International Journal of Modelling and Simulation

ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/tjms20

Under-frequency load shedding of microgrid

systems: a review

T. Madiba, R.C. Bansal, N.T. Mbungu, M. Bettayeb, R.M. Naidoo & M.W. Siti

To cite this article: T. Madiba, R.C. Bansal, N.T. Mbungu, M. Bettayeb, R.M. Naidoo & M.W. Siti
(2022) Under-frequency load shedding of microgrid systems: a review, International Journal of
Modelling and Simulation, 42:4, 653-679, DOI: 10.1080/02286203.2021.1964061

To link to this article: https://doi.org/10.1080/02286203.2021.1964061

Published online: 20 Oct 2021.

Submit your article to this journal

Article views: 1025

View related articles

View Crossmark data

Citing articles: 27 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=tjms20
INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION
2022, VOL. 42, NO. 4, 653–679
https://doi.org/10.1080/02286203.2021.1964061

ARTICLE

Under-frequency load shedding of microgrid systems: a review

T. Madibaa, R.C. Bansalb,c, N.T. Mbungub,c,d, M. Bettayebb,e, R.M. Naidooc and M.W. Sitid
a
Department of Electro-mechanic, Polytechnic Faculty, University of Lubumbashi, Lubumbashi, DRC; bDepartment of Electrical Engineering,
University of Sharjah, Sharjah, United Arab Emirates; cDepartment of Electrical, Electronic and Computer Engineering, University of Pretoria,
Pretoria, South Africa; dDepartment of Electrical Engineering, Tshwane University of Technology, Pretoria, South Africa; eCenter of Excellence
in Intelligent Engineering Systems (CEIES), King Abdulaziz University, Jeddah, Saudi Arabia

ABSTRACT ARTICLE HISTORY

Everywhere, typical network systems are challenging two opposite situations, mainly caused by Received 4 April 2021
suppliers and consumers. The provision of new power quality resources and an increase in Accepted 1 August 2021
electricity consumption are the main problems in satisfying the daily need and the rising number KEYWORDS
of critical base loads. Because of the quick and unforeseen increase in consumers and failure to Blackouts problems;
construct new micro-energy power plants in distribution systems in general, optimal control computational methods;
should be implemented to assure the distribution system’s proper operation. The application of distributed energy resources;
adequate computational load shedding techniques is necessary for this approach. This paper load shedding techniques;
reviews microgrids’ work, a considerable number of computational intelligence methods, and microgrid; optimal control;
different classes of load shedding techniques applied to date in the distribution systems. utility grid
A conclusion has been drawn given several unforeseen blackout problems occurring in major
electrical distribution companies worldwide. The maximum use must be made of renewable
energy resources, mainly when there are load shedding and blackout problems.

1. Introduction
A microgrid (MG) aims to optimally coordinate distrib­
uted energy resources (DER) and energy consumption. blackouts was an incorrect operation of protective sys­
This is seen as an interconnection of diverse power tems, especially distance relays [6]. Over the last dec­
components from the generation to the power con­ ades, several severe blackouts worldwide have caused
sumption that can operate within defined electrical con­ economic loss and had a social impact [7].
straints [1–3]. The optimal operation is guaranteed by MG’s operation is based on two principal modes,
the fashion conceptual structure of a controllable entity which are respectively grid-connected and islanded
that respects the electrical grid. An MG is also seen as mode. In grid-connected mode, the MG is intercon­
a localised grouping of power sources to supply the nected with a power system and can exchange its
loads. This electrical system of an MG usually operates, power with the primary grid. While in islanded
and it is synchronised with the conventional centralised mode, the MG operates with only its energy
power grid. However, it is essential to notice that the resources and no interconnection with the grid sys­
MG can disconnect and operate independently as phy­ tem. The study is based on a literature survey of
sical or economic conditions dictate. A small-scale elec­ various control variables that guarantee an MG’s
trical system near the end-users can be considered an effectiveness. As in most case scenarios, an MG con­
MG system [4,5]. Small-capacity hydro-units, biogas, tains diverse DERs; controlling frequency, voltage
diesel generator (DG), photovoltaic (PV), and wind and stability of the electrical system are principal
power systems constitute various power resources to parameters to deal with an MG’s operation [8].
be implemented in an MG. Each of these energy Implementing renewable energy resources (RERs)
resources can be set independently or combined to into the conventional power network is one of the
form a hybrid power supply. It is necessary to notice challenging problems to be considered. An MG sys­
that through an MG system, the network congestion, tem provides a higher level of flexibility and relia­
line losses and system operation cost due to lines are bility to the power network performance. This is
reduced, ensuring higher energy efficiency to the net­ because it offers the opportunity of operating in
work. A factor that played an important role in recent both islanded and grid-connected modes. Besides,

CONTACT R.C. Bansal [email protected] P. O. Box 27272 Sharjah, UAE, Tel: +971 65050971, - ©University of Sharjah. Sharjah, U.A.E.
© 2021 Informa UK Limited, trading as Taylor & Francis Group
654 T. MADIBA ET AL.

different MG components can physically be neared network system so that load shedding actions are
or distributed geographically to one another [9,10]. avoided to improve the system’s reliability.
Aspects affecting islanded mode are inefficient
design or lack of an under-frequency load shedding In MG power distribution, some of the significant
(UFLS) system, lack of an under-voltage load shed­ power shortage problems are based on reserve source
ding (UVLS) system and inefficient vegetation man­ insufficiency. Additionally, the non-provision of excess
agement strategies. Besides, smart MG provides power generated into the MG system creates an imbal­
intelligent technology and approaches that effectively ance in the MG distribution system in future additional
deal with various policy and regulatory barriers, loads. The effects caused by the inefficiency of the power
resulting from cost-effective, power quality, reliability, supply in the MG system under load shedding are the
and environmental challenges [5]. The intelligent power system’s unbalancing. The only way of solving
computational methods are based on the optimal con­ this kind of issue is by shedding some non-critical loads
trol approaches implemented with appropriated so that the system can be running under balancing
mathematical modelling equations and programming mode [15]. Also, shedding some of the loads is not
tools so that UFLS problems are solved with regards to always the right solution because of reducing the pro­
keeping the frequency at an acceptable value required ductivity and rentability of the significant critical con­
by the power system standard [11–13]. In assumption, sumers in the MG system.
this is still not yet the favourable solution; loads are For this reason, this research has focused on analys­
still going to be shed, which means the MG distribu­ ing different optimal solutions which permitted the
tion system’s profitability is not secured guaranteed. continuation operating mode of the MG system. It is
Thus, the need to apply RER as an additional supply essential to notice that it may be better to implement an
reserve is significant [14]. Once the additional supply optimal intelligent method to activate at a specific time
source has been injected into the MG system, excess horizon. This is based on the incursion of additional
power generated will be available to solve UFLS pro­ sources of DERs into the MG system so that the level of
blems. As an impact, no loads will be shed in case of customer need from the demand side of the MG power
such problems. system can be met. Due to these reasons, considered as
Recently, several research works have focused on lousy side effect impacts in the MG system, it is impor­
improving the MG UFLS scheme’s operating mode in tant to stabilise frequency and compensation with
which some UFLS methods were applied in the MG to excess power generated much needed for the MG dis­
solve this under frequency problem. This work assesses tribution system.
different proposed optimal control method strategies to This paper reviews the UFLS work done in MG. It is
address the UFLS scheme of MG, and this method will organised into four main sections: (2) the MG system,
be based on real-time techniques to improve the UFLS (3) frequency regulation and frequency response analy­
problem in MG. Besides, the benefit of implementing sis in MGs, (4) the influence of the contribution of RERs
this optimal approach can be summarised as follows: in load shedding, (5) load shedding intelligent compu­
tational techniques, and (6) conclusion.
● Providing a stable and reliable optimal control
system performance for MG UFLS scheme in
response to supply and load demand. 2. Microgrid system
● Make sure that the balance between the supply and
2.1 Principle of the microgrid system
demand side is accurately achieved
● Increase the rentability and profitability of the The connection of an MG to the grid must be at the
microgrid UFLS scheme same voltage level. The point of standard coupling
● Reduce the payback time and improve the lifecycle (PCC) guarantees the synchronisation and connection
cost of all significant expensive equipment involved of different MG energy resources with the primary grid.
in the design of MG systems such as RERs compo­ During diverse problems due to voltage instability and
nents, generators, power transformers and their other connection challenges, the PCC disconnects the
associated equipment MG to the grid. This disconnection can automatically or
● Control carefully all frequency fluctuations, which manually be done by a switch, leading to an islanded
is considered the state variable to be as closed as to operation model of an MG [14]. It is essential to notice
zero (Δf 0). that the primary grid connects the distributed power
● In case of power shortage, excess power generated supply with the diverse energy consumer sectors.
must be at an acceptable level to balance the However, during the critical time, which requires
 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 655

repairing some power grid components, this affects the infrastructure, changing to smoother methods to supply
power grid’s stakeholders. Therefore, an MG enters into humanitarian services to the poor [20]. An additional
the game to cover these challenges. An MG connected to power source will always help solve the problem for
the utility grid offers operating possibilities. These aim these typical MG systems to avoid load-shedding
to boost the grid capability, resilience, and reliability difficulties.
and then guarantee the distribution system operator
(DSO) capability to supply the consumer during crises
2.2 Components of a microgrid system
such as storms or power outages or other major reasons
[16]. It should also be essential to notice an MG for An MG does not only backup the power grid during
power sources DGs, including energy storage systems emergency scenarios but can also be used to minimise
and RERs (Wind and PV). The operation of an MG the energy costs or boost up the local resource in the
aims to make it function indefinitely. Therefore, the presence of demand growth. An MG provides some
system requirement needs a robust management strat­ excellent opportunities to the world’s communities,
egy that can handle the system uncertainty of diverse including being more energy-independent and environ­
components of an MG [5]. The MG systems are custo­ mentally friendly [21–23]. Researchers [24] have repre­
mised into four types, which are end-users MG, remote sented a typical MG with its components, which are
MG (rMG), utility MG (µgrids) and virtual MG (vMGs). listed as follows: the biomass generator system, hydro
End-user and utility MG are often seen as the same pump system, micro-hydropower system, standby gen­
system. This can be called customer MGs (µgrids), and erator, PV system, wind turbine generator (WTG) and
they are self-governed and usually downstream of the battery bank storage system in case of power short­
a single PCC. This type of MGs enhances the DGs’ age. Besides, the consumer side is characterised by the
benefits [12] and coordinates the power system outages load, which is not constant according to its usage.
[17]. The load-shedding strategy is considered one of Because of the variation of the load, the MG system is
the techniques to solve the power system’s unbalancing always unbalanced.
problems. The µgrids or community MGs are fashioned
into the regulation of the DSO. Therefore, MGs must
2.3 Impact of renewable energy resources in
comply with the power system codes. The difference
microgrids
between the utility grid and µgrids resides in the regu­
latory and business model perspective. The power short­ RERs contain several environmentally friendly
age is seen as the main principal issue of the µgrids. This resources independent of the sun, and these can be
is a problem that can be resolved by load-shedding classified as follows: biomass, hydropower, solar energy,
methods [18]. vMGs contain multiple sites of DER. tidal and wind energy. Due to diverse current challenges
The control coordination of vMGs operates as a single caused by the secure power supply and climate change,
entity to manage the power flow on the electrical system. several policies are being implemented to increase
This system can function under the operating mode of power from RERs. The implementation of power plants
the MG system. The load-shedding problems and solu­ based on RERs presents some environmental problems,
tions can be considered in the small-scale basic to bal­ which are essential to avoid unexpected events. An
ance vMGs. intensive assessment of various current and potential
Remote MGs are unable to function in grid- environmental issues derived from RERs’ power pro­
connected mode. In 2017, the variation in terms of duction helps make excellent policies that minimise the
their total power can be ranged from 349 MW to 1.1 overall cost of renewable energy exploitation. It should
GW, and this capacity can hold a total revenue of more be necessary to note that the RERs are climate-friendly
than 10 USD billion [19]. In rMGs, the main technical and can enhance energy security and independence.
challenges are maintaining the balance of active power However, the RERs contain some limitations relative
(frequency control) and reactive power (voltage con­ to low energy density and conversion technologies’
trol). Implementing a hybrid power system that inte­ low efficiency for their exploitations. These constraints
grates DERs, battery energy storage (BES), diesel can exclusively or even primarily render the RERs
generators, and inverters system can secure power in implantation worldwide impracticable. RERs yield
remote areas where there is no possibility of being con­ more unsatisfactory performance in some exploitations
nected to a utility grid [18]. This requires appropriate compared to conventional power plants such as fossil
support of batteries, damp loads to maintain active fuel and nuclear [14].
power and reactive power compensators. Remote MGs RERs are currently considered as efficient power
also serve as the anchors of new, appropriate scale sources to provide sustainable and clean energy.
656 T. MADIBA ET AL.

However, a large renewable power generator con­ primarily inductive transmission lines, are not
nected to the utility grid comes with diverse opera­ practically acceptable for MGs. Subsequently,
tional issues that can disturb the distribution network models ought to be changed [4,30].
operators (DNOs) [25]. The severity of these problems ● Low inertia: This is the factor to be considered
depends directly on the applied technologies of RERs, seriously, primarily when synchronous generators
geographic area of the system implementation and are used in the MG system. The MG presents a low-
penetration level. Therefore, studying the possible inertia feature distinct from bulk power systems,
impacts of large power generation from RERs on the where many synchronous generators are used to
DNO side and analysing their components can provide back the inertia issues from DER. When there is
feasible resolutions that meet all interconnection a significant component of power electronic devices,
requirements and other issues before implementation. this phenomenon is evident [31]. In the case of
The power quality problems associated with RERs are stand-alone operation, when an appropriated con­
characterised by the enriched current, safety protec­ trol strategy is not considered, the consequence will
tion, under-voltage or over-voltage, output power fluc­ be a severe fluctuation of frequency, which can
tuation, harmonics distortion and frequency cause a lot of damage to the MG system [28,32].
fluctuation [26]. This section aims to discuss possible ● Uncertainty: This is the negative side of the MG
technical problems caused by high penetration hybrid system’s lifecycle period because the running
RERs in distribution systems. period is directly proportional to its productivity
and payback time of the invested capital. The
MG system’s operation is subject to uncertainty,
2.4 Challenges of microgrids
such as its configuration’s economical and reli­
There are several operational challenges to be able operation. The load profile and weather
addressed in the design of control and protection forecast are two essential factors that make this
schemes to ensure that the present levels of reliability coordination more challenging in isolated MGs.
are maintained by MG systems combined with inte­ The critically unbalanced MG system and ran­
grated DER units, which are not pointedly affected, domly higher component failure rates require
and that the potential benefits of the DG are fully a complex challenge throughout an unlimited
harnessed [27]. The most significant challenges in time horizon. This uncertainty is higher than
MG control and protection can be listed as: those in bulk power systems due to the reduced
energy demand and positively associated varia­
● Bidirectional power flows: The implementation of tions of available power resources [33,34].
distributed energy generation (DEG) units in the
DSO levels can origin some reverse power flow The subsequent sub-section gives an overview of the
problems that may cause diverse complications control architecture’s operational method in the MG
such as fault current distribution, protection coor­ power system because of the challenges listed above.
dination issue, undesirable power flow profile and
voltage control problem [5].
2.5 Control architecture of the MG systems
● Stability issues: The MG can operate properly only
when the stability issue has been addressed. For Regarding the architecture of MG control or any control
this, the MG system application of integration of formulation, two different methods can be considered,
the control scheme of DG units may create local namely a centralised or decentralised approach. Fully
oscillations. Therefore, a ride through small- centralised control is based on a considerable amount of
disturbance and stability assessment is required. information transmitted between all units involved, and
Besides, the transfer of activities between the stand- from there, the decision is taken and made available at
alone and grid-connected modes of an MG’s only one point. This will present significant implemen­
operation may be the leading cause of transient tation problems since interconnected power systems
stability [28,29]. usually contain vast geographical regions and involve
● Modelling: This is the critical point of an MG an immense number of units. A completely centralised
system, as the operational principle is always control scheme is presently seen as an impracticable
based on and depends on its mathematical model. solution. Besides, using decentralised control, its local
For that, many characteristics in the traditional controller controls without knowledge of the outside
system, such as the constant power loads, preva­ position of others [35]. The hierarchical control block
lence of three-phase balanced conditions and diagram of MGs creates a good agreement between
 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 657

decentralised and centralised control schemes. This is represents an overview of the control architecture often
based on three control levels, which are primary, sec­ applied in MG systems connected to the primary grid.
ondary, and tertiary control schemes. The following section introduces frequency regula­
Primary control aims to design and focused on the tion and frequency response analysis in the MG power
satisfaction of conditions such as frequency and voltage system to address the challenges listed above.
stability. This structure also provides the set points for
a lower controller, namely the current and voltage con­
trol loops of DERs. These internal control schemes are 3. Frequency regulation and frequency
frequently described as the zero-level control system of response analysis in microgrids
MGs. Secondary control typically has a sampling time of
3.1 Importance of frequency regulation and
seconds to minutes, which supports the primary and
response for MG systems
secondary control scheme’s dissociated dynamics and
simplifies their specific design strategies. The setpoint of The power quality depends on diverse factors [36]. The
primary control is provided through the secondary con­ frequency oscillation is one of those factors. The fre­
trol level. This provides a centralised controller, and it quency fluctuation is often a function of an imbalance
aims to restore the MG frequency and voltage and between power production and power consumption. In
compensates for the system deviations that derive small-scale implementation of an MG with RER, the
from primary control. Secondary control aims to com­ frequency oscillation can be negligible compared with
ply with power quality requirements, such as the accep­ a large-scale system. The increase in severity of this type
table voltage level at critical buses. The tertiary control of problem can be observed when there is a substantial
scheme is the final control level. This control structure amount of RER penetration level into the system.
considers different economic aspects of the MG’s opti­ Therefore, the ride through capability of the synchro­
mal operational standpoint. The sampling time can be nous machine is an important parameter to consider.
set from minutes to hours and coordinate the power This avoids any chance of winding speed and genera­
flow between the MG and the primary grid [5]. Figure 1 tor’s damage [37,38]. The stability of operating MG

Microgrid 1 Microgrid 2 Microgrid 3 Microgrid N

Microgrid Network

Primary Control

Main Grid
Secondary Control

Tertiary Control

Figure 1. Control architecture of MG system [24].

658 T. MADIBA ET AL.

depends on its ability to autonomously handle the sys­ integration at each second. The secondary reserve oper­
tem frequency with small or sizeable intermittent RER ates in minutes’ time scale to deal with governor
integration. The smart grids technology based on response in which the primary reserve could not solve
demand response (DR) provides an opportunity to and introduce automatic generation control regulation.
deal with an MG frequency control [22] effectively. This strategy looks to guarantee the ESS SOC recovery
and dispatch some load if necessary. The tertiary reserve
scheme manages the load flowing with a dispatch strat­
3.2 Frequency regulation and response analysis
egy derived from the secondary reserve on an hourly
with application of renewable energy sources
basis. This last reserve approach will still carry-on dis­
With increasing generated power from RESs, investigat­ patching load flowing and integrate a short-term capa­
ing their impact on the electrical system and control city based on the day ahead method for a complex
schemes is of great interest. The frequency regulation system that lasts a few hours or days.
and frequency response as a significant part of the load In [47], an optimisation strategy is developed to
shedding concept and discusses existing power flow participate in inertia and an MG’s primary frequency
control methods into MGs before applying load shed­ control response. This system is based on leveraging
ding solutions can be implemented in the MGs to solve a reduced-order model. The design model contains the
UFLS problems. In [39], the methodology is presented, droop coefficients, frequency dynamic, and synthetic
stating that accidents in power systems are always inertia of DER to design the system performance index
unforeseen, and they have significant effects and disrupt to meet all time-domain optimisation requirements.
people’s lives. When the power network fails to supply The system formulation looks to deal with the frequency
adequate energy to customers, both the system’s fre­ overshoot and steady-state regulation. Thus, a classical
quency and rotating speed will drop to a lower level economic dispatch method is also implemented to guar­
and even violate the acceptable constraint, which ham­ antee the DERs can share the energy injections for
pers the system stability, particularly in an islanded inertia frequency and primary frequency response in
power network [40]. proportion to their power ratings. In [48], the primary
The integration of the RERs at low scale installed frequency control is ensured by a model predictive
power with low or medium voltage are increasing control (MPC). The system presents a strategy that can
worldwide. The MG system is an excellent platform solve diverse frequency stability problems. These are
that guarantees the system coordination for different often caused by the high-power loss of ultra-voltage
DERs’ integration while ensuring the control scheme faults and high integration levels of RERs. Therefore,
and providing the full benefits of RERs. When an MG an active frequency regulation (AFR) based on the MPC
operates at a low voltage system, it can play a controller scheme is designed to handle and coordinate different
cluster of DERs [40–45]. Due to diverse, challenging bulk power system constraints and problems. This is,
climate problems, the integration of RERs in the sector thus, seen as an AFR framework, which can combine
is seen as one solution to reduce the carbon gas emission both off-line and online applications. The off-line AFR
in the atmosphere. The generated power from DERs analyses from the main steps are cluster to form random
comes with several instability issues that negatively scenarios, grade emergency levels, formulate control
affect the power network. Therefore, the MG concept strategies and control decision tables, and determine
is introduced to tackle RERs penetration at different control parameters and system sensitivity. The online
levels [46,47]. application looks: To first monitor operation state in
Shim et al. [45] developed a harmonious integrating real-time and match typical scenario on-line; To judge
strategy to coordinate the faster-acting energy storage the match degree between current operation state and
system (ESS) and generators. The presented method typical scenario; And then, control with the updated
aims to regulate the system frequency by controlling control decision table after disturbance if the last step
its reserves in the power grid during the high penetra­ is yes, or modify control decision table of typical sce­
tion of renewable generation. Therefore, a droop control nario, get new one if the last step is no, and compute the
and state of charge (SOC) feedback framework is yes to close the online instruction.
designed to unify the system frequency with DERs. An intensive assessment of the impact and depen­
This control scheme operates at the primary and sec­ dency of DER integration is presented in [24,47]. The
ondary frequency control level. The time scales of fre­ designed model aims to cover some of the main chal­
quency control are based on different controller levels of lenges of wind power penetration into the utility grid.
the MG system. The primary reserve looks to deal with Some of those problems that this developed model looks
the inertial and governor response that supports ESS to solve are other fluctuations from wind power
 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 659

integration that affect the system variable frequency play an essential role in MGs’ operational process,
deviation and decrease frequencies stability, which which can be connected to the utility grid or operate
depends on reducing inertia and regulation capability. in a stand-alone mode where critical loads are not
This research study aims to provide a framework that supplied from other sources. For a grid-connected MG
can determine the maximum wind power penetration system, the principal network operates in active-reactive
level for a specified frequency deviation limit. Tow con­ power (PQ) mode. While for an islanded MG, the utility
trol areas interconnected by tie-line are analysed. In grid runs in voltage-frequency (V-F) mode. It is essen­
[49], an optimal control method for two interconnected tial to notice that MG’s operating mode is determined in
areas is developed. The proposed strategy of this model function of their specific information, including current,
applied a load frequency control strategy. This method voltage, power and frequency [52].
used the ESS dynamic to create an optimisation technic For unique MG, diversity control methods are
that can ensure the system frequency regulation. applied to stabilise the output signal. Generally, PQ
Consumers’ choice is the main objective of the control is used in grid-connected operating mode, and
upcoming DG in an area where a new power system droop control is incorporated with the V-F controller in
needs to be implemented [50]. A variable speed resolu­ islanded mode. Combining diverse control approaches
tion has been proposed to optimise a cluster of small is applied to guarantee the MG system’s operation sta­
water turbines incorporated into the medium voltage bility in each given mode. This method can sponta­
AC grid configuration. Applying an AC/DC-DC/AC neously create a take-off between two operating modes
boundary is required for the system to run at of MG [53]. To monitor the MG properly, it is essential
a variable speed even when connected to to apply the voltage-current (V–I) measurement in real-
a conventional 50 Hz power grid. Thus, the droop and time environments. This can communicate with all
inverse droop control presented in Figure 2 help admin­ power system components to transfer the information
ister the active and reactive power flows in both MG transfer for analysis, including examining the transient
operation mode types. DG units are usually coupled to state, operation control and evaluation of the power
a PCC system with isolating power transformers. It can quality in the electrical network [54].
observe some DC components with injected currents of During a given operation time of MG, DG units can
more than 5% compared to the rated output current be observed with a different expected output signal of
[51]. The precondition is regularly assured by adjusting active power. Some DG units can also cause this con­
the dual-loop controller’s parameters without virtual flicting event to supply reactive power while others are
impedance [52]. consumers. Decoupage of active and reactive power is
With smart grid technologies, generator controls can an excellent solution to this type of problem [55]. In
be achieved, and the MG can be connected to the utility [56], a solution is proposed to turn output impedance
grid at the substation level via a PCC. The converters inductive. It is shown that introducing virtual

Figure 2. Power frequency droop control strategy [24].

660 T. MADIBA ET AL.

impedance is the best possible option to resolve this power efficiency, and the droop control law improves
problem. An innovative droop control personifies the the power-sharing capability [62]. In the previous two
idea of equivalent control. Based on this thought, DG in decades, the conventional droop control law has been
MG is corresponding to a virtual generator with virtual revised to enhance power distribution and MGs’ solid­
impedance. The assumption and the practicability of the ity. MGs’ small-signal solidity fades at higher droop
innovative methods are developed, and the theory of the gains, even though it is immune to other factors, includ­
virtual impedance control strategy is implemented ing controller gains and tie-line impedance [63]. The
in [57]. development of the operational principle of virtual
The combination of hierarchical control shown in resistance is done in [64]. Supplementary droop and
Figure 3 contains three different modes, namely inverter adaptive feed-forward compensation can be utilised to
mode control, self-configuration control and the use of stabilise the electrical network under high droop gains.
MG central control (MGCC). An MGCC method opti­ The impact of high integration of various DG technol­
mises the operation mode in the MG. This requires ogies on the electrical system’s transient stability is
a data exchange between the microsource, load and examined in [65]. Synchronverters and synchronous
central control. The method utilises virtual synchronous virtual machines can be used as control methods to
generator arithmetic to manage the ESS, operating as adjust the frequency deviation problems in MG power
a synchronous generator to maintain the voltage and systems [66]. A constant value regulates rising inertia
frequency stability [58]. practically in the inverter control techniques scores in
In an MG system, the MGCC resides at the centre of decreasing the maximum rotor speed deviation for the
MG architecture, acting as the brain of the entire system nearby source [52] and inverter output during power
that monitors and controls the operations over the change [67].
supervisory control data acquisition network, imple­ For management on the demand side, it can be
menting information and communication technologies. assumed that two independent generating circuits sup­
MGCC facilitates all the functions of sources and loads ply the consumers. Thus, one can supply critical loads.
control to realise the balance between the load and This circuit must be disconnected in the case of fre­
generation sides [59]. Power allocation among DGs in quency dropped below a specified value (such as
an MG is made feasible by applying control techniques 42.5 Hz). The other supplies the non-critical loads.
such as a droop control system [60]. This is implemen­ The intelligent load management controller (ILCM) is
ted to control DGs’ frequency and voltage with a power used in this scenario to deliver a frequency response.
electronic component [61]. Nevertheless, the droop The ILCM looks to define the system frequency where
control methods apply to high-voltage MGs to improve the energy demand’s coordination is autonomous for

Figure 3. Inverter mode control concept [24].

 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 661

every sampling time of 15 minutes in communication In [69], the model and several data conversion stages
with the MGCC. This arrangement is shown in of a digital frequency relay are presented. The relay
Figure 4 [68]. covers two frequency conditions, which are over and
The operational principle of the ILMC is based on the under frequency. The digital relay has an advantage over
non-critical load demand Pnc and the shed loadPshed ; it the traditional relay based on the electromechanical
depends on the ON-OFF times of the under-frequency relay. The digital relay provides high accuracy and
relays. The control block system is characterised by the speed compared to the electromechanical relay. A full-
system of equations (1) described below [24]: power system blackout can be observed during the sig­
nificant frequency variations.
� Parizad et al. [70] present a smart approach to coor­
Pnc tripped time ofunder frequency relay
Pshed ¼ dinate the shedding of demand and power generation.
0 Not tripped time ofunder frequency relay
ðT The procedure is tested using an islanded network for
ΔEus ¼ Pshed dt an intelligent power management system. The devel­
0 oped model looks to stabilise the power system. It
(1) should be necessary to notice that this procedure deal
with under-voltage and frequency load shedding.
The unserved energy is evaluated at the end of a cycle of Therefore, it has been observed that the implemented
15 minutes, and it must use the previous value to update design model guarantees the normal operating condi­
the new value for a new cycle. tions that can be affected by the voltage and frequency of

Figure 4. Load control management concept [24].

662 T. MADIBA ET AL.

the electrical network. This system also maximises the control strategy combining a hybrid system’s operation
available power while ensuring the safety and reliability containing components such as a battery energy storage
of the electrical power operation. The architecture of system (BESS), diesel generator, dump load, and wind
this load shedding processes provides a calculation generator in an MG is evaluated. The control scheme is
opportunity that prioritises some loads to coordinate divided into two main tasks in the developed configuration
the load shedding system’s trigger signal. In [71], an approach model: selecting a suitable management compo­
improved strategy is presented to deal with the under- nent and offering frequency regulation [74]. A general
frequency load shedding condition. This scheme detects controller selects the component to function. Therefore,
the power deficit and selects load during the shedding the frequency regulation can be coordinated through an
process. Continuous monitoring achieves the over­ independent specific controller. The suggested control
shooting signal of the second frequency derivative of scheme is assessed on an MG power system with combined
the centre of inertia. This proposed strategy deals with kinds of power generation. The proposed control scheme
three main modules: total load shed amount estimation operates efficiently to monitor each component’s power
due to first power shortage, optimum load shedding, flow optimally and regulate the system frequency. Besides,
and new power shortage estimation. the penetration of RERs is guarantee within a stable power
Several countries have been adopted different topol­ network [75]. The coordination of stabilising mechanisms
ogies and structure in the function of system priorities without power mismatch estimation is permitted by
of the MG system. Developed countries such as Canada, deploying a global control scheme [76]. This controller
Europe, Japan, and the USA create more active research was designed with the feature that deals with the IF-
into the MG system that can provide an opportunity to THEN inference engine. The high-order mathematical
deal with load shedding. Several practical projects and models are suitable for modelling and testing this type of
laboratory capabilities have been designed, and some control scheme of MG frequency response. This proves the
research works are in progress [72]. The future smart variation of the frequency has been controlled between the
grid technologies provide a well-structured plug-and- constraint limits of Δf = ± 0.005 Hz. Figure 5 shows the
play penetration of MG. This system is connected via categorised relationship amongst the control layers.
committed highways for the exchange of command, In [43], the approach to carry out a categorised con­
data, and power. trol scheme adept at challenging many distributed
micro-sources is assessed. It has been observed that
efficient coordination of MG’s control structure offers
3.3 Particularity of frequency response in the the possibility to stability the system frequency. The
islanded MG systems setpoint adjustment instructions can send the informa­
In [73], it is studied that islanded MGs can be self-sufficient tion to diverse MG components and regulate the system
in terms of controlling the system frequency and voltage. frequency from various system disturbances, such as
This independent control scheme is caused by the islanded islanding and load changes. It has been observed that
operation mode of this type of MG. A frequency regulation the implementation of a hierarchical control scheme can

Figure 5. Global control layers relationship concept [5,24].

 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 663

easily control demand-side management. On the other Thus, for each time horizon time, Ts, the frequency
side, the central autonomous management controller error (frated – f) and the frequency error inte­
cannot directly access the system information through gralðKp þ KI 1s Þare applied to establish the additional
the data relative to micro-sources in every MG con­ power ΔP [77].
nected to the grid. Therefore, hierarchical control offers In [78], 1 MW BESS is integrated into the grid,
a flexible and cost-effective approach to efficiently man­ allowing Zurich’s power grid configuration. This devel­
age the electrical network with various MGs and high opment is based on the description in aspect and illus­
integration levels of RERs [41]. Figure 6 suggests tration of the structure’s flexibility to deliver different
a flowchart of the hierarchical control scheme. advantages to the distribution power grid. The assess­
As can be seen in the flowchart, the procedure is ment of this research demonstrates that the integration
a closed-loop control system, runs once each sampling of such a large amount of the BESS helps regulate the
time, Ts, and is characterised by equation (2) as follows: system frequency and can participate in the Swiss ancil­
�� � � lary service markets improvement. The data analysis
1 proved that the operation of this system is in islanded
ΔP ¼ Kp þ KI � ðfrated f Þ : (2)
s operation mode. This can virtually guarantee the quality
of power flow and provide a suitable frequency and
voltage regulation for a commercial MG setting. An
optimisation strategy for BESS to regulate the frequency
system using the primary control approach includes
dead-band charging [79]. The SOC of battery is used
to adjust the system during offset [80] or by changing
average tactic as developed by Borsche’s method to
calculate offset power as

1 X k � �
poff ðk þ dÞ ¼ presp ðiÞ þ ploss ðiÞ (3)
a i¼k¼a

where a is the averaging period, d is a delay, ploss is the

sum of all power losses in the battery system, and presp is
the primary frequency control power. The principal
concept is that the offset varies slowly to do not interfere
with frequency control performance while the actual
response to the grid frequency stays instantaneous [81].
In [82], a study on remote power networks due to
the absence of external support shows that low iner­
tia constants provide inadequate frequency regulation
strategies. Consequently, after an active power mis­
match into the system, frequency expeditions are
severe. Conventionally, it was found that synchro­
nous generators provided severe contribution for
a better active power and frequency control scheme.
The load shedding system is also seen as an emer­
gency solution to contribute to the system’s perfor­
mance and avoid the network’s frequency collapse.
Renewable energy provides a prospect to the power-
frequency control in some grid codes [83]. This type
of control feature is often implemented in isolated
power networks with considerable wind potential.
The wind power contribution in regulating the sys­
tem frequency has been evaluated through
a curtailment control scheme at the wind power
level [84]. The frequency regulation of the DG is
Figure 6. Flowchart of a control scheme for the electrical net­ characterised by the system of equations (4).
work with several MGs and high integration levels of DG [24].
664 T. MADIBA ET AL.

�
dΔω
¼ ΔPm2HΔPe friendly, and reliable power supply during different
dt (4)
Δf ¼ ΔP � R demand conditions [90]. The optimal design of this
energy mix is considered one of the main issues to
In steady-state, the frequency deviation is proportion­ deal with the hybrid power system’s performance
ally determined in function of the power imbalance in requirement. Several research studies have developed
the system ΔP, and the equivalent droops constant R of in this field. This concept has provided
the power system. It should be essential to note that a comprehensive assessment of such optimum design
a large steady-state frequency error can be observed approaches that can substantially contribute to large
during low regulation capability in a power system [85]. renewable energy integration by improving the systems’
economy [91].
DER offers diverse technologies in a low-scale energy
4. Influence of contribution of renewable
system. This mixes different energy resources, both
energy sources in load shedding
renewable and conventional, that can be close to the
RERs can contribute to the improvement of the electri­ end-users. The principal challenge of this type of tech­
cal network operation during load shedding problems. nologies is the coordination of several control and
This can also be observed in the operation of the MG manipulated variables of the designed system. The
system. DG deployment, either in centralisation [86] or energy mix model’s optimum energy management
decentralised energy, is generated or stored by a range of reduces both the energy cost and carbon dioxide emis­
small grid-connected DER [87]. Most of the conven­ sions significantly [92].
tional generation power components are centralised and A novel methodology to solve to resolve the fre­
are often transmitted in long distance. For better effi­ quency is presented in [93]. This strategy uses a single-
ciency of the power system, this system must be con­ phase high-frequency AC, which provides a better plat­
nected to others’ energy generation to create hybrid form to integrate DERs. The MG performance has been
power systems. detailed to hold the electrical network’s power quality
In [88], a hybrid power system that combined PV through the applied power flow analysis. It is necessary
and wind power are illustrated. This system aims to to note that the system model is developed under active
increase the power supply from the RERs. The optimal filtering. This model is also called a unified power qual­
strategy maximises wind power during the winter and ity conditioner and universal active power line condi­
solar power generation during the other summer. It was tioner. The advantage of the proposed model is its
observed that the hybrid power system often yields ability to operate under intelligent behaviour that man­
greater economic and environmental profits compared ages the energy flow on the MG system and reduces the
to single power generation operated in isolation. Ref overall operating cost.
[24]. gives a brief idea of how the control system will In [94], a computing strategy based on the block
avoid load shedding issues in power shortage. diagram of the UFLS scheme is proposed. The strategy
RESs are increasingly being considered as the main combines the scenario generation process and mixed-
choices in supply-side planning for MGs [88]. MG sys­ integer linear programming model. The generation
tem configuration concentrates on the definition, ideal model consists of dealing with the uncertainties extrac­
design, initiation, planning, operation, and sizing of tion of RERs (wind and solar generation), load fluctua­
a hybrid to minimise the operating and lifecycle cost tions and generation deficiency. Three optimal
while guaranteeing diverse environmental problems. scenarios have been developed. These are Monte Carlo
Four different study scenarios are assessed, including Simulation, roulette wheel mechanism and scenario
diesel generator only, fully renewable generation, hybrid reduction algorithm. The second based on mixed-
diesel renewable energy and MG connected to the grid. integer linear programming formulation aims to com­
The results are based on the objectives to evaluate and pute each scenario by first providing a control signal to
assess their economics, environmental and operational the relay model. The relay structure coordinates the
performance. Besides, an assessment of the systems is timing and operation logic of the control scheme
done to establish the economic aspect for the grid- under generation frequency that can hold the system
connected operation of an MG power system. constraints. The objective of this scheme is to compute
Ref [89]. explains how RERs such as hydro-based the system performance index. The system computation
energies, solar and wind-solar, are environmentally block diagram is tested with an IEEE 39-bus electrical
friendly and can be used more widely. Thus, energy network. It was shown that the proposed model deals
mix based on RERs with conventional power resources effectively with the different uncertainty of power gen­
can provide a more economical, environmentally eration and demand fluctuations.
 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 665

In [95], an adaptive load shedding model is presented Table 1. Under frequency minimum requirements [24].
to deal with the electrical network’s synthetic inertia Stage Number Under frequency limits Minimum time
response. This is because the conventional load shed­ 1 > 49:4Hz Continuous operation
2 � 49:4Hz 3 minutes
ding technic cannot work under low inertia systems. � 48:4Hz 30 seconds
The proposed model aims to control the UFLS for 3 � 47:8Hz 7.5 seconds
� 47:3Hz 45 cycles
high penetration of the RERs and ESS. This research � 47Hz Instantaneous trip
work provides the more precise power deficit computa­
tion for renewable energy system while ensuring an
adaptation of RER variation and SOC of battery during decrease power demands into the electrical network
load shedding conditions. expressly. For suitable UFLS program applications, fre­
Gu et al. [96] present an adaptive decentralised UFLS quency is a reliable indicator of generation deviancy or
model for an islanded intelligent MG. The designed overload conditions. Load shedding can be realised by
model looks to stabilise the frequency of the electrical under frequency relays. This sends an instruction signal
network. Therefore, a peer-to-peer controller scheme is to the circuit breakers when the system frequency can­
applied for DGs. The control scheme operates under the not meet the relay’s frequency setting. The tripping is
smart grid technologies using the power line commu­ optimally coordinated with several stages. This com­
nication technology. The optimal control of the pro­ prises a certain amount of power demand until the
posed adaptive decentralised UFLS is based on the system frequency will be re-established. For an electrical
multi-agent system that effectively communicates the grid of 50 Hz, most utilities opt 49.3 Hz. Thus, the first
system constraints and different neighbouring agents. target frequency stage and the last stage are located
In [97], an adaptive, UFLS-based optimal design between the range values of 48.5 ≤ frated ≤ 48.9 [99]. It
scheme is presented. The proposed model coordinates should be necessary to notice that the load shedding
the operational uncertainty by applying smart grids based on the system frequency has a disadvantage. This
technologies. The demand response (DR) approach can cause some unwanted trip to the power grid. The
through the advance metering infrastructure (AMI) is power grid’s optimal operation consists of coordinating
used to formulate the mixed-integer linear program­ the system outage that categorises the demand based on
ming model. The developed framework applied the power generation [100]. During the operational process
Internet of things technologies to create a smart grid of the unwanted trip, commonly known as over-
environment. This network manages the power flow tripping, an overshoot of frequency has occurred. The
from the generation to the consumption of the energy reason is that the power network cannot recover fast
optimally. enough between steps of tripping. Therefore, this can
lead to excessive tripping because of the lack of use of
the frequency rate [101]. Table 1 shows typical under-
5. Load shedding and computational frequency protection settings for turbine-generator units
intelligence techniques [102]. In [103], it is studied how much damage and
reduction in generation system output performance are
5.1 Requirements of UFLS system
caused by under frequency problems; under frequency
Load shedding occurs when insufficient electricity degrades all components of a power system, mainly
generation cannot stratify the power demand from generators and turbines. Thus, loss of generation may
all end-users; it could then be required to supply lead to a cascading loss of generating units due to
some areas and/or particular consumers into the excessive frequency drop. Load shedding is the last
network. It should be noted that load shedding expected protection possibility against cascading outages
differs from the power outage caused by faults or [104].
other unexpected events during the power grid’s Finally, before the load-shedding technique can be
steady operation. It is also necessary to stabilise applied, some of the organised electrical companies
possible electricity generation with its power use gas and hydropower plant options as energy
demand. Load shedding is applied to cover the resource reserves, contracted and voluntary options
power system’s performance when the power system with specific large customers to decrease their
equilibrium is required. There is no other procedure demand. If all these applied technic have been
that can deal with this optimal requirement of the exhausted and power demand is not still met, the
electric grid operation [98]. last possibility will be the load-shedding option
UFLS is seen as a possibility to manage various applied to the power system to solve the pro­
control schemes. The purpose of this structure is to blem [105].
666 T. MADIBA ET AL.

5.2 Classification and types of load shedding control variables, the LPM orders the loads in SPS in
scheme (LSS) increasing order of priority and passes this list to the
In an extensive scale power system application, the combi­ control action module (CAM). The CAM holds numerous
nation of UFLS and UVLS is well coordinated to monitor circuit breakers (CBs) and bus transfers and switches status
overload in the electrical network [106]. Figure 7 presents from the dynamic database. The load (Pshed) is considered
the block diagram for a given load shedding scheme. This the system input, which the CAM also holds. Using these
scheme presumes that some external optimal control control variables, the CAM optimally defines which loads
application will provide the amount of load to be shed. to shed and the lowest number of control actions that will
The load prioritisation module (LPM) uses the various de-energise [107,108].
systems’ electrical parameters, the critical nature of loads, The setting up of a computation structure always
and the control variable’s demand level. These features depends on the functionality of the power grid. The
come from the active database. At any provided time, an modelling of the system assists in determining critical
integrated circuit may perform one mission or elements that greatly influence system performance.
a combining mission to coordinate the LSS. Taking these A combination of the fossil-fired turbine and hydro-

Figure 7. Block diagram of load shedding scheme [24].

Figure 8. Block diagram of the speed governing system [24].

 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 667

turbine units assists in assessing the power plant model. ● The reduction of the load amount to be subtracted
This configuration is sub-divided into many respective under the load shedding period
subsystems, namely the steam turbine, primary fre­ ● The optimisation at the lower value of the overall
quency response, over-speed protection control (OPC) competitive cost of the UFLS strategy.
logic, LSS and the general structure of the system
[107,109]. Figure 8 presents the general structure of The existing methods used for developing UFLS strate­
the control system. This reflects the basis of modelling gies contain three varieties. These are the main version
the primary subsystem [110]. The system model of the of the system control: traditional, half-adaptive and self-
hydro-turbine and its governing system are more adaptive ones [114]. Besides these versions, there is
detailed in [111]. When the generator’s rotor speed another human intelligent version group called compu­
reaches about 103% of its rated value, the OPC system tational intelligence techniques. Figure 9 represents the
will automatically send out a signal to shut down the categories of load shedding techniques associated with
main stop valve and reheat the stop valve to avoid the operation principle, which includes a block of intel­
a further rise in the turbine velocity [112]. In the rotat­ ligent computational techniques [115]. The classified
ing mass and load damping module, the parameter methods are explained in the sub-sections as follows:
damping coefficient must range within the average to
keep the system’s stability [113]. 5.2.1 Traditional or conventional load shedding
In general, UFLS is regarded as the last option for schemes
solving severe frequency decline in network structures. The conventional load shedding scheme is a cheap and
Indeed, when power systems are exposed to massive straightforward method. This scheme’s advantages are
disturbances or terrible power deficiencies, the capabil­ lack of information on the magnitude of disturbances,
ity to sustain the power equilibrium and reduce the inability to estimate the power imbalance’s actual value,
fluctuation of frequency proportionally depends on the shedding the fixed percentage of loads during distur­
capability and configuration of the UFLS strategy bances, and not suitable to implement in modern and
applied. Therefore, a good UFLS model is based on the complex power network [116]. This traditional load
following requirements: shedding can be implemented using frequency relay-
based load shedding, programmable logic controller-
● The ability to restrain the frequency decline based load shedding, under-voltage relay-based load
● The optimal minimisation of the time spent to shedding, and breaker interlock load shedding [117].
recover frequency, and avoidance of frequency In [118], novel UFLS schemes that deal with diverges
overshoot or hovering challenges of conventional UFLS technic is presented.

Figure 9. Types of load shedding strategies [24].

668 T. MADIBA ET AL.

The proposed model optimally utilises the monitoring reliable and robust scheme. This scheme’s only disadvan­
directional relays, power flow through feeders and tage is it cannot perform the optimum load shedding due
RERs. The DG consider in this model is based on the to variation and frequency behaviour [112]. The adaptive
wind turbine and PV system. This scheme’s objective is scheme for load shedding can be contained the rate of
to coordinate the selection optimally of the feeder dur­ change of frequency-based Load Shedding and SCADA
ing the load shedding process. Based on the load to Based Load Shedding [123]. The adaptive approach has
supply, the DG disconnection is minimised during the been improved and considered as a yet more practical
load shedding by considering the system frequency algorithm. Current research works have focused mainly
response. on improvements and enhancements of the UFLS scheme
The traditional load shedding method does not offer without DGs. This is mainly based on estimating the load
the optimal solution and cannot efficiently cover the amount to be shed, load shedding priority and power
modern power grid problem. The load shedding stages imbalance between consumption and generation [124].
are implemented according to the first and second set­ In [115], an adaptive combinational load shedding
ting values of frequency to shed the required loads. method is presented to improve diverse challenges from
During the continuous decrease of the system fre­ the traditional technique. The designed model aims to
quency, the load shedding process can be activated to stabilise the power system during severe disturbances
support the stability of the power system until the due to load shedding. The advantage of the proposed
restoration of frequency at its expected value. The tradi­ model is its ability to coordinate both UFLS and UVLS
tional load shedding offers the possibility to determine while considering the electrical network’s frequency
the amount of load to shed. Traditional load shedding decline and voltage stability margin. The adaptive load
schemes deal with shedding a given percentage of load shedding strategy is seen as the advanced model of the
whenever the frequency falls below a frequency setting traditional load shedding scheme.
value. The frequency setting values and the amount of The application of UFLS schemes can enhance the
load to be shed for each step are determined off-line, system frequency dynamic response and improve the
based on experience and the simulations [115]. The stability of the power network. This is one of the reasons
traditional or conventional LSS systems are not particu­ see several research works have been developed in these
larly good, and accurate methods are required to solve fields. UFLS schemes can offer the power grid process
power system and blackout problems in the distribution proper guideline at each step to follow and specific load
system. to be shed [73]. This strategy creates the step forward that
the power system planning should follow to guarantee
5.2.2 Half-adaptive load shedding schemes the stability of the electrical network that can derive from
A half-adaptive load shedding scheme is designed, mod­ system disturbance, frequency drop and overload. The
elled, implemented and simulated in a power system conventional load shedding, several power systems, cover
simulator for engineers. In [119], the comparison vast regions, and between most of them, there is a weak
between the adapted UFLS and the conventional UFLS coupling of the supply and load sides [125]. The system
based on real-time synchronised frequency measure­ frequency is a continuously changing variable that
ment is studied. A 10-generator 39-bus system for the depends on the consumption and generation of the
conventional system designed for a 33% generation out­ energy flow in the power grid. The load–generation
age is tested. The application of the adapted method of imbalance should be known in advance; for this purpose,
UFLS proves that smaller amounts of load are shed the traditional UFLS scheme design is based on the
compared to the conventional system [120]. The con­ assumed contingency. The control scheme is matched
ventional UFLS system considers the overload calcula­ to a specific condition. Therefore, the UFLS scheme
tion, estimation of the number of load shedding steps, cannot match all the operating conditions. Another
calculating the amount of load to be shed and the relay kind of LLS is based on frequency declining [126].
settings factor [121]. The adaptive UFLS system con­ A method of load shedding is introduced by
siders the frequency variation at all buses and its design a nonlinear optimisation model, considering long-term
based on disturbance magnitude-location, selection of voltage stability. Even though a nonlinear optimisation
load bus, total load to be shed, and rate of change of model can resolve the problem accurately, this method
frequency (ROCOF) each bus [122]. takes a long time to yield practical results. Another
method, more suitable in practice, has been proposed
5.2.3 Adaptive load shedding schemes as a multistep linearisation technique to transform the
The adaptive load shedding technique offers the advan­ nonlinear optimisation model to an approximated line­
tages of accurate determination of the load-shedding, arised model, which can solve the load shedding
 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 669

problem under contingency conditions at a fast speed dynamic programming using the particle swarm opti­
while retaining acceptable accuracy [127]. misation method to estimate the suitable distribution of
renewable mobile power stations. This is a modified
5.2.4 Computational intelligent techniques optimal strategy that can guarantee a permanent search
The intelligent computation techniques provide the space in the presence of discrete load shedding policies.
optimum of the load shedding scheme and efficiently The methods were robust in handling technical and
applied in modern and complex power systems. The operational limits and assessing different MG scenarios,
disadvantage of this system is not guaranteed to adapt including DG capacity, integration of EVs and load
with a combination of frequency and voltage instabil­ shedding based on urgency groups.
ities in the system [116]. Computational intelligent load A simple computational algorithm is presented in
shedding can be implemented using a genetic algorithm [130] to assess the load shedding problem. The system
and neural networks for UFLS System [117]. Some of model studies the emergency scenarios during the elec­
the essential computational intelligent techniques are trical network stress of holding an optimal power flow
summarised and discussed in the following subsubsec­ solution. This intelligent algorithm combines two sub-
tion. The discussions of these techniques are based on problems based on optimal power flow (OPF). This
the mathematical modelling of the grid systems and consists of applying linear programming based OPF
algorithms. for both sub-problems. The first aims to restore the
solvability, and then the second look to improve the
5.2.4.1 Global and local computational intelligent voltage stability margin (VSM). The first sub-problem,
techniques. In [128], a comprehensive understanding which based on resorting to solvability, presents
of cyber and physical interaction is presented. The smart a design contrast. This is the fact of using an objective
grid technologies are the right framework environment function-based voltage magnitude of weak bused
to model different power system problems that may lead instead of a performance index based on restoring
to blackout and loss of revenue. When it is not well power flow solvability. The second sub-problem based
coordinated, the load shedding may lead to the electrical on VSM applies a comprehensive conventional load
network’s diverse failure events, such as fault propaga­ shedding strategy. This scheme considers both the eco­
tion, cascading failure, etc. Failure prevention due to nomic and technical impacts of load shedding while the
such events is one of the principal objectives of devel­ VSM limits are linearised to be added into the linear
oping intelligent computational techniques. The com­ programming method based on OPF [131]. The advan­
puter networks are considered a cyber-domain in the tage of this strategy is its ability to form a global com­
smart grid environment. The cyber domain is efficiently putational intelligent method.
developed upon the physical domain that contains all
physical infrastructure of the power grid. Therefore, 5.2.4.2 Sensitive computational intelligent technique.
a communication collaboration between the cyber and The fuzzy logic control (FLC) is the advantage to be
physical domain creates excellent interactions during applied in any power system network. During the load-
the load shedding process by stopping fault propagation sensitive change to the MG load, the FLC-based load
into the power grid while shedding a computed amount shedding provides an optimal MG operation section
of load. The smart grid environment’s global load shed­ [52]. It has been observed a traditional UFLS strategy
ding strategy introduces optimal coordination of shed­ works effectively in the interconnect grid, but it may not
ding actions and failure probability. The computational operate well when dealing with an islanded MG [132].
intelligence provides an opportunity for failure preven­ Therefore, a novel, FLC based UFLS scheme is devel­
tion design. oped to cover this gap. The strategy looks to coordinate
A potential solution that can help the power system the system frequency, load prioritisation and rate of
prevent the increasing number of dangerous disruptive change of frequency. This scheme designs a load shed
events and meet their novel reliability conditions is controller module combining with FLC load shedding.
presented in [129]. This solution is considered as new The FLC monitor the system and measure the amount
opportunities that facilitate the integration of renewable of load to shed while the load shed controller stabilises
mobile power stations. This structure offers the MGs the the system frequency. It also should be necessary to
possibility to effectively operate in both modes, namely notice that the FLC based load shedding aims to deter­
islanded and grid-connected mode. The computational mine the power imbalance and the load shedding con­
intelligence based on the energy management strategy is trol module sheds the required while stabilise the system
developed to create a scheduled islanded MG system frequency [133]. The suggested method is justified by
operation. The system model specifically designs employing various event-based and response-based
670 T. MADIBA ET AL.

scenarios. The system outcomes have shown that the 5.2.4.3 Islanding detection methods. MG’s successful
designed scheme successfully sheds an optimal number operation in the distribution network is one of the
of loads while stabilising the frequency. PSCAD soft­ essential features, and this depends on MG’s accuracy
ware was used to model the proposed FLC-based load and fast islanding detection of DG. In [136], several
shedding. This consists of developing a script that can techniques of islanding detection have been presented.
handle the FLC behaviour to deal with the UFLS These methods are function of communication net­
scheme. Two inputs and one output are considered as work, passive, active and hybrid strategy. It can be
the control variables and manipulated variable of the observed designing these methods in isolation presents
system design. The frequency (f) and its rate of change some shortcomings that lead to inaccuracies in the
(df/dt) are the system’s inputs. The volume of load shed islanding detection of DG. Therefore, computational
(Lshed) is the system output. FLC determines the number intelligence techniques are a robust and flexible method
of the load to be shed in the function of the input values. to deal with these types of problems. The advantage of
Figure 10 presents the FLC-based load shedding model, comprehensive computational intelligence techniques
which includes fuzzification, rule base, inference presented in this Subsection 5.2.4 describes the strate­
mechanism and defuzzification steps [134]. gies applied for islanding detection of DG. Therefore,
A Newton-based dynamic algorithm has reformulated a comparison of computational intelligence-based tech­
an optimal load shedding strategy that considers generator niques in system accuracy over current techniques offers
control impacts and frequency and voltage characteristics. some information to better design and determine the
An emotional problem is devised with power generation as excellent electrical network method [137]. The types of
a dependent variable [110]. The system frequency must be islanding detection techniques are represented in Table
maintained at its constant value under normal operating 2 and associated operating principles and characteristics
conditions. During a given disturbance into the power [138].
grid, such as a rapid rise in the demand side or an unpre­ An islanding detection approach applies the fre­
dicted outage of a generator or a transmission line, the quency rate change of reactive power (dq/dt) to identify
causing power mismatch between the demand and gen­ islanding [139]. The approach contains two stages. The
eration can cause a variation in the system frequency [135]. first stage detects islanding for large power differences

Table 2. Existing islanding detection methods and their characteristics [24,138].

Parameter Remote Passive Active Hybrid
Principle Communication Controlling changes Injecting frequency Combined
between DG-utility in the system signals/applying passive-active
parameter method
Non-detection None Large Small Very small
Detection time Short Short Long Long
Power quality No degradation No degradation High degradation Minimal degradation
Reliability High Low High High
Cost High Low Low Low

Figure 10. FLLSC block diagram [24].

 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 671

between DG and total power demand. Nevertheless, for shedding and guarantee the system frequency stability
smaller power differences, where the value of dq/dt may without any overshoot [144]. The system performance
be insufficient to discriminate between islanding and index meets its requirement to create optimal load-
other events, the second stage is performant. This is shedding behaviour of the power system under UFLS.
based on a load-connected strategy (LCS), and this Three principal modules assist in modelling this techni­
strategy helps avoid this shortcoming. Equation (5) que. These are the central inertia frequency calculator
describes the explanation of each stage as follows: module (COIFCM), load shed amount calculator mod­
ule (LSACM) and optimal load shedding module
● The first stage is characterised by equation five as (OLSM). The spinning reserve, islanded, and grid-
� � � � � connected operating mode of this combined system
dq dq1 dq2 dqn are characterised by a system of equations (8) – (10) as
Mq5 � ; Mq5 ¼ max ; ; . . . ::
dt min dt1 dt2 dtn follows [145]:
0 < t < 0:1s (5) 8
>
> fgrid ! ½LSACM�Grid connected Mode
where n, and [dq/dt]island are the sample reading >
< PN
Hi Fi
horizon and the islanding event’s threshold value. COIFM , i¼1 (8)
>
> ½ f COI ¼ PN �Islanded Mode
The value dq/dt is set to respond to a significant >
: Hi
mismatch between the power consumption and genera­ i¼1

tion. For large power mismatch, the energy demand is

8 N
detected when Mq5 < [dq/dt]island. > P PN
>
> ½ GCiðMaxÞ PGi �Spinning Reserve
>
< i¼1 i¼1
● For the second stage, the LCS is initiated for 15 cycles LSACM , PN
when the system is characterised by equation five and >
> ΔP ¼ ½2 Hfni � dfdtCOI �Power balance
>
> i¼1
:
islanding is detected when equation six is reached. ½TGL ¼ ðPDG þ Pgrid Þ PLoad �Grid connected
� � � � (9)
dq dq
< Mq5 < (6)
dt min dt island �
C ¼ 2N 1 P
� � OLSM , (10)
dq AbsðerrorÞ i ¼ ΔP Pi Combinaition
Mq15 > 0:1 < t < 0:4 s (7)
dt LCS where fn is the rated frequency, fgrid is the grid fre­
where [dq/dt]island is the minimum value for detecting quency, and fCOI is the centre of inertia frequency, ΔPis
the LCS scheme [140]. The practical feasibility is shown the power imbalance, Hi is the inertia constant of the ith
in Figure 11. generator (seconds) and GCi and PGi are the maximum
This method reduces the cost by replacing the CB generation capacity of generated power of the ith DG.
with the solenoid-operated contactor and, for the case of AbsðerrorÞ i is the error of the ith combination, C is the
P
a micro-hydropower system, controlling the frequency combination and Pi Combinaition is the sum of the
and voltage of the induction generator [141] and induc­ active power of the ith combinations [146].
tion generator controller (IGC) [142]. This method has The delay time, which is a function of the calculation,
been used worldwide, such as in Africa, Asia and communication and circuit breaker operation time, is
Europe, ranging from 1 kW to 30 kW. IGCs have been supposed to be five cycles according to the practical
realised in England, Indonesia, Nepal, the Philippines, considerations. The test system considered is part of
and Sri Lanka [143]. an existing 11 kV Malaysian distribution grid. It consists
of hybrid DG resources with three DG units, two mini-
hydro DG units and one biomass DG unit.
5.2.4.4 Combination of computational intelligent tech­
niques. A combination of random and fixed priority
loads to manage a UFLS problem has been applied in 5.2.4.5 Practical cases experimented within some of the
[24]. A non-optimum load shedding pattern has been countries in the world. A UFLS scheme used by Tenaga
observed when all distribution systems are placed with Nasional Berhad (TNB) from Malaysia was reviewed,
fixed priority. A developed model with load priority and modifications and improvements were made to
based on random and fixed priority offers the flexibility reflect the system’s current changes, making the scheme
opportunity that achieves optimal load shedding. This more up to date. Moreover, a power system’s safe opera­
methodology’s various assessment shows that the pro­ tion requires that system frequency be kept within
posed technique can optimally manage the load a specified range, reviewed in [147]. The increasing
672 T. MADIBA ET AL.

inverter are also reviewed [149]. Several combina­

tion methodologies have been proposed. These
methods serve to design MGs’ controller at all oper­
ating levels, including primary, secondary, global
and emergency controls. For better monitoring of
an MG operation, the control algorithms must be
robust, smart, and optimal to manage the system’s
power flow [150]. The load shedding scheme is
sustained by primary and secondary levels of
power generation control. The primary control
adjusts the manipulated variable of the DG units,
which depend on their rating power. The secondary
control loos to deal with system frequency restora­
tion after any change in the demand side. Particle
swarm technique-based optimisation is developed
and presented to define BESS’s optimal size with
an LSS [151]. A novel load shedding approach that
considers the power stability with the economy in
load scaling for a distribution structure with DG
such as CHP, small wind and PV is presented in
[152]. This scheme classifies the power demand into
two groups (heavy and light groups). The technical
model uses a ranking that is based on willingness to
pay, frequency threshold and ROCOF.
The operating conditions of the MG are dynamically
altered by both the load and the intermittent nature of
the RER. This study presents distributed cooperative
control using a multi-agent system to uphold such
MGs’ voltage and frequency stability. Electric gird is
often susceptible to diverse damages. This vulnerability
can be seen during some disasters, and the downstream
MG may be disconnected from the power grid for a long
time. A methodology that combines two-stage load
shedding and wind power to minimise the islanding
duration time while maintaining the system frequency-
stable is developed [153]. UFLS maintains the power
system in a stable condition when a severe disturbance
occurs. The method is demonstrated using an MG
testbed with the gas-turbine, wind-turbine generator,
and solar PV power [154]. Different approaches are
evaluated for improving the system frequency response,
Figure 11. Flow chart of the islanding detection method [141].
including applying UFLS and increasing generator iner­
tia [155]. To restore the stable operation of the MG after
the transient fault disturbance, an adaptive LSS is pro­
posed to disconnect the proper amount of non-essential
integration of renewable energies in the power system, load according to the feeder loading and wind speed to
stressing the need for a thorough analysis of defence prevent the tripping of the wind generator by under-
plans’ adequacy, reviewed in [148]. frequency constraint. A proposition of an MG emer­
MG development concerns focus on technology gency control strategy has been described in [156]. The
generation, MG architecture, power electronics, method aims to stratify the power demand behaviour
control systems and protection systems. Various based on their importance and frequency regulation.
technological developments related to the MG sys­ Therefore, this scheme optimises the MG system to
tem and its case study development using a grid-tie guarantee system performance.
 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 673

5.2.5 Contribution to the existing intelligent In the power system reality, it is suggested that research
techniques on the design, application and use of computational
An appropriate control scheme of the hybrid systems techniques and islanding detection strategies must be
should be designed to keep the MG system operating carefully considered at the highest level of thinking so
continuously. The strategy must operate effectively with that the suggested optimal control actions in MG sys­
the specific type of energy resources from diesel genera­ tems will produce excellent results in case of load shed­
tion and RERs. It has been observed that the fluctuation ding problems. It is concluded that a suitable method of
due to this type of power resources can affect the pro­ solving blackout problems, in general, will be to hold
ductivity of the power system due to the frequency and better energy resources in reserve so that when there is
voltage stability that they may cause. During the distur­ an urgent need for energy generated, the power will still
bance of the power system caused by these phenomena respond to customers’ satisfaction. As authors, after this
(under-voltage and under-frequency of an MG), the review process regarding the near future, we will advise
system required an additional power source to augment the engineering world not to shed normal or sensitive
the electrical network’s operating flexibility and reliabil­ loads when the MG is in a UFLS scheme because this
ity. In no additional power source, load shedding is will cause a reduction in the productivity of significant
considered one of the best options to guarantee the companies. As a solution, we propose controlling the
MG system’s balancing aspect. Therefore, monitoring MG UFLS scheme to use energy resources reserves to
and controlling the MG frequency and voltage stability avoid interruption of rentability, especially for the
can avoid unstable situations during a disturbance on industrial sector. When loads are not shaded, this
the power generation. RER effectively contributes to impacts savings of power and money, which are the
guarantee this type of problem. As the RERs often main aspects of the economy. To achieve successful
offer their generation peak during this day, it is results, it is necessary to do proper mathematical mod­
requested to have them connected with ESS so that the elling of MG UFLS schemes by taking into considera­
stored energy can boost MG’s optimal operation. tion the power as a control variable and the frequency as
The future recommendations of the MG’s effective a state variable, ensuring that the additional power from
operation, both off-grid and grid-connected mode, will the new energy resources will not disturb the existing
be desirable to offer a great reserve of RERs. It should be MG system in terms of variation of the network para­
better to provide a cost-effective backup. PV and micro- meters. Besides, the use of renewable energy must be
hydropower systems are suitable to cover these roles in maximised so that the main supply will not imply cost
some scenarios where the needs of supplementing the for the existing power system. In this case, the main
existing power supplies and minimising diesel genera­ problem will be controlling and adjusting the variation
tion’s operating cost are required. A diesel generator can between the frequency from the existing power system
also solve the UFLS problem. This is an emergency and that from the additional MG system intelligently.
solution that is not often optimal. It increases the Even networks with DGs will fail because of rapid fre­
MG’s operating cost, and it has a negative impact on quency oscillations resulting from a mismatch between
the environment due to carbon emission. Therefore, consumption and generation. We strongly believe that it
RERs are environmentally friendly solutions that satisfy is necessary to implement an intelligent method to solve
different stakeholders. load shedding problems, which remains key to solving
problems in the power system that faces blackouts and
power shortages.
6. Conclusion
This paper reviews the UFLS methods applied in MGs,
which are still facing challenges of power shortages. The Disclosure statement
absolved blackouts in the power sector have recently
No potential conflict of interest was reported by the author(s).
been questioned around the world. The question was
based on the reliability of conventional and adaptive
load shedding schemes that can prevent unexpected Notes on contributors
power outages. To tackle this problem, reliable methods
are essential to offer an accurate and fast load shedding Dr T. Madiba has more than 20 years’ experience in the
Electro-Mechanical Engineering field, lecturing at the
scheme to avoid breakdown into the power network.
Universities, Validating and integrating the life cycle
This article discusses four main groups of under- plans for the plant, Rotating Machines, Piping Systems,
frequency systems: the conventional, semi-adaptive, Pumps, Power transformers, Generators, Power System,
adaptive, and computational methods applied in MGs. commercial and industrial project and Research and
674 T. MADIBA ET AL.

Development in Energy and Environment industry (heavy Governance Committee of University of Sharjah. He has
current). On an academic level, he has got a PhD Electrical published over 350 journal and conference papers in the
Engineering from University of Pretoria, after receiving a fields of control, optimization, energy and signal proces­
double degree which contains an international program of sing. He has also supervised over 50 M. Sc. and Ph. D.
Master of Sciences and Master of Technology Cum laude students. He contributed to the design and development of
both in Power Electrical Engineering from F’SATI-ESIEE several curricula for Universities in the Gulf countries,
and the Tshwane University of Technology. He is Asia and Africa. His recent research interest is in process
appointed an Associate Professor at Electromechanical control, fractional dynamics and control, soft computing,
Department of the University of Lubumbashi and he is renewable energies and engineering education.
working as Managing Consultant with South African and
Dr Raj M. Naidoo has more than 20 years of teaching,
DRC companies. He is registered as a Professional
research, and industrial experience. He received his Ph.D. in
Engineer with the Engineering Council of South Africa.
electrical engineering from the University of Cape Town,
Dr Professor Ramesh C. Bansal has more than 25 years of South Africa, in 2008. He is currently an associate Professor
teaching, research, academic leadership, and industrial experi­ in the Department of Electrical, Electronic, and Computer
ence. Currently, he is a Professor in EE Department, Engineering at the University of Pretoria, South Africa. He
University of Sharjah (UAE) and extraordinary professor at was previously the CEO and founder of Enermatics Energy
University of Pretoria. In previous postings, he was Professor and a board member at Stellenbosch Wind Energy
and Group head (Power) in EEC Department, University of Technologies. He has worked and consulted to several blue
Pretoria (South Africa) and worked with the University of chip companies. He is a director at EWYZE. He has published
Queensland (Australia); University of the South Pacific (Fiji); many research articles in various journals and conferences
BITS Pilani (India); and Civil Construction Wing, All India and has diversified research interests in the areas of the
Radio. Prof. Bansal has published over 325 journal articles, smart grid, renewable energy, and power systems.
conference papers, books, books chapters. He has Google
Dr Mukwanga W. Siti received the B.Sc. and M.Sc. degrees in
citations of 12,000 and h-index 50. He has supervised 25 Ph.
electrical engineering from the University of Pretoria,
D. and 4 Post Docs. He is an Editor/AE of IET-RPG, Tech.
Pretoria, South Africa, in 2000 and 2002, respectively. and
Eco. Smart Grids, and Sust. Energy, and Electric Power Comp.
the Ph.D. degree from University of Kwazulu Natal, Durban,
and Systems. He is a Fellow, and CP Engg IET-UK, and Fellow
South Africa, in 2015. After joining in 2004, he is currently a
Institution of Engineers (India). His research interests include
Senior Lecturer with the Electrical Engineering Department,
renewable energy, power systems, and smart grid.
Tshwane University of Technology, Pretoria North, and South
Dr Nsilulu T. Mbungu is an engineer and researcher with Africa. He is inventor of 2 patents, and has authored more
more than 10 years’ working experience which includes teach­ than 50 refereed papers, a monograph, and one chapter. His
ing and research, consulting services, maintenance and engi­ research interests include power system control, smart grid
neering project management. He holds a Bachelor of Civil and renewable energy.
Engineer in Electromechanical Engineering from the
University of Lubumbashi, the Democratic Republic of
Congo (DRC). He also has a BEng Honors, Master and PhD References
of Engineering in Electrical Engineering from the University
of Pretoria (UP), South Africa. His vision is based on cutting [1] Jirdehi MA, Tabar VS, Ghassemzadeh S, et al. Different
edge, multi and transdisciplinary research for sustainable aspects of microgrid management: a comprehensive
development goals that support future focus. This consists of review. J Energy Storage. 2020;30:101457.
diversified and interdisciplinary research interests in Artificial [2] Bakar NNA, Hassan MY, Sulaima MF, et al. Microgrid
Intelligence and Machine learning in the energy system, and load shedding scheme during islanded mode: a
Control System, Electrical Machines, Energy and Renewable review. Renew Sustain Energy Rev. 2017;71:161–169.
Energy, Electromechanics, Power Systems, Power Electronics, [3] Mehta S, Basak P. A comprehensive review on control
Optimization System, Internet of things technologies and techniques for stability improvement in microgrids.
Smart Grids Technologies. Int Trans Electr Energy Syst. 2021;31(4):1–28.
Dr Professor Maamar Bettayeb received the B.S., M.S., and [4] Lone SA, Mufti MD. Modelling and simulation of a
Ph.D. degrees in Electrical Engineering from University of stand-alone hybrid power generation system incorpor­
Southern California, Los Angeles, in 1976, 1978 and 1981, ating redox flow battery storage system. Int J Model
respectively. He worked as a Research Scientist at the Simul. 2008;28(3):337–346.
Bellaire Research Center at Shell Oil Development [5] Mbungu NT, Naidoo RM, Bansal RC, et al. Overview of
Company, Houston, Texas, USA. From 1982 to 1988, He the optimal smart energy coordination for microgrid
directed the Instrumentation and Control Laboratory of applications. IEEE Access. 2019;7:163063–163084.
High Commission for Research in Algeria. In 1988, He [6] Andersson G, Donalek P, Farmer R, et al. Causes of the
joined the Electrical Engineering Department at King Fahd 2003 major grid blackouts in North America Europe,
University of Petroleum and Minerals, Dhahran, Saudi and recommended means to improve system dynamic
Arabia. He has been Professor at University of Sharjah performance. IEEE Trans Power Syst. 2005 Nov;20
UAE since August 2000. He is the Vice Chancellor for (4):1922–1928. .
Research and Graduate Studies at University of Sharjah, [7] Qiang L, Lin Z, Ke G. Review on the dynamic charac­
starting September 2014. He is also the Chair of the IT teristics of micro-grid system. In: Proceedings of the
 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 675

7th IEEE Conference on Industrial Electronics and system using model predictive control design. IET
Applications, Singapore; 2012. p. 2069–2074. Renew Power Gener. 2017 Dec;11(14):1760–1768.
[8] Madiba T, Bansal RC, Justo JJ, et al. Optimal control [22] Tungadio DH, Bansal RC, Siti MW, et al. Predictive
system of under frequency load shedding in microgrid active power control of two interconnected microgrids.
system with renewable energy resources. In: FM Rabiul Technol Econ Smart Grids Sustain Energy. 2018 Dec;3
Islam, Kabir Al Mamun, Maung Than Oo Amanullah, (1):1–15.
editors. Smart energy grid design for island countries. [23] Kabir MN, Mishra Y, Bansal RC. Probabilistic load flow
Cham, Switzerland: Springer; 2017. p. 71–96. for distribution systems with uncertain PV generation.
[9] Hung DQ, Mithulananthan N, Bansal RC. Integration Appl Energy. 2016 Feb;163:343–351.
of PV and BES units in commercial distribution sys­ [24] Madiba T. Optimal control of microgrid under fre­
tems considering energy loss and voltage stability. Appl quency load [PhD Thesis]. University of Pretoria; 2018.
Energy. 2014 Jan;113:1162–1170. [25] Mohseni M, Islam SM. Review of international grid
[10] Mbungu NT, Bansal RC, Naidoo RM. Smart energy codes for wind power integration: diversity, technology
coordination of autonomous residential home. IET and a case for global standard. Renew Sust Energ Rev.
Smart Grid. 2019 Sep;2(3):336–346. 2012;16(6):3876–3890.
[11] Acosta MN, Andrade M, Vazquez E, et al., [26] Adnan M, Tariq M, Zhou Z, et al. Load flow balancing
Improvement of the frequency response indicators by and transient stability analysis in renewable integrated
optimal UFLS scheme settings. In: IEEE International power grids. Int J Electr Power Energy Syst. 2019
Symposium on Industrial Electronics. Delft, Jan;104:744–771.
Netherlands: IEEE 29th International Symposium on [27] George A, Boxiong S, Arowo M, et al. Review of solar
Industrial Electronics (ISIE); June 17-19; 2020. p. energy development in Kenya: opportunities and
1250–1255. challenges. Renewable Energy Focus. 2019;29:123–140.
[12] Alhelou H, Hamedani-Golshan M, Njenda T, et al. [28] Bevrani H, Golpîra H, Messina AR, et al. Power system
Wide-area measurement system-based optimal frequency control: an updated review of current solu­
multi-stage under-frequency load-shedding in inter­ tions and new challenges. Electr Power Syst Res.
connected smart power systems using evolutionary 2021;194(February):107114.
computing techniques. Appl Sci. 2019 Feb;9(3):508. [29] Barker PP, De Mello RW. Determining the impact of
[13] Sigrist L, Rouco L, Echavarren FM. A review of the state distributed generation on power systems: part 1 - Radial
of the art of UFLS schemes for isolated power systems. distribution systems. Proc IEEE Power Eng Soc Transm
Int J Electr Power Energy Syst. 2018;99:525–539. Distrib Conf. 2000;3:1645–1656.
[14] Mbungu NT, Naidoo RM, Bansal RC, et al. An overview [30] Naidoo R, Pillay P, Visser J, et al. An adaptive method
of renewable energy resources and grid integration for of symmetrical component estimation. Electr Power
commercial building applications. J Energy Storage. Syst Res. 2018 May;158:45–55.
2020;29:101385. [31] Bansal RC, Bhatti TS, Kothari DP. A novel mathema­
[15] Madiba T, Bansal RC. Optimal load-shedding control tical modelling of induction generator for reactive
of a microgrid power system. Electr Power Compon power control of isolated hybrid power systems.
Syst. 2018 Apr;46(7):768–787. Int J Model Simul. 2004;24(1):1–7.
[16] Sechilariu M, Wang B, Locment F. Building-integrated [32] Strachan N, Farrell A. Emissions from distributed vs.
microgrid: advanced local energy management for centralized generation: the importance of system
forthcoming smart power grid communication. performance. Energy Policy. 2006 Nov;34(17):2677–2689.
Energy Build. 2013 Apr;59:236–243. [33] Kumar D, Mathur HD, Bhanot S, et al. Forecasting of
[17] Kruger GR, Naidoo RM, Bansal RC, et al., “Distribution solar and wind power using LSTM RNN for load fre­
grid model of high-density residential housing clusters quency control in isolated microgrid. Int J Model
incorporating embedded generation. In: IEEE Simul. 2021;41(4):1–13.
Industrial Electronics Conference,Singapore; Oct 18-21; [34] Shea JJ. Distributed generation-the power paradigm for
2020, pp. 2969–2974. the new millennium [Book Review]. IEEE Electr Insul
[18] Abdelaziz MMA, Farag HE. An enhanced supervisory Mag. 2003;19(3):44.
control for islanded microgrid systems. IEEE Trans [35] Narbel PA, Hansen JP, Lien JR. Energy technologies
Smart Grid. 2016 Jul;7(4):1941–1943. and economics. Cham, Switzerland: Springer
[19] Chen G, Li Z, Liu Z. A distributed solution of economic International Publishing; 2014.
dispatch problem in islanded microgrid systems. In: [36] Mbungu NT, Bansal RC, Naidoo RM, et al.
Proceedings of the 28th Chinese Control and Decision Discriminatory protection analysis of three-phase asyn­
Conference, Yinchuan, China: CCDC; 2016. p. chronous motors during power disturbances. Electr
6804–6809. Power Compon Syst. 2019 Mar;47(4–5):431–443.
[20] Azizi SM, Khajehoddin SA. Robust load frequency con­ [37] Musasa K, Nwulu NI, Gitau MN, et al. Review on DC
trol in islanded microgrid systems using μ-synthesis collection grids for offshore wind farms with
and D-K iteration. In: 10th Annual International high-voltage DC transmission system. IET Power
Systems Conference, SysCon – Proceedings. Orlando, Electron. 2017 Dec;10(15):2104–2115.
FL, USA: Annual IEEE Systems Conference (SysCon); [38] Bansal R. Handbook of distributed generation: electric
2016. power technologies, economics and environmental
[21] Mbungu NT, Naidoo R, Bansal RC, et al. Optimisation impacts. Cham, Switzerland: Springer International
of grid connected hybrid photovoltaic-wind-battery Publishing; 2017.
676 T. MADIBA ET AL.

[39] Wan Y, Ela E, Orwig K. Development of an equivalent stability based load shedding scheme for self-healing
wind plant power-curve. Golden, CO (United States): of power system. IEEE Trans Smart Grid. 2011 Jun;2
National Renewable Energy Lab. (NREL); 2010. (2):221–230.
[40] El Chaar L, Lamont LA, Elzein N. Wind energy tech­ [55] Kumar V, Joshi RR, Bansal RC. Experimental evalua­
nology - Industrial update. Detroit, MI, USA: IEEE tion of matrix converter for wind energy conversion
Power and Energy Society General Meeting; 2011. system under various abnormal conditions. Int J Renew
[41] Mbungu NT, Bansal RC, Naidoo RM, et al. A dynamic Energy Res. 2014;4(1):15-22.
energy management system using smart metering. Appl [56] Rudez U, Mihalic R. Analysis of underfrequency load
Energy. 2020 Dec;280:115990. shedding using a frequency gradient. IEEE Trans Power
[42] Mbungu NT, Bansal RC, Naidoo RM, et al. Deliv. 2011 Apr;26(2):565–575.
Modelling of a smart hybrid water-power system [57] Ning X, Jiang H, Zhu X. An improved under fre­
for freshwater and energy supply network. In: 1st quency load shedding scheme in multi-machine
International Conference on Power, Control and power system based on the rate of change of
Computing Technologies, ICPC2T, Raipur, India; frequency. APAP - Proc Int Conf Adv Power Syst
2020. p. 6–9. Autom Prot. 2011;1:775–780.
[43] Lasseter RH. MicroGrids. In: IEEE Power Engineering [58] Chistyakov Y, Kholodova E, Netreba K, et al. Combined
Society Winter Meeting. Conference Proceedings,New central and local control of reactive power in electrical
York, NY, USA, vol. 1, pp. 305–308. grids with distributed generation. In: EEE International
[44] Thomas R, Van Zyl S, Naidoo RM, et al. Recloser based Energy Conference and Exhibition, ENERGYCON,
energy exposure assessment of a distribution network. Florence, Italy; 2012. p. 325–330.
J Energy South Africa. 2019;30(4):41–50. [59] Cheng J, Li S, Wu Z, et al. Analysis of power decoupling
[45] Shim JW, Verbic G, Zhang N, et al. Harmonious inte­ mechanism for droop control with virtual inductance
gration of faster-acting energy storage systems into in a microgrid. Autom Electr Power Syst. 2012 Mar;36
frequency control reserves in power grid with high (7):27–32.
renewable generation. IEEE Trans Power Syst. 2018 [60] Xiao H, Luo A, Shuai Z, et al. An improved control
Nov;33(6):6193–6205. method for multiple bidirectional power converters in
[46] Mbungu NT, Bansal RC, Naidoo R, et al. An optimal hybrid AC/DC microgrid. IEEE Trans Smart Grid. 2016
energy management system for a commercial building Jan;7(1):340–347.
with renewable energy generation under real-time elec­ [61] Sigrist L, Egido I, Rouco L. Principles of a centralized
tricity prices. Sustain Cities Soc. 2018 Aug;41:392–404. UFLS scheme for small isolated power systems. IEEE
[47] Guggilam SS, Zhao C, Dall Anese E, et al. Optimizing Trans Power Syst. 2013;28(2):1779–1786.
der participation in inertial and primary-frequency [62] Song Z, Liu J, Liu Y, et al. A method for the design of
response. IEEE Trans Power Syst. 2018 Sep;33 UFLS scheme with dynamic correction. Energy Power
(5):5194–5205. Eng. 2013;5(04):442–447.
[48] Jin C, Li W, Shen J, et al. Active frequency response [63] Taylor CW, Cresap RL, Nassief FR. Northwest power
based on model predictive control for bulk power pool transient stability and load shedding controls for
system. IEEE Trans Power Syst. 2019 Jul;34 generation-load imbalances. IEEE Trans Power Appar
(4):3002–3013. Syst. 1981;PAS-100(7):3486–3495.
[49] Siti MW, Tungadio DH, Sun Y, et al. Optimal frequency [64] Doolla S, Bhatti TS, Bansal RC. Load frequency control
deviations control in microgrid interconnected of an isolated small hydro power plant using multi-pipe
systems. IET Renew Power Gener. 2019 Oct;13 scheme. Electr Power Compon Syst. 2011 Jan;39
(13):2376–2382. (1):46–63.
[50] Mashhour E, Moghaddas-Tafreshi SM. A review on [65] Hippert HS, Pedreira CE, Souza RC. Neural networks
operation of micro grids and virtual power plants in for short-term load forecasting: a review and
the power markets. In: ICAST - 2nd International evaluation. IEEE Trans Power Syst. 2001 Feb;16
Conference on Adaptive Science and Technology, (1):44–55.
Accra, Ghana; 2009. p. 273–277. [66] Nogales FJ, Contreras J, Conejo AJ, et al. Forecasting
[51] Wang J, Wang Z, Xu L, et al. A summary of applications next-day electricity prices by time series models. IEEE
of D-FACTS on microgrid. In: Asia-Pacific Power and Trans Power Syst. 2002 May;17(2):342–348.
Energy Engineering Conference, APPEEC, Shanghai, [67] Sonkar SP, Lal VN, Singh RK. Three-phase
China; 2012. quasi-Z source inverters with regulated multiple AC
[52] Laghari JA, Mokhlis H, Bakar AHA, et al. Application outputs for microgrid applications and three-phase
of computational intelligence techniques for load shed­ residential load. IET Power Electron. 2020 Aug;13
ding in power systems: a review. Energy Convers (11):2222–2235.
Manag. 2013 Nov;75:130–140. [68] Rao YS, Laxmi AJ, Kazeminehad M. Modeling and
[53] Kunwar A, Bansal RC. Stability analysis of power sys­ control of hybrid photovoltaic-wind energy conver­
tem with doubly fed induction generator-based wind sion system. Int J Adv Eng Technol. 2012;3
farm connected through series flexible AC transmission (2):192–201.
SYSTEM device compensated line. Int J Ambient [69] Balasubramanian G, Singaravelu S. Fuzzy logic based
Energy. 2014 Jul;35(3):118–131. controller for a stand-alone hybrid generation system
[54] Seethalekshmi K, Singh SN, Srivastava SC. using wind and photovoltaic energy. Int J Adv Eng
A synchrophasor assisted frequency and voltage Technol. 2012;668(2):668–679.
 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 677

[70] Parizad A, Khoshkhoo H, Dehghan S, et al. An intelli­ [85] Singh D, Misra RK, Singh D. Effect of load models in
gent load and generation shedding procedure in an distributed generation planning. IEEE Trans Power
islanded network using a smart power management Syst. 2007 Nov;22(4):2204–2212.
system. IEEE Proc Smart Grid Conf SGC. [86] Ghosh A, Joshi A. A new approach to load balancing
2018;2018:1–12. and power factor correction in power distribution
[71] Jallad J, Mekhilef S, Mokhlis H, et al. Improved UFLS system. IEEE Trans Power Deliv. 2000;15(1):417–422.
with consideration of power deficit during shedding [87] Zhong QC, Weiss G. Synchronverters: inverters that
process and flexible load selection. IET Renewable mimic synchronous generators. IEEE Trans Ind
Power Gener. 2018;12(5):565–575. Electron. 2011 Apr;58(4):1259–1267.
[72] Lu X, Guerrero JM, Sun K, et al. An improved droop [88] Visscher K, De Haan SWH. Virtual synchronous
control method for dc microgrids based on low band­ machines (VSG’S) for frequency stabilisation in future
width communication with dc bus voltage restoration grids with a significant share of decentralized genera­
and enhanced current sharing accuracy. IEEE Trans tion. CIRED Semin: SmartGrids Distrib. 2008;2008
Power Electron. 2014;29(4):1800–1812. (12380):82.
[73] Govindaraj T, Hemalatha D. Dynamic reactive power [89] Zhong QC, Weiss G. Static synchronous generators for
control of islanded microgrid using IPFC. Int J Innov distributed generation and renewable energy. In: IEEE/
Res Electr Electron Instrum Control Eng. 2014;2 PES Power Systems Conference and Exposition, PSCE,
(1):2321–5526. Seattle, WA, USA; 2009.
[74] Wang C, Nehrir MH. Analytical approaches for optimal [90] Hu J, Zhu J, Dorrell DG, et al. Virtual flux droop
placement of distributed generation sources in power method - A new control strategy of inverters in
systems. IEEE Trans Power Syst. 2004 Nov;19 microgrids. IEEE Trans Power Electron. 2014;29
(4):2068–2076. (9):4704–4711.
[75] Hung DQ, Mithulananthan N, Bansal RC. Analytical [91] Viawan FA, Karlsson D, Sannino A, et al. Protection
expressions for DG allocation in primary distribution scheme for meshed distribution systems with high
networks. IEEE Trans Energy Convers. 2010 Sep;25 penetration of distributed generation. In: Power
(3):814–820. Systems Conference: Advanced Metering, Protection,
[76] Hien NC, Mithulananthan N, Bansal RC. Location and Control, Communication and Distributed Resources,
sizing of distributed generation units for loadabilty PSC,Clemson, SC, USA; 2006. p. 99–104.
enhancement in primary feeder. IEEE Syst J. 2013;7 [92] Di Somma M, Graditi G, Heydarian-Forushani E, et al.
(4):797–806. Stochastic optimal scheduling of distributed energy
[77] Lasseter R, Akhil A, Marnay C, et al. The CERTS resources with renewables considering economic and
MicroGrid concept. White Pap Transm Reliab environmental aspects. Renewable Energy. 2018
Program, Off Power Technol US Dep Energy. 2002;2 Feb;116:272–287.
(3):30. [93] Battistelli C, Baringo L, Conejo AJ. Optimal energy
[78] Lopes JAP, Moreira CL, Madureira AG. Defining con­ management of small electric energy systems including
trol strategies for microgrids islanded operation. IEEE V2G facilities and renewable energy sources. Electr
Trans Power Syst. 2006 May;21(2):916–924. Power Syst Res. 2012 Nov;92:50–59.
[79] Zeng Z, Yang H, Zhao R. Study on small signal stability [94] Rafinia A, Moshtagh J, Rezaei N. Stochastic optimal
of microgrids: a review and a new approach. Renew robust design of a new multi-stage under-frequency
Sust Energ Rev. 2011;15(9):4818–4828. load shedding system considering renewable energy
[80] Pogaku N, Prodanović M, Green TC. Modeling, analy­ sources. Int J Electr Power Energy Syst. 2020
sis and testing of autonomous operation of an Jun;118:105735.
inverter-based microgrid. IEEE Trans Power Electron. [95] Silva SS, Assis TML. Adaptive underfrequency load
2007 Mar;22(2):613–625. shedding in systems with renewable energy sources
[81] Majumder R, Ghosh A, Ledwich G, et al. Load sharing and storage capability. Electr Power Syst Res. 2020
and power quality enhanced operation of a distributed Dec;189:106747.
microgrid. IET Renew Power Gener. 2009;3(2):109. [96] Gu W, Liu W, Zhu J, et al. Adaptive decentralized
[82] Kroutikova N, Hernandez-Aramburo CA, Green TC. under-frequency load shedding for Islanded smart dis­
State-space model of grid-connected inverters under tribution networks. IEEE Trans Sustainable Energy.
current control mode. IET Electr Power Appl. 2007;1 2014;5(3):886–895. .
(3):329. [97] Rafinia A, Rezaei N, Moshtagh J. Optimal design of an
[83] Yao W, Chen M, Matas J, et al. Design and analysis of adaptive under-frequency load shedding scheme in
the droop control method for parallel inverters consid­ smart grids considering operational uncertainties.
ering the impact of the complex impedance on the Int J Electr Power Energy Syst. 2020 Oct;121:106137.
power sharing. IEEE Trans Ind Electron. 2011;58 [98] Shun L, Qingfen L, Jiali W. Dynamic optimization of
(2):576–588. adaptive under-frequency load shedding based on
[84] Vandoorn TL, De Kooning JDM, Meersman B, et al. WAMS. In: Proceedings of IEEE Information
Automatic power-sharing modification of P/V droop Technology, Networking, Electronic and Automation
controllers in low-voltage resistive microgrids. IEEE Control Conference, ITNEC,Chongqing, China; 2016.
Trans Power Deliv. 2012;27(4):2318–2325. p. 920–926.
678 T. MADIBA ET AL.

[99] Maouedj R, Mammeri A, Draou MD, et al. dispatching to benefit system operations. IEEE Trans
Performance evaluation of hybrid photovoltaic-wind Power Syst. 2012 May;27(2):840–848.
power systems. Energy Procedia. 2014;50:797–807. [114] Terzija VV, Koglin H-J. Adaptive underfrequency load
[100] Atwa YM, El-Saadany EF, Salama MMA, et al. Optimal shedding integrated with a frequency estimation
renewable resources mix for distribution system energy numerical algorithm. IEE Proc - Gener Transm
loss minimization. IEEE Trans Power Syst. 2010 Feb;25 Distrib. 2002 Nov;149(6):713.
(1):360–370. [115] Girgis AA, Mathure S. Application of active power
[101] Joorabian M, Ajodani M, Baghdadi M. A method for sensitivity to frequency and voltage variations on load
voltage regulation in distribution network equipped shedding. Electr Power Syst Res. 2010 Mar;80
with OLTC transformers and DG units. In: Asia- (3):306–310.
Pacific Power and Energy Engineering Conference, [116] Sapari NM, Mokhlis H, Laghari JA, et al. Application of
APPEEC,Chengdu, China; 2010. load shedding schemes for distribution network con­
[102] Soroudi A, Ehsan M, Zareipour H. A practical nected with distributed generation: a review. Renew
eco-environmental distribution network planning Sust Energ Rev. 2018;82:858–867.
model including fuel cells and non-renewable distrib­ [117] Lakra P, Kirar M. Load shedding techniques for system
uted energy resources. Renewable Energy. 2011 Jan;36 with cogeneration: a review. An Int J. 2015;4(3):83-96.
(1):179–188. [118] Das K, Nitsas A, Altin M, et al. Improved load-shedding
[103] Niknam T. A new approach based on ant colony opti­ scheme considering distributed generation. IEEE Trans
mization for daily Volt/Var control in distribution net­ Power Deliv. 2017 Feb;32(1):515–524.
works considering distributed generators. Energy [119] Hirodontis S, Li H, Crossley PA. Load shedding in
Convers Manag. 2008 Dec;49(12):3417–3424. a distribution network. In: 1st International
[104] Jabr RA, Pal BC. Ordinal optimisation approach for Conference on Sustainable Power Generation and
locating and sizing of distributed generation. IET Supply, SUPERGEN,Nanjing, China; 2009.
Gener Transm Distrib. 2009 Aug;3(8):713–723. [120] Nechadi E, Harmas MN, Hamzaoui A, et al. A new robust
[105] Soroudi A, Ehsan M. A distribution network expansion adaptive fuzzy sliding mode power system stabilizer.
planning model considering distributed generation Int J Electr Power Energy Syst. 2012 Nov;42(1):1–7.
options and techo-economical issues. Energy. 2010 [121] Robitaille M, Agbossou K, Doumbia ML. Modeling of
Aug;35(8):3364–3374. an islanding protection method for a hybrid renewable
[106] Ochoa LF, Dent CJ, Harrison GP. Distribution network distributed generator. Can Conf Electr Comput Eng.
capacity assessment: variable DG and active networks. 2005;2005:1477–1481.
IEEE Trans Power Syst. 2010 Feb;25(1):87–95. [122] Fathi M, Bevrani H. Statistical cooperative power dis­
[107] Yang J, Fletcher JE, O’Reilly J. Permanent magnet syn­ patching in interconnected microgrids. IEEE Trans
chronous generator based wind power generation sys­ Sustainable Energy. 2013;4(3):586–593.
tem fault protection schemes. In: IET Conference on [123] Srinu P, Venkata NR, Harshitha TD. Islanding scheme
Renewable Power Generation (RPG), Edinburgh; 2011. and auto load shedding to protect power system.
p. 14. Int J Comput Electr Eng. 2013;1(4):508–512.
[108] El-Metwally MM, El-Sobki MS, Attia HA, et al. Priority [124] Falahi M, Butler-Purry K, Ehsani M. Dynamic reactive
ranking of industrial loads and application of Demand power control of islanded microgrids. IEEE Trans
Side Management technique. In: Eleventh International Power Syst. 2013;28(4):3649–3657.
Middle East Power Systems Conference,El-Minia, Egypt; [125] Dai X, Zhang K, Zhang T, et al. ANN generalised
2006. inversion control of turbo-generator governor. IEE
[109] Fang H, Chen L, Dlakavu N, et al. Basic modeling and Proc - Gener Transm Distrib. 2004 May;151(3):327.
simulation tool for analysis of hydraulic transients in [126] Ajjarapu V, Christy C. The continuation power flow:
hydroelectric power plants. IEEE Trans Energy a tool for steady state voltage stability analysis. IEEE
Convers. 2008;23(3):834–841. Trans Power Syst. 1992;7(1):416–423.
[110] Sigrist L, Egido I, Rouco L. A method for the design of [127] Terzija VV. Adaptive underfrequency load shedding
UFLS schemes of small isolated power systems. IEEE based on the magnitude of the disturbance estimation.
Trans Power Syst. 2012 May;27(2):951–958. IEEE Trans Power Syst. 2006 Aug;21(3):1260–1266.
[111] Yu HY, Bansal RC, Dong ZY. Fast computation of the [128] Lu Z, Wei M, Lu X. How they interact? Understanding
maximum wind penetration based on frequency cyber and physical interactions against fault propaga­
response in small isolated power systems. Appl tion in smart grid. Proc - IEEE INFOCOM. Atlanta,
Energy. 2014 Jan;113:648–659. GA, USA: IEEE INFOCOM 2017 - IEEE Conference on
[112] Du Z, Wei J, Hu H, et al. An optimization model of Computer Communications; 2017.
under frequency load shedding in interconnected AC/ [129] Monteiro JR, Rodrigues YR, Monteiro MR, et al.
DC power system considering emergency generation Intelligent RMPS allocation for microgrids support
control. In: Int. Conf. on Power System Technology: during scheduled islanded operation,”. IEEE Access.
towards Green, Efficient and Smart Power System. 2020;8:117946–117960.
Chengdu, China: IEEE International Conference on [130] Fernandes TSP, Lenzi JR, Mikilita MA. Load shedding
Power System Technology; 2014. p. 626–632. strategies using optimal load flow with relaxation of
[113] Argiento R, Faranda R, Pievatolo A, et al. Distributed restrictions. IEEE Trans Power Syst. 2008 May;23
interruptible load shedding and micro-generator (2):712–718.
 INTERNATIONAL JOURNAL OF MODELLING AND SIMULATION 679

[131] Nakawiro W, Erlich I. Voltage security assessment and Available from: https://ieeexplore-ieee-org.uplib.idm.
control system using a hybrid intelligent method. In: oclc.org/document/4299516 .
EEE Bucharest PowerTech: Innovative Ideas Toward [144] Zong Y, Kullmann D, Thavlov A, et al. Application of
the Electrical Grid of the Future, Bucharest, Romania; model predictive control for active load management in
2009. a distributed power system with high wind penetration.
[132] Seidi Khorramabadi S, Bakhshai A. Critic-based IEEE Trans Smart Grid. 2012;3(2):1055–1062.
self-tuning pi structure for active and reactive power [145] Mohd Zin AA, Mohd Hafiz H, Aziz MS. A review of
control of VSCs in microgrid systems. IEEE Trans under-frequency load shedding scheme on TNB
Smart Grid. 2015 Jan;6(1):92–103. system. In: National Power and Energy Conference,
[133] El Moursi M, Joos G, Abbey C. A secondary voltage PECon – Proceedings,Kuala Lumpur, Malaysia; 2004.
control strategy for transmission level interconnection p. 170–174.
of wind generation. IEEE Trans Power Electron. 2008 [146] Mohd Zin AA, Mohd Hafiz H, Wong WK. Static and
May;23(3):1178–1190. dynamic under-frequency load shedding: a
[134] Laghari JA, Mokhlis H, Halim Abu Bakar AB, et al. An comparison. Int Conf Power Syst Technol
intelligent under frequency load shedding scheme for POWERCON. 2004;1:941–945.
islanded distribution network. In: EEE International [147] De Boeck S, Van Hertem D. Under frequency load
Power Engineering and Optimization Conference, shedding schemes in systems with high PV penetration:
PEOCO - Conference Proceedings,Melaka, Malaysia; impact and improvements. Eindhoven, Netherlands:
2012. p. 40–45. IEEE Eindhoven PowerTech PowerTech. 2015.
[135] Sanaye-Pasand M, Seyedi H. New centralised adaptive [148] Raghami A, Ameli MT, Hamzeh M. Primary and sec­
load-shedding algorithms to mitigate power system ondary frequency control in an autonomous microgrid
blackouts. IET Gener Transm Distrib. 2009 Jan;3 supported by a load-shedding strategy. In: 4th Annual
(1):99–114. International Power Electronics, Drive Systems and
[136] Wu M, Lu DDC. An active damping method for stabi­ Technologies Conference,Tehran, Iran; 2013. p.
lization of cascaded connected two stage converter sys­ 282–287.
tems with constant power loads in DC microgrids. In: [149] Kerdphol T, Qudaih Y, Mitani Y. Battery energy storage
Proceedings - IEEE International Symposium on Circuits system size optimization in microgrid using particle
and Systems,Melbourne, VIC, Australia; 2014. p. swarm optimization. In: IEEE PES Innovative Smart
2664–2667. Grid Technologies Conference Europe,Istanbul, Turkey;
[137] Padron S, Hernandez M, Falcon A. Reducing 2015. p. 1–6.
under-frequency load shedding in isolated power sys­ [150] Zhang H, Lai CS, Lai LL. A novel load shedding strategy
tems using neural networks. Gran Canaria: a case study. for distribution systems with distributed generations.
IEEE Trans Power Syst. 2016 Jan;31(1):63–71. In: IEEE PES Innovative Smart Grid Technologies
[138] Mahat P, Chen Z, Bak-Jensen B. Underfrequency load Conference Europe,Istanbul, Turkey; 2015. p. 1–6.
shedding for an islanded distribution system with dis­ [151] Xu Y, Li H, Tolbert LM. Inverter-based microgrid control
tributed generators. IEEE Trans Power Deliv. 2010 and stable islanding transition. In: IEEE Energy Conversion
Apr;25(2):911–918. Congress and Exposition, ECCE,Raleigh, NC, USA; 2012. p.
[139] Laghari JA, Mokhlis H, Karimi M, et al. A new 2374–2380.
under-frequency load shedding technique based on [152] Leng D, Polmai S. Experiment on distributed coopera­
combination of fixed and random priority of loads for tive control with multi-agent system for a single-phase
smart grid applications. IEEE Trans Power Syst. 2015 microgrid. In: 12th International Conference on
Sep;30(5):2507–2515. Electrical Engineering/ Electronics,Computer,
[140] I. 1547. Standard for interconnecting distributed Telecommunications and Information Technology,Hua
resources with electric power systems - IEEE standard; Hin, Thailand; 2015.
2003. [153] Chen CS, Hsu CT, Lin CH, et al. Design of tie line
[141] Teoh CW, Tan WY. An overview of islanding detection tripping and load shedding scheme for distribution
methods in photovoltaic systems. World Acad Sci Eng microgrid systems with wind power generation. In:
Technol Eng Int J Electr Comput. 2011;5(1):1341–1349. Industrial and Commercial Power Systems Technical
[142] Warren C New; General Electric Company. Load shed­ Conference,Newport Beach, CA, USA Newport Beach,
ding, load restoration and generator protection . . . - CA, USA; 2011.
Google scholar. Philadelphia, Pa. : General Electric [154] Xing S. Microgrid emergency control based on the
Company, Power Systems Management; 1974. [cited stratified controllable load shedding optimization. IET
2021 Feb 21]. Available from: https://scholar.google. Conf Publ. 2012;2012(611 CP):59.
com/scholar?hl=fr≈sdt=0%2C5&q=Load+shedding% [155] Jamuna K, Swarup KS. Multi-objective biogeogra­
2C+load+restoration+and+generator+protection phy based optimization for optimal PMU
+using+solid+state+and+electromechanical+under placement. Appl Soft Comput J. 2012 May;12
+frequency+relays&btnG= . (5):1503–1510.
[143] IEEE guide for the application of protective relays used [156] Sivasubramani S, Swarup KS. Multiagent based differ­
for abnormal frequency load shedding and restoration - ential evolution approach to optimal power flow. Appl
IEEE standard. C37.117-; 2007 [cited 2021 Feb 21]. Soft Comput J. 2012 Feb;12(2):735–740.