Electrical Power and Energy Systems 114 (2020) 105384

Contents lists available at ScienceDirect

Electrical Power and Energy Systems

journal homepage: www.elsevier.com/locate/ijepes

Performance evaluation of impedance-based synchronous generator out-of- T

step protection in the presence of unified power flow controller
⁎
Seyed Rasoul Hosseini, Mehdi Karrari , Hossein Askarian Abyaneh
Aboorayhan Building, Depatrment of Electrical Engineering, Amirkabir University of Technology, Tehran, Iran

A R T I C LE I N FO A B S T R A C T

Keywords: Flexible ac transmission system (FACTS) controllers change the transmission line voltage and current signals
Out-of-step protection measured with protection relays. These devices affect the operation of impedance-based relays and can lead to
Unified power flow controller relays under/over-reaching. This study investigates the effects of a unified power flow controller (UPFC) as one
FACTS of the most important shunt-series compensators on the synchronous generator out-of-step (OOS) protection. A
Adaptive protection
detailed model of the UPFC is used for the purpose of the study and its performance has been simulated in three
operational modes of the shunt-series compensator including static compensator (STATCOM), static synchronous
series compensator (SSSC) and UPFC. The results show that all the three modes of the UPFC change the im-
pedance trajectory and disrupt the diagnosis of the OOS protection. Moreover, an analytical approach is de-
veloped to eliminate the negative effects of the devices. The approach only requires the UPFC voltage and
current synchro-phasor data to calculate the modified impedance. The results show that the presented method
properly eliminates the negative effects on the OOS relay and prevents it from under-reaching in all the UPFC
operation modes. The most important advantage of the modified algorithm is that it can be generalized to other
types of FACTS devices.

1. Introduction
Power system stability is an important issue in the power system
operation. In the normal operation of the power system all synchronous
machines are in step with each other and operate at the same average
speed. However, when faults occur in the system, the resulting dis-
turbances cause oscillations in machine rotor angles and consequent
power flow swing in the system. Depending on the severity of the dis-
turbance, a power swing may evolve into a stable or an unstable power
swing. In the unstable power swing, the system cannot return to syn-
chronous operation. The unstable power swing and machine loss of
synchronism are called out-of-step (OOS) condition [1]. The OOS
condition causes cyclic variations in the currents and voltages of the
effected machine, which are damaging the machine. Moreover, the
condition may endanger the stability of the entire system [2]. To avoid
such damages, a reliable generator protection scheme is necessary to
detect the OOS condition. The conventional detection function of the
OOS protection is based on analyzing the measured impedance at the R-
X plane [3,4]. Some traditional methods of power swing and OOS de-
tection such as concentric characteristic and blinder schemes, Rdot
scheme, continuous impedance calculation, swing-center voltage were
summarized in [5]. Among these methods, the blinder schemes have
been more considered. The main advantage of this method is its sim-
plicity. New approaches such as distributed dynamic state estimator
[6,7], synchrophasor-based out-of-step relaying [8,9] and frequency
based out-of-step protection [10] have been presented in recent studies.
The main features of these methods are compatibility and strength
against changes in power system configuration. However, these ap-
proaches are complicated and have a high computational requirement.
One of the main problems with respect to the impedance-based tech-
niques is the change of measured impedance due to the large dis-
turbances [11] and the using of compensators in the transmission line.
Many publications considered the distance relay performance for
different fault types/point with various FACTS devices [12–24]. An-
other studies investigated the effects of FACTS devices on the loss of
excitation (LOE) protection of synchronous generator [25–29]. Results
of these studies showed that the presence of FACTS devices causes a
substantial delay in the performance of LOE relay. Only few articles
have been published on distance relay performance for a compensated
line during a power swing condition [30,31]. The authors in [30] si-
mulated and compared the performance of two power swing detection
algorithms with and without series capacitors. In [31], the impedance
seen by a distance relay during power swing conditions in a compen-
sated line with UPFC was investigated but any modified approach was

⁎
Corresponding author.
E-mail address: karrari@aut.ac.ir (M. Karrari).

https://doi.org/10.1016/j.ijepes.2019.105384
Received 2 November 2018; Received in revised form 23 May 2019; Accepted 24 June 2019
Available online 11 July 2019
0142-0615/ © 2019 Elsevier Ltd. All rights reserved.
S.R. Hosseini, et al.
not proposed.
In this study, the impact of UPFC is analyzed on the synchronous
generator OOS protection. It should be noted; that the power swing
characteristic/logic of distance relay is different from generator OOS
characteristic/logic and must be evaluate, separately. According to
authors’ knowledge this issue has not been addressed in the literatures.
To this end, a gate turn-off thyristor (GTO)-based detailed model of
UPFC is used and different operation modes of the UPFC including
shunt (STATCOM), series (SSSC) and shunt-series (UPFC) mode have
been considered in the simulations. The main contributions of this
study are:
1. The measured impedance by OOS relay is extracted based on gen-
erator internal voltage and external system voltages in the presence
and absence of UPFC.
2. The performance of the OOS relay in the presence of UPFC during
power swing condition is studied and the impact of SSSC, STATCOM
and UPFC is investigated.
3. An modified approach is developed to correct the impedance de-
viation caused by the UPFC
4. The performance of modified approach is validated by simulation
results
During power swing condition the impedance loci seen by the OOS
relay without UPFC are circles [32]. The presence of UPFC changes the
impedance path and causes the measured impedance deviate from
circle Trajectory. Due to this deviation, it is possible; the OOS condition
not to be detected and causes under-reaching of the OOS relay. Using
the analytical investigation and simulation, a new modified approach is
proposed to correct the impedance path. The performance of proposed
approach is validated by the simulation results. Since the proposed
approach considers the effects of all the UPFC operation modes on the
OOS relay operation, it proves to be highly useful in power system
protections.
The rest of this paper is organized as follows: Section 2 describes the
generator impedance-based OOS protection. Section 3 introduces the
principles about the UPFC performance and describes the different
operating modes of the UPFC. In Section 4, the modified impedance-
based OOS protection is presented. Section 5 presents the simulation
results of the loss of synchronism in the presence of the UPFC. Section 6
investigates the performance of the modified approach and the results
are addressed and analyzed. Finally, Section 7 concludes the study.
2. Synchronous generator OOS protection
During power swing in the network, OOS generators must be rapidly
isolated from the power system not only to prevent damage to the
generator, turbine and step up transformer, but also to prevent in-
stability from spreading to other portions of the system. There are
several schemes available for OOS protection, most of which detect a
loss of synchronism by analyzing the measured impedance at the R-X
plane. Two commonly used approaches to OOS protection are the mho
element and the blinder scheme [1].
In order to assess the relay operation, the system shown in Fig. 1(a)
is considered. In this system, the CT and VT of the relay are placed in
the generator terminal and after step up transformer. In this figure, IR ,
VR , E1, E2 and VB1 are the relay current and voltage phasors, the internal
voltage of generator, external system voltage and step up transformer
output bus bar voltage, respectively. The impedance Z1 is the sum of the
generator and transformer reactance and Z2 is the sum of the trans-
mission line and external system impedances. The thevenin equivalent
of the system is at right side of the Bus B1. Fig. 1(b) shows the block
diagram of the impedance-based OOS relay. First, the voltage and
current are sampled using the VT and CT. Then, the samples pass
through the low pass filter to remove high frequency components. After
analog-to-digital conversion, the voltage and current phasors are
2
Electrical Power and Energy Systems 114 (2020) 105384
calculated with the full cycle discrete Fourier transform (FCDFT). The
measured apparent impedance of the relay is obtained by dividing the
voltage phasor into the current phasor. Finally, the impedance path is
obtained by plotting the apparent impedance on the real-image plane.
In this study, we used the blinder scheme with the characteristic of
Siemens 7UM62 protection relay shown in Fig. 2 [33]. The character-
istic boundaries are determined by the setting parameters impedances
Za, Zb, Zc and Zd, which are given in Table 1. In this table xd' , XTR and
Xsys are transient reactance, transformer reactance and system re-
actance, respectively. Moreover, δCRIT is the critical rotor angle and
must be determine by stability analysis such as equal area criterion and
its range is 120–150 degrees. Based on Siemens 7UM62 manual
δCRIT = 120 is chosen. The polygon is symmetrical around its vertical
axis. The power swing polygon is divided into two parts. Characteristic
1 represents the lower section of the rectangle and Characteristic 2
covers the upper hatched area. Depending on the electrical center of the
power swing, or in the vicinity of the power station, the impedance
trajectory passes from the Characteristic 1 or the Characteristic 2.
Due to the symmetrical nature of the power swing, the first condi-
tion for power swing detection is that the positive sequence component
of the current exceeds an adjustable limit I1 while the negative se-
quence current remains below an adjustable value I2. Additionally,
detection of an OOS condition requires that the impedance trajectory
enters a characteristic from one side and exits from the opposite side
(Cases 1 and 2). On the other hand, if the impedance trajectory enters
and exits the same side, in this case, the power swing tends to stabilize
(Cases 3 and 4).
The logic diagram of the out-of-step protection of 7UM62 is shown
in Fig. 3 [33]. When an out-of-step condition is recognized, i.e. when
the impedance vector has passed through a power swing characteristic,
an annunciation is issued which also identifies the crossed character-
istic. Additionally, a counter n1 (for characteristic 1) or n2 (for char-
acteristic 2) is incremented. Out-of-step protection pickup is activated
when a counter reaches to 1. A further out-of-step indication is set for
an adjustable indication time period, each time a counter is in-
cremented. After an adjustable holding time, the pickup resets to zero.
The holding time is started a new each time a counter is incremented. A
trip command is issued when the number of power swing polygon
crossings has reached a selectable number. This command is main-
tained for at least the set time THOLDING. The minimum trip command
duration Tmin TRIPCOM. does not start until the pickup has reset.
3. Principles of the unified power flow controller (UPFC)
performance
The UPFC is a FACTS device that is primarily used for power-flow
control and voltage regulation in transmission networks. In addition, it
can be used to damp low-frequency dynamic oscillations (e.g., inter-
area oscillations and subsynchronous resonance) and enhance transient
stability. Fig. 4(a) presents a single-line diagram of a UPFC with its
main components. The UPFC is composed of two shunt and series
converters with coupling transformers and a common DC link between
the converters. The simplified single-line diagram of the UPFC is shown
in Fig. 4(b). Analytically, the UPFC can be considered as a voltage
source with variable amplitude and phase angle. The corresponding
vector diagram is shown in Fig. 4(c).
The UPFC is composed of three inter-connect subcircuits including
series, shunt and DC link subcircuits. The series converter injects a
voltage in series with the line which, in turn, modifies the voltage at
terminal 2(V2 ). According to the familiar Eqs. (1) and (2), the active and
reactive power at the sending-end can be affected by controlling the
magnitude and angle of V2 . In this way, by injecting Vinj , V2 can take any
values inside the circle of Fig. 3(c). The radius of the circle is given by
the maximum magnitude of the injected voltage (Vinj _max ).
S.R. Hosseini, et al. Electrical Power and Energy Systems 114 (2020) 105384

CT
Z1 IR Z2
VR
VT

OOS Relay

Im(Z)
(a)
Output of CT Phasor V
Positive Re(Z)
Outout of VT A/D Sequence I

Analog to
LPF FCDFT
Digital
(b)
Fig. 1. (a) The Equivalent system for studying OOS protection; (b) The OOS relay block diagram.

Fig. 2. Generator OOS protection using a blinder scheme.
angle values, active power is needed and, the shunt subcircuit provides
active power required by the series subcircuit. In the UPFC mode, the
series injected voltage angle can take any values in the 0 ≤ ρ ≤ 2π
range. The UPFC controllable region is obtained by keeping the injected
voltage to its maximum value (Vinj _max ) and varying its phase angle from
zero to 360 degrees. By plotting P and Q obtained from (1) and (2) for
different values of ρ , the controllable region of the UPFC is obtained as
an elliptic. Fig. 6 shows the UPFC controllable region with Vinj _max = 0.1
pu in the single machine infinite bus (SMIB) test system discussed in [1]
(Fig. 5). This system consists of four 555 MVA generators, which are
modeled with an equivalent generator. This equivalent generator is
connected to the infinite bus through the step-up transformer and the
double-circuit transmission line. Detail information about this is given
in the Appendix.
Before the UPFC installation, the natural active and reactive power
flows in line 1 (j0.5) were 1300 MW and 400 MVAr, respectively. By
UPFC installation the maximum active and reactive power flows
π π
reaches to 1697 MW and 950 MVAr with ρ = 2 and ρ = 4 ., respec-
3π
tively. Also, the minimum power flows reaches to 1050 MW in ρ = 2
3π
and 165 MVAr in ρ = 4
.

Table 1
4. The modified impedance-based out-of-step protection approach
The setting parameters of the OOS relay polygon.
Parameter Setting Fig. 7 illustrates the single-line diagram of the single-machine in-
finite bus system after installing UPFC at the beginning of the trans-
Za (x d' + XTR + Xsys ) / 2
Za ≈ &δCRIT = 120 mission line between the buses B1 and B2. This system consists of a
tan(δCRIT / 2)
Zb synchronous generator connected to an infinite bus by the transmission
xd'
Zc (0.7–0.9) XTR
line. At any time step, the series and shunt parts of UPFC are modelled
Zd XTR + αXsys &α ≈ 0.25 − 0.30 by series impedance Zse and shunt admittance Ysh, respectively. Clearly,
after the installation of UPFC, the measured impedance by the OOS
relay will change, and this change may cause relay maloperation. In this
V2 Vr section, through an analytical approach, the difference between the
P= sinδ
X (1) measured impedance before and after the installation of UPFC is ad-
dressed.
V2
Q= (V2 − Vr cosδ ) andV2 = V1 + Vinj Before the installation of UPFC, the system is similar to Fig. 1(a) and
X (2)
the impedance measured with the OOS relay can be obtained by fol-
The magnitudes and phase-angle difference between Vinj and I2 de- lowing (3)–(5).
termine the active and reactive powers required by the series subcircuit
to inject in the line in order to produce the desired Vinj The UPFC could E1 − E2
IR =
be operated at three SSSC, STATCOM and UPFC modes. In the SSSC Z1 + Z2 (3)
mode, only the series converter is connected to the system and Vinj is
always injected in quadrature with I2 to avoid active power consump-
VR = E1 − Z1 IR (4)
tion/injection. In the case of a UPFC, since Vinj can take any arbitrary

3
S.R. Hosseini, et al. Electrical Power and Energy Systems 114 (2020) 105384

3512 T-SIGNAL
5067
O/S Char. 1

3511 T-HOLDING
5069
Release
UL1 O/S det. Char. 1
UL2
n=0 5071
UL3 U1 Z= U1
I1
Increment
n=n+1 O/S TRIP Char. 1

REP. CHAR. 1 3509

Characteristic 1 Trip Matrix
3512 T-SIGNAL
IL1 I1> 5068
IL2
IL3 I1 O/S Char. 2
I2 &
I2< 3511 T-HOLDING
5070
Release
O/S det. Char. 2
Tmin TRIP CMD
n=0 5071
Increment
n=n+1 O/S TRIP Char. 2

5053 REP. CHAR. 1 3509 5062

Characteristic 2
> O/S BLOCKED O/S BLOCKED

Fig. 3. Logic diagram of the OOS protection [33].

VR E Z + E2 Z1 conversion of the UPFC embedded system to a new SMIB system. This

ZR0 = = 1 2
IR E1 − E2 (5) conversion is necessary because the SMIB test system is used in the
theoretical basis of the OOS protection [1,33].
The impedance ZR0 is the measured impedance by OOS relay before In Fig. 8 (b), the thevenin impedance (Zth) and thevenin voltage
the installation of UPFC. (Eth ) can be calculated by (6) and (7), respectively.
By installing the UPFC at the beginning of the transmission line, the
equivalent circuit of the right side of relay can be considered as the-
venin equivalent circuit (see Fig. 8(a) and (b)). Fig. 8 shows the

PS , QS Pr , Qr
UPFC V2
Ess
E V1 Vr Er
Zs Ise ZL Zr

Ish IL
Sending End Shunt Receiving End
DC Series
Converter Link Converter

DC DC
AC AC

Vinj VL2
(a) V2

V1 Vinj V2 VL2 Vr V1
I2
jX
I2

Vr
(b) (c)
Fig. 4. The working principle of the UPFC: (a) the one-line representation of the UPFC; (b) the simplified single-line diagram; and (c) the corresponding vector
diagram.

4
S.R. Hosseini, et al. Electrical Power and Energy Systems 114 (2020) 105384

HT EB UPFC
LT Bu1 Bu2
Trans. UPFC
4×555 j0.5
Infinite
MVA Bus
j0.15
Et j0.93
OOS Relay
F
P
Q

Fig. 5. Single machine infinite bus system.

(a) (b)

4
2
2
7
4

V 0 OOS Relay
inj 3
4
3
(c)
2
Fig. 8. (a) The network side of Fig. 7; (b) The equivalent circuit of the network
side of Fig. 7; (c) The equivalent network after installing UPFC.
5
4
Zse E 1 Z2 E 1 + Z1 E2
ZR = +
(Ysh (Z2 + Zse ) + 1) E1 − E2 (Ysh (Z2 + Zse ) + 1) E1 − E2

Fig. 6. The controllable region of the UPFC installed at the SMIB system. By multiplying the sides of the above equation in
[(Ysh (Z2 + Zse ) + 1) E1 − E2], the above equation can be written as fol-
lows:

ZR [(Ysh (Z2 + Zse ) + 1) E1 − E2] − Zse E 1 = Z2 E 1 + Z1 E2

Now, by dividing the sides of the above equation into E1 − E2 the

following equation is obtained:

[(Ysh (Z2 + Zse ) + 1) E1 − E2 ] E1 Z E + Z1 E2

ZR − Zse = 2 1
E1 − E2 E1 − E2 E1 − E2

According to the Eq. (5), the right side of above equation is equal to
the measured impedance before installation of UPFC (ZR0 ). Therefore,
the modified impedance is obtained as follows:

Fig. 7. The single-line diagram of the studied power system. [(Ysh (Z2 + Zse ) + 1) E1 − E2 ] E1
ZR0 = ZR − Zse
E1 − E2 E1 − E2 (10)
(Zse + Z2 ) V − VB2
Zth = (Z2 + Zse )||Ysh = &Ysh VB1 = Ish, Zse = B1 In the STATCOM operation mode, Zse → 0 and the modified im-
Ysh (Z2 + Zse ) + 1 Ise pedance obtains as follows:
(6)
E1
1 ZR0 = ZR (1 + Ysh Z2 )
Eth = E2 E1 − E2 (11)
Ysh (Z2 + Zse ) + 1 (7)
Moreover, for the SSSC operation modeYsh → 0
By comparing Fig. 8(c) and Fig. 1(a), according to the Eq. (5), the
measured impedance after the UPFC installation (ZR ) is obtained as E1
ZR0 = ZR − Zse
follows: E1 − E2 (12)

VR E Z + Eth Z1 Eq. (10) was obtained by assumption the sampling point of the
ZR = = 1 th
IR E1 − Eth (8) voltage and current at the beginning of the transmission line and after
the step up transformer. If the sampling point placed at the generator's
the procedure of obtaining the measured impedance is followed by bus and before the transformer, with similar procedure, the modified
the substitution of the corresponding Zth and Eth values from (6) and (7) impedance is obtained from the following equation.
into (8) as follows:
E1 E2 E1
E1 ( Y
(Zse + Z2 )
) + (Y
1
E2 ) Z1 ZR0=(ZR − Ztr )[Ysh (Z2 + Zse ) + 1] − ZR − Zse
ZR = sh (Z2 + Zse ) + 1 sh (Z2 + Zse ) + 1
⇒ E1 − E2 E1 − E2 E1 − E2
1
E1 − ( Y E2 ) (9) (13)
sh (Z2 + Zse ) + 1

It is worth mentioning that the UPFC is a non-linear element and it’s

(Z2 + Zse ) E 1 + Z1 E2
ZR = ⇒ modelling with Zse and Ysh in the above equations does not mean linear
(Ysh (Z2 + Zse ) + 1) E1 − E2 assumptions. These calculations should be done at any time step. In

5
S.R. Hosseini, et al. Electrical Power and Energy Systems 114 (2020) 105384

other words, the series and shunt part of the UPFC are modeled as time
variant impedance and admittance, respectively. According the
substitution theorem this modelling is approvable, because the bran-
ches voltage and current do not change.

5. The out-of-step (OOS) protection in the presence of the unified

power flow controller (UPFC)

The SMIB test system discussed in Section 3 was used as a case study
to evaluate the performance of the OOS relay function in the presence
of the UPFC. The UPFC is embedded at the beginning of the one circuit
of the transmission line as shown in Fig. 5. The UPFC consists of two Fig. 10. Rotor angle variations during the OOS condition.
200-MVA, three-level, 48-pulse GTO-based converters with one being
connected to the shunt at the bus HT and the other being connected in
series between the buses Bu1 and Bu2. The shunt and series converters
can exchange power through a DC bus. Detailed information of the
system is given in appendix. The UPFC was used to control the active
and reactive powers flowing through this circuit while controlling
voltage at the bus HT.
First, we assumed that the UPFC was not in grid and the active and
reactive powers naturally flew in the lines. The natural active and re-
active power flow in the first line (UPFC embedded in Fig. 5) were
1300 MW and 415 MVAr, respectively. A three-phase to ground short
circuit fault occurred at the beginning of the second line (the point F) at
t = 1 s, and at t = 1.1 s, the second line was disconnected from the
network. The impedance path between the time interval t = 1.1 to
t = 1.72 is shown in Fig. 9. After the fault was cleared, the generator
accelerated and the δ angle increased. This situation caused the im-
pedance path to moves the left side and leave the left side of char-
acteristic 1 at t = 1.67 s. Rotor angle variations during the fault interval
and after that are shown in Fig. 10. It can be observed that after clearing
the fault, the rotor angle was not stable and fluctuated between 180∘ and
− 180∘ , indicating that the synchronism was lost. When the syn-
chronism was lost, the voltage and current were oscillating and the
measured impedance moved in circular path. Two rounds of impedance
paths are shown in Fig. 11. Fig. 11. Next rounds of the impedance trajectory.
By adding the UPFC to the system, the dynamic response of the
system is changed. The UPFC converters can be operated in three series,
shunt and series-shunt modes. In the series mode, the UPFC operates as
a SSSC and controlling injected voltage while keeping the injected
voltage in quadrature with current. The effect of the SSSC operation
mode has been investigated on the detection of the OOS relay in the
same fault conditions. Fig. 12 shows the comparison of the measured
impedance path for different injection voltages. By comparing Figs. 12

Im(Z)

Electrical Center
t=1.72 t=1.43 t=1.11
t=1.67 t=1.5

Re(Z)

Fig. 12. The measured impedance trajectory in the presence of the SSSC.

and 9, it is clear that the SSSC caused the impedance path deviations. It
also changed the intersection point of the impedance path with the
imaginary axis and OOS characteristics. As shown in Fig. 12, the im-
pedance path entered Characteristic 1 from right side and crossed the
imaginary axis inside Characteristic 1. This intersection point defined
Fig. 9. The impedance locus after the fault at the time interval t = 1.11 to
the characteristic selection of the OOS function. In this situation, to
t = 1.72.

6
S.R. Hosseini, et al.
Fig. 13. Next rounds of the impedance trajectory in the presence of the SSSC.
detect the OOS, the impedance path must pass through the left side of
Characteristic 1. However, the SSSC dynamic response changed the
direction of the impedance path upwards, left Characteristic 1 from the
upper side and disrupted the OOS detection function. As shown in
Fig. 12, increasing the percentage of the injected voltage increased the
impedance path deviation. Due to the intersection point of the ima-
ginary axis inside Characteristic 1, with the passage of the impedance
trajectory from the left side of Characteristic 2 (Vinj = 0.25, 0.05 and
0.075 pu), the OOS was not detected.
In the presence of SSSC (FACT devices), the shape of the impedance
path is not round that is shown in Fig. 13 for Vinj = 0.1 pu.
The shunt mode operates as a STATCOM and controlling voltage at
the bus Bu1. The STATCOM is operating in two modes of voltage reg-
ulation (VR) and reactive power (Var) control. The voltage regulation
mode maintains the voltage at a set point by generating positive or
negative reactive power. In the Var control mode, according to the set
point, the positive or negative reactive power is generated by
STATCOM.
Similar to the SSSC mode, the STATCOM mode also changes the
direction of the impedance after the same fault condition, as discussed
in the beginning of this section. Fig. 14 shows the comparison of the
measured impedance path for different STATCOM operation modes.
Given the low capacity of the STATCOM compared to the system ca-
pacity, the change in the operating point has almost identical results.
Fig. 14. The measured impedance trajectory in the presence of STATCOM.
7
Electrical Power and Energy Systems 114 (2020) 105384
However, the zoomed-out image shows that at the zero and negative
operation points, the amount of the impedance path diversion is larger.
Moreover, the STATCOM in the voltage control mode has a similar
response to the Var control mode with the Qref = 1 pu. This is due to
the low voltage of Bu1 (0.95 pu). Therefore, the STATCOM works with
full capacity (Q = 1 pu) and increases the voltage up to 0.97pu.
In the series-shunt (UPFC) mode, the shunt converter operates as a
STATCOM. It controls the voltage of the Bu1 bus by controlling the
absorbed or generated reactive power while allowing active power
transfer to the series converter through the DC bus. Fig. 15(a) shows the
independent active and reactive power flow control at the transmission
line 1. Initially, the active and reactive power was set at 6.5 pu
(1300 MW) and 2 pu (400 MVAr), respectively (natural power flow).
Then, at t = 0.4 s and t = 0.7 s, the reactive and active power reference
increased to 7 pu and 3 pu, respectively; the line 1power flow followed
this change. The power flows of the transmission lines are shown in
Fig. 15(b). It is clear that by increasing line 1 power flow, the line 2
power flow decreases.
Thus, the UPFC changes the natural power flow of the line. In the
event of a fault, the UPFC attempts to maintain the active and reactive
power flows at the regulated set points, and due to the great increase of
the controller errors, the controller could not regulate the powers at the
set points. In this condition, the dynamic response of the UPFC changes
the voltage and current profiles, and subsequently changes the mea-
sured impedance at the generator terminal. To illustrate this issue, the
effect of the UPFC was investigated on the detection of the OOS relay in
the same fault conditions. Before the fault occurred, the active and
reactive set points were 6.5 and 2 pu, respectively (Per unit is based on
the UPFC base power). The impedance trajectory after fault clearing
between t = 1.46 s and t = 2.67 s is shown in Fig. 16. The first round of
the impedance path crosses the imaginary axis inside the characteristic
2. Therefore, the characteristic 2 is selected as OOS operation char-
acteristic. Then, the path enters into Characteristic 1 from upper side
and exits from the left side. Since the characteristics of the exit point
and the OOS function are not identical, the OOS is not detected. For the
next rounds, the situation is worse as the impedance path crosses the
imaginary axis outside the Characteristic 2 and the OOS function is not
enabled.
According to changes at network topology after occurrence of the
fault or the UPFC operation point, the effect of UPFC may be different.
In Fig. 17, the measured impedance path is shown in three operating
points of UPFC. It is clear that by increasing the active and reactive
power reference (Pref = 7 pu and Qref = 2.5 pu), the impedance de-
viation increased.
6. Performance evaluation of modified approach
In this section the performance of the modified impedance-based
approach presented in the section four is investigated. The modified
OOS function block diagram is illustrated in Fig. 18. The modification
process is based on the equations obtained for the modified impedance
in different modes of the UPFC. It should be noted that the im-
plementation of this corrective method requires the UPFC voltage and
current phasor data. For the UPFC installed in the power plant sub-
station, these data are available and can be easily used in the correction
algorithm. If the UPFC is installed far away from the power plant, then,
the voltage and current synchronous phasors must be sent to the power
plant through the communication channel. Nowadays, with the advent
of the wide-area measurement system (WAMS), this possibility is pro-
vided by phasor measurement units. For comparison of the presented
modification results and the conventional OOS function, the same fault
condition discussed in the previous section is repeated and the relay
operation is investigated.
First, the UPFC operated in the SSSC mode and its impact is tested
on the OOS protection. Fig. 18 shows the modified impedance path
after the fault in different injected voltages. It is observed that the
S.R. Hosseini, et al. Electrical Power and Energy Systems 114 (2020) 105384

Fig. 15. The active and reactive power flow: (a) the UPFC response to change of Pref and Qref, (b) the active and reactive power flow of the transmission lines 1 and 2.

modified impedance path enters into the Characteristic 1 from the right
side and exits from its left side. This method corrects the impedance
path deviation in Fig. 11 and resolves the problem of non-detection of
the OOS relay. Although the modification solves the problem of the
impedance path diversion from Characteristic 1, however, in compar-
ison with Fig. 8, this path is uneven and has deviations. This deviation
is due to the changes in the E1 and E2 values after installing the UPFC.
In the Eqs. (10), (11) and (12) the modified path is obtained by em-
bedding the values of Zth and Eth in Eq. (8). However, this correction has
good results and modifies the impedance path for other UPFC function
modules, correctly.
t=2.67 The modified impedance path in the presence of SSSC is shown in
t=1.46
Fig. 19. By comparing this figure with Fig. 12, it is clear that the
modified approach detect OOS condition, properly. Also, the compar-
ison of modified and measured impedance paths in the STATCOM mode
is shown in Fig. 20. It is clear that in this mode, the correction of im-
pedance leads to the correct diagnosis of the OOS. The modified im-
pedance path in the presence of the UPFC between t = 1.46 s and
Fig. 16. The measured impedance trajectory in the presence of the UPFC.
t = 2.67 s is shown in Fig. 21. The active and reactive set points were
6.5 and 2 pu, respectively. It can be seen that the measured impedance
path in Fig. 16 was corrected in this figure. Moreover, the next rounds
of the measured impedance path in Fig. 16, which have a worse con-
dition, have been improved in the modified impedance path. In general,
it can be stated that using the modified OOS algorithm, the loss of
synchronism is detected in the presence of the UPFC without any pro-
blems in all the operating modes.
Table 2 shows the results of the OOS relay operation in the pre-
sence/absence of the UPFC. In this table, the fault occurring and
clearing times (t fault , tclear ), the crossing time of the rotor angle from 180
degree (tδ = 180 ) and the detection time of the relay (tdet ) are presented.
Here, the ‘ND’ stands for not detected. The values of the tδ = 180 column
indicates that the stability of the system increases in the SSSC mode
until it stabilizes in the tclear = 1.1 s. for Vinj = 0.75 pu. Therefore, the
fault clearing time is considered 1.11 s.
Without the UPFC, the OOS protection function detected the loss of
synchronism at t = 1.67 s. The presence of the UPFC causes the OOS not
to be detected by this function in the all modes. However, the modified
function, detects the OOS in the all operation modes, but the detection
times (DTs) are different. In the SSSC mode, the DT is greater than
1.67 s (t > 1.67). This is due to the further stability of the system. In
Fig. 17. The measured impedance path deviation in different operation points the STATCOM mode, the DT time increases for the positive reactive
of the UPFC. power and decreases for the negative reactive power. After the fault

8
S.R. Hosseini, et al.
Fig. 18. The modified OOS relay block diagram.
Fig. 19. The modified impedance path in the presence of SSSC.
Fig. 20. Comparison of the measured and modified impedance paths in the
presence of STATCOM.
clearing, the positive reactive power injection improves the stability
with the STATCOM, while the negative reactive power injection
9
Electrical Power and Energy Systems 114 (2020) 105384
Fig. 21. The modified impedance path in the presence of UPFC.
decreases it. The UPFC mode also decreases the detection time and
stability. This issue intensifies by increasing the line loading.
It worth to mentioning, the proposed method is robust during stable
power swing. Because of the existence of internal voltage (E1) and ex-
ternal system voltage (E2) in the modified equation, this method has
correct detection in the all stable and unstable PSW conditions. The
modified impedance trajectory under stable power swing is shown in
Fig. 22. It is clear that the impedance trajectory is outside of char-
acteristics and the OOS relay does not pickup in this condition. The
corresponding rotor angle of stable power swing is also shown in
Fig. 23.
7. Conclusion
In this study, the effect of the UPFC has been investigated on the
impedance-based OOS protection function, and a novel modification
has been proposed. The UPFC was studied in three operation modes
including series (SSSC), shunt (STATCOM) and shunt-series (UPFC),
and its effects have been investigated. In the presence of UPFC, when
the fault duration is greater than the critical fault clearing time, the
generator would be unstable and the OOS relay operation would be
affected by the UPFC. In this situation, all the operation modes of the
UPFC caused the OOS not to be detected by the relay. In the SSSC mode,
S.R. Hosseini, et al. Electrical Power and Energy Systems 114 (2020) 105384

Table 2
The comparison of important times of the OOS relay.
UPFC MODE Impedance type t fault (s) tclear (s) tδ = 180 (s) tdet (s)

No FACTS Measured 1 1.1 1.5 1.67

SSSC Vinj = 0.025 pu Measured 1 1.1 1.6165 ND
Modified 1 1.1 1.755
Vinj = 0.05 pu Measured 1 1.1 1.711 ND
Modified 1 1.1 1.855
Vinj = 0.075 pu Measured 1 1.11 1.625 ND
Modified 1 1.11 1.77
Vinj = 0.1 pu Measured 1 1.11 1.7085 ND
Modified 1 1.11 1.865
STATCOM Qref = 1 pu Measured 1 1.1 1.5215 ND
Modified 1 1.1 1.675
Qref = 0 pu Measured 1 1.1 1.473 ND
Modified 1 1.1 1.602
Qref = −1 pu Measured 1 1.1 1.4423 ND
Modified 1 1.1 1.565
Vref = 1 pu Measured 1 1.1 1.521 ND
Modified 1 1.1 1.675
UPFC Pref = 1200 MW Measured 1 1.1 1.4515 ND
Qref = 300 MVAr Modified 1 1.1 1.565
Pref = 1300 MW Measured 1 1.1 1.427 ND
Qref = 400 MVAr Modified 1 1.1 1.545
Pref = 1400 MW Measured 1 1.1 1.384 ND
Qref = 500 MVAr Modified 1 1.1 1.495

improved; however under this condition, the amount of impedance

deviation is increased. The both operation modes of STATCOM (voltage
regulation and reactive power control mode) have almost same effect
on the measured impedance and lead the relay to under reaching. The
operation point of the UPFC also, affected the impedance path, and this
effect is increased by increasing the loading. To solve these problems,
modified approach is used to eliminate the effect of the UPFC. This
approach requires the UPFC voltage and current phasor data to be sent
to the relay location via the communication channel. The results de-
monstrate that the proposed modified method improves the accuracy of
the OOS impedance-based function and helps to the correct diagnosis
when the UPFC is in service. According to this study, it is essential to
consider the absence or presence of UPFC when the OOS characteristic
is setting. Also further studies should be done to investigate the impact
of other FACTS devices On the OOS protection.

Declaration of Competing Interest

Fig. 22. The modified impedance trajectory during stable power swing.
The authors declared that there is no conflict of interest.

Fig. 23. The rotor angle of generator during stable power swing.

by increasing the injected voltage, the stability of the system is

Appendix

A list of some sample system parameters are given in below:

See Table 3.

10
S.R. Hosseini, et al.
Table 3
Generator parameter.
Parameter Value (pu)
Xd 1.81
Xq 1.76
Xd' 0.30
Xq' 0.65
Xd'' 0.23
Xq'' 0.25
Xl 0.15
Transformer: ST = 2220 MVA, 24/500 kV, R1 = 0.002, L1 = 0, R2 = 0.002, XL2 = 0.15 (p.u.), Rm = Lm = 500 p.u.,
D1/Yg winding connection.
UPFC:
Shunt controller (STATCOM): nominal power = ± 200 MVA, four phase shifting transformers: ST = 200/4 MVA and 6/
15 kV, DC link capacitors: Cp = Cm = 2500F, Droop = 0.01, Voltage controller: Kp = 5, Ki = 4. Current controller:
Kp = 5, Ki = 40.
Series controller (SSSC): nominal power = ± 200 MVA, maximum injection voltage: 10%, four phase shifting trans-
formers: ST = 200/4 MVA and 0.6/0.6 kV, DC link capacitors: Cp = Cm = 2500F, Droop = 0.01, Voltage controller:
Kp = 5, Ki = 4. Current controller: Kp = 5, Ki = 40.
References
[1] Kundur P. Power system stability and control. New York: Mc-Grow Hill; 1994.
[2] Alinezhad B, Kargar HK. Out-of-step protection based on equal area criterion. IEEE
Trans Power Del 2017;32(2):968–77.
[3] Anderson PM. Protective schemes for stability enhancement, in power system
protection. New York: IEEE Press; 1998.
[4] Reimert D. “Loss of Synchronism,” protective relaying for power generation sys-
tems. New York: CRC Press Taylor & Francis Group; 2006.
[5] IEEE. Power swing and out-of-step considerations on transmission lines. IEEE Power
System Relaying Committee Working Group D6; 2005.
[6] Farantatos E, Cokkinides GJ, Meliopoulos AP. A predictive generator out-of-step
protection and transient stability monitoring scheme enabled by a distributed dy-
namic state estimator. IEEE Trans Power Del 2016;31(4):1826–35.
[7] Cui Y, Kavasseri RG, Brahma SM. Dynamic state estimation assisted out-of-step
detection for generators using angular difference. IEEE Trans Power Del 2017;32:3.
[8] Ivankovic I, Kuzle I, Holjevac N. Multifunctional WAMPAC system concept for out-
of-step protection based on synchrophasor measurements. Int J Electr Power Energy
Syst 2017:77–87.
[9] Ivankovic I, Kuzle I, Holjevac N. Wide area information-based transmission system
centralized out-of-step protection scheme. Energies 2017;10:633.
[10] Zhang S, Zhang Y. Characteristic analysis and calculation of frequencies of voltages
in out-of-step oscillation power system and a frequency-based out-of-step protec-
tion. IEEE Trans Power Syst 2019;34:1.
[11] Sorrentino E, Navas G, Orea E. Effect of an additional large disturbance during
power swings on the impedance seen by the out-of-step blocking function. Int J
Electr Power Energy Syst 2018;99:79–84.
[12] Chaoying C, Yong C, Xuefei D, Liyi C. Study of protective relay's behaviors in
Unified Power Flow Controller of FACTS. In: Proc IEE fourth international con-
ference on advances in power system control, operation and management,
APSCOM-97, 2; 1997. p. 632–7.
[13] Dawei F, Chengxue Z, Zhijian H, Wei W. The effects of flexible AC transmission
system device on protective relay. In: Proc IEEE international conference on power
system technology, POWERCON 2002, 4; 2002. p. 2608–11.
[14] Dash PK, Pradhan AK, Panda G, Liew AC. Digital protection of power transmission
lines in the presence of series connected FACTS devices. IEEE Trans Power Deliv
2000;15(1):38–43.
[15] Dash PK, Pradhan AK, Panda G, Liew AC. Adaptive relay setting for flexible AC
transmission systems (FACTS). IEEE Trans. Power Del. 2000;15(1):38–43.
[16] El-Arroudi K, Joos G, McGillis DT. Operation of impedance protection relays with
the STATCOM. IEEE Trans Power Del 2002;17(2):381–7.
[17] Khederzadeh M. The impact of FACTS device on digital multifunctional protective
relays. In: Proc IEEE/PES transmission and distribution conf and exhib 2002: Asia
11
Electrical Power and Energy Systems 114 (2020) 105384
Parameter Value (pu)
Ra 0.003
'
Tdo 8.0
'
Tqo 1.0
''
Tdo 0.03
''
Tqo 0.07
H 3.5
KD 0
Pacific, vol. 3, Oct. 6–10; 2002. p. 2043–8.
[18] Zhou X, Wang H, Aggarwal RK, Beaumont P. Performance evaluation of a distance
relay as applied to a transmission system with UPFC. IEEE Trans Power Deliv
2006;21–3:1137–47.
[19] Khederzadeh M. UPFC operating characteristics impact on transmission line dis-
tance protection. In: Proc 2008 IEEE/PES general meeting; 20-24 July 2008. p. 1–6.
[20] Seethalekshmi K, Singh SN, Srivastava SC. Adaptive distance relaying scheme in
presence of UPFC using WAMS. In: Proc 2009 IEEE/PES power systems conference
and exposition, PES'09; 15-18 March 2009. p. 1–6.
[21] Pandey SP, Tripathy M. Impact of UPFC on distance relay: a case study. In: Proc
IEEE int conf power, control embedded syst, Nov. 29–Dec. 1; 2010. p. 1–4.
[22] Jamali S, Kazemi A, Shateri H. Modified distance protection in presence of UPFC on
a transmission line. In: Proc 9th IET int conf AC DC power transm, Oct. 19–21;
2010. p. 1–6.
[23] Khederzadeh M, Ghorbani A. STATCOM modeling impacts on performance eva-
luation of distance protection of transmission lines. Eur Trans Elect Power Eng
2011;21(8):2063–79.
[24] Khederzadeh M, Ghorbani A. Impact of VSC-based multiline FACTS controllers on
distance protection of transmission lines. IEEE Trans Power Del 2012;27(1):32–9.
[25] Elsamahy M, Faried SO, Sidhu T. Impact of midpoint STATCOM on generator loss of
excitation protection. IEEE Trans Power Del 2014;29(2):724–32.
[26] Ghorbani A, Mozafari B, Soleymani S, Ranjbar AM. Operation of synchronous
generator LOE protection in the presence of shuntFACTS. Electr Power Syst Res
2015;119:178–86.
[27] Ghorbani A, Soleymani S, Mozafari B. A PMU-based LOE protection of synchronous
generator in the presence of GIPFC. IEEE Trans Power Del Apr. 2016;31(2):551–8.
[28] Raghavendra K, Nangrani SP, Bhat SS. Modeling of operation of loss of excitation
relay in presence of shunt FACTS devices. In: 6th International conference on power
systems; 2016. p. 1–6.
[29] Yaghobi H. A new adaptive impedance-based LOE protection of synchronous gen-
erator in the presence of STATCOM. IEEE Trans Power Del Dec.
2017;32(6):2489–99.
[30] Moravej Z, Pazoki M, Khederzadeh M. Impact of UPFC on power swing character-
istic and distance relay behavior. IEEE Trans Power Del 2014;29(1):261–8.
[31] Esmaeilian A, Ghaderi A, Tasdighi M, Rouhani A. Evaluation and performance
comparison of power swing detection algorithms in presence of series compensation
on transmission lines. In: Proc 10th int conf environment elect eng, May 8–11;
2011. p. 1842–8.
[32] Poage FC, Streifus CA, MacGregor DM, George EE. Performance requirements for
relays on unusually long transmission lines. AIEE Trans 1943;62(6):275–83.
[33] Multifunctional Machine Protection 7UM62 Manual, 4.6 ed. Siemens Co.; 2010. p.
165-74.