IEEE Trans Power Deliv 2010;25:817.
[46] Mahamedi, B. A new power swing blocking function based on wavelet [57] Lobos, T, Rezmer, J, Koglin, HJ. Analysis of power system transients using
transform. In: Proceedings of the 2nd international conference on electric wavelets and prony method. In: Proceedings of the IEEE porto power tech;
power and energy conversion systems (EPECS); 2011. p. 16. 2001. p. 4.
[47] Dubey, R, Samantaray, SR, Tripathy, A, Babu, BC, Ehtesham, M. Wavelet based [58] Cho KR, Kang YC, Kim SS, Park JK, Kang SH, Kim KH. An ANN based approach
energy function for symmetrical fault detection during power swing. In: to improve the speed of a differential equation based distance relaying
Proceedings of the students conference on engineering and systems (SCES); algorithm. IEEE Trans Power Deliv 1999;14:34957.
2012. p. 16. [59] Vaidya, AP, Venikar, PA. ANN based distance protection of long transmission
[48] Brahma SM. Distance relay with out-of-step blocking function using wavelet lines by considering the effect of fault resistance. In: Proceedings of the 2012
international conference on advances in engineering, science and manage-
transform. IEEE Trans Power Deliv 2007;22:13606.
ment (ICAESM); 2012. p. 5904.
[49] Mohamad NZ, Abidin AF. S-Transform based technique to detect high
[60] Feilat, EA, Al-Tallaq, K. A new approach for distance protection using articial
resistance symmetrical fault during power swing. In: Proceedings of the IEEE
neural network. In: Proceedings of the UPEC 39th international universities
8th international colloquium on signal processing and its applications (CSPA);
power engineering conference, Vol. 1; 2004. p. 4737.
2012. p. 282286.
[61] Esmaeilian, A, Astinfeshan, S. A novel power swing detection algorithm using
[50] Mohamad NZ, Abidin AF. A new technique for high resistance fault detection
adaptive neuro fuzzy technique. In: Proceedings of the international con-
during power swing for distance relay. AASRI Procedia 2012;2:505.
ference on electrical engineering and informatics (ICEEI); 2011. p. 16.
[51] Mohamad, NZ, Abidin, AF, Munim, WNWA. A new tracking method of
[62] Sanaye-Pasand M, Jafarian P. An adaptive decision logic to enhance distance
symmetrical fault during power swing based on S-transform. In: Proceedings protection of transmission lines. IEEE Trans Power Deliv 2011;26:213444.
of the IEEE international power engineering and optimization conference [63] Zadeh HK, Li Z. A novel power swing blocking scheme using adaptive neuro-
(PEDCO) Melaka, Malaysia; 2012. p. 1416. fuzzy inference system. Electr Power Syst Res 2008;78:113846.
[52] Abidin, AF, Mohamed, A, Shareef, H. Power swing detection for correct [64] Kampeerawat, W., Buangam, W., Chusanapiputt, S. New swing-blocking
distance relay operation using S-transform and neural networks. In: Proceed- methods for digital distance protection using support vector machine. In:
ings of the 2011 5th International power engineering and optimization Proceedings of the international conference on power system technology
conference (PEOCO); 2011. p. 27984. (POWERCON); 2010. p. 16.
[53] Samantaray, SR, Dubey, RK, Tripathy, LN, Babu, BC. Spectral energy function for [65] Vapnik V. The nature of statistical learning theory. New York: Springer; 1995.
fault detection during power swing. In: Proceedings of the international [66] Kampeerawat, W, Buangam, W, Chusanapiputt, S. The swing-blocking meth-
conference on energy, automation, and signal (ICEAS); 2011. p. 16. ods for digital distance protection based on wavelet packet transform and
[54] Mohamad, NZ, Abidin, AF, Munim, WNWA, Yahya, M. Symmetrical fault support vector machine. In: Proceedings of the Asia-pacic power and energy
detection technique during power swing based on S-transform. In: engineering conference (APPEEC); 2012. p. 14.