sustainability

Article
A Reactive Power Compensation Strategy for Voltage
Stability Challenges in the Korean Power System
with Dynamic Loads
Yunhwan Lee and Hwachang Song *
Department of Electrical and Information Engineering, Seoul National University of Science and Technology,
Seoul 01811, Korea; [email protected]
* Correspondence: [email protected]; Tel.: +82-2-970-6403




Received: 30 November 2018; Accepted: 7 January 2019; Published: 10 January 2019


Abstract: This paper describes a methodology and specifics for technical studies on fault-induced
delayed voltage recovery (FIDVR) mitigation to ensure power system reliability. Optimal locations of
the dynamic volts-ampere-reactive (VAR) sources are determined for addressing the FIDVR issues in
the voltage stability analysis and assessment methodology. We propose a voltage stability analysis
method for planning dynamic VAR sources for bettering electric power transmission systems under
contingency conditions. A time-domain dynamic simulation is performed to assess short-term
voltage stability. While conducting dynamic simulations, sensitivity analysis is performed to assess
the need for dynamic VAR sources. This study focuses on a reactive power compensation strategy to
determine system voltage recovery performance by optimal flexible alternating current transmission
system (FACTS) placement in a metropolitan region. The objective of this study is to determine the
optimal installation of dynamic VAR sources while satisfying the requirements of voltage stability
margin and transient voltage dip under a set of criteria. New insights are presented on the effect of
FACTS controls on the reactive power compensation, which supports voltage recovery. The main
features of the proposed method are (i) the development based on a load model for FIDVR, (ii) the
use of sensitivity analysis of the network to the variations of the IM load, (iii) the establishment of
the control function and compensation strategy to maintain the voltage of system within criteria
limits, and (iv) the use of the sensitivity analysis based on branch parameterization for unsolvable
cases. Case studies on the Korean power system validated the performance of the proposed strategy,
showing that it effectively installed FACTS under contingency scenarios.

Keywords: fault induced delayed voltage recovery; reactive power compensation; sensitivity analysis;
transient stability; voltage criteria; voltage stability analysis

1. Introduction
Modern power systems are becoming increasingly complex and dynamic owing to an increased
use of induction motor (IM) loads. The increasing usage of these loads with low inertia characteristics
make power systems more vulnerable to short-term voltage stability (STVS). Such problems might be
aggravated by the growing use of single-phase heating, ventilation, and air conditioner (HVAC) units
with high-efficiency and low-inertia motors. When a fault occurs on the load side, where there is a
high proportion of IM loads, a large power vacancy is induced and can cause a widespread voltage to
drop. Additionally, the demand for reactive power increases significantly at large IM loads. Under
these conditions, the local dynamic volts-ampere-reactive (VAR) reserves are insufficient and reactive
power support cannot be delivered; therefore, the power system may experience fault induced delayed
voltage recovery (FIDVR) or voltage instability [1–4].

Sustainability 2019, 11, 326; doi:10.3390/su11020326 www.mdpi.com/journal/sustainability

Sustainability 2019, 11, 326 2 of 19

FIDVR is a phenomenon caused by the dynamic behavior of constant-torque IM loads. When a

motor stalls following a severe fault, the reactive power requirements might be 5 to 8 times higher than
normal. The unexpected increase in the reactive power requirement may prevent voltage recovery in
a local area power system. Furthermore, when a fault occurs during peak load conditions, a highly
inductive power system load may not be able to maintain the voltage at the accepted levels [5–8].
To prevent catastrophic situations due to FIDVR, appropriate controlling actions are required to
mitigate these conditions. Control solutions to alleviate FIDVR can be classified into demand-side and
supply-side solutions [5]. The demand-side solution uses a protection system to rapidly disconnect
IM loads under low-voltage conditions and limit the amount of load affected by low voltages [9].
Supply-side solutions focus on mitigating FIDVR threats and reinforcing power networks with
dynamic VAR compensation devices, such as a static VAR compensator (SVC) and a static synchronous
compensator (STATCOM) [10]. Voltage source converter devices, such as STATCOM, are efficient in
controlling the voltage and can provide cost-effective voltage support to boost the local bus voltage
during FIDVR events. Therefore, it is necessary to develop cost-effective reactive power compensation
strategies that enhance STVS [1,2]. Various studies have been carried out to analyze the optimal
placement of dynamic VAR sources. The method in [11] evaluates the cost and benefits of flexible
alternating current transmission system (FACTS) device installation, while taking voltage stability
into consideration. Reference [12] describes the optimal placement of dynamic VAR sources using
post-contingency trajectory sensitivity analysis. In [13], the locations of dynamic VAR sources are
determined based on modal analysis after considering the input signal for the supplementary controller
of FACTS devices. Similarly, the methodology in [14] determines the optimal locations and sizing
of FACTS devices using genetic algorithms. Tiwari et al. adopted a multi-contingency constrained
reactive power planning method by decomposing the optimization problem [15]. Furthermore, several
studies were conducted to alleviate the STVS problem using sensitivity analysis. In [16,17], a trajectory
sensitivity analysis method is proposed to mitigate voltage instability using dynamic reactive power
resources. References [18–21] used sensitivity as a performance measure for VAR sources to satisfy the
requirements of voltage stability margin and transient voltage dip. In [22], a sensitivity index is applied
to determine the most influential locations of dynamic VAR support under certain contingencies.
The method in [23] proposed a new model for optimal management and control of the distribution
system grid based on the sensitivity analysis of branch power flow equations. However, most of the
existing studies only considered steady-state analysis when allocating dynamic VAR sources to resolve
STVS issues. Thus the methods might not capture the full performance of the dynamic VAR sources,
not considering the impact of IMs to transient voltage dynamics.
This paper presents an advanced method for dynamic VAR source compensation control schemes
with FACTS to mitigate the STVS and FIDVR problems of power systems. This method is for the
decision making on optimal operations and placement of dynamic reactive power compensation in
terms of voltage stability, based on the sensitivity analysis information from the proposed methods.
The main features of the proposed method are (i) the development based on a load model for FIDVR,
(ii) the use of sensitivity analysis of the network to the variations of the IM load, (iii) the establishment
of the control function and compensation strategy to maintain the voltage of system within criteria
limits, and (iv) the use of the sensitivity analysis based on branch parameterization for unsolvable
cases. Of them, the last feature is very important because of the security assessments of the power
system reflecting the various contingencies that can be performed in terms of supporting reactive
power compensation planning in order to guarantee secure system operations. When weak network
conditions are combined with significant penetration of the IM load at fault clearing, the network
is prone to FIDVR. Therefore, the objective of this study is to develop appropriate reactive power
compensation strategies to resolve FIDVR problems. This study employed active power–voltage (PV)
and reactive power–voltage (QV) analyses to identify the system margin in terms of voltage stability.
In addition, the optimal allocation of dynamic VAR sources using a version of branch-parameter
continuation power flow (BCPF) method [24] is applied. Using the methodology, it is possible to
Sustainability 2018, 10, x FOR PEER REVIEW 3 of 19

sources using a version of branch-parameter continuation power flow (BCPF) method [24] is applied.
Using the 2019,
Sustainability methodology,
11, 326 it is possible to derive the results of sensitivity analysis for restoring reactive
3 of 19
power under severe outage conditions. Furthermore, time domain analysis is conducted for dynamic
stability monitoring in accordance with the power system planning standards. Case studies
derive the results
conducted on theofworst
sensitivity analysis
conceivable for restoring
contingency reactive
at the peakpower
of the under
wintersevere
seasonoutage conditions.
considered in the
Furthermore,
Korean power system operation planning, which is threatened by serious FIDVR complications.with
time domain analysis is conducted for dynamic stability monitoring in accordance This
the power system planning standards. Case studies conducted on the worst conceivable
situation belongs to rare but extremely high impact events conceivable in system operations; one contingency
at the peak
example ofofit the
is awinter
loss ofseason
765 kV considered in thetransmission
double-circuit Korean power system
lines in a operation
Korean power planning, which
system. The
isproposed
threatened by serious FIDVR complications. This situation belongs to rare
strategy includes the coordination of the performance of transient stability with but extremely high
an
impact events conceivable in system operations; one example of it is a loss
adequate countermeasure to help make the grid resilient against catastrophic contingency. Inof 765 kV double-circuit
transmission lines in asimulation
addition, extensive Korean power system.
studies areThe proposed to
performed strategy includes
investigate itsthe coordinationand
performance of the
to
performance of transient stability with an adequate countermeasure
demonstrate its effectiveness in resolving the problem of STVS. to help make the grid resilient
against catastrophic contingency. In addition, extensive simulation studies are performed to investigate
its performance andfor
2. Methodologies to Countermeasures
demonstrate its effectiveness in resolving
against Short-Term the problem
Voltage Stabilityof Problems
STVS.

2.2.1.
Methodologies for Countermeasures
Load Model Development Metodology against Short-Term Voltage Stability Problems
2.1. Load
LoadModel Developmentexert
characteristics Metodology
a significant influence on power system stability. However, load
modeling
Load is complicated because
characteristics a typical load
exert a significant bus, represented
influence on power in stability
system studies,However,
stability. is composed of
load
a large number of devices. In addition, the composition changes depending on many factors,
modeling is complicated because a typical load bus, represented in stability studies, is composed of a
including time, weather conditions, and state of the economy. Even if the load composition is known
large number of devices. In addition, the composition changes depending on many factors, including
exactly, it would be difficult to represent each individual component, as there are millions of such
time, weather conditions, and state of the economy. Even if the load composition is known exactly,
components in the total load supplied by a power system. Therefore, load representation in system
it would be difficult to represent each individual component, as there are millions of such components
studies is based on a considerable amount of simplification [25–29]. This subsection describes the
in the total load supplied by a power system. Therefore, load representation in system studies is
fundamental concepts related to the development of the load model used in this study. For the
based on a considerable amount of simplification [25–29]. This subsection describes the fundamental
purpose of this study, a more detailed dynamic load representation of the Korean power system was
concepts related to the development of the load model used in this study. For the purpose of this study,
generated. Detailed information on the actual load compositions was obtained from the national
a more detailed dynamic load representation of the Korean power system was generated. Detailed
statistical information database [30]. The recent change in the peak load in winter is shown in Figure
information on the actual load compositions was obtained from the national statistical information
1.
database [30]. The recent change in the peak load in winter is shown in Figure 1.

Figure1.1.Induction
Figure Inductionmotor
motorloads
loadsproportion
proportionat
atwinter
winterpeak
peakcondition.
condition.

In
InFigure
Figure1,1,one
onecancannotice
noticeananincrease
increaseininseasonal
seasonal demand
demand due due to
to an
an increase
increase in
in winter
winter loads;
loads;
further,
further,the
theannual
annualvariation
variationininpower
powerdemand
demandisiscontinuously
continuouslyincreasing.
increasing.This
Thisincrease
increaseininelectricity
electricity
consumption
consumption is attributed
attributedtotothetherise
rise
in in
gasgas prices.
prices. An increase
An increase in severe
in severe cold conditions
cold conditions due to due to
climate
climate
changechange and changes
and changes in lifestyle
in lifestyle are contributing
are contributing to thetoincrease
the increase in residential
in residential power
power equipment.
equipment. In
In addition,
addition, duetotoadvancements
due advancementsininthe theservice
serviceindustry,
industry, the
the demand
demand for load in the commercial
commercialclass
class
isisrapidly
rapidlyincreasing.
increasing.In Inparticular,
particular,ininthe
thecase
caseof
ofwinter
winterpeak
peakload,
load,the
theproportion
proportionof ofcommercial
commercialloadload
exceeded
exceededthatthatof
ofthe
theindustrial
industrialload.
load.
Generally, IM loads operate in a constant torque mode depending on the voltage of the connected
nodes [2,9]. As soon as the voltage magnitude is reduced, IM loads attempt to slow down. If a low
voltage persists for a certain period of time, the IM loads might be stalled or stopped. However, when
Sustainability 2018, 10, x FOR PEER REVIEW 4 of 19

Generally,
Sustainability IM326loads operate in a constant torque mode depending on the voltage of4 the
2019, 11, of 19
connected nodes [2,9]. As soon as the voltage magnitude is reduced, IM loads attempt to slow down.
If a low voltage persists for a certain period of time, the IM loads might be stalled or stopped.
the voltagewhen
However, beginsthetovoltage
recover,begins
the IMto load restarts,
recover, thedraws
IM loadlarge reactivedraws
restarts, powerlarge
and current
reactivefrompowerthe and
grid,
resulting in a delayed voltage recovery. Another typical load that causes load
current from the grid, resulting in a delayed voltage recovery. Another typical load that causes loadrecovery is the heating
load based
recovery on thermostat-controlled
is the heating load based onresistance. Thermostatically
thermostat-controlled controlled
resistance. loads are self-restoring
Thermostatically controlledin
nature. If these electric heating loads constitute a substantial share of the load,
loads are self-restoring in nature. If these electric heating loads constitute a substantial share especially in winter, the
of the
system might be significantly affected. In the case of a heating load with a thermostat,
load, especially in winter, the system might be significantly affected. In the case of a heating load the load might
increase
with even if tapthe
a thermostat, blocking is performed.
load might increase These
even ifcharacteristics
tap blocking might be the main
is performed. These reason for voltage
characteristics
instability [29].
might be the main reason for voltage instability [29].
AA key
key factor
factor that
that results
results inin FIDVR
FIDVR inin power
power systems
systemsisis the
the nature
nature ofof the
the IM
IM loads
loads inin the
the
transmission and distribution systems. The performance characteristics of
transmission and distribution systems. The performance characteristics of these loads are changingthese loads are changing
dueto
due totechnological
technologicaladvances.
advances.AAlargelargethree-phase
three-phaseinduction
inductionmotor
motor(TPIM)
(TPIM)andandsmall
smallsingle-phase
single-phase
inductionmotor
induction motor(SPIM)
(SPIM)are areconnected
connecteddirectly
directlytotothe
theelectrical
electricalnetwork.
network.These
Thesetypes
typesofofIM IMloads
loadsare
are
expected to be a major component of the aggregate load due to economic reasons.
expected to be a major component of the aggregate load due to economic reasons. Figure 2 shows the Figure 2 shows the
structureof
structure ofthe
theload
loadmodel
modelused
usedin inthis
thisstudy.
study.

Modelstructure
Figure2.2.Model
Figure structurefor
forload
loadbuses.
buses.

InFigure
In Figure2,2,the
theimpedance
impedanceand andload
loadtap
tapchanger
changersettings
settingswere
werebased
basedonontypical
typicalvalues
valuesininthethe
available data. For the IM loads, two different motor models were used. One of the motor loads is ais
available data. For the IM loads, two different motor models were used. One of the motor loads
a SPIM,
SPIM, which
which hashas a low
a low inertia
inertia andand responds
responds instantaneously
instantaneously to voltage
to voltage changes.
changes. AnotherAnother
motormotor
load
isload is a TPIM;
a TPIM; they they are commonly
are commonly usedused in industry
in industry to drive
to drive processes.
processes. TheThe ZIPload
ZIP loadrepresents
representsstatic
static
loads (e.g., constant impedance, constant current, and constant power loads)
loads (e.g., constant impedance, constant current, and constant power loads) [29,31]. The areas of [29,31]. The areas of
particularconcern
particular concernin inthis
thisstudy
study(metropolitan
(metropolitanareaareaofofthe
theKorean
Koreanpeninsula)
peninsula)are
areprimarily
primarilyresidential
residential
and commercial. The primary sources for power demand consist of the TPIM
and commercial. The primary sources for power demand consist of the TPIM (commercial class) (commercial class)andand
SPIM (residential class). In order to better represent the power consumption characteristics
SPIM (residential class). In order to better represent the power consumption characteristics of IM of IM loads,
it is necessary
loads, to take to
it is necessary advantage of the power
take advantage of theconsumption characteristics
power consumption under the most
characteristics undersevere
the mostpeak
conditions, in terms of voltage stability.
severe peak conditions, in terms of voltage stability.
2.2. Reactive Power Support Devices
2.2. Reactive Power Support Devices
As large quantities of IM loads exert a significant influence on the stability of the power systems,
As large quantities of IM loads exert a significant influence on the stability of the power systems,
the requirements for reactive power support have become important. Reactive power plays a major
the requirements for reactive power support have become important. Reactive power plays a major
role in the transmission of power systems. During IM reacceleration after a fault, when high current
role in the transmission of power systems. During IM reacceleration after a fault, when high current
is drawn, a high reactive power support is required for a few seconds to provide a fast response [7].
is drawn, a high reactive power support is required for a few seconds to provide a fast response [7].
FACTS technologies are available in different forms, such as SVC, STATCOM, thyristor-controlled
FACTS technologies are available in different forms, such as SVC, STATCOM, thyristor-controlled
series capacitor (TCSC), and unified power flow controllers (UPFC). Among them, SVC and STATCOM
series capacitor (TCSC), and unified power flow controllers (UPFC). Among them, SVC and
are categorized as dynamic VAR compensation resources that rapidly inject or absorb VARs to support
STATCOM are categorized as dynamic VAR compensation resources that rapidly inject or absorb
power system voltage immediately after system disturbances. The STATCOM technology increases
VARs to support power system voltage immediately after system disturbances. The STATCOM
voltage stability by providing dynamic control and compensation of the system voltage. The advantage
technology increases voltage stability by providing dynamic control and compensation of the system
of STATCOM is that its compensating current is not dependent on the system voltage level at the
voltage. The advantage of STATCOM is that its compensating current is not dependent on the system
point of common coupling (PCC) when the voltage drops; it operates at full capacity. Utilizing the
STATCOM switching control allows for faster control response when compared to SVC and improved
Sustainability 2018, 10, x FOR PEER REVIEW 5 of 19

voltage level at the point of common coupling (PCC) when the voltage drops; it operates at full
Sustainability 2019, 11, 326 5 of 19
capacity. Utilizing the STATCOM switching control allows for faster control response when
compared to SVC and improved performance in power systems. In addition, STATCOM is more
resilient
performanceto changes
in power in power system
systems. conditions,
In addition, such asisvariations
STATCOM in harmonic
more resilient levels,
to changes and requires
in power system
smaller outdoor
conditions, suchequipment
as variations when in compared to SVC and
harmonic levels, [32–34].
requires smaller outdoor equipment when
As mentioned
compared earlier, STATCOM regulates the voltage at its PCC by controlling the amount of
to SVC [32–34].
reactive power thatearlier,
As mentioned is injected
STATCOMor absorbed
regulatesfrom the power
the voltage at itssystem
PCC by through its voltage-source
controlling the amount of
converter.
reactive power When theis system
that injectedvoltage is high,
or absorbed fromSTATCOM absorbsthrough
the power system reactiveitspower (inductive
voltage-source mode),
converter.
while
Whenthe thedevice
system generates
voltage and injects
is high, reactive power
STATCOM absorbsinto the system
reactive power when the system
(inductive voltage
mode), is low
while the
(capacitive
device generates mode).andFigure 3 represents
injects reactive the the
power into voltage-current
system when the (V-I) characteristics
system voltage is lowof (capacitive
SVC and
STATCOM.
mode). Figure 3 represents the voltage-current (V-I) characteristics of SVC and STATCOM.

Figure3.3.V-I
Figure V-Icharacteristic
characteristicofofthe
thestatic
staticVAR
VARcompensator
compensatorand
andthe
thestatic
staticsynchronous compensator.
synchronous compensator.

In Figure 3, one can notice that outside the linear operating range, STATCOM, and SVC act
In Figure 3, one can notice that outside the linear operating range, STATCOM, and SVC act
differently. SVC and STATCOM perform differently when the control limits of system voltage are
differently. SVC and STATCOM perform differently when the control limits of system voltage are
reached. Both V-I curves are regulated with a slope when operating within limits. In the control range,
reached. Both V-I curves are regulated with a slope when operating within limits. In the control range,
a droop characteristic is used to avoid excessive control and to coordinate with other voltage control
a droop characteristic is used to avoid excessive control and to coordinate with other voltage control
equipment. STATCOM has symmetrical inductive and capacitive reactive power ranges, but a fixed or
equipment. STATCOM has symmetrical inductive and capacitive reactive power ranges, but a fixed
switched shunt compensation can be used to bias the output range as desired. Additionally, STATCOM
or switched shunt compensation can be used to bias the output range as desired. Additionally,
results in a better voltage support in low-voltage conditions. Therefore, an increase in the reactive
STATCOM results in a better voltage support in low-voltage conditions. Therefore, an increase in the
power support of STATCOM quickly compensates for the disturbance-enhanced voltage stability
reactive power support of STATCOM quickly compensates for the disturbance-enhanced voltage
margin [35,36]. The main purpose of STATCOM is to continuously regulate and support voltage under
stability margin [35,36]. The main purpose of STATCOM is to continuously regulate and support
normal and transient conditions in the Korean power system. Dynamic reactive support is needed so
voltage under normal and transient conditions in the Korean power system. Dynamic reactive
that the transmission system remains stable in the metropolitan area for alleviating FIDVR.
support is needed so that the transmission system remains stable in the metropolitan area for
alleviating
2.3. VoltageFIDVR.
Stability Analysis and Assessment
In thisStability
2.3. Voltage section,Analysis
we describe the configuration of a methodology for voltage stability analysis.
and Assessment
The analysis method is based on previously developed techniques [20,24,28,29]; however, it goes
In this
further section, we
to undertake describedynamic
extensive the configuration of a methodology
reactive power compensationfor voltage
based stability analysis.
on characteristic The
IM loads.
analysis method is based on previously developed techniques [20,24,28,29]; however, it
The two main features of this method are the use of steady-state analysis and time domain simulations.goes further
to undertake
Initially, extensive
PV analysis wasdynamic reactive
performed power
to assess thecompensation based
voltage stability on characteristic
margin of the Korean IM loads.
power The
system
two
undermain features
various of this method
contingencies. PV are the use
analysis of be
can steady-state analysisthe
used to establish andactual
time domain simulations.
system load-serving
Initially, PV analysis was performed to assess the voltage stability margin of the
capability by plotting voltage versus increasing load. This method is used as a screening tool for
Korean power
system under various contingencies. PV analysis can be used to establish the actual system
determining the voltage stability limit of the system based on the criteria of the Western Electricity
load-
serving capability by plotting voltage versus increasing load. This method is used as a screening tool
Coordinating Council (WECC) [37]. In the WECC, criteria are presented as a procedure available for
for determining the voltage stability limit of the system based on the criteria of the Western Electricity
determining voltage stability limit under a fixed load condition. To consider uncertainty, the required
margins for performance levels are specified; in this paper, the minimum required margins on the
Sustainability
Coordinating2018, 10,Council
x FOR PEER REVIEW
(WECC) [37]. 6 of 19for
In the WECC, criteria are presented as a procedure available
determining voltage stability limit under a fixed load condition. To consider uncertainty, the required
Coordinating Council (WECC) [37]. In the WECC, criteria are presented as a procedure available for
margins for performance levels are specified; in this paper, the minimum required margins on the
determining voltage
Sustainability 2019, 11,stability
326 limit under a fixed load condition. To consider uncertainty, the required
6 of 19
WECC voltage stability criteria are used. In addition, a procedure for evaluating the secure operating
margins for performance levels are specified; in this paper, the minimum required margins on the
limits of voltage stability is established. Figure 4 shows the PV curves for margin evaluation at each
WECC voltage stability criteria are used. In addition, a procedure for evaluating the secure operating
performance
WECC voltage level.
stability criteria are used. In addition, a procedure for evaluating the secure operating
limits of voltage stability is established. Figure 4 shows the PV curves for margin evaluation at each
limits of voltage stability is established. Figure 4 shows the PV curves for margin evaluation at each
performance level.
performance level.

Figure 4. PV curve for margin evaluation.

Figure 4. PV curve for margin evaluation.
As shown in Figure 4, based Figure on4. PV
thecurve
WECC for margin evaluation.
criteria, two cases of voltage stability analysis are
considered.
As shown One incanFigure
notice4,that basedtheonmaximum
the WECC operating
criteria, point on the
two cases of Pvoltage
axis must haveanalysis
stability a MW margin
are
As shown in Figure 4, based on the WECC criteria, two cases of voltage stability analysis are
equal to or greater
considered. One can than noticethe thatvalues measured
the maximum from point
operating the nose
on thepoint
P axisof thehave
must PV acurve for each
MW margin
considered.
equal to One can notice thatvalues
the maximum operating point on of
thetheP axis mustforhave a MW margin
performance level. In
or greater than this
the study, itmeasured
was fixedfromthatthethenose
minimum
point power PVmargin
curve foreach
performance
performancelevel
equallevel.
to or greater
Inisthis than thefixedvalues measured from the margin
nose point of the PV levelcurveA, for each
A, which thestudy, it was
contingency, must that
be the minimum
greater than 5%. power for performance
For performance level B, the minimum which is
margin
performance
the level. Inmust
contingency, this bestudy,
greaterit was
thanfixed
5%. that
For the minimum
performance power
level B, margin
the minimum for performance
margin level
should be
should be greater than 2.5% [37]. It is to be noted that the performance level should not cause voltage
A, which is the
greater thancontingency,
2.5% must benoted
greater
thatthan
the 5%. For performance levelnot
B, cause
the minimum margin
collapse within the [37].
limits It of
is to
thebestudy. performance level should voltage collapse
should be greater
within the limitsthanof 2.5%
the we [37]. It is to be noted that the performance level should not cause voltage
study.
In the next step, employed QV analysis to determine the control actions necessary to
collapse within
In the the
next limits
step, weof the study.QV analysis to determine the control actions necessary to overcome
employed
overcome violations of the voltage criteria. The purpose of this analysis is to determine the region of
In the next
violations step,
of the we employed
voltage criteria. QV
The analysis this
purpose to determine thedetermine
control actions necessary to
reactive deficiency in terms of reactive power of for theanalysis is to
severe contingencies the region
determined ofthrough
reactive PV
overcome violations
deficiency in termsof the voltagepower
of reactive criteria.
forThe
the purpose of this analysis
severe contingencies is to determine
determined throughthePVregion of
analysis.
analysis. Additionally, QV analysis is used to identify the required amount of reactive power
reactive deficiencyQV
Additionally, in analysis
terms ofisreactiveused to power
identifyfor
thethe severeamount
required contingencies
of reactivedetermined through PV
power compensation.
compensation. Figure 5 represents the QV curve for determining the reactive power margin.
FigureAdditionally,
analysis. 5 represents the QVQVanalysis
curve forisdetermining the reactive
used to identify power margin.
the required amount of reactive power
compensation. Figure 5 represents the QV curve for determining the reactive power margin.

Figure
Figure 5.
5. Concept
Concept of QV
QV curve
curveand
andreactive
reactivepower
power margin.
margin.

Figure 5. Concept of QV curve and reactive power margin.

 Sustainability 2018, 10, x FOR PEER REVIEW 7 of 19
Sustainability 2019, 11, 326 7 of 19
From Figure 5, it can be inferred that the reactive power margin can be calculated using the QV
curve, which defines the amount of reactive power that can be consumed by the bus. When a fault
From Figure 5, it can be inferred that the reactive power margin can be calculated using the QV
occurs, the reactive power margin of the bus should be maintained at a level sufficient to meet the
curve, which defines the amount of reactive power that can be consumed by the bus. When a fault
criteria. Therefore, it is possible to calculate the reactive power margin on each bus for serious faults,
occurs, the reactive power margin of the bus should be maintained at a level sufficient to meet the
so that the bus or area with a low reactive power margin can be determined as the reactive power
criteria. Therefore, it is possible to calculate the reactive power margin on each bus for serious faults,
vulnerable site. To this end, a VQVI margin vulnerability index can be used to identify vulnerable
so that the bus or area with a low reactive power margin can be determined as the reactive power
areas. VQVI considers the changes in the reactive power margin before and after the contingency and
vulnerable site. To this end, a VQVI margin vulnerability index can be used to identify vulnerable
can be expressed as follows.
areas. VQVI considers the changes in the reactive power margin before and after the contingency and
can be expressed as follows. 𝑝𝑜𝑠𝑡 − 𝑐𝑜𝑛𝑡𝑖𝑛𝑔𝑒𝑛𝑐𝑦 𝑟𝑒𝑎𝑐𝑡𝑖𝑣𝑒 𝑝𝑜𝑤𝑒𝑟 𝑚𝑎𝑟𝑔𝑖𝑛
𝑉𝑄𝑉𝐼 = (1)
𝑝𝑟𝑒 − 𝑐𝑜𝑛𝑡𝑖𝑛𝑔𝑒𝑛𝑐𝑦 𝑟𝑒𝑎𝑐𝑡𝑖𝑣𝑒 𝑝𝑜𝑤𝑒𝑟 𝑚𝑎𝑟𝑔𝑖𝑛
post − contingency reactive power margin
VQV I = (1)
pre − contingency reactive power margin
When a severe contingency occurs, it might be difficult to analyze systems using general power
flowWhen
equations as contingency
a severe the solutionsoccurs,
are not valid after
it might a contingency.
be difficult to analyzeTherefore, a branch-parameter
systems using general power
continuation
flow equationspower
as the flow method
solutions is used
are not validtoafter
the arestore power Therefore,
contingency. flow solvability in unsolvable
a branch-parameter
continuation power flow method is used to the restore power flow solvability in by
contingencies based on a continuous power flow. The branch parameter 𝑌 is multiplied the series
unsolvable
contingencies based on a continuous power flow. The branch parameter Y is multipliedthe
and shunt admittances of the branch. The BCPF can trace the power flow solution, in which by state
the
series and shunt admittances of the branch. The BCPF can trace the power flow solution, in which the𝑌
of the branch moves from a pre-contingency state to a post-contingency state by changing the
parameter
state from zero
of the branch movestofrom onea [24]. This method
pre-contingency represents
state the YV analysis
to a post-contingency described
state by changinginthethis
Y
investigation. Figure 6 shows two cases of YV curves with contingencies.
parameter from zero to one [24]. This method represents the YV analysis described in this investigation.
Figure 6 shows two cases of YV curves with contingencies.

Figure 6. Stable and unstable cases of YV curve.

Figure 6. Stable and unstable cases of YV curve.
In Figure 6, one can notice that the branch parameter Y = 0 is in service and Y = 1 is out of service.
In Figure 6, one can notice that the branch parameter 𝑌 = 0 is in service and 𝑌 = 1 is out of
These branch parameter variation curves represent stable and unstable cases, respectively. If the value
service. These branch parameter variation curves represent stable and unstable cases, respectively. If
of Y cannot reach one, as described in case 2, a control action is needed for power flow convergence.
the value of 𝑌 cannot reach one, as described in case 2, a control action is needed for power flow
To determine the sensitivity of adequate injection VAR for reactive power compensation strategies is
convergence. To determine the sensitivity of adequate injection VAR for reactive power
calculated at the nose point in the YV curve. Sensitivity information is divided into active power and
compensation strategies is calculated at the nose point in the YV curve. Sensitivity information is
reactive power. This study utilizes sensitivity to reactive power, which is influenced by changes in the
divided into active power and reactive power. This study utilizes sensitivity to reactive power, which
voltage. The sensitivity can be calculated as
is influenced by changes in the voltage. The sensitivity can be calculated as
∂Qi
Participation Factor = 𝜕𝑄 (2)
𝑃𝑎𝑟𝑡𝑖𝑐𝑖𝑝𝑎𝑡𝑖𝑜𝑛 𝐹𝑎𝑐𝑡𝑜𝑟 =∂Y (2)
𝜕𝑌
where Qi is the reactive power at load bus and Y is the branch parameter. With this information,
the most efficient load bus for restoring the voltage can be selected. The reactive power reserves of the
Sustainability 2018, 10, x FOR PEER REVIEW 8 of 19
Sustainability 2019, 11, 326 8 of 19
where 𝑄 is the reactive power at load bus and 𝑌 is the branch parameter. With this information,
the most efficient load bus for restoring the voltage can be selected. The reactive power reserves of
VAR compensation resources are determined at the point of convergence of the power flow equations.
the VAR compensation resources are determined at the point of convergence of the power flow
Thus, if the remaining flows can be shifted to other branches using reactive power compensation,
equations. Thus, if the remaining flows can be shifted to other branches using reactive power
the power system might be restored.
compensation, the power system might be restored.
In the unsolvable case, the YV analysis can trace the power flow solution depends on the variations
In the unsolvable case, the YV analysis can trace the power flow solution depends on the
of transmission line parameter via continuation when a power flow solution does not exist after the
variations of transmission line parameter via continuation when a power flow solution does not exist
branch outage [24]. The parameterization of the branch can be depicted as a π-equivalent circuit,
after the branch outage [24]. The parameterization of the branch can be depicted as a π-equivalent
as shown in Figure 7.
circuit, as shown in Figure 7.

Figure 7. Representation of π-equivalent

π-equivalent circuit
circuit with
with branch
branch parameter.
parameter.

Figure
Figure 77indicates
indicatesthat parameterλ 𝜆is is
thatparameter incorporated
incorporated into a power
into a power flow equation
flow equation to implement the
to implement
branch parameter
the branch variation.
parameter If theIfbranch
variation. of bus
the branch ofi bus 𝑖 in
to j is 𝑗 isoutage
to an in an state,
outage thestate,
active and
the reactive
active and
power flow
reactive fromflow
power busfromi to bus 𝑖 tobe
busj can bus 𝑗 can be
n  h∗ ∗ io
Pij 𝑃Vi , 𝑉V,j 𝑉
, λ, 𝜆 ==((1
1 −−λ𝜆) |Vi||2 𝐺G∗ii∗++|𝑉
) |𝑉 |V|i |𝑉
Vj 𝐺 G cos(𝜃 ) + 𝐵 ∗ ∗
sin(𝜃 )


ij cos ( θ ij ) + Bij sin ( θ ij ) (3)
(3)

∗  h∗
Qij 𝑄Vi , 𝑉V,j𝑉 , 𝜆
=
= ((1 −λ𝜆) −|V|i |2𝐵Bii∗++|𝑉
) −|𝑉 |V|i |𝑉Vj 𝐺 Gij∗sin(𝜃
sin(θ)ij− 𝐵∗B∗cos(𝜃 )
n io
1− )− (4)
, λ ij cos( θij ) (4)

where G𝐺ii∗∗,,G𝐺ij∗ ∗, ,and

where andBij∗𝐵, ∗are thethe
, are admittances
admittances of of
thethe
removed
removed branch.
branch.AsAs
thethe
continuation
continuationparameter
parameter λ
𝜆 increases, the system gradually closes to the power system with branch outages from pre-
increases, the system gradually closes to the power system with branch outages from pre-contingency
systems.
contingency When λ is zero,
systems. When 𝜆 is zero,
the original power flow equations
the original power are
flowobtained.
equations When one, the𝜆
λ reaches When
are obtained.
new power
reaches one,flow the equations
new power thatflow
represent the network
equations with the
that represent branch
the totally
network removed
with is obtained.
the branch totally
This method
removed comprisesThis
is obtained. a predictor
methodand a corrector,
comprises and the BCPF
a predictor and uses an augmented
a corrector, and theJacobian matrix.
BCPF uses an
The augmented
augmented Jacobian Jacobian
matrix.usedThein the BCPF canJacobian
augmented be expressed asthe BCPF can be expressed as
used in
 
⁄𝜕𝛿 ∂P
𝜕𝑃T /∂δ
∂P ⁄𝜕𝑣 ∂P
𝜕𝑃T /∂v ⁄𝜕𝑌
𝜕𝑃T /∂Y
J A 𝐽= = ⁄𝜕𝛿 ∂Q
𝜕𝑄T /∂δ ⁄𝜕𝑣 ∂Q
𝜕𝑄T /∂v ⁄𝜕𝑌 
𝜕𝑄T /∂Y (5)
(5)
 
 ∂Q
e𝑒k

P𝑃T ∈∈ R
𝑅n−1 , ,Q𝑄 ∈∈R𝑅n−npv−1
T

where P𝑃T and

where andQ𝑄 T
representthe
represent theinjection
injectionofofthe
theactive
activeand
andreactive
reactivepower,
power, respectively.
respectively. Meanwhile,
Meanwhile, ek
𝑒 chosen
is to remove
is chosen to removethethe
ill-conditioned
ill-conditioned problem
problem from
fromthethe
singularity. Further,
singularity. Further,npv𝑛𝑝𝑣
is the number
is the of
number
voltage-controlled
of voltage-controlledbuses. As shown
buses. in Equation
As shown (5), ∂P(5),
in Equation 𝜕𝑃∂Q
T /∂Y, ⁄𝜕𝑌
T , 𝜕𝑄 ⁄𝜕𝑌 ,k and 𝑒 are augmented
/∂Y, and e are augmented from the
original
from thepower flow
original Jacobian
power flowmatrix.
Jacobian matrix.
Using these characteristics,
Using these characteristics, thethe
BCPF can can
BCPF determine whether
determine the cause
whether the ofcause
the divergence problem
of the divergence
is unsolvable. The following is a sensitivity analysis procedure for applying the
problem is unsolvable. The following is a sensitivity analysis procedure for applying the BCPF BCPF application to
obtain a solution.
application to obtain a solution.
Step
Step1:1: Solve
Solvethe
thepower
powerflow
flowand
andperform
performthe
theYV
YV analysis
analysis for
for specific
specific contingencies
contingencies that
that are
are
determined
determined by by a preliminary
a preliminary study.
study.
Sustainability 2019, 11, 326 9 of 19

Sustainability 2018, 10, x FOR PEER REVIEW 9 of 19

Step 2: The BCPF is applied to select the unsolvable contingencies in the list.
Step
Step 3: 2: The BCPF
Determine is applied
the reactive to select
power the unsolvable
reserves contingencies
and reference voltages ofinthe
thebuses
list. for converging.
Step 3: Determine the reactive power reserves and reference voltages of the
Step 4: In order to find effective buses for reactive power compensation, a sensitivity buses foranalysis
converging.
is
Step 4: In order to
then performed. find effective buses for reactive power compensation, a sensitivity analysis is
then performed.
Step 5: Perform the sensitivity analysis of each buses where the FACTS devices are installed.
Step 5: Perform the sensitivity analysis of each buses where the FACTS devices are installed.
Load
Loadbuses
buses with
withlarge
largeparticipation
participationfactor
factorvalues
values have
have more
more influence
influence inin contributing
contributing totothe
thevoltage
voltage instability.
instability. Consequently,
Consequently, thesethese buses
buses can be regarded as best candidate
can be regarded as best candidate location forfor
location
determining
determining thethe
injection to enhance
injection to enhancesystem voltage
system stability.
voltage stability.
Finally, this study attempted to determine
Finally, this study attempted to determine the mostthe most severe contingency
severe andand
contingency voltage-vulnerable
voltage-vulnerable
areas through static analysis and detailed time simulations for verifying
areas through static analysis and detailed time simulations for verifying the reduction the reduction in in
FIDVR.
FIDVR.
Usually, transient load characteristics are more sensitive to voltage than steady-state
Usually, transient load characteristics are more sensitive to voltage than steady-state load load characteristics.
To address FIDVR To
characteristics. issues, in this
address study,issues,
FIDVR the power system
in this study,dynamic
the powerperformance is evaluated
system dynamic against is
performance
typical transient
evaluated voltage
against criteria
typical according
transient to thecriteria
voltage WECCaccording
and NERCtoplanning
the WECCstandards.
and NERCThe criteria
planning
imposed on the transient voltage dip are summarized in the following sections
standards. The criteria imposed on the transient voltage dip are summarized in the following and will be used to
sections
illustrate the proposed reactive power planning approach. Figure 8 shows the voltage
and will be used to illustrate the proposed reactive power planning approach. Figure 8 shows the performance
parameters with respect parameters
voltage performance to the transient
withvoltage
respectdip criteria
to the [37]. voltage dip criteria [37].
transient

Figure 8. Post-fault
Figure voltage
8. Post-fault performance
voltage criteria
performance for for
criteria loadload
bus.bus.

As shown in Figure 8, the post-fault transient voltage performance is expected to meet the
As shown in Figure 8, the post-fault transient voltage performance is expected to meet the
following criteria.
following criteria.
• • TheThe
post-fault transient
post-fault voltage
transient dipdip
voltage or overshoot should
or overshoot notnot
should exceed
exceed25% at the
25% load
at the buses
load or 30%
buses or 30%
at the generator buses; it should not exceed 20% for more than 20 cycles at the load buses.
at the generator buses; it should not exceed 20% for more than 20 cycles at the load buses.
• • TheThe
post-transient voltage
post-transient deviation
voltage should
deviation notnot
should exceed 5% 5%
exceed at any bus.bus.
at any
From
From Figure
Figure 8, the
8, the post-fault
post-fault transient
transient voltage
voltage performance
performance is expected
is expected to meet
to meet criteria,
criteria, namely:
namely:
FACTS is an effective technology to mitigate transient voltage dip by providing
FACTS is an effective technology to mitigate transient voltage dip by providing dynamic VAR support dynamic VAR
forsupport
reactive for reactive
power power compensation.
compensation. This study
This study focuses focuses on the
on identifying identifying the most
most effective effective
locations
locations and determining the capacity of FACTS so that the transient voltage dip
and determining the capacity of FACTS so that the transient voltage dip criteria are satisfied criteria are satisfied
for
for mitigating FIDVR. To accomplish this task, analysis of the transient voltage recovery
mitigating FIDVR. To accomplish this task, analysis of the transient voltage recovery of the Korean of the Korean
power
power system
system waswas conducted
conducted according
according to the
to the above
above described
described standards.
standards. Figure
Figure 9 shows
9 shows a flowchart
a flowchart
of the employed voltage stability analysis and assessment
of the employed voltage stability analysis and assessment method. method.
Sustainability 2019, 11, 326 10 of 19
Sustainability 2018, 10, x FOR PEER REVIEW 10 of 19

Figure 9.
Figure 9. Flowchart
Flowchart of
of the
the voltage
voltage stability
stability analysis
analysis and
and assessment
assessment method.
method.

3.
3. Results
Results and
and Discussion
Discussion
This
This section
section describes
describes thethe results
results of
of voltage
voltage stability
stability analysis
analysis in in the
the Korean
Korean power
power system.
system.
Steady-state
Steady-state andand dynamic
dynamic studies
studies were
were performed
performed basedbased on on the
the proposed
proposed methodology.
methodology. The The Korean
Korean
power system has the following distinct features: (i) concentration of
power system has the following distinct features: (i) concentration of load demand in theload demand in the metropolitan
region; (ii) major
metropolitan generating
region; (ii) majorplants located plants
generating in nonmetropolitan regions; and (iii)
located in nonmetropolitan environmental
regions; and (iii)
regulations
environmental regulations and land for constructing new facilities, such as transmission lines.these
and land for constructing new facilities, such as transmission lines. Due to Due
geographical characteristics,
to these geographical six interface
characteristics, sixlines are operated
interface lines areto operated
supply power to the power
to supply metropolitan
to the
area from non-metropolitan
metropolitan areas. System operators
area from non-metropolitan in the Korean
areas. System operators power system
in the are power
Korean alwayssystem
concernedare
with the lack of a reactive power reserve in the metropolitan area for an interface
always concerned with the lack of a reactive power reserve in the metropolitan area for an interface flow [38]. Under these
circumstances,
flow [38]. Under when a serious
these fault occurs
circumstances, whenin interface
a serious lines,
faulta region
occurs with a high proportion
in interface of IMwith
lines, a region loadsa
might significantly
high proportion increase
of IM loads the VAR
might demand during
significantly increase voltage
the VAR recovery.
demand However, externalrecovery.
during voltage reactive
power cannot be sufficiently delivered to the IM load-concentrated region
However, external reactive power cannot be sufficiently delivered to the IM load-concentrated regionthrough long-distance
transmission lines. Astransmission
through long-distance a result, the lines.
system Asmight
a result,experience
the system STVS
mightproblems.
experience Unless
STVSimmediate
problems.
countermeasures
Unless immediate countermeasures are applied, this leads to a FIDVR phenomenon. Figure 10voltage
are applied, this leads to a FIDVR phenomenon. Figure 10 shows a bus shows
profile withoutprofile
a bus voltage a countermeasure.
without a countermeasure.
After the fault occurs, the voltages of the load buses drop immediately. Without a countermeasure
scheme, the bus voltage cannot be recovered to its normal value after fault clearance. From Figure 10,
the following observations can be made:

(1) The IM loads absorbed much more reactive power after the fault. Consequently, voltage instability
occurs following a severe fault.
(2) Motor speed recovery is closely related to voltage recovery. If the speed cannot be recovered after
clearing the fault, the motors will absorb much of the dynamic VAR.
(3) The slip characteristics depend on the mechanical and electrical torques. When the mechanical
and electrical torque do not equal each other, the slip moves to 1 as the equilibrium point
is exceeded.

(a) (b)
flow [38]. Under these circumstances, when a serious fault occurs in interface lines, a region with a
high proportion of IM loads might significantly increase the VAR demand during voltage recovery.
However, external reactive power cannot be sufficiently delivered to the IM load-concentrated region
through long-distance transmission lines. As a result, the system might experience STVS problems.
Unless immediate countermeasures are applied, this leads to a FIDVR phenomenon. Figure 10 shows
Sustainability 2019, 11, 326 11 of 19
a bus voltage profile without a countermeasure.

Sustainability 2018, 10, x FOR PEER REVIEW 11 of 19

Figure 10. Responses of bus voltages under the most severe contingency. (a) Bus voltage magnitude.
(b) Induction motor speed & slip.

After the fault occurs, the voltages of the load buses drop immediately. Without a
countermeasure scheme, the bus voltage cannot be recovered to its normal value after fault clearance.
From Figure 10, the following observations can be made:
(1) The IM loads absorbed much more reactive power after the fault. Consequently, voltage
instability occurs following a severe fault.
(2) Motor speed recovery is closely related to voltage recovery. If the speed cannot be recovered
after clearing the fault,
(a) the motors will absorb much of the dynamic VAR. (b)
(3) The slip characteristics depend on the mechanical and electrical torques. When the mechanical
Figure 10. Responses of bus voltages under the most severe contingency. (a) Bus voltage magnitude.
and electrical torque do not equal each other, the slip moves to 1 as the equilibrium point is
(b) Induction motor speed & slip.
exceeded.
Here,
Here,thethecurrent
currentunder
undervoltage
voltageload
loadshedding
shedding(UVLS)
(UVLS)scheme
schemeoperated
operatedin inthe
theKorean
Koreanpower
power
system is reviewed. The UVLS scheme is a demand-side solution that accomplishes
system is reviewed. The UVLS scheme is a demand-side solution that accomplishes the stabilization the stabilization of
the system
of the system by by
shedding
shedding thethe
pre-defined
pre-defined loads in in
loads thethe
case ofof
case aa
severe
severecontingency
contingencyasasan anevent-based
event-based
special
special protection scheme (SPS). However, the current UVLS system has a time delay betweenthe
protection scheme (SPS). However, the current UVLS system has a time delay between the
activation
activationandandthe
theactual
actualload
loadinterruption.
interruption.WhenWhenthe theload
loadshedding
sheddingisisdelayed,
delayed,ititmight
mightnot
notalleviate
alleviate
the
theFIDVR
FIDVR phenomenon.
phenomenon. In In addition,
addition,the theUVLS
UVLSsystem
systemisisforced
forcedtoto shed
shed thethe pre-defined
pre-defined loads
loads as
as the
the last resort, possibly resulting in an inconvenience for the end-use consumer. Therefore,
last resort, possibly resulting in an inconvenience for the end-use consumer. Therefore, in this paper, in this
paper, a scheme
a scheme is proposed
is proposed to utilize
to utilize thethe dynamic
dynamic reactivepower
reactive powerresources
resourcesto to mitigate
mitigate the
the FIDVR
FIDVR
phenomenon.
phenomenon. The comparison results of the bus voltage recovery for an existing scheme andthe
The comparison results of the bus voltage recovery for an existing scheme and the
proposed
proposedscheme
schemeareareshown
shownininFigure
Figure11.11.

Figure 11.Comparison
Figure11. Comparisonof
ofvoltage
voltagerecovery
recoveryresults
resultswith
withexisting
existingand
andproposed
proposedschemes.
schemes.

Figure 11 indicates that the FDIVR phenomenon occurs even when the existing scheme is applied.
Figure 11 indicates that the FDIVR phenomenon occurs even when the existing scheme is
This is attributed to the reactive power consumption of the motor load being increased dramatically
applied. This is attributed to the reactive power consumption of the motor load being increased
as the voltage level falls to the limit value in the case of delaying the load shedding caused by the
dramatically as the voltage level falls to the limit value in the case of delaying the load shedding
communication delay. On the other hand, it can be seen that the voltage is recovered when the
caused by the communication delay. On the other hand, it can be seen that the voltage is recovered
proposed scheme is applied. It is noted that FACTS could have a positive effect on delayed voltage
when the proposed scheme is applied. It is noted that FACTS could have a positive effect on delayed
recovery owing to the induction motor power consumption characteristics. This paper presents the
voltage recovery owing to the induction motor power consumption characteristics. This paper
presents the analysis associated with the design and application of the dynamic reactive power
control strategy in case studies based on the practical system. It is desirable that FACTS devices need
be installed in metropolitan areas with adequate amount of reactive compensation, because the
FACTS system has a small footprint and requires less outdoor equipment. The placement of the most
important buses, in terms of voltage stability, are determined through the proposed methodology.
Sustainability 2019, 11, 326 12 of 19

analysis associated with the design and application of the dynamic reactive power control strategy
in case studies based on the practical system. It is desirable that FACTS devices need be installed in
metropolitan areas with adequate amount of reactive compensation, because the FACTS system has
a small footprint and requires less outdoor equipment. The placement of the most important buses,
in terms of voltage stability, are determined through the proposed methodology. Furthermore, the
system load condition corresponding to peak winter is applied in the case studies. As for the IM load,
the above-mentioned (in Section 2) load model is adopted. In the simulation, IMs are used to represent
inductive loads. For FIDVR events, it is assumed that IM loads have a composition proportion of
30% in the dynamic analysis based on the statistical system [30]. In this simulation, two case studies,
solvable and unsolvable, are considered. The results of the two cases are described in detail in the
following subsections.

3.1. Case 1: Performance Level A

First of all, in order to assess the vulnerable areas, a contingency list was derived using PV
analysis. In the WECC criteria, the voltage stability analysis included in the PV and QV analysis.
Through the using concept, this paper reports a study of voltage stability limit in the metropolitan
region and the neighboring region in the Korean power system [38]. In the PV analysis associated with
this study, the minimum active power margin for performance level A shall be greater than 5% under
the contingency. The credible contingency sets to be applied in this simulation are all double circuit
outages of 345 kV and 765 kV lines. Among them, the results of the six worst contingencies are shown
in Table 1.

Table 1. Active power margin of contingency list.

Contingency List Line (# of Circuits) % Margin Performance Level Voltage Level

#1 1200–5010 (2) 5.87 A 765 [kV]
#2 4010–6030 (2) 459 B 765 [kV]
#3 1300–3300 (2) 5.22 A 345 [kV]
#4 1400–1500 (2) 6.04 A 345 [kV]
#5 2400–4600 (2) 7.08 A 345 [kV]
#6 2500–4750 (2) 7.44 A 345 [kV]

From Table 1, it can be seen that the voltage stability limit of % margin corresponding to No. 2
in the contingency list corresponds to performance level B; therefore, it might be regarded as the
most serious failure. The remaining situations correspond to performance level A; the % margin of
contingency No. 3 was the lowest. Therefore, No. 3 was considered as a solvable case and simulated.
Further, reactive power reserve calculation for No. 3 was performed based on the VQVI margin
vulnerability index. The results are shown in Table 2. These results are related to the capability of
maintaining voltage stability on individual load buses and they can be used to identify vulnerable
areas in the system.

Table 2. Results of the reactive power margin with contingency No. 3.

Contingency List Ranking Bus No. V-Q margin Vulnerability Index

1 1400 0.02575
2 1300 0.02749
3 1700 0.03528
4 2500 0.03867
5 3300 0.05437
# 3 (1300–3300)
6 4700 0.07713
7 4500 0.07997
8 1200 0.08049
9 4100 0.08847
10 3600 0.09412
Sustainability 2019, 11, 326 13 of 19

Sustainability 2018, 10, x FOR PEER REVIEW 13 of 19

From the data in Table 2, one can infer that the northern metropolitan area includes vulnerable
routeInlines
sites. No. between
3 outage, 1300 and 3300.
the critical Figure 12that
contingencies shows the shows
influence the diagram
the stability of the metropolitan
of the northern area are the
region.
route lines between 1300 and 3300. Figure 12 shows the shows the diagram of the metropolitan region.

Figure12.
Figure 12. One-line
One-line diagram
diagram of
of metropolitan
metropolitanregion
regionin
inKorean
Koreanpower
powersystem.
system.

As
As shown
shown in in Figure
Figure 12,
12, the
themetropolitan
metropolitan regionregion isislargely
largelydivided
dividedinto intoeastern,
eastern, western,
western, and and
northern
northernareas.
areas.InInaddition,
addition,generators
generatorsare areaggregated
aggregatedin inthe
thewestern
westernarea.
area.In Insubstation
substation1300,
1300,active
active
and
andreactive
reactivepowers
powersare aresupplied
suppliedfrom from31503150C/CC/C before faults
faults (i.e.,
(i.e., outages
outages between
between 1300 1300 andand 3300,
3300,
occur.
occur.When
When thethe fault
faultoccurs,
occurs, the the1300
1300substation
substation wouldwould be be at
atthe
theend
endof ofaaload.
load.ForForthis
thisreason,
reason,the the
1300
1300 substation
substation might
might have
have problems
problems of ofinsufficient
insufficientreactive
reactivepower
powersupply.
supply. Thus,
Thus, this
this area
area needs
needs
reactive
reactivepower
powersources
sourcesbecause
becauseplenty
plentyof ofpower
powerisissupplied
suppliedthrough
throughthe theinterface
interfaceline
lineininthe
theeastern
eastern
and
andsouthern
southernareas.
areas.Thus,
Thus,ititisisconcluded
concludedthat thatFACTS
FACTSdevices
devicesshould
shouldbe beinstalled
installedatat1400,
1400,1300, 1700,
1300,1700,
2500,
2500,and
and3300
3300totocounter
counterthetheFIDVR
FIDVRphenomenon.
phenomenon.
Using
Using the above
above results,
results, aatransient
transientstability
stabilityanalysis
analysis waswas performed
performed to verify
to verify the STVS.
the STVS. The
The objective
objective of the
of the dynamic
dynamic simulation
simulation is toisestablish
to establish appropriate
appropriate dynamic
dynamic VAR VAR supports
supports to satisfy
to satisfy the
the voltage
voltage criteria.
criteria. FACTS
FACTS device
device placement
placement determinedbased
determined basedon onthe
theQVQVanalysis
analysisisisshown
shown in in Table
Table 2.
Ten
Tencandidate
candidatebuses
buseswere
werechosen
chosento toinstall
installthe
theFACTS
FACTSdevices,
devices,which
whichwerewereinstalled
installed ininthe
theorder
orderofof
ranking
ranking totoobserve
observe whether
whether thethe voltage
voltage waswas recovered
recovered or or not.
not. As
As the
the typical
typical size
size of
of aa FACTS
FACTSdevice device
installed
installedininthe
thetransmission
transmissionlevel levelisis300
300MVAR,
MVAR, thethe
dynamic
dynamic VAR VARsupport
supportsource capacity
source limitlimit
capacity is setis
at 300±300
set±at MVAR.MVAR.The post-fault
The post-fault voltage trajectories
voltage of theofbuses
trajectories are shown
the buses in Figure
are shown 13. 13.
in Figure
Sustainability 2019, 11, 326 14 of 19
Sustainability 2018, 10, x FOR PEER REVIEW 14 of 19

Figure 13. Post-fault voltage trajectories of buses.

Figure 13. Post-fault voltage trajectories of buses.
From Figure 11, one can notice that the transient voltage criteria are satisfied when three FACTS
From Figure 11, one can notice that the transient voltage criteria are satisfied when three FACTS
devices are installed. If only one or two FACTS devices are installed, the FIDVR event caused by an
devices are installed. If only one or two FACTS devices are installed, the FIDVR event caused by an
insufficient reactive power support cannot be resolved. Therefore, it was decided that the FACTS
insufficient reactive power support cannot be resolved. Therefore, it was decided that the FACTS
devices should be installed at 1400, 1300, and 1700 to mitigate the FIDVR phenomenon.
devices should be installed at 1400, 1300, and 1700 to mitigate the FIDVR phenomenon.
3.2. Case 2: Performance Level B
3.2. Case 2: Performance Level B
The proposed reactive power compensation strategy has also been applied to No.2 on the
The proposed
contingency reactive power
list in performance levelcompensation
B. As mentioned strategy
earlier,has also been No.2
contingency applied to No.2theonmost
represents the
contingency list in performance level B. As mentioned earlier, contingency No.2
severe outage. For performance level B, corresponding to the bus section representing contingency, represents the most
severe
the outage.margin
minimum For performance level B,than
should be greater corresponding to the bus
2.5%. This means thatsection representing
a 2.5% margin on thecontingency,
PV analysis
is required for the worst-case performance level B disturbance cases. The 2.5% is on
the minimum margin should be greater than 2.5%. This means that a 2.5% margin the PV
based analysis
on the load
level measured at the voltage collapse point. Therefore, the proposed method based on BCPF load
is required for the worst-case performance level B disturbance cases. The 2.5% is based on the was
level measured at the voltage collapse point. Therefore, the proposed method based
applied to verify the unsolvable case. To identify effective buses for reactive power compensation, on BCPF was
a sensitivity analysis was then performed at the critical point of the YV curve using the participationa
applied to verify the unsolvable case. To identify effective buses for reactive power compensation,
sensitivity
factor analysis
calculated forwas then buses.
the load performed at the critical
The results point of
of sensitivity the YVare
analysis curve
shownusing the participation
in Table 3.
factor calculated for the load buses. The results of sensitivity analysis are shown in Table 3.
Table 3. Results of a sensitivity analysis with contingency No. 2.
Table 3. Results of a sensitivity analysis with contingency No. 2.
Contingency List Ranking Bus No. Participation Factor
Contingency List Ranking 1 Bus2700
No. Participation
0.0153 Factor
2 4500 0.0148
1 3 2700
2500 0.0147 0.0153
2 4 4700
4500 0.0139 0.0148
5 4650 0.0137
# 2 (4010–6030) 3 2500 0.0147
6 4600 0.0136
4 7 4700
3600 0.0131 0.0139
5 8 4650
4400 0.0126 0.0137
# 2 (4010–6030)
6 9 4800
4600 0.0125 0.0136
7 10 4200
3600 0.0123 0.0131
8 4400 0.0126
From Table 3, one can observe 9 that the eastern4800
metropolitan area consists0.0125of vulnerable sites.
In the Korean power system, when 10 the most severe 4200
765 kV transmission line0.0123fault occurs, power
flow takes a roundabout way to the nearby lines. When a large amount of power is concentrated
in theFrom
powerTable 3, one can observe
transmission that the
line, voltage eastern metropolitan
instability area consists
occurs. In addition, of vulnerable
the dynamic naturesites.
of IMIn
the Korean
loads is the power
same assystem,
imposingwhen the most
a heavy severe
demand for765 kV transmission
fast-responding line fault
reactive power occurs, powerunder
resources flow
takes a roundabout
low-voltage wayThis
situations. to the nearby
leads lines. When
to secondary a large
effects, amount
such of power
as FIDVR, which is further
concentrated in the
exacerbates
power
the transmission
problem. line, voltage
The transient instability
voltage dip and occurs.
slow In addition,
voltage the dynamic
recovery issues nature
might of beIM loads is the
addressed by
same as imposing
fast-responding a heavy
dynamic demand
VAR for fast-responding
resources. Candidate buses reactive power resources
are considered under
for selecting the low-voltage
locations of
situations. This leads to secondary effects, such as FIDVR, which further exacerbates the problem.
The transient voltage dip and slow voltage recovery issues might be addressed by fast-responding
Sustainability 2019,
Sustainability 2018, 11,
10, 326
x FOR PEER REVIEW 15
15 of 19
19

dynamic VAR resources. Candidate buses are considered for selecting the locations of reactive power
reactive power in
compensation compensation in order
order to finalize the to finalize
space and the space and geographical
geographical location
location to install to install
FACTS FACTS
devices. For
devices. For dynamic assessments, the above-mentioned performance criteria should
dynamic assessments, the above-mentioned performance criteria should be followed. Figure 14 be followed.
Figure
shows 14theshows theresponses
voltage voltage responses of buses
of buses under under contingency
contingency conditions
conditions after installation.
after FACTS FACTS installation.

Figure
Figure 14. Post-fault voltage
14. Post-fault voltage response
response of
of buses
buses with
with flexible
flexible alternating
alternating current
current transmission
transmission system.
system.

It can be seen in Figure 14 that the time domain simulations are run by applying a fault. To address
It can be seen in Figure 14 that the time domain simulations are run by applying a fault. To
the delayed voltage recovery caused by IM stalling under post-fault voltage trajectories, the voltage
address the delayed voltage recovery caused by IM stalling under post-fault voltage trajectories, the
criteria are met using five FACTS devices. When four or less FACTS devices are installed, the FIDVR
voltage criteria are met using five FACTS devices. When four or less FACTS devices are installed, the
phenomenon cannot be mitigated due to insufficient VAR support. Therefore, it was concluded
FIDVR phenomenon cannot be mitigated due to insufficient VAR support. Therefore, it was
that the FACTS devices should be installed at 2700, 4500, 2500, 4700, and 4650 to mitigate the
concluded that the FACTS devices should be installed at 2700, 4500, 2500, 4700, and 4650 to mitigate
FIDVR phenomenon.
the FIDVR phenomenon.
3.3. Discussion
3.3. Discussion
The results of the stability analysis identified the areas vulnerable to FIDVR and the candidate
The results of the stability analysis identified the areas vulnerable to FIDVR and the candidate
buses for most effective actions have been proved; the participation factors evaluated could provide
buses for most effective actions have been proved; the participation factors evaluated could provide
the valuable information on the adequate FACTS application against FIDVR. In addition, from the
the valuable information on the adequate FACTS application against FIDVR. In addition, from the
bus and branch participations, remedial actions to improve the overall system voltage stability could
bus and branch participations, remedial actions to improve the overall system voltage stability could
be decided. The above-mentioned results can be summarized in Table 4. From these results, one can
be decided. The above-mentioned results can be summarized in Table 4. From these results, one can
notice that the number and placement of FACTS devices to quickly recover voltage depends on the
notice that the number and placement of FACTS devices to quickly recover voltage depends on the
performance level. In addition, dynamic voltage recovery times for two FACTS implementation cases
performance level. In addition, dynamic voltage recovery times for two FACTS implementation cases
are obtained for the comparison purposes.
are obtained for the comparison purposes.
Table 4. Comparison results of Case 1 and Case 2.
Table 4. Comparison results of Case 1 and Case 2.
Fault Induced Delayed
Injection Buses Voltage
Voltage Recovery Time
Recovery Fault Induced Delayed
# of FACTS Voltage Recovery Mitigate?
# of Injection Buses
Case 1 Case 2 Time
Case 1 Case 2 Voltage Recovery
Case 1 Mitigate?
Case 2
FACTS
Case 1 Case 2 Case 1 Case 2 Case 1 Case 2
One 1400 2700 2.3416 s 2.5171 s No No
One
Two 1400
1400, 1300 2700
2700, 4500 2.3416 s
2.2141 s 2.5171 s
2.3984 s NoNo No
No
Two
Three 1300 1700
1400, 1300, 2700, 4500
2700, 4500, 2500 2.2141 s s 2.3984
0.4578 s s
2.2419 NoYes No
No
Four
Three 1400, 1300, 1700 2700,
2700, 4500,2500
4500, 2500, 4700 0.4578 s 1.6877
2.2419 s s Yes No
No
Five 2700, 4500, 2500, 4700, 4650 0.3975 s Yes
Four 2700, 4500, 2500, 4700 1.6877 s No
2700, 4500, 2500, 4700,
Five 0.3975 s Yes
In Table 4, it is indicated that when 4650the system voltage is restored up to 0.8 p.u. within 2 s, and that
FIDVR can be properly mitigated. For Case 1, the voltage was recovered within 2 s when three FACTS
were In Table 4,
applied. it Case
For is indicated that when
2, the voltage wasthe system within
recovered voltage2 is restored
s when fourup to 0.8 were
FACTS p.u. within 2 s, and
equipped, but
that FIDVR can be properly mitigated. For Case 1, the voltage was recovered within 2 s when
the transient voltage dip criteria were not satisfied, and hence FDIVR cannot be adequately alleviated. three
FACTS
It can bewere applied.
seen that a fiveFor Caseapplication
FACTS 2, the voltage was recovered
is required within
to mitigate 2 s when four FACTS were
the FIDVR.
equipped, but the transient voltage dip criteria were not satisfied, and hence FDIVR cannot be
adequately alleviated. It can be seen that a five FACTS application is required to mitigate the FIDVR.
 Sustainability 2018, 10, x FOR PEER REVIEW 16 of 19
Sustainability 2019, 11, 326 16 of 19

In Case 1, when the fault occurs, it is seen that reactive power cannot be supplied to the northern
metropolitan
In Case 1, area
whenfrom 3150occurs,
the fault CC. For this
it is seenreason, it is effective
that reactive power to injectbeFACTS
cannot supplieddirectly
to the into 1400.
northern
FACTS should also be installed on buses 1300 and 1700 that supply power to the
metropolitan area from 3150 CC. For this reason, it is effective to inject FACTS directly into 1400. metropolitan region.
In Caseshould
FACTS 2, it isalso
seenbethat FACTS
installed oninjection
buses 1300 directly on buses
and 1700 4700 and
that supply 4650toisthe
power effective. In addition,
metropolitan region.it
could be confirmed that it is effective to inject FACTS at 2700, 4500, and 2500
In Case 2, it is seen that FACTS injection directly on buses 4700 and 4650 is effective. In addition,to increase the supply
itofcould
reactive power in the
be confirmed thateastern area. It is
it is effective toseen
injectthat the system
FACTS voltages
at 2700, 4500, andcan2500
be successfully
to increase maintained
the supply
against this voltage instability scenario by five FACTS placed by the proposed methodology.
of reactive power in the eastern area. It is seen that the system voltages can be successfully maintained Figure
against this voltage instability scenario by five FACTS placed by the proposed methodology. Figuretwo
15 shows the FACTS locations for improved voltage stability in a metropolitan region for the 15
cases.
shows the FACTS locations for improved voltage stability in a metropolitan region for the two cases.

Figure15.
Figure 15.Representation
Representationof
ofFACTS
FACTSinjection
injectionlocation
locationininmetropolitan
metropolitanregion.
region.

Figure15
Figure 15shows
showsthat
thatthe
theFACTS
FACTScontrol
controlisisreasonably
reasonablyperformed
performedtotohelp
helpmaintain
maintainproper
propervoltage
voltage
regulationin
regulation inmetropolitan
metropolitanareas,
areas,especially
especiallyforforthe
thecritical
criticalcontingency.
contingency.As Asindicated
indicatedby
bythe
theabove
above
results, the information of bus participations can help determine the placements of FACTS,
results, the information of bus participations can help determine the placements of FACTS, and the and the
branchparticipations
branch participationscan
canbebeused
usedtotoidentify
identifycandidate
candidatelines
linesfor
forseries
seriescompensations.
compensations.This
Thishas
hasbeen
been
demonstratedininactual
demonstrated actualsystem
systemstudies,
studies,asasininthe
thediscussion.
discussion.

4.4.Conclusions
Conclusions
This
Thispaper
paperpresents
presentsaanewnewframework
frameworkon onthe
thereactive
reactivepower
powercompensation
compensationstrategy
strategydesigned
designed
based
based on voltage stability analysis. The framework adopts the conventional analysis of activeand
on voltage stability analysis. The framework adopts the conventional analysis of active and
reactive
reactivepower
powermargins
marginsto todecide
decidethethecritical
criticalcontingencies
contingenciesand and the
the areas
areas of
of reactive
reactive power
power deficiency.
deficiency.
This
Thispaper
paperadopts
adoptsnotnotonly
onlysteady-state
steady-stateanalysis
analysisbut
butalso
alsodynamic
dynamictime-domain
time-domainsimulations
simulationsfor forthe
the
adequate amount of dynamic reactive power support. There also might be a limitation
adequate amount of dynamic reactive power support. There also might be a limitation of only of only covering
solvable
coveringcases using
solvable theusing
cases conventional analysis,analysis,
the conventional but thisbutpaper adoptsadopts
this paper the sensitivity analysis
the sensitivity for
analysis
unsolvable cases using the branch parameterized continuation power flow. For
for unsolvable cases using the branch parameterized continuation power flow. For adequate adequate compensation
strategies,
compensationvulnerable buses
strategies, as compensation
vulnerable buses aslocations must belocations
compensation carefullymust
chosen
be in terms ofchosen
carefully voltagein
stability.
terms ofThe casestability.
voltage study results demonstrated
The case study resultsthatdemonstrated
the compensation level
that the decided by the
compensation proposed
level decided
method could mitigate the FIDVR caused by IM loads. In addition, the proposed
by the proposed method could mitigate the FIDVR caused by IM loads. In addition, the proposed method might be
applied
methodfor determining
might be appliedthefor
optimal numberthe
determining and placement
optimal numberof FACTS devices toofresolve
and placement FACTSthe FIDVR.
devices to
resolve the FIDVR. By the performance assessment, the alleviating impacts of dynamic VAR
Sustainability 2019, 11, 326 17 of 19

By the performance assessment, the alleviating impacts of dynamic VAR compensation on the FIDVR
phenomenon can be observed. Consequently, the dynamic support of reactive power could enhance
the STVS. The contributions of this paper can be summarized as follows:

• This paper proposed an approach to deciding the appropriate reactive power compensation,
considering the FIDVR problem. The proposed scheme can provide proper remedial actions in a
reliable manner for various post-disturbance conditions.
• The approach proposed in this paper solved the optimal dynamic VAR allocation problem by
systematically interfacing power system time-domain simulation.
• This strategy might further help explore and deploy more availability options when the desired
injection amount fits VAR resource options for unsolvable cases.

Author Contributions: Y.L. conceived and designed the research, conducted static and dynamic simulations,
and wrote the paper. H.S. improved the theoretical part and supervised the research.
Funding: This research received no external funding.
Acknowledgments: This research was supported by the Human Resources Development of the Korea Institute of
Energy Technology Evaluation and Planning (KETEP) grant funded by the Korea government Ministry of Trade,
Industry & Energy (20174030201840).
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Paramasivam, M.; Salloum, A.; Ajjarapu, V.; Vittal, V.; Bhatt, N.B.; Liu, S. Dynamic optimization based
reactive power planning to mitigate slow voltage recovery and short term voltage instability. IEEE Trans.
Power Syst. 2013, 28, 3865–3873. [CrossRef]
2. Glavic, M.; Novosel, D.; Heredia, E.; Kosterev, D.; Salazar, A.; Habibi-Ashrafi, F.; Donnelly, M. See it fast to
keep calm: Real-time voltage control under stressed conditions. IEEE Power Energy Mag. 2012, 10, 43–55.
[CrossRef]
3. Williams, B.R.; Schmus, W.R.; Dawson, D.C. Transmission voltage recovery delayed by stalled air conditioner
compressors. IEEE Trans. Power Syst. 1992, 7, 1173–1181. [CrossRef]
4. Taylor, L.Y.; Hsu, S.M. Transmission voltage recovery following a fault event in the Metro Atlanta area.
In Proceedings of the Power Engineering Society Summer Meeting, Seattle, WA, USA, 16–20 July 2000;
pp. 537–542.
5. North American Electric Reliability Corporation. Technical Reference Paper Fault-Induced Delayed Voltage
Recovery; North American Electric Reliability Corporation: Princeton, NJ, USA, 2009.
6. Alaqeel, T.A.; Almohaimeed, S.A.; Suryanarayanan, S. A review of air conditioning motor loads stalling on
voltage recovery in the saudi electric grid. In Proceedings of the Power Symposium (NAPS), Morgantown,
WV, USA, 17–19 September 2017; pp. 1–6.
7. De Leon, J.D.; Taylor, C.W. Understanding and solving short-term voltage stability problems. In Proceedings
of the Power Engineering Society Summer Meeting, Chicago, IL, USA, 21–25 July 2002; Volume 2, pp. 745–752.
8. Lu, N.; Yang, B.; Huang, Z.; Bravo, R. The system impact of air conditioner under-voltage protection schemes.
In Proceedings of the Power Systems Conference and Exposition, Seattle, WA, USA, 15–18 March 2009;
pp. 1–8.
9. Shaffer, J.W. Air conditioner response to transmission faults. IEEE Trans. Power Syst. 1997, 12, 614–621.
[CrossRef]
10. Bai, H.; Ajjarapu, V. A novel online load shedding strategy for mitigating fault-induced delayed voltage
recovery. IEEE Trans. Power Syst. 2011, 26, 294–304. [CrossRef]
11. Wibowo, R.S.; Yorino, N.; Eghbal, M.; Zoka, Y.; Sasaki, Y. FACTS devices allocation with control coordination
considering congestion relief and voltage stability. IEEE Trans. Power Syst. 2011, 26, 2302–2310. [CrossRef]
12. Sapkota, B.; Vittal, V. Dynamic VAr planning in a large power system using trajectory sensitivities. IEEE Trans.
Power Syst. 2010, 25, 461–469. [CrossRef]
13. Farsangi, M.M.; Nezamabadi-Pour, H.; Song, Y.H.; Lee, K.Y. Placement of SVCs and selection of stabilizing
signals in power systems. IEEE Trans. Power Syst. 2007, 22, 1061–1071. [CrossRef]
Sustainability 2019, 11, 326 18 of 19

14. Ghahremani, E.; Kamwa, I. Optimal placement of multiple-type FACTS devices to maximize power system
loadability using a generic graphical user interface. IEEE Trans. Power Syst. 2013, 28, 764–778. [CrossRef]
15. Tiwari, A.; Ajjarapu, V. A computer package for multi-contingency constrained reactive power planning.
In Proceedings of the Power & Energy General Society Meeting, Denver, CO, USA, 26–30 July 2015; pp. 1–5.
16. Huang, W.; Sun, K.; Qi, J.; Xu, Y. A new approach to optimization of dynamic reactive power sources
addressing FIDVR issues. In Proceedings of the 2014 IEEE/PES Power Systems Conference and Exposition,
National Harbor, MD, USA, 27–31 July 2014; pp. 1–5.
17. Huang, W.; Sun, K.; Qi, J.; Xu, Y. Voronoi diagram based optimization of dynamic reactive power sources.
In Proceedings of the Power & Energy Society General Meeting, Denver, CO, USA, 26–30 July 2015; pp. 1–5.
18. Wildenhues, S.; Rueda, J.L.; Erlich, I. Optimal allocation and sizing of dynamic var sources using heuristic
optimization. IEEE Trans. Power Syst. 2015, 30, 2538–2546. [CrossRef]
19. Xu, Y.; Dong, Z.Y.; Meng, K.; Yao, W.F.; Zhang, R.; Wong, K.P. Multi-objective dynamic VAR planning against
short-term voltage instability using a decomposition-based evolutionary algorithm. IEEE Trans. Power Syst.
2014, 29, 2813–2822. [CrossRef]
20. Liu, H.; Krishnan, V.; McCalley, J.D.; Chowdhury, A. Optimal planning of static and dynamic reactive power
resources. IET Gener. Transm. Distr. 2014, 8, 1916–1927. [CrossRef]
21. Meliopoulos, A.S.; Cokkinides, G.; Stefopoulos, G. Voltage stability and voltage recovery: Load dynamics
and dynamic VAR sources. In Proceedings of the Power Systems Conference and Exposition, Atlanta, GA,
USA, 29 October–1 November 2006; pp. 124–131.
22. Tiwari, A.; Ajjarapu, V. Optimal allocation of dynamic VAR support using mixed integer dynamic
optimization. IEEE Trans. Power Syst. 2011, 26, 305–314. [CrossRef]
23. Di Fazio, A.R.; Russo, M.; Valeri, S.; De Santis, M. Sensitivity-based model of low voltage distribution
systems with distributed energy resources. Energies. 2016, 9, 801. [CrossRef]
24. Seo, S.S.; Kang, S.G.; Lee, B.J.; Kim, T.K.; Song, H.C. Determination of reactive power compensation
considering large disturbances for power flow solvability in the korean power system. J. Electr. Eng. Technol.
2011, 6, 147–153. [CrossRef]
25. Hamzah, I.A.; Yasin, J.A. Static Var compensators (SVC) required to solve the problem of delayed voltage
recovery following faults in the power system of the Saudi electricity company, western region (SEC-WR).
In Proceedings of the IEEE Bologna Power Tech Conference, Bologna, Italy, 23–26 June 2003; Volume 4.
26. Kundur, P.; Paserba, J.; Ajjarapu, V.; Andersson, G.; Bose, A.; Canizares, C.; Van Cutsem, T. Definition and
classification of power system stability. IEEE Trans. Power Syst. 2004, 19, 1387–1401.
27. Hill, D.J. Nonlinear dynamic load models with recovery for voltage stability studies. IEEE Trans. Power Syst.
1993, 8, 166–176. [CrossRef]
28. Kundur, P.; Balu, N.J.; Lauby, M.G. Power System Stability and Control; McGraw-Hill: New York, NY, USA,
1994; Volume 7.
29. Van Cutsem, T.; Vournas, C. Voltage Stability of Electric Power Systems; Springer Science & Business Media:
New York, NY, USA, 1998; Volume 441.
30. Electric Power Statistics Information System. Available online: http://epsis.kpx.or.kr/epsisnew/ (accessed
on 2 November 2018).
31. Lu, N.; Xie, Y.; Huang, Z.; Puyleart, F.; Yang, S. Load component database of household appliances and
small office equipment. In Proceedings of the Power and Energy Society General Meeting-Conversion and
Delivery of Electrical Energy in the 21st Century, Pittsburgh, PA, USA, 20–24 July 2008; pp. 1–5.
32. Paserba, J.J.; Leonard, D.J.; Miller, N.W.; Naumann, S.T.; Lauby, M.G.; Sener, F.P. Coordination of a distribution
level continuously controlled compensation device with existing substation equipment for long term var
management. IEEE Trans. Power Syst. 1994, 9, 1034–1040. [CrossRef]
33. Paramalingam, J.; Nakamura, F.; Matsuda, A.; Yamanaka, D.; Tsuchiya, T. Application of FACTS Devices for
a Dynamic Power System within the USA. In Proceedings of the International Power Electronics Conference,
Niigata, Japan, 20–24 May 2018; pp. 2329–2334.
34. Larsen, E.V.; Miller, N.W.; Nilsson, S.L.; Lindgren, S.R. Benefits of GTO-based compensation systems for
electric utility applications. IEEE Trans. Power Deliv. 1992, 7, 2056–2064. [CrossRef]
35. Hingorani, N.G.; Gyugyi, L.; El-Hawary, M. Understanding FACTS: Concepts and Technology of Flexible AC
Transmission Systems; IEEE Press: New York, NY, USA, 2000.
Sustainability 2019, 11, 326 19 of 19

36. Pourbeik, P.; Koessler, R.J.; Quaintance, W.; Wong, W. Performing comprehensive voltage stability studies for
the determination of optimal location, size and type of reactive compensation. In Proceedings of the Power
Engineering Society General Meeting, Montreal, QC, Canada, 18–22 June 2006; pp. 1–6.
37. Western Electricity Coordinating Council. Voltage Stability Criteria, Undervoltage Load Shedding Strategy,
and Reactive Power Reserve Monitoring Methodology; Reactive Power Reserve Work Group Technical Studies
Subcommittee Western Electricity Coordinating Council: San Diego, CA, USA, 2007. Available online:
http://www.wecc.biz/main.html (accessed on 15 November 2018).
38. Lee, B.; Song, H.; Kwon, S.H.; Jang, G.; Kim, J.H.; Ajjarapu, V. A study on determination of interface flow
limits in the KEPCO system using modified continuation power flow (MCPF). IEEE Trans. Power Syst. 2002,
17, 557–564.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).