Professional Documents
Culture Documents
Liwei Wanga
a
School of Engineering, University of British Columbia, Kelowna, BC V1V1V8 Canada
b
Institute of Electrical Systems and Energy, Federal University of Itajuba, Itajuba, MG 37500 Brazil
Keywords: Voltage stability margins of modernbulk power systems are significantly affected by renewable energy sources
Flexible resources (RESs). The intermittency of these units can constantly move the critical location driving the system to voltage
Voltage stability collapse along different coherence regions. Traditional approaches were not designed to deal with this scenario,
Variable renewable generation typically providing a centralized actuation based on a system-wide perspective. In this sense, this paper proposes
a novel area-based outlook to take advantage of the new possibilities enabled by power systems distributed
controllable resources, e.g. flexible resources, to tackle this critical operational challenge. For this sake, a novel
area-based sensitivity index, exploring both the network-wide sensitivity and the local characteristics of voltage
collapse is proposed to determine the most effective buses for voltage support and their respective capability of
increasing the system's load margin. Comparative case studies indicate the superior capability of the proposed
direct method to deal with the new perspectives of modern power systems.
1. Introduction flexible resources to improve several network aspects (e.g., unit com-
mitment, peak shaving, and load shifting) has been recently in-
Securing a safe voltage stability margin is one of the most critical vestigated in the literature [5,6]. In this perspective, this paper revisits
aspects in guaranteeing the reliability of power systems [1]. Tradi- the voltage stability problem proposing a novel adaptive area-based
tionally, the required voltage stability margins were achieved using voltage stability index that can be used along with flexible resources to
VAR compensating devices; e.g., static VAR compensators (SVC) [2]. ensure a secure voltage stability margin under varying renewable en-
These components were typically allocated at the network's most cri- ergy sources (RES). The proposed index avoids overcompensation in
tical nodes as identified by a system-wide sensitivity analysis [3]. specific system regions, and consequently early saturation of the vol-
However, this can lead to overcompensation in specific system regions; tage stability margin improvement capacity, by adaptively determining
as the identified critical nodes tend to be close to one another [3]. the network most critical cores and areas considering the changes in
Moreover, as power systems move towards an increased penetration of RESs generation. To this end, first a network-wide sensitivity analysis is
renewable energy sources (RESs), these approaches may no longer be performed to provide a general perspective of the system voltage sta-
efficient, since the variable and uncertain generation of such resources bility. Next, the local characteristics of voltage collapse are explored,
can significantly affect the system static stability margins [2–4] and can leading to the calculation of the proposed area-based sensitivity index.
continuously move the starting point of voltage collapse across different This index can directly determine the most efficient locations and their
regions of the system. respective capability to improve the system voltage stability. Flexible
On another avenue, recent developments in monitoring and control resources are then used at these locations to provide the required vol-
have been leading to a pronounced improvement in power system tage support.
flexible resources capacity [5], including the use of energy storage
systems (ESS), distributed generation (DG),electric vehicles (EVs) and
demand response (DR) to provide load flexibility. The use of such
Corresponding author.
⁎
E-mail addresses: mairaribas@unifei.edu.br (M.R. Monteiro), yuri.rodrigues@ubc.ca (Y.R. Rodrigues), morad.abdelaziz@ubc.ca (M. Abdelaziz),
zambroni@unifei.edu.br (A.C.Z. de Souza), liwei.wang@ubc.ca (L. Wang).
https://doi.org/10.1016/j.epsr.2020.106384
Received 1 November 2019; Received in revised form 11 March 2020; Accepted 15 April 2020
Available online 16 May 2020
0378-7796/ © 2020 Elsevier B.V. All rights reserved.
M.R. Monteiro, et al. Electric Power Systems Research 186 (2020) 106384
2. Literature review probabilistic flexibility is used to avoid RES power curtailment. Re-
ference [25] presents the application of different emerging flexible re-
Voltage stability has been continuously under investigation within sources as sources of scheduled energy reserve in the electricity market.
the power system industry and academic community given the mag- Further, a new power system expansion planning approach is proposed
nitude of its consequences and direct association to major blackout in [26] considering the participation of flexible resources aimed at the
events worldwide [2]. A large body of research work has been dedi- reduction of new assets acquisition.
cated to the assessment of power system voltage stability, including Moreover, flexible resources are especially important to voltage
seminal works as continuation power flow methods associated with P-V stability support as modern power system move toward large integra-
curve analysis [7], L-index [8], singular value of the load flow Jacobian tion of variable renewable resources, which can lead to significant re-
[9], tangent vector [10] and voltage collapse proximity indicator ductions in the system voltage stability margin and variations in the
(VCPI) [11]. In addition, new indexes were recently proposed con- critical location responsible for driving the system to voltage collapse.
sidering the novelpower systems perspectives. The work developed in In [27] curtailable loads are considered as active market participants in
[12] presents a new voltage stability index based on power flow di- heavily loaded systems seeking to ensure power system voltage stabi-
vergence and decay of voltage magnitude. The study in [13] proposes a lity. The work developed in [28] proposes the application of DR to
new P-index able to indicate the steady-state distance to voltage col- support voltage control actions, while considering the system voltage
lapse, as well as dynamic voltage stability. Next, in [14] submodular stability preservation. Next, the work proposed in [29] expands the
optimization is explored as a tool to prevent voltage instability. Whilst perspective presented in [28], featuring optimal voltage regulation with
[15] seeks to mitigate the risk voltage stability taking advantage of the DR participation for both primary and secondary control levels. Still,
increase observability provided by PMUs. the current state-of-art falls short in the proposal of a method able to
Based on these works, several methods have been proposed seeking take advantage of these distributed controllable resources to ensure
the improvement of power systems’ voltage stability margins through secure voltage stability margins. This perspective is depicted in Table 1,
the insertion of compensation devices. In [16] a strategy based on the presenting a comparison between available works in the literature and
bifurcation method is used to determine the optimal location, control the proposed approach.
and appropriate sizes of SVCs and thyristor-controlled series compen-
sation (TCSC) to avoid voltage collapse. The work in [3] uses the 3. Proposed approach
system-wide voltage stability perspective provided by the tangent
vector to determine the most critical locations for reactive power Based on the major opportunities enabled by flexible resources and
compensation. Heuristic solutions based on genetic algorithm are ex- the new voltage stability requirements of modern power systems, this
plored by [17] and [18] considering transmission lines loadability paper proposes a novel area-based voltage stability index. The proposed
constraint and the L-index as optimization criteria. These works seek to index seeks to provide a direct measure for flexible resources dispatch,
adequately determine the locations and capacities for the installations with the objective of guaranteeing reliable voltage stability margins for
of flexible AC transmission systems (FACTS), and superconducting modern bulk power systems with significant penetration of RES gen-
magnetic energy storage (SMES). Multi-objective programming ap- eration. For this, the proposed approach is composed of four main steps,
proaches for static and dynamic (short-term) voltage stability im- respectively 1) voltage stability margin determination, 2) system-wide
provement were developed in [19] and [2]. These approaches em- sensitivity, 3) local area-based sensitivity, and 4) proposed index deri-
ployed VCPI for quantifying the system voltage stability seeking the vation. To perform these steps, knowledge of the system current oper-
adequate placement of static synchronous compensators (STATCOMs) ating condition, i.e. loading, generation and system parameters, and
and multi-stage planning for aged SVC equipment retirement. forecasted direction of load increase are required. Using this informa-
More recently, new opportunities due to bulk power system ex- tion, first the system voltage stability margin is assessed employing the
pansion/modernization have been drawing significant attention and continuation method. Based on this result, a system wide-sensitivity is
motivated important novel studies proposed in the literature. These derived indicating the susceptibility of each bus to lead the system to
works take advantage of the significant penetration of RESs and ex- voltage collapse. From this overall perspective, multiple coherence
pansion of HVDC systems to improve power systems voltage stability areas are determined and local area-based sensitivities obtained. This
margin without the requirement of new compensators installation. In information is used in the development of the proposed sensitivity
[20] the reactive power produced by grid-connected variable-speed index, providing a direct multi-area perspective for flexible resources
wind generators was used to enhance steady-state voltage stability dispatch.
margin. A new perspective considering both active and reactive power To validate the proposed approach, comparative simulation case
injections of photovoltaic systems was proposed in [21] to provide studies are carried out in Matlab considering the IEEE 118-bus test
dynamic voltage support to short-term voltage stability. Additional system with modifications to reflect the progress of modern power
control actions are developed in [22] to utilize VSCeHVDC for global systems toward large penetration of RESs and flexible resources. In this
enhancement of steady-state voltage stability. Although interesting and environment, the proposed approach is stressed with several case stu-
with meaningful results, the above-mentioned methods somewhat de- dies respectively divided in: 1) Validation and performance evaluation,
tract the original design of these systems and may not be of great ap- 2) Sensitivity analysis 3) Comparative analysis between flexible re-
preciation in actual operative reality. Consequently, their applicability sources and RESs operation under voltage control for voltage stability
is limited, as many operators would prefer to run these units at unity support. These case studies showcase the need for voltage stability
power factor, given that active power production is the one typically support in systems with significant penetration of RESs, while depicting
rewarded [23]. the superior ability of the proposed approach to ensure safe operative
These possible limitations are avoidable using power system flexible conditions. The advantages provided by the proposed approach in
resources capacity [5]. These resources can have their operation comparison to the state-of-art are threefold and described as follows:
adapted to better accommodate power system requirements without
compromising their core values design. Moreover, differently from the • Newoutlook for voltage stability support of modern power system
above-mentionedapplications that may have their actions limited by the based on flexible resources: the proposed area-based design enables
requirements of the actual site at which the equipment are installed. the distributed potential of flexible resources to fulfill the sounding
Flexible resources are widespread in the network, allowing for their use requirements of modern power systems voltage stability. Different
at the most effective locations to support various power system op- from other approaches that would require new infrastructures or
eration and planning requirements [24–29].In [24] a new assessment of detract the original design of the existing explored system, as e.g.
2
M.R. Monteiro, et al. Electric Power Systems Research 186 (2020) 106384
×
×
×
×
×
×
×
are meant for this kind of requirement.
•
×
Prevention of early saturation in voltage stability margin improve-
×
×
×
×
×
ment: The identification of critical cores and system partitioning
allow the proposed index to determine the most effective buses for
voltage stability support along different coherent areas of the net-
Iterative
solution
method
×
driving the system to voltage collapse would prevent local over-
based solution
margin improvement.
method
reserve
energy
collpase.
×
with flexible resources, the procedure for the system voltage stability
RES curtailment
margin
Secure
×
×
×
×
and respectively denote the sets of nodes with generation and loads,
being possible that a node is contained in both sets, re-
stability
Voltage
×
×
×
×
×
generic node i.
Pi = PiL ·(1 i, flex ), 0 <1 ki (1a)
resources
i, flex
Flexible
2
DR
×
×
×
×
×
Vi Vi
Pi L = Pi L 0 . i p · + i p · + i p
Vi0 Vi0
RES
(1b)
×
×
×
×
×
×
node i, constant current (I) and power (P) shares are defined by ip
and ip , being the participation coefficients subjected to the relation
Load
{( i · + i · + i · ) = 1: [ i · , i · , i · ] 3| = [0, 1]}, i .
×
×
×
×
×
×
×
tric vehicles.
Table 1
3
M.R. Monteiro, et al. Electric Power Systems Research 186 (2020) 106384
time-shifting of specific groups of interruptible loads due to operation representing the respective power system that may no longer be stable
requirements [28]. This perspective is following modeled for a generic to the current operating condition, i.e. it may lead to the merging of two
node iby the respective model depicted in [33], equilibrium points with an abrupt crossing of eigenvalues from the
right-half plane to the left-half plane side. It should be noted that the
Pi, DR = pia , = [pi1 , …, pi Ai ]
Ai
literature has consistently demonstrated that voltage collapse is mainly
a Ai (2)
connected to SNBs and LIBs [37-39].
where Pi, 〈DR〉 is the total nodal i demand response capacity, Ai is set of In this perspective, a widely used approach in the literature to de-
interruptible/adjustable loads connected at node i, pia denotes the termine power system's voltage stability margin, i.e. the distance be-
corresponding load demand for each load a ∈ Ai. tween the current operating point to the point of voltage collapse, or
Following, distributed generation enables the injection of specific SNB/LIB point, is based on the continuation method. This method is
controllable contributions of power to the grid. These contributions are divided into two steps: predictor and corrector. The first step is re-
limited by their current dispatching power and respective rated gen- sponsible for increasing the network load and generation based on a
eration capacity [34]. direction of load increase [ ] .This direction is assumed to be
aknownquantity, which is a reasonable assumption provided adequate
{Pi, : Pimin Pi, Pimax
, D G , (Pi, )2 + (Qi, )2 (Simax 2
, DG ) }
DG , DG DG DG DG
load forecast [39–40]. In case that a reliable load forecast is not
(3) available, a worst-case direction can be used to ensure that the system
where Pi, 〈DG〉 and Qi, 〈DG〉 represent the unit current active and reactive presents a satisfactory distance to voltage collapse [41]. Procedures to
power generation, Pimin , DG and Pi, DG denote minimum and maximum
max the determine the worst-case direction are detailed depicted in [7–42].
active power generation, and Simax, DG denotes the DG rated power.
In addition, equipment as EVs may perform both above-mentioned Pi,[ ] + j ·Qi,[ ] = Pi (·)
p0 + j·Qi p0 · [ ], : PV + PQ
4.3. Voltage stability margin where Y[ ] denotes the tangent vector sensitivity calculated in an equi-
librium condition at a generic iteration τ, Dx f |p[ ] is the Jacobian matrix,
Let each transmission node contained in the set be classified is a scalar gain used to control the direction increase and
under one of the three operative modes: PV, PQ or Vθ, respectively f / |p[ ] denotes the initially injected reactive and active powers.
composed in subsets PV , PQ and V . The power system network To reduce the expensive computational effort as the system ap-
can be represented by a generic nonlinear model: proaches voltage collapse,stopping criteria can be included to terminate
the continuation process. The stopping criteria denoted by ξ is a voltage
0 = f (x , ) (5) collapse index that can anticipate the vanishing eigenvalue, tending to
where f : × n is a nonlinear function with space dimension
n zero as the bifurcation is approached. It is directly determined from (8)
n = (| PV | + 2·| PQ |) , | · | is an operator that returns the cardinality of for a given operating point having a negligible calculation cost [43].
a generic set, x: =[ i, …, | PV |+| PQ |, Vi , …, V| PQ |]T n| i de- Reference [43] also addresses that employing the CRIC method can
notes the system state variables, i represents a generic transmission reduce computational burden, turning the continuation method eligible
node, is a particular parameter that moves the system from one for load margins calculation.
equilibrium point p0 to another p[τ], where represents the con- = Y[ T ·D f · Y[
] x p[ ] ] (8)
tinuation iteration, and 0 n
represents a vector of all zeros.
The system voltage stability can be studied by driving the system
from a known initially stable operating condition up to instability. This 4.4. Proposed area-based voltage stability support
instability condition is defined as bifurcation point and can be divided
into two main category groups. The first group represents instability Based on the voltage collapse information the proposed area-based
conditions due to the merging of two equilibrium points leading to a sensitivity index is developed. For this sake, a novel methodology
null eigenvalue, i.e. characterizing either a saddle-node (SNB), tran- composed of three main stages is proposed: 1) system-wide perspective
scritical (TB) or pitchfork bifurcation (PB); or by the crossing of the for the definition of most critical buses; 2) area-based sensitivity pro-
imaginary axis by a pair of complex conjugate eigenvalues, re- viding the characterization of critical cores and coherence areas; and 3)
presenting a Hopf bifurcation (HP) [37]. The second group is related to area-based voltage stability index responsible to determine the most
power system operative limits, especially generators reactive power effective buses and their respective contribution for voltage stability
limits, being defined as limited induced bifurcations (LIB). This type of support.
bifurcation occurs due to violations in generators limits, leading to First, a system-wide sensitivity is obtained indicating the suscept-
changes in these units operating mode from constant active power and ibility of each bus to lead the system to voltage collapse. This sensitivity
voltage mode, i.e. PV mode, to constant active and reactive power is achieved through the calculation of the tangent vector at the vicinity
mode, i.e. PQ mode, [38]. This action leads to a new set of equations of the voltage collapse point denoted by Y#: = dx /d |p p# ; reference on
4
M.R. Monteiro, et al. Electric Power Systems Research 186 (2020) 106384
T
Y# = :[Y#PV , Y#PQ] (9a)
T
Y#PQ = Y 1,#, …, Y PQ ,#
, V
Y1,#, …, YV PQ ,#
(9b)
YPQ PQ
,# : =rank (abs (Y# )) (10a)
T
YPQ =: YPQ PQ
,1 , …, Y ,2·
,# PQ Fig. 1. P-V curves for a system-wide approach.
(10b)
B† = {( n†): †
= PQ
YiPQ PQ
, n = Y ,n, i [1, 2· PQ ], n [1, ]}
n i
In this perspective, the development of voltage stability support
(10c) strategies based on system-wide approaches may no longer be efficient
where rank↓( · ) is a function that returns the descending rank of a to ensure safe operating conditions, as power systems move toward new
generic vector, YPQ,# is the ranked criticality index, i
PQ
corresponds to voltage stability requirements due to the increasing penetration of
the buses that make up Y# , and n denotes the ranked buses.
PQ † RESs. This new perspective requires a significant capacity for voltage
From this process, the buses classified as the most effective for stability margin improvement and ability to effectively tackle constant
voltage stability support are typically contained in the same neigh- shifts in the starting point of voltage collapse across different regions of
borhood of the node responsible for driving the system to voltage col- the system. Seeking to accomplish these challenges and avoid the pre-
lapse, leading to a concentrated actuation. However, bulk power sys- viously mentioned limitations, a second stage is developed based on the
tems usually contain more than one coherence area; i.e., a region of the previous system-wide sensitivity. The proposed approach identifies
systemcomprising buses with similar behavior, here denoted by the sets local critical cores for voltage collapse and divide the system into co-
m , where m [1, ] indicates the area index and denotes the herence areas denoted by the sets m . The use of coherence areas
total number of areas. In this sense, bearing in mind that the voltage is proven as an effective solution to improve voltage related actions due
collapse is a local phenomenon, the characterization of the buses sen- to these phenomena local characteristics, e.g. disturbance propagation
sibility from a system-wide perspective in bulk power systems en- and security regions applications [44-45], however the proposed per-
vironments would not render the expected effect, as the centralized spective and application have not yet been explored. The definition of
action in the system most critical region, 1, would not produce critical cores avoids the centralized compensation performed by the
meaningful improvements in other coherence regions, i.e. system-wide approach, allowing a greater range for the system voltage
P VS †
m P VS m m 1, where P VS m are re- stability margin improvement, ζ. Moreover, as compensation becomes
presentations of the system voltage stability when each respective co- distributed throughout the network, it is possible to continually tackle
herence area m assumes the condition to drive the system to voltage the development of voltage collapse even when a change in the region
collapse; superscript † denotes the system response after system-wide conducive to this incidence is imposed.
based compensation. To obtain P VS m in case that the area under This feature is clearly shown inFig. 2. The proposed area-based
analysis is not the most critical region originally leading the system to strategy can improve all coherence areas voltage stability margin, i.e.
voltage collapse, i.e. m|m 1, compensations are performed in the P VS ‡
m P VS m m [1, ]. Therefore, allowing the
more critical region(s) until the voltage collapse condition is shifted to achievement of the required/desired voltage stability margin im-
the region m of interest. provement, | = 0 , even for scenarios demanding large im-
In this perspective, system-wide based approaches would fail in case provements in the system loading margin. The system responses after
that a significant improvement in the system load margin is required. the proposed area-based voltage stability support are denoted by the
The realization of a large compensation would render a sub-
stantialincrease in the voltage stability margin of the critical region,
1 . However, under a system perspective, this action would
†
P VS
only move the starting point of voltage collapse to the subsequent most
critical region, m|m 1, which has not been improved by this com-
pensation. This leads to an early saturation in the actual improvement
capacity of the system voltage stability by the subsequent most critical-
area voltage collapse condition, i.e. P
2 , where P VS
† †
VS S† = P VS 2 |P VS 2 P VS S†
depicts the actual system P-V curve after system-wide compensation. As
a result, the system would most likely not be able to achieve the desired
load margin improvement, ζ, where the respective lacking margin is
denoted by χ.
This conditionis clearly depicted in Fig. 1. It shows the voltage
stability profile of a bulk power system with several coherence areas
before and after the system-wide compensation. Where symbol opS
denotes the system operating point, and ɛ% is the required secure Fig. 2. P-V curves for the proposed area-based approach.
margin between the operative point and the point of voltage collapse.
This value is established by specific regulations as [1].
5
M.R. Monteiro, et al. Electric Power Systems Research 186 (2020) 106384
bus and the previously associated buses contained in the same area.
Discussions about the determination of the partitioning parameter are
omitted due to space constraints, the reader is referred to [10] where a
comprehensive analysis is developed indicating that φ should be set
within the range of 0.5 to 0.75. This process is performed for all but the
Vθ bus, V , which is manually associated with the respective co-
herency area at the end of the process. Next, the respective buses col-
lected in m are extracted from the set of available buses for network
partitioning and the process repeated. After completing the first area
division, the next critical core, m‡ , is determined as the most effective
PQ-bus that is not contained in the previous area(s) division and so on.
This procedure is repeated until each bus is allocated in a respective
area. The process that identifies the coherence areas
{ m, m [1, ]} around their respective critical cores
{ m‡ , m [1, ]} , is described as follows:
T
PQ
Ym,#: = Y i,#, …, Y PQ , YVi,#, …, YV PQ , i PQ
m
m ,# m ,#
(11a)
PQ
Y ‡m,#: =max (abs (Ym,#)) (11b)
m = {( ‡
m ):
‡
m = i
PQ
YiPQ ‡
, m,# = Y m,#, i [1, 2· PQ ]} (11c)
m+1 = m m (11e)
where m , m PQ
and Ym ,# respectively denote the sets of all buses, PQ-
PQ
buses and criticality index of flexible PQ-buses that are not associated to M,( )
‡ ‡ ‡ i,
any area. Y ‡m,# represents the criticality index of the critical core m‡ . M,( ) = ( i, ): i, = · M
M=1 M,( ) I = 1 I,( )
m represents the buses collected in a respective coherence area m ,
while βi denotes the bus corresponding to the criticality index Yi,# . The Yi,( ) YM,( ), i [1, M ], M [1, ]} (12b)
initial conditions are given by 1 = , 1 = , 1PQ = PQ . The re-
spective set collecting all critical cores is defined by : ={ ‡
m, · , …, ‡
, · }, m [1, ] (12c)
B‡: ={ 1‡, …, ‡ } PQ
By using the critical cores and coherence areas knowledge, a novel where Yi, 〈∇〉 represents the criticality of a generic bus i , Γm is the
area-based voltage stability index is obtained for each bus. This index, number of buses used for compensation in area m , Ym, 〈∇〉 and
denoted by Ψi, ∇, is composed of two terms. The first term represents
‡
m, are sets containing the buses criticality and proposed index of a
the area's contribution based on the density of critical spots, while the respective area m , and Λ represents the total system flexibility
second term denotes the respective contribution from the bus based on usage, i.e. i Pi DR + Pi DG + Pi EV = .
its criticality information. This means that if one area has a very critical A detailed description of the process to obtain the proposed area-
location prone to voltage collapse but possesses a low density of critical based voltage stability index is illustrated in Fig. 3.
buses, the total compensation in this area will not be performed in
excess, as compensating this very critical spot renders an overall im- 5. Results and discussions
provement of the area's voltage stability. On the contrary, if an area has
a high density of critical points, additional compensation will be re- This section depicts the requirement for area-based voltage stability
quired, as the compensation of the single most critical location will not support as power systems move toward large penetration of RESs. Here
lead to an actual improvement of the area voltage stability profile. The comparative case studies are developed to showcase the improvements
proposed index ‡i, and the flexibility contribution of each bus, de- provided by the proposed approach for modern power systems’ voltage
noted by the set Ψ〈∇〉, are represented by the following equations stability support, in contrast to traditional direct system-wide strategies
and successive linear programming (SLP). For this purpose, case studies
Ym, = {(Yi, ): Yi, = Yi,#/Y‡m,#, m, i [1, m], m [1, ]}
i
are developed employing the IEEE 118-bus system modified to re-
(12a) present grid developments, where power systems are likely to
6
M.R. Monteiro, et al. Electric Power Systems Research 186 (2020) 106384
with low generation states can meaningfully affect the system voltage ment was not translated to the overall system stability margin expressed
stability margin and even lead the system to operating conditions be- by P VS S†. The centralized actuation in the most critical buses is
yond the voltage stability limit. This scenario requires effective voltage not able to improve other regions voltage stability,
support actions to ensure a reliable system operation, which is explored P VS †
m P VS m, m 1| m [1, 3]. In contrast, it
by the performed case studies considering flexible resources usage to leads to the overcompensation of the previously most critical region of
ensuresecure voltage stability margins in the absence of RESs con- the system, 1, leading to the change of the location responsible to
tribution. In this perspective, three main case studies are developed to drive the system to voltage collapse to the next most critical region, i.e.
demonstrate the proposed approach effectiveness: 1) Validation and 2 . Therefore, limiting the system voltage stability margin improve-
performance evaluation; 2) Sensitivity analysis; and 3) Comparative ment by the subsequent most critical area voltage collapse condition,
2 , which did not have its load margin sig-
†
analysis between flexible resources and RESs operating under voltage P VS S†=P VS
control for voltage stability support. nificantly affected by the performed compensation, i.e.
2 . In this perspective, the system improved
†
P VS 2 P VS
5.1. Validation and performance evaluation stability margin after voltage stability support based on system-wide
perspective is denoted by = 1.01 |P VS S†. The obtained condition
The first case study highlights modern power systems significant is still lower than the WECC safe operating margin criteria of 105%,
need for voltage stability support and the superior ability of the pro- remaining 4% of the required voltage stability margin to be fulfilled,
posed area-based approach to deal with the new system operative i.e. = 4% . This result would inevitably demand costly system expan-
reality and requirements. For this, comparisons with a traditional direct sions and installation of new infrastructures to reach the regulation
system-wide approach and SLP [46] are performed. The comparative criteria [1].
system-wide approach is based on the sensitivity of the tangent vector. Seeking to fulfill the secure voltage stability margin requirement
This technique was proposed in [10] and employed in [3] to determine and avoid costly system upgrades, the proposed area-based voltage
the set of candidate buses for voltage stability support. For this, four stability index is employed, being the results depicted in Fig. 4(b). From
sub-cases are developed, respectively: Case 1 - System behavior without Fig. 4(b) one can observe that the flexible resources usage based on the
voltage stability support; Case 2 - System behavior with flexible re- proposed area-based approach leads to an overall improvement in the
sources support based on traditional direct system-wide strategy; Case 3 system voltage stability margin. In this scenario, all system coherence
- System behavior with flexible resources support based on the pro- areas have achieved the voltage stability margin requirement by WECC,
posed direct area-based voltage stability index; Case 4 - System beha- i.e. 1.05 |P VS ‡
m m [1, 3]. As well, after performed
vior with flexible resources support based on successive linear pro- the proposed area-based voltage stability support, the system most
gramming. critical region was shifted to the coherence area 3,
3 . However, differently from the system-wide
‡
The respective voltage stability profile of the system for Cases 1–3 P VS S‡=P VS
are depicted in Fig. 4(a)-(b) where P VS S depicts the actual system strategy, the proposed approach distributed action in the network cri-
P-V curve, P VS tical cores can satisfactorily address this issue, overcoming the early
m are representations of the system voltage sta-
bility when each respective coherence area m assume the condition to saturation faced by the system-wide strategies and leading to a system
drive the system to voltage collapse. The system responses without voltage stability margin of = 1.05 |P VS S‡ . The obtained results
voltage stability support (Case 1) are denoted by the representations are due to the distributed actions in the network most critical cores,
without superscript, while the system behavior after the provision of which allows for a greater range of improvement to the system voltage
voltage stability support based on system-wide (Case 2) and the pro- stability margin, ensuring the required load margin improvement of
posed area-based approach (Case 3) are respectively denoted by the = 7%| = 0 with a flexibility usage of = 2.5 p.u..
superscripts † and ‡; op S indicates the system operating condition,ζ Following, a new case study is developed seeking to verify the
describes the necessary improvement in the system loading margin and proposed direct approach ability to determine the system most ade-
χ represents the residual voltage stability margin to ensure the system quate allocation for flexible resources support in comparison with SLP.
7
M.R. Monteiro, et al. Electric Power Systems Research 186 (2020) 106384
Fig. 4. P-V curves for the system before and after voltage stability support based on (a) system-wide strategy; (b) proposed area-based index.
The obtained results present significant similarity between the pro- Table 4
posed direct approach and SLP, i.e. 67% of the selected buses for Flexibility usage outlook for Λ=2.5 p.u.a.
flexible resources usage are common for both methods. Moreover, an Case 2: Traditional system-wide method
equivalent level of voltage stability margin improvement is achieve by
both methods using the same amount of flexibility, i.e. = 1.05 for Bus 109 108 106 101 94 102 95 93 86
both Case 3 and Case 4 with a flexibility usage of = 2.5 Ψi, ∇ (p.u.) 0.53 0.51 0.46 0.19 0.17 0.17 0.16 0.16 0.15
Case 3: Proposed area-based index, m = 3, [1, 3]
p.u..Theseresults showcase the proposed direct approach ability to
m
Bus i 109 108 106 41 39 44 33 35 117
meaningfully improve the system load margin at a significantly reduced Ψi, ∇ (p.u.) 0.29 0.28 0.25 0.29 0.29 0.22 0.29 0.29 0.29
computational cost. Eliminating the computational burn imposed by Case 4: Successive Linear Programming
the successive iterations necessary for gathering feedback information Bus i 109 108 106 41 39 117 2 3 13
required by optimization technics. Ψi, ∇ (p.u.) 0.25 0.24 0.21 0.53 0.35 0.24 0.32 0.21 0.07
Table 3
Voltage stability margin for different case studies. Next, the sensitivity of voltage stability margin with respect to
Load margin Max. Load (p.u.) Λ (p.u.) flexible resources usage are investigated and illustrated in Fig. 5. For
this, a strategy similar to the one presented in [29]is employed, in
a
Case 1 0.98 72.9 – which successive variations in the use of flexibility are performed and
Case 2 1.01 75.1 2.5
the voltage stability margin of the system is analyzed. As one may
Case 3 1.05 78.0 2.5
Case 4 1.05 78.3 2.5 observe, the proposed approach sustains significant improvements in
the system voltage stability margin with the increasing usage of flex-
a
op S = 74.4 p.u. ibility. Moreover, a large improvement in the operational range is
8
M.R. Monteiro, et al. Electric Power Systems Research 186 (2020) 106384
verified. This results is expected given the proposed area-based index necessary secure load margin established by WECC regulation, i.e.
distributed action in the network critical cores. In this perspective, all 105%. This perspective is significantly enhanced considering the use of
system coherence areas are compensated considering the criticality of flexible resources. As depicted by P V S‡ in Fig. 6, with the same
the respective region, therefore avoiding the overcompensation in a 8.3 p.u. of flexible resources usage, one can achieve a system voltage
single region and the consequent inefficiency in the overall improve- stability margin of 110%. This result significantly surpass WECC op-
ment of system voltage stability margin. For comparison sake, a tradi- erational requirements and represents an improvement of 42% in
tional system-wide approach is presented, where one can observe a comparison with RESs support. Demonstrating that flexible resources
rapidly saturation with a flexibility usage of = 0.4 due to its cen- are more effective than RESs operating under voltage control to im-
tralized compensation in the most critical nodes. prove the system voltage stability. These results are expected as flexible
In addition, it should be noted that local voltage stability sensitivity resources dispatch are performed from the system most critical nodes
metrics are embedded in the proposed approach to define the respective under voltage stability perspective, whereas RESs contribution are
candidate buses. In this sense, the selected buses for flexible resource conditioned to the location of the unit installation. These locations are
usage depicted in Table 4, represent the system most locally sensible typically subjected to the availability of potential capacity for the re-
buses under voltage stability perspective. spective development, e.g. wind generators, and may not be the most
adequate location for voltage stability support.
Still, a combined operation of RESs under voltage control and
5.3. Flexible resources and RES operating under voltage control flexible resources can significantly improve the system voltage stability
and lead to a moreefficient use of flexible resources. In this case, RESs
mathsize="6pt"This section seeks to assess flexible resources ability action take priority over flexible resources, which are used to fulfill the
to improve voltage stability in comparison, and in association, with RES remaining gap to ensure secure voltage stability margins. This com-
generation under voltage control. Different flexible resources are con- bined operation leads to meaningful benefits as flexible resources usage
sidered, including dispatchable DGs of rated power {15,8,10,14}MVA is significantly reduced. A comparison with Case 2 in Section 4.1 in-
connected in buses {109,108,41,33}; EVs with level 2 chargers pre- dicates that by combining the use of flexible resources and RESs under
senting charging/discharging capacity of 11 kW associated in the re- voltage control, a reduction of 56% in the total application of flexible
spective amounts, i.e. number of EVs, {250, 250, 380; 330, 460, 290, resources is achieved. Moreover, this reduction represents a decrease of
150, 240, 190} at selected buses for flexibility support, i.e. buses {109, 73% in DR usage when considering multiple sources of flexibility,
108, 106, 41, 39, 44 33, 35 117}; as well as DR capacity as necessary. which can significantly avoid possible curtailments requirements. The
The dispatch of flexible resources is based on priority rank seeking to detailed load margin is illustrated by P V S‡ in Fig. 6, while the
maximize demand supplying, i.e. DGs are the first to be used, followed contribution of each group is depicted in Table 6.
by EVs, and in case that additional flexibility is necessary DR is applied.
Given the paper goal of improving power systems voltage stability, 6. Conclusions
renewables operating under voltage control have their reactive power
focused on load margin enhancement [47], Traditional methods may no longer be efficient for securing safe
i.e. f = Cv·[ i max(0, |Vi | V max ) + i
max(0, |Vi | V min)], where voltage stability margin for power systems moving toward large pene-
each unit has the following local setting tration of intermittent RESs. In this paper, a novel area-based outlook
Q = Constr(Q^ + · P , S max 2 (P^ + P )2 )|V V max ,
i i i i i i i i using flexible resources is proposed to tackle this new power system
Constr(x , x max ) = {x , |x| x max ; ( |x| )·x max , otherwise } , wheref denotes
x
perspective. The proposed approach can overcome the early saturation
the objective function, Cv is the penalty term for voltage violation, Vmax in the improvement of power system voltage stability margin faced by
and Vmin denote the maximum and minimum operational voltage limits, traditional centralized strategies that may endanger the achievement of
Simax is the unit rated power capacity, αi is the slope coefficient and Q^i secure operative conditions for modern power systems. For this, the
system most critical cores and their respective coherence areas are
and P^ are reactive and active power setpoints.
i
identified. Based on this information a novel distributed voltage stabi-
Here three comparative case studies are presented: 1) RESs oper-
lity index is developed. The results indicate that the application of the
ating under voltage control, S* ; 2) Sole use of flexible resources, S‡ ; 3)
proposed index to identify effective buses for voltage support and their
Combined use of flexible resources and RESs operating under voltage
respective ability of increasing the system load margin, in association
control, S*,‡. The respective results are depicted in Fig. 6.
with flexible resources, is able to greatly improve and secure safe static
As one may observe, RESs operation under voltage control are able
voltage stability margins. The main contributions of this work can be
to improve the system load margin up to 103% using a combined total
summarized as follows:
of 8.3 p.u.. Still, their contribution is not sufficient for ensuring the
9
M.R. Monteiro, et al. Electric Power Systems Research 186 (2020) 106384
Table 6
Individual contributions for combined operation of flexible resources and RESs under voltage control for voltage stability margin improvement.
RES Bus i 106 108 101 79 78 118 43 35 37
Qg (p.u.) 1.5 1.0 0.7 0.4 1.5 1.4 0.2 0.5 1.1
Flexible resources Bus i 109 108 106 41 39 44 33 35 117
Ψi, ∇ (p.u.) 0.13 0.12 0.11 0.13 0.13 0.1 0.13 0.13 0.13
DGs 100% 67% – 77% – – 100% – –
EVs – 33% 76% 23% 78% 64% – 41% 32%
DR – – 24% – 22% 36% – 59% 68%
Software, Validation, Writing - original draft, Writing - review & optimization for voltage control, IEEE Trans. Power Syst 33 (1) (2018) 502–513.
editing. Morad Abdelaziz: Conceptualization, Data curation, Formal [15] S.C. Chevalier, P.D. Hines, Mitigating the risk of voltage collapse using statistical
measures from PMU data, IEEE Trans. Power Syst 34 (1) (2019) 120–128.
analysis, Funding acquisition, Investigation, Methodology, Resources, [16] C.A. Cañizares, Z.T. Faur, Analysis of SVC and TCSC controllers in voltage collapse,
Software, Validation, Writing - original draft, Writing - review & IEEE Trans. Power Syst 14 (1) (1999) 158–165.
editing. A.C. Zambroni de Souza: Conceptualization, Data curation, [17] S. Gerbex, R. Cherkaoui, A.J. Germond, Optimal location of multi-type FACTS de-
vices in a power system by means of genetic algorithms, IEEE Trans. Power Syst 16
Formal analysis, Funding acquisition, Investigation, Methodology, (3) (2001) 537–544.
Resources, Software, Validation, Writing - original draft, Writing - re- [18] X. Huang, G. Zhang, L. Xiao, Optimal location of SMES for improving power system
view & editing. Liwei Wang: Conceptualization, Data curation, Formal voltage stability, IEEE Trans. Appl. Supercond 20 (3) (2010) 1316–1319.
[19] Y. Xu, Z.Y. Dong, C. Xiao, R. Zhang, K.P. Wong, Optimal placement of static com-
analysis, Funding acquisition, Investigation, Methodology, Resources, pensators for multi-objective voltage stability enhancement of power systems, IET
Software, Validation, Writing - original draft, Writing - review & Gener. Transm.Distrib 9 (15) (2015) 2144–2151.
editing. [20] V.S. Kumar, K.K. Reddy, D. Thukaram, Coordination of reactive power in grid-
connected wind farms for voltage stability enhancement, IEEE Trans. Power Syst 29
(5) (2014) 2381–2390.
Declaration of Competing Interest [21] K. Kawabe, Y. Ota, A. Yokoyama, K. Tanaka, Novel dynamic voltage support cap-
ability of photovoltaic systems for improvement of short-term voltage stability in
power systems, IEEE Trans. Power Syst 32 (3) (2017) 1796–1804.
The authors declare that they have no known competing financial [22] O.A. Urquidez, L. Xie, Singular value sensitivity based optimal control of embedded
interests or personal relationships that could have appeared to influ- VSC-HVDC for steady-state voltage stability enhancement, IEEE Trans. Power Syst
ence the work reported in this paper. 31 (1) (2016) 216–225.
[23] M. Kayikci, J.V. Milanovic, Reactive power control strategies for DFIG-based plants,
IEEE Trans. Ener. Conv 22 (2) (2007) 389–396.
Acknowledgements [24] Z. Lu, H. Li, Y. Qiao, Probabilistic flexibility evaluation for power system planning
considering its association with renewable power curtailment, IEEE Trans. Power
Syst 33 (3) (2018) 3285–3295.
This work was partially supported by CAPES, CNPQ, and NSERC [25] E. H.-Forushani, M.E.H. Golshan, M. S.-khah, P. Siano, Optimal operation of
(Canada). The author Maíra M. Monteiro especially thanks CAPES and emerging flexible resources considering sub-hourly flexible ramp product, IEEE
CNPQ – Sandwich Doctorate Abroad/ Process n° 201128/2018/7, and Trans. Sustain. Ener 9 (2) (2018) 916–929.
[26] H. Hamidpour, J. Aghaei, S. Pirouzi, S. Dehghan, T. Niknam, Flexible, reliable, and
the author Yuri R. Rodrigues especially thanks CAPES Notice No. 18/ renewable power system resource expansion planning considering energy storage
2016 of the Full Doctoral Program Abroad/ Process n° 88881.128399/ systems and demand response programs, IET Renew. Power Gener. 13 (11) (2019)
2016-01. 1862–1872.
[27] A. Jalali, M.S. Sepasian, M.K.S.-E. Eslami, Undisruptive load curtailment scheme to
ensure voltage stability margin, IET Gener.Transm.Distrib 13 (9) (2019)
References 1509–1519.
[28] A. Rabiee, A. Soroudi, B. M.-ivatloo, M. Parniani, Corrective voltage control scheme
[1] Reactive Reserve Working Group (RRWG), “Guide to WECC/NERC planning stan- considering demand response and stochastic wind power, IEEE Trans. Power Syst.
dards I.D: Voltage support and reactive power,” Western Electricity Coordinating 29 (6) (2014) 2965–2973.
Council, Mar. 2006. [29] A. Rabiee, S.M. M.-Bonab, M. Parniani, I. Kamwa, Optimal cost of voltage security
[2] J. Liu, Y. Xu, Z.Y. Dong, K.P. Wong, Retirement-driven dynamic VAR planning for control using voltage dependent load models in presence of demand response, IEEE
voltage stability enhancement of power systems with high-level wind power, IEEE Trans. Smart Grid 10 (3) (2019) 2383–2395.
Trans. Power Syst 33 (2) (2018) 2282–2291. [30] B.P. Bhattarai, I.C. Mendaza, K.S. Myers, B. B.-Jensen, S. Paudyal, Optimum ag-
[3] A.C. Zambroni de Souza, L.M. Honório, G. L. Torres, Increasing the loadability of gregation and control of spatially distributed flexible resources in smart grid, IEEE
power systems through optimal-local-control actions, IEEE Trans. Power Syst. 19 Trans. Smart Grid 9 (5) (2018) 5311–5322.
(1) (2004) 188–194. [31] A. Akrami, M. Doostizadeh, F. Aminifar, Power system flexibility: an overview of
[4] M. Bakhtvar, A. Keane, Allocation of wind capacity subject to long term voltage emergence to evolution, J. Mod. Power Syst. Clean Energy 7 (5) (2019) 987–1007.
stability constraints, IEEE Trans. Power Syst 31 (3) (2016) 2404–2414. [32] IEA-PVPS Task 14, Flexible resources for flexible transmission system operation,
[5] J. Kazempour, B.F. Hobbs, Value of flexible resources, virtual bidding, and self- Int. Ener. Agency Photovoltaic Power Syst. Programme (2017) Milan, October.
scheduling in two-settlement electricity markets with wind generation-Part I: [33] A. M.-Rad, V.W.S. Wong, J. Jatskevich, R. Schober, A. L.-Garcia, Autonomous de-
principles and competitive model, IEEE Trans. Power Syst 33 (1) (2018) 749–759. mand-side management based on game-theoretic energy consumption scheduling
[6] H. M.-Rad, Optimal demand bidding for time-shiftable loads, IEEE Trans. Power for the future smart grid, IEEE Trans. Smart Grid 1 (3) (2010) 320–331.
Syst 30 (2) (2014) 939–951. [34] E. Lannoye, D. Flynn, M. O'Malley, Evaluation of power system flexibility, IEEE
[7] T. Van Cutsem, C. Vournas, Voltage Stability of Electric Power Systems, Springer, Trans. Power Syst 27 (2) (2012) 922–931.
Boston, 1998. [35] P. Palensky, E. Widl, M. Stifter, A. Elsheikh, Modeling intelligent energy systems:
[8] P. Kessel, H. Glavitsch, Estimating the voltage stability of a power system, IEEE co-simulation platform for validating flexible-demand EV charging management,
Trans. Power De 1 (3) (1986) 346–354. IEEE Trans. Smart Grid 4 (4) (2013) 1939–1947.
[9] P. Sauer, M.A. Pai, Power system steady-state stability and the load-flow Jacobian, [36] Y.R. Rodrigues, A.C. Zambroni de Souza, P.F. Ribeiro, An inclusive methodology for
IEEE Trans. Power Syst. 5 (4) (1990) 1374–1383. plug-in electrical vehicle operation with G2V and V2G in smart microgrid en-
[10] A.C. Zambroni de Souza, C.A. Cañizares, V.H. Quintana, New techniques to speed vironments, Int. J. of Electric Power &Ener. Syst 102 (2018) 312–323.
up voltage collapse computations using tangent vectors, IEEE Trans. Power Syst 12 [37] W. Rosehart, C.A. Canizares, V.H. Quintana, Effect of detailed power system models
(3) (1997) 1380–1387. in traditional and voltage stability constrained optimal power flow problems, IEEE
[11] M. Moghavvemi, O. Faruque, Real-time contingency evaluation and ranking tech- Power Engineering Review 22 (12) (2002) 59.
nique, IEEE Proc. Gener. Transm.Distrib. 145 (5) (1998) 517–524. [38] W. Rosehart, C. Canizares, V. Quintana, Optimal power flow incorporating voltage
[12] W.P.J. Philippe, S. Eftekharnejad, P.K. Ghosh, Development of a new voltage sta- collapse constraints, Proc. IEEE PES Summer Meeting. Conference Proceedings, 2
bility index and its implementation considering voltage, measurement uncertainty, Edmonton, 1999, pp. 820–825.
IET Gener. Transm.Distrib 13 (18) (2019) 4011–4020. [39] W.D. Rosehart, C.A. Canizares, V.H. Quintana, Multiobjective optimal power flows
[13] M. Kamel, A.A. Karrar, A.H. Eltom, Development and application of a new voltage to evaluate voltage security costs in power networks, IEEE Trans. Power Syst 18 (2)
stability index for on-line monitoring and shedding, IEEE Trans. Power Syst 33 (2) (2003) 578–587.
(2018) 1231–1241. [40] F. Alvarado, I. Dobson, Y. Hu, Computation of closest bifurcations in power systems,
[14] Z. Liu, A. Clark, P. Lee, L. Bushnell, D. Kirschen, R. Poovendran, Submodular IEEE Trans. Power Syst 9 (2) (1994) 918–928.
[41] G. Chen, D.J. Hill, X. Yu, Bifurcation Control, Theory and Applications, Springer,
10
M.R. Monteiro, et al. Electric Power Systems Research 186 (2020) 106384
Berlin, 2003. [45] M.R. Monteiro, G.F. Alvarenga, Y.R. Rodrigues, A.C.Z. de Souza, B.I.L. Lopes,
[42] M.M.A. Abdelaziz, E.F. El-Saadany, Determination of worst case loading margin of M.C. Passaro, M. Abdelaziz, Network partitioning in coherent areas of static voltage
droop-controlled islanded microgrids, Int. Conf. Electric Power and Ener. Conv. stability applied to security region enhancement, Int. J. of Electric Power &Ener.
Syst, Istanbul, 2013, pp. 1–6. Syst 117 (2020) 105623.
[43] F.W. Mohn, A.C.Z. de Souza, Tracing PV and QV curves with the help of a CRIC [46] E. Riccietti, S. Bellavia, S. Sello, Sequential linear programming and particle swarm
continuation method, IEEE Trans. Power Syst 21 (3) (2006) 1115–1122. optimization for the optimization of energy districts, Engineering Optimization 51
[44] H. Mehrjerdi, S. Lefebvre, M. Saad, D. Asber, A decentralized control of partitioned (1) (2019) 84–100.
power networks for voltage regulation and prevention against disturbance propa- [47] R.A. Jabr, Robust Volt/VAr control with photovoltaics, IEEE Trans. Power Syst 34
gation, IEEE Trans. Power Syst 28 (2) (2013) 1461–1469. (3) (2019) 2401–2408.
11