Electric Power Systems Research 226 (2024) 109945

Contents lists available at ScienceDirect

Electric Power Systems Research

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

A state of art review on the opportunities in automatic generation control of

hybrid power system
Ritu Verma a, *, Suresh K. Gawre a, N.P. Patidar a, Sonali Nandanwar b, c
a
Department of Electrical Engineering, Maulana Azad National Institute of Technology Bhopal, India
b
Department of Electrical and Electronics Engineering, IES College of Technology, Bhopal, India
c
Department of Electrical and Electronics Engineering, IES University, Bhopal, India

A R T I C L E I N F O A B S T R A C T

Keywords: Load and generation must be balanced for the system to be stable, which implies frequency deviation must be
Genetic Algorithm (GA) minimal. Whenever load perturbation takes place in the network, frequency diverges from its nominal value and
Hybrid Power System (HPS) unplanned exchange of power leads to unfortunate consequences. This will cause the mal-operation of electrical
Electric Vehicle (EV)
equipment such as change in speed, low efficiency, vibrations, harmonics, inaccuracy etc. Automatic Generation
Photo-Voltaic (PV)
Teaching-learning based optimization (TLBO)
Control (AGC) provides a signal to alter the actual output power of multiple generators within a certain range
based on changes in system frequency. For this purpose, several control techniques were employed in the system.
This paper focuses on various approaches in the domain of automatic generation control for microgrid based
hybrid power system. AI based techniques shows progressive results as compared to other classical and cascaded
approaches. Machine learning and deep learning based controlling schemes has also been studied and study
shows that these techniques provides loftier performance in comparison of other AI methods due to the avail­
ability of large number of trained data sets of controlled variable. Further massive data of controlled variables
can be used to restore the system after cyber-attack. It is also addressed how to integrate electric vehicles with
the power grid for Load Frequency Control. At the end of this article, some research gap such as role of demand
response, protection and control issues, resilient controllers for cyber threats etc. are also addressed which will be
useful for researcher working in this area.

1. Introduction implementation of solar photovoltaic (PV) and Wind system is to ensure

power in the islanded area along with grid integration. Since the
With the rapid increase in the load demand of consumers, it has been installation of solar panels are more flexible and can be installed on
required that power generation should also be increased. Conventional domestic regions. Moreover, PV array and wind power plants are mainly
energy sources always be the major generation entity in the power deployed in microgrids as well as in Distributed Generations (DG) [4].
sector. Since these sources are exhaustible in nature and their require­ Since the power generation from these sources have stochastic behavior
ment for power generation increases day by day, availability of different and can be utilized to meet the local demand of consumer as well as
sources should be explored [1]. Renewable energy sources are power injection into the grid [5]. Due to the intermittent nature, these
non-exhaustible, free of cost and provides safe environment with less sources faces several technical and economical hurdles. Microgrid is a
pollution, also becomes very important to meet the demands of elec­ local group of electricity users which utilize the existing infrastructure
trical loads. Now-a-days, the foremost clean energy sources are solar, and defined as a system which is able to manage its operation whether to
wind and biogas. As the availability of these sources depend on nature, operate autonomously or grid connected and have seamless transition i.
power generation varies accordingly [2,3]. Moreover, contribution of e. connect and decouple from the utility for power trading and supply of
power from solar and wind has increased in the energy sector. The basic ancillary services [6]. These are installed with several distributed

Abbreviations: AGC, automatic generation control; LFC, load frequency control; MG, microgrid; DER, distributed energy resources; AI, artificial intelligence; PID,
proportional derivative controller; DEG, diesel engine generator; PSO, particle swarm optimization; HPS, hybrid power system; GWO, Grey Wolf optimization; DFIG,
doubly fed induction generator; ML, machine learning; DL, deep learning; AVR, automatic voltage regulation.
* Corresponding author.
E-mail address: ritumits3@gmail.com (R. Verma).

https://doi.org/10.1016/j.epsr.2023.109945
Received 20 October 2022; Received in revised form 7 October 2023; Accepted 11 October 2023
Available online 20 October 2023
0378-7796/© 2023 Elsevier B.V. All rights reserved.
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

generations focused mainly on renewable sources and energy storage idea about the need of constant frequency. Implementation of AGC
system along with a synchronous diesel generator [7]. Due to the schemes in microgrid connected Hybrid Power System is described in
perforation of renewable power in to the grid, system instability section V. Section VI shows the study of various control techniques in the
occurred. Instability in the system is an important aspect of controlling field of AGC. Final conclusions are given in section VIII along with a
microgrid. With the increasing development in the power network, discussion in Section VII.
various controlling techniques has been developed to face the technical
issues occurred in the system. These technical issues emerges in the 2. Overview of microgrid
system led to instability such as rotor angle, voltage and frequency
instability presented in Fig. 1. According to the description given by IEEE in [13], Microgrid is
When a physical disruption occurs in the system, the system loses its described as a lot of integrated loads and DERs that acts as a sole unit in
synchronism, which causes rotor angle stability, which can be reduced reference to power grid under specified electrical limits and it is shown
by using power system stabilizers, thyristor excitors, etc. [8]. The in Fig. 2. Microgrid allows seamless changeover among off grid mode
disruption at the network interface also caused voltage instability, and and on-grid mode and the point at which the transition takes place is
the system failed to keep the voltage at a constant level. Maintaining the called Point of Interconnection (POI). Microgrid is also classified of
load-generation equilibrium is necessary for the system’s frequency having a control system which is able to manage automatically, by
regulation. The stability of the system is badly impacted by any integrating generation, energy storage and controllable loads. Microgrid
load-generation imbalance or by irregular loading. When a sudden load control system(MCS) controls the microgrid connection and disconnec­
change occurs due to an increase or decrease in the amount of power tion from the utility grid for energy exchange and provision of auxiliary
needed, power generation needs to be adjusted appropriately. However, services without any voltage and frequency fluctuation. It also in­
because the load generation balance is not maintained in a timely corporates the function of microgrid energy management system.
manner, frequency fluctuation occurs, which makes the system unstable A Microgrid includes following components [14]:
[9]. To resolve this problem in a PV system, the Maximum Power Point
Tracking (MPPT) method is used to make sure that PV modules produce (i) An equipment which connects the microgrid with the main grid at
the most electricity possible. The MPPT method is very effective and POI.
desirable because it guarantees maximum power output even when (ii) A system which includes microgrid energy management system.
there are system disturbances. Additionally, the system’s rotational (iii) Local distribution systems like transformer, switchgear etc.
inertia has decreased as DG penetration in the utility connected via (iv) Distributed energy resources and loads within the microgrid
power convertors has grown. As a result, Synthetic Inertia Control (SIC), boundaries.
which injects an additional power component whenever there is a fre­ (v) Control elements like microgrid controller, sensors communica­
quency disturbance event, must be activated for frequency regulation tion network etc.
[10]. Some energy storage systems and a hybridization of systems-based
control approach have also been deployed alongside these strategies. MG serves several objectives subjected to the prime purpose of their
Matching generation and demand is accomplished through Automatic demand. Distributed energy resources (DERs) are commonly perceived
Generation Control (AGC), which allows the system to operate effort­ as a mechanism for effectively managing the reliability of energy supply
lessly and continuously [11]. When load changes suddenly, generation within a grid during contingencies. Additionally, they serve to enable
adjusted automatically while maintaining frequency and voltage sta­ localized optimization of energy supply by exerting control over the
bility of the system is Known as AGC [12]. distribution of energy resources. Microgrids are also a way to get elec­
The prime benefaction of this literature is delineated below: tricity to remote areas, to use green energy as a systematic way to get
electricity to rural areas, and to make the supply more reliable and
1 Deals with the various optimization techniques employed for mini­ secure for end users [14]. A microgrid or distribution system of a grid
mizing frequency deviation and active power regulation. may encounter a number of difficulties as a result of the deployment of
2 Various issues such as sensitivity analysis, load perturbation, optimal DER, including problems with fault prevention and dynamic control.
power flow etc. in renewable based hybrid microgrid systems are However, there are several problems that are more frequent in small
addressed. grids when it comes to protection and management. These problems
3 AI based control schemes are focused throughout the literature. include bidirectional power flow that causes voltage excursions outside
4 Paper also provides some emerging research gap in the field of of acceptable bounds, fault current supplied from multiple sources, loss
electric vehicle integration to the grid. of synchronism between multiple sources when a fault occurs, a po­
5 Role of Demand side management in AGC is also addressed. tential limit on the magnitude of the fault current, a lower inertia or
lower primary time constant, frequent changes in operational configu­
The prime purpose of AGC is to conserve the frequency in a specified ration brought on by economic optimization, and the intermittent na­
minimal range along with acquiring tie-line power in multi area system ture of distributed renewable generators. Also, according to IEEE
at some specified values. This paper is categorize in following manner, standards, an electrical system is consider to be a microgrid , if it possess
section II microgrid systems, its mode of operation, section III shows the three main features (i) specified electrical limits (ii)microgrid control
concept of AGC in conventional AGC schemes. Section IV gives us brief system and (iii) critical load which is less than the installed capacity else
microgrid disconnected from main grid and act as standalone system
and serve local loads [15]. Microgrid is classified on account of their
connection and named as grid connected (on-grid) systems shown in
Fig. 3 and islanded mode (off-grid) shown in Fig. 4.
Whenever microgrid is associated to the utility, local operator are
bounded in such a way that they can not regulate either voltage or
frequency due the grid standards, such systems are called grid connected
system or on-grid system. Energy storage needed for on-grid system is
less and they introduce voltage, frequency and harmonic disturbance in
the system. In this mode, the microgrid feeds to both local loads and the
grid.. This mode is also called non autonomous mode [15]. On the other
Fig. 1. Types of instability. hand ,when microgrid operates in islanded mode i.e. they are not

2
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

Fig. 2. Microgrid structure.

Fig. 3. On grid PV system.

Fig. 4. Off grid (OR) stand alone PV system.

connected with the grid either due to planned transition, unplanned follows strict guidelines drew on voltage and frequency of MG. Islanded
transition or due to black out, they are called stand alone system or off microgrids are now studied commonly due to increased load demand
grid systems. During islanded mode, operators are free to regulate from industries and urban areas [17]. Non conventional sources with
voltage and frequency with in specified electrical boundaries. Off- grid large transients are more often used power source for islanded micro­
systems is in need of adequate storage proficiency in regards to control grids and the fluctuations caused due to renewable will leads to system
power alteration owing to irregular power generation and serves local black out. An islanded microgrid basically consist of a source, load, and
loads. This mode is also called autonomous mode [16]. When the a controllable energy storage. Storage system is preferred as per the
microgrid runs in island mode and has numerous converter-based gen­ necessity of microgrid. As power production is nil during night time,
erators (CBGs), certain problems like low inertia, low fault current etc. energy system should be assembled to serve the local loads. And these
get exacerbated. Because of these problems conventional protection and storage should counter quick , so as to abolish the power fluctuations of
control methods might not be adequate for microgrids. It might be PV system and can withstand the grid during night. Energy storage
necessary to use protection systems that are different from the usual system are categorized as (i) Short term energy storage mainly deals
ones. Protection mechanisms may occasionally need to be dynamically with voltage fluctuation (ii) long term deals with frequency deviation
changed dependent on the microgrid’s operational state. and energy management during transients [18]. DG’s of microgrid are
Microgrid acquire power from DER’s like solar, wind systems or DEG generally PV, wind which possess intermittent behavior and suffer se­
and possess an inverter which specifies voltage and frequency i.e. they vere fluctuations due to climate effects. Power ramp rates needs to be

3
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

controlled within a specified limit, while reducing the chance of The main purpose of AGC should be known by reviewing an isolated
destruction when suddenly clouds clear out over a solar power station. system where load is fulfilled by a single generating unit. Frequency
Variation in the flow of power at certain rate is termed as ramp or slew varies with the load deviation and can be easily seen in the character­
rate. Energy storage is in need of ramp rate to stabilize solar based MG istics of governor droop magnitude and load frequency. When power
and this rate maintains the microgrid voltage within a specified range. demand of the system increases, governor tries to increase its speed,
Its concept is simple and assist in minimizing the size of storage required hence the generation increases (act as a primary control) and meets the
for the PV network. Ramp rate control for a solar panel is given as [19] required demand. As a result the change in frequency will also settle
down to the zero. But after this phenomenon, a steady state frequency
dPPV ΔP
R(t) = = (1) error still persists, which is further diminished by secondary control
dt ΔT′
action where a controller is employed for the governor which retains the
Where, system frequency in its specified range of operation. This phenomenon
adjusts the reference point of speed and regulates the frequency by
R(t) = instant ramp rate making zero error. While imposing AGC mechanism in multi area sys­
PPV = instant power generated by PV panel tem, power exchange takes place through two ways i.e. flat tie line and
flat frequency control. Frequency should be kept constant regardless of
When ramp rate reached beyond a specified limit, power output tie line power in any case.
needs to be regulated. Under such conditions energy storage systems Real power transfer under normal operating condition for a two area
such as battery, super capacitors etc. or generating units such as fuel system is [1]:
cells are integrated to overcome these fluctuations and maintain smooth
generation. P12 =
|E1 ||E2 |
sinδ12 (2)
X12
3. Automatic generation control Where, E1 &E2 are sending end and receiving end excitation voltages.

In the present competitive market, diminishing frequency deviation X12 = X1 + Xtie + X2 (3)
in an interconnected system is a difficult task. Modification in stability
δ12 = δ1 − δ2 (4)
services mainly focuses on reliability, better control and efficiency.
These modern controls fulfill the utility and user’s requirement of sys­ Small change in tie-line flow,
tem operation and control. Installation of large number of generators
dP12
increases in proportion of power demand. When the systems are inter­ ΔP12 = .Δδ12 (5)
dδ12
connected, area having maximum power generation will share the load
demand of other area. These areas are called regulated areas and linked = PS Δδ12 (6)
with the tie lines [20]. For system stability, total power generation in a
system should be a resultant sum of demand and losses occurred in the Where,
system during operation. Violation of this case leads to the frequency dP12
deviation as well as irregularity in tie line power exchange. In order to PS = = PS (Δδ1 − Δδ2 ) (7)
dδ12
maintain this condition, a control scheme is used known as Automatic
Generation Control and Fig. 5 shows a basic AGC scheme in the system. 4. Requirement of constant frequency
For economical operation, AGC is designed to prolong frequency of the
grid at or near the minimal value i.e. 50 Hz by diminishing the deviation Some authors studied different controlling schemes of microgrid
occurred due to load perturbation and maintaining the supply between such as generation excitation control, load tripping, reactive power
control areas. The prime purpose of AGC schemes is given below [21]; compensation, fault clearance technique etc. Some of these schemes are
always active at the point of generation and act as a real time power
(i) Maintaining power balance in the existing system. balancing. When there is a disturbance occurred in the load generation
(ii) Keeping operating limits in the range of generators and tie line balance, system stability degrades effectively. Kundur [3] addresses real
limits. and phantom power control in view of the network as an entity and also
(iii) Frequency deviation should be zero. described the modeling and characteristics of the equipment used for the
control purpose. Author depicts that frequency deviation is an index in

Fig. 5. Basic scheme of automatic generation control [1].

4
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

load generation balance in a defined area. System operator maintains changer setting. At control centers, optimum operating point is
the energy equilibrium by conserving the system frequency in the determined by the economic load dispatch method. It is a slowest
minimal range of operation. control action in generator.
Main function of AGC are listed below:
Tertiary control maintains generator power output due to economic
(i) When any large power imbalance occurred in the system, it ap­ reason. In steady state, generation change due to load change may not be
plies drastic emergency control i.e. primary or secondary control economical. Since different generators have different cost of power
while maintaining load generation balance. generation, a system operator desired to regulate the power generated in
(ii) Synchronize and strongly coupled synchronous generators should several units for maximum economic profit. This generation adjustment
be employed in control areas. is done manually and often known as tertiary control.
(iii) Frequency deviation and power deviation caused in control area AGC operation is mainly carried out by two system inputs,
due to power imbalance should be zero.
(i) System frequency
A major goal in the development of LFC is regulating the generation (ii) Net in or out tie line power
and crossline power between multiarea systems. In order to maintain
scheduled power exchanges, each region should posses controlled gen­ A complete schematic diagram of all control loops incorporated in
eration [22]. Below are the main control loops of AGC. the network is shown in Fig. 7. In two area system, power exchange
between two control areas occurred. Let these two areas be A1 andA2 .
Change in system frequency is,
(1) Primary AGC loop: Primary loop act at an initial step. Whenever
Δω = ω − ω0 (10)
the demand of user varies, i.e. more electrical energy is required
at the output of generator, as a result the turbine governor set Where,
tries to adjust its speed while sustaining the power between user
and mains. ω = system frequency
(2) Secondary AGC loop: A steady state error is found in the system ω0 = nominal frequency
frequency following the execution of primary control loop. In
secondary AGC control, controller is employed (which is integral Change in net tie line power flow is,
controller) in the system which adjusts speed governor setting by
ΔPnetint = Pnetint − Pnetsched (11)
doing fine frequency adjustment. As this loop has large response
time, this loop acts slowly in comparison to primary loop and a Where,
schematic of these are shown in Fig. 6. Controller action is
∫ Pnetint is the difference between power leaving in area 1 and 2
ΔPe = − KI Δf . dt (8) Pnetsched is power scheduled contractually.

Due to load generation mismatch between two areas, various sce­

narios occurred in the system. Scenario of power imbalance is given in
Where, Table 1 below [23–25]:
From the above scenario it has been observed that governor sets the
KI is integral controller gain required control action. Area control error is evaluated by
Pre specified tie line power and specified frequency,
After taking Laplace, ACE = − ΔPnetint − Δω.B (12)

ΔPe (s) =
− KI
Δf (s) (9) Where,
S
B= frequency bias factor in MW/Hz

(1) Tertiary AGC loop : After the execution of both the above loops ,
a third loop comes into action which make sure that load is
economically shared among all the generators in the system. For
economic load dispatch, tertiary loop adjusts the generator speed

Fig. 6. Control loops of AGC [1]. Fig. 7. Frequency control loops of AGC [5].

5
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

Table 1 function is used. Author considered a solar and wind based microgrid
Different scenarios. incorporated HPS shown in Fig. 8. SMES system is also used as an energy
Frequency in Tie line power in Cause Action required storage system as it provides fast response, higher efficiency etc. [25].
area 1 area 2 Influence of battery storage on isolated HPS is addressed in [27]. A
Δf < 0 ΔPnetint < 0 Increase in load Area 1 generation to simulation has been done in order to analyze both unvarying and dy­
of area 1 be increased namic system stability. Author chooses GA to tune the controller
ΔPnetint > 0 Increase in load Area 2 generation to parameter. Load frequency analysis of microgrid incorporated HPS
of area 2 be increased using fuel cell was done by Singh et al. [28] and simulation results shows
Δf > 0 ΔPnetint < 0 Decrease in load Area 2 generation to
of area 2 be decreased
that Diesel Engine Generator(DEG) improves frequency regulation.
ΔPnetint > 0 Decrease in load Area 1 generation to Author [29] developed a Fuzzy Logic Controller (FLC) designed for
of area 1 be decreased AGC in an islanded wind diesel network whose parameter are basically
optimized using Bee Colony Optimization(BCO) and it has been
observed that FLC is too sturdy as compared to other. Several optimi­
5. AGC schemes in hybrid power system
zation algorithm such as PSO, GWO, BCO has been utilized for opti­
mizing different controller paraments. An optimization problem is
A HPS is explicated as the amalgamation of more than one renewable
formulated for PID controller and solved using Krill heard (KH) algo­
generation sources for producing power and it is seen that wind and
rithm [30]. Krill heard algorithm is also a nature inclined technique
solar are more prominent source of power due to their reliability and low
which utilizes the krill herding motions. PID controller is optimized by
cost operation. Automatic generation control allows signal to all the
employing KH technique for minimizing frequency deviation in micro­
generators of a specified area for regulating real power output which
grid system. PID is a conventional controller also called three term
changes due to system frequency variation. Several LFC control methods
controller shown in Fig. 9 contains control action of proportional(P),
and control strategies were given by researchers over a long time for the
integral (I) and derivative (D) controllers. As P controller is utilized for
regulation of frequency as it is a foremost factor for power system
latest inaccuracies, I controller for former faults where as D controller is
operation and control. Several literatures were proposed to understand
utilized in determining the error variation rate.
various control schemes, as these are developed to improve LFC
These controllers are mostly used in industrial control system and
mechanism.
can be used as P, I, PI, PD, PID for dissimilar control applications. In this
paper PID controller reduces steady state error while improving dy­
namic response. Simulation results obtained shows undershoot, oscil­
5.1. Advanced optimization method in HPS
lation amplitude etc. Exploration of PID controller is given in Table 2.
Apart from these conventional controller, there are some non con­
Singh et al. [23] shows the affect of Superconducting Magnetic En­
ventional controllers such as fuzzy controller, novel fuzzy controller is
ergy Storage System (SMES) on microgrid integrated HPS and this
also exists. The foremost dissimilarity between conventional and non
microgrid is mainly based on PV and wind power. Due to sudden load
conventional controller is that mathematical model of the process is
change and variable nature of generation , frequency deviation occurs
required in designing a conventional controller. A novel fuzzy logic
which has been mitigated by the combined effect of SMES and Grey Wolf
controller was proposed in [31] which ensures the effortless jump over
Optimization (GWO) optimized PI controller. Pratibha and Bhavni [26]
from on- grid to islanded mode, while retaining in solidity of the system.
proposed an AGC mechanism utilized restructured network amidst
Diesel engine generators(DEG) are most commonly used distributed
incorporation of wind power generation. Wind generator integration
generation technology due to their reliability in plug and play operation
minimizes the total system inertia during disturbances, as the wind unit
of microgrids. During grid connected mode, designed controller per­
provides no inertia and disconnects the system from the mains. Optimal
forms its task by operating PV at maximum power output and DEG at
speed control parameters based on a doubly fed induction generator
their optimal point. Similarly, in off grid or isolated mode, controller
(DFIG). Speed control variables of Doubly Fed Induction Generator
tracks the frequency of the microgrid and support the system stability.
(DFIG) deployed wind turbine has been obtained by utilizing PSO which
DEG and PV are co ordinately controlled. Here the DG’s are operated at
inflates the involvement of DFIG in the frequency control action. PSO is
their maximum power output, so as to reduce the consumption from the
a computing approach which fine -tunes an issue by repetitively
DEG. Proposed controller is tested and verified under certain operating
improving a candidate solution with respect to the specified quality
condition and model is presented in Fig. 10. Result obtained indicates
standard. PSO is a metaheuristic technique which employs the concept
that the controller renders very good performance under real time
of swarm intelligence and efficient to solve any mathematical
environment.
complexity.
Silva et al. [33] depicts that energy storage are also beneficial in the
Bhongade et al. [24] developed a Proportional Integral (PI)
area of power system stability by focusing on several merits of BESS like
controller having gain KP and KI are tuned by adopting GWO algorithm.
power quality improvement, load frequency control, load sharing etc.
For optimum response, Integral Time Square Error (ITSE) objective

Fig. 8. Model of AGC applied microgrid incorporated HPS [23].

6
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

Fig. 9. Block diagram of PID controller [30].

conventional sources. Researcher shows the experiment performed on

Table 2 microgrid based on solar and wind in an islanded mode. Author [35]
Performance of PID controller.
discuss about the modeling of microgrid incorporated HPS and their
Controller Rise Time Settling Overshoot Steady State scope in future as renewable are clean energy sources and helps in
Parameter Time Error
serving local as well as industrial load (when connected to the grid.).
KP Reduced Shoot up Minor Reduced Also the emerging advancements of smart grids for load side manage­
deviation ment and their control strategies has been done by [36]. Maiti et al. [37]
KI Reduced Shoot up Shoot up Eliminated
KD Minor Reduced Reduced Minor
presents the PSO optimized Fractional Order PID (FOPID) controller. For
deviation deviation optimum response, ITAE objective function is used. A fractional PID
controller shown in Fig. 11 is an addition of classical PID controller as it
provides increased flexibility in controller design in regards of standard
Electric vehicle is also an emerging technology in system functioning PID controller. These controllers are less sensitive for parameter change
and are used in regulating the microgrid frequency. Whenever load in a control system. Transfer function of FOPID controller is ,
variation takes place, how electric vehicle action take care of the power
required is shown in some literature. Padhy et al. [32] shows a plug in GC (s) =
U(s) KI
= KP + λ + KD Sμ (λ, μ > 0) (13)
vehicle connected hybrid system which optimizes a cascaded PI-PD E(s) S
controller using GWO. Reformed GWO tuned cascaded PI- PD Where,
controller is designed in support of LFC in the presence of plug in EV.
Developed approach is applied to tune controller gains in the vicinity of KP = gain of proportional controller
uncertainties like governor dead zone, rate limits etc. Researcher also KI = gain of integral controller
proposed hierarchical control strategies like device level control, system KD = gain of derivative controller
level control in the field of energy storage involved in HPS. As genera­
tion from renewables have stochastic behavior, instability occurred in 5.2. Application of control strategies
the system, which is diminished by optimizing a PID controller. PID
controller gains are tuned using GWO. Microgrid is a combination of Khamis et al. [38] presents the microgrid incorporated HPS con­
several DG’s and Energy storage systems, which supplies power to local taining solar, battery, fuel cell linked with the utility using grid tied
utility while improving power quality. Power produced from DER’s like inverter. An experiment has been performed on this system which shows
solar, wind etc. is assemble in DC bus via power regeneration system. that voltage and frequency deviates with load change. As fast as the load
Benefits of microgrid is utilized in [34] and due to the advancement of demand varies these factors also varies in microgrid based system.
electrical entity in need of more power generation from non Different control strategies and application of power electronic

Fig. 10. . Schematic of DEG-PV hybrid AC microgrid [32].

7
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

Fig. 11. Block diagram of FOPID [37].

converters for frequency regulation has seen. Author [39] emphasis on (iii) Control using soft computing and intelligent techniques
hybrid power system using battery energy storage. It has been shown
that for frequency regulation, battery energy storage provides good re­ System stability is enhanced by incorporating microgrid generation
sults and is used most widely in the power system. Some researchers sources and several methods have been proposed. Bhatti et al. [47]
[40] also focuses on the HVDC system and provides an optimum solution shows the LFC scheme in multi area system in which a PID controller is
for power flow with less cost and presents a solution for optimal power employed and their parameters are optimized with Integral Square Error
flow. Literature [41] also shows that these strategies are efficient when (ISE) and Integral Time Average Error (ITAE) using genetic algorithm
wind farms are integrated with grid and also proposed some advanced for primary control based optimal scheduling on generation side. LFC
control methods. At device level control, hybrid storage control strategy mechanism of interconnected system containing wind power generation
firmed on Port Controlled Hamiltonian (PCH) is designed to achieve has been developed in [48]. As power generation from wind is variable
better output performance. Similarly, at system level control a multi­ in nature, frequency fluctuation occurs which are mitigated using AGC
level power allocation and fuzzy logic is proposed to obtain efficient technique.
wind integrated grid. Literature [49] shows that author developed a method named Dy­
namic Gain Tuning Control (DGTC)for wind generating sources involved
6. AGC control schemes system so as to automatically calculate the parameters of PI controller
under disturbances occurring system. Similarly [50] shows the imple­
Frequency deviations occurred in microgrid due to the disturbance of mentation of AGC in wind integrated multi area system by describing an
load generation balance and this imbalance is mainly due to the smart algorithm for active power balance control. Solar and wind based
grids relying on renewable energy sources as they are variable in nature microgrid incorporated HPS by designing a Biogeography-Based Opti­
[42]. Several control strategies like conventional control, robust control, mization (BBO) based PID controller has proposed by Rehman [51].
adaptive control etc. have been proposed to overcome this issue. Con­ Output power from microgrid is nature dependent, its efficiency has to
ventional methods or classical control methods are usually graphical be improved by applying MPPT method in order to extricate maximal
design based and proposed in the domain of frequency. Several con­ power from PV panel.
trollers have been employed in the system , so as to perpetuate the Author [52] demonstrated a new fractional order fuzzy PID
system stability. Following are the classification of various control controller and examined it on 3 composition of multi area multiple
schemes: source AGC system. Proposed controller is tuned using BFOA. Study
shows that proposed controller gives superior outcomes in contrast with
(i) Classical Control other optimized controller. But the literature mentioned have not dis­
cussed the impact of MPPT in the area of AGC. Thus researchers [53,54]
Classical control mainly focuses on root locus, Nyquist and bode focuses on AI based control approaches like fuzzy logic, ANN etc. for
plots. John et al. [43] shows that classical control methods are simple as load frequency regulation. Author considered load perturbation in
compare to others and can be easily implemented in the existing sys­ hydro thermal interconnected system. Studies shows that AI techniques
tems. These methods are used in AGC and primarily developed for are time consuming, complex designing and inefficient for the system.
nominal operation without any constraint breach and while changing Computational intelligence method such as GA, Bacterial Foraging
any operating point ,they are ineffective. Due to the use of Single Input Optimization Algorithms (BFOA) etc. is a combination of artificial in­
Single Output methods (SISO), classical controls are rarely used in telligence and optimization techniques presented by Pandey et al. [55].
present scenario. Literature [56] focused on obtaining tune parameters of integral
controller by applying trajectory insensitivities and these parameters are
(ii) Robust Control Schemes tuned using gradient type N-R algorithm. Recently a technique is pro­
posed in [57], named Biogeography-Based Optimization (BBO). This
Researchers in [44,45] proposed several techniques viz. H∞ control, method optimize the gain parameter of a unique controller called 3
mixed control ( H∞ / H2) and μ analysis and synthesis which provides Degree Of Freedom ID Controller (3DOF-ID Controller) simultaneously
solidity and better attributes in opposition of unpredictability and and its structure is shown in Fig. 12.
change in load demand. Author works in the deregulated environment Along with this, it also optimizes parameters of electric governor in a
under bilateral policy and proposed these control techniques to develop two area system. Parameters are tuned using Integral Square Error (ISE)
a distributed vigorous AGC scheme. In this approach, H2 approach re­ as objective function. This method also calculated certain uncertainties
duces tracking error and linear-quadratic cost of the control inputs, and like Governor Dead Band (GDB), Generation Rate Constraints (GRC) etc.
H∞ control ensures control- loop system stability in the presence of by taking 1% step load change in consideration.
control limits and non linearities. Pan et al. [46] developed a FOPID
controller applied for AGC scheme in distributed energy system. Author (iv) Other control techniques
consider a solar, wind, fuel cell , aqua electrolyzer based HPS and gains
of controller is optimized by implementing these methods. Author [58] studies a new method which is Decentralized Sliding

8
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

Fig. 12. Structure of 3-degree of freedom controller [57].

Control Method (SMC called Sliding Mode Control) implied for LFC in an
interconnected system having uncertainties. In [59] author depicts the
behavior of SMC controller applied in multiarea interconnected AGC
scheme in de regulated environment and result shows that its perfor­
mance is better and robust as certain factors such as GRC, GDB etc. will
remain unaffected from this technique. Some researcher [60,61] focuses
on AGC scheme applied in thermal, solar and gas based multi area
interconnected system communing several uncertainties such as GDB,
Fig. 13. Auxiliary control logic [67].
GRC etc. For non linearity like GDB, error between the calculated and
ideal output of generator is sum up to the controller output for rejection.
Similarly for GRC, difference of real and ideal response is supplied back generator fed by the existing governor control. Bio-diesel based syn­
to the controllers observer for estimation and elimination. chronous generator is considered as a power source of microgrid whose
Mohamed et al. [62] presents decentralized Model Predictive Control frequency deviation is minimized using proposed control logic. Author
(MPC) for multi area system. This paper shows that this method is not [68] depicts that the control logic acts faster than the existing governor
well suited in co ordinating each subsystem as there is no any kind of controls and performance has been compared. Fig. 14 shows the modi­
interactive information between them. Recently paper [63] explains a fied system in which the proposed control logic is employed in AGC
Distributed Model Predictive Control (DPMC) while working under loop.
deregulated environment. Authors worked on a multi area system where
user demand fluctuates in sight of nonlinearities and glitches. Simula­ (v) Hybrid Intelligent control
tion has been done and shows that this control is beneficial in deregu­
lated environment. DPMC has certain disadvantage like it involves more Some researchers [69,70] uses Linear Matrix Inequalities (LMI) and
computational power and cost. Ma et al. [63] proposed an outline of genetic algorithms to evolve a sturdy PI controller in LFC schemes due to
distributed control system for large deregulated system. This outline load perturbation. A GA based PI control for a 3 area multi source system
breaks the larger system into small subsystems and controller is con­ has also studied in [71]. This study also helps in determining frequency
nected to each subsystem locally. These controllers transfers data be­ bias factor. This is mainly depends on primary control action and
tween each other using a communication bus. damping coefficient in a control zone. Hosseini et al. [72] proposed a
Researchers [64,65] focus on dealing with the strategy to control the LFC control for a deregulated system based on PSO and Adaptive Arti­
missed sampling data caused due to the time delay occurred while ficial Neural Fuzzy Inference System (ANFIS) for optimal scheduling of
interacting with different subsystem. These researchers presents a MPC gain. ANFIS is investigated by communing uncertainties like GRC in a
based AGC scheme in an interconnected system, as this method is very control area and determines DB. Simulation has been done and result
popular to handle multivariable and constrained system. Some re­ obtained shows AGC mechanism is completely obeyed as both the ob­
searchers also explained MPC control applied in the interconnected jectives like frequency regulation and no tie line error in the control
system having large renewable penetration along with the existing time areas is obtained. Real time solutions has obtained as adaptive control
delay. Recently a Quasi Oppositional Harmony Search Algorithm gains are updated in accordance with the perturbation of load. This
(QOHS-PID) in a five domain system operating in a restructured envi­ method shows good results but the PSO involve takes large computation
ronment is proposed by Shive [66]. These areas consist of certain power time to optimize controller which seems ineffective for real time prac­
generating unit like reheat thermal, hydro, gas units and non linearities tice. To take care of this limitation Debbarma [73] proposes a Bacterial
like GDB,GRC were tested. Gains of controller i.e. KP , KI , KD are tuned Foraging optimization(BFO)Technique to perform two functions simul­
using QOHS algorithm while taking Integral Square Error (ISE) objective taneously such as optimizing non integer controller gain and speed
function. Multi area , power system is simulated and test result shows adjustment variable in multi area system. BFO is also a nature inspired
that this algorithm stabilize the frequency regulation, profiles of algorithm which utilizes the pasturing action of E.coli bacteria. It is
scheduled tie line power and also regulate the power from GENCO’S to a
rated value under physical constraints. Several soft computing tech­
niques or artificial intelligent techniques are also investigated under
deregulated environment for control purpose. Mishra et al. [67]
designed an auxiliary control system to mitigate the frequency variation
in a microgrid. Author uses an additional signal in the LFC loop shown in
Fig. 13, to prolong frequency by regulating the generator power output.
Proposed logic is tested on standard IEEE 13,33 and 69 bus distribution
system , for its performance under certain conditions.
Control logic is the power imbalance results because of the later
alteration in user demand and modified mechanical input from the
Fig. 14. Modified system based on auxiliary control logic [68].

9
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

simple and easy to use with a limitation of poor convergence in complex tuned and compared using several AI methods such as genetic algorithm,
optimization problems. PSO etc. Study shows that GOA optimized PDF plus (1+ PI ) imparts
Researchers like Li et al. [74] depicts the problem of DMPC in non superior response in comparison with other PI, PID controller in respect
linear system restricted with communication delays. Author proposed a of low stability time and maximum overshoot. The efficacy of proposed
DMPC scheme by utilizing a robust constraint in which a time lag is controller is tested in power system containing FACTS devices by doing
involved to analyze the system security. This paper also shows that sensitivity analysis under several load perturbation and parameter
achieving stability is a problem faced in non linear systems because of variations.
the complexity in computation and certain limitation in bandwidth In [81–83] authors developed a FOPID controller with filter coeffi­
communication. In [75] Author proposed a novel bi-objective optimi­ cient to analyze the influence of realistic HVDC tie line in addition to
zation model of real-time AGC dispatch (BOAD) approach for maxi­ realistic dish sterling solar thermal system (RDSTS). Proposed controller
mizing the technical and economical aspects simultaneously of is employed to study the active power regulation in multiarea under de
independent system operator when they allocate real time generation regulated environment. Three areas containing thermal RDSTS system
command to distinctive generation sources. For such purpose author having different GRC rates are modelled in which each area have dis­
utilizes a non dominating sorted GA-II to achieve high quality pareto tribution and generation companies. A novel cascaded controller is
front for real time AGC consign. Proposed approach minimizes mileage proposed which involves the properties of FOPDN and FOPIDN and
payment of frequency regulation market as well as total generation controllers gains are simultaneously tuned employing a new meta­
variation. Literature study shows that coordinated optimization in AGC heuristic algorithm termed as crow search algorithm. Response of sug­
dispatch while using model predictive control as well as machine gested work is analyzed with the response of PID, FOPID etc. and it is
learning method are considered as key methods. In this BOAD based represented as oscillation amplitude, overshoot , stability time. In de
approach computation time is large and due to this delay real time regulated environment , policy schemes like Poolco, bilateral are
optimization is not achieved. Due to this a research gap is obtained contemplated and analysis shows that the effect of proposed method on
focusing on improvising the speed of optimization. LFC in RDSTS and RHVDC tie line is superior than other cascaded
In [76] author proposed a cascaded proportional integral FO- PID controller in de regulated environment as well. Dynamic performance of
controller for minimizing frequency deviation in hybrid systems. the considered system is tested under deregulated scenario. Khamari
Controller gains are tuned using novel Gorilla Troops Optimizer (GTO) et al. [84] proposed a hybrid TLBO based Tilt-Integral-Derivative (TID)
as it is not earlier implemented in microgrids . Author considered 2 area controller for load frequency regulation under de-regulated scenario.
system containing diesel, wind ,solar PV system along with energy TLBO method is metaheuristic population based algorithm which does
storage system.. Efficiency and robustness of proposed controller is not use specific parameters for reducing the time duration to produce
collated with primary FOPID controller whose gains are tuned using efficient output and stable convergence. This method mimics the
GTO ,GWO,PSO etc. under distinct structures like step and dynamic load teaching nature of teacher in classroom and its impact on the nature of
change, variation in source power etc. An integral absolute error student.
objective function is used in PI FOPID controller which improves the Author considered two area four unit system initially and the pro­
fitness function compared to other structure controller and provide posed controller is tested under MATLAB/ SIMULINK environment.
better transient response thereby diminishing the frequency deviation. Output response of proposed approach is compared with GA, differential
Since two area system is considered frequency regulation in area1 , evolution techniques optimized PI, PID controller under same scenario.
area2 and tie line power is observed under different cases. GTO tuned Result shows that the proposed approach is superior and the result is
cascaded PI FOPID controller shows high superiority and improved given in the form of settling time, overshoot, undershoot etc. After that
dynamic results when subjected to different operating conditions in the author considered two area six unit system and each area is incor­
hybrid system [77]. Model is simulated in Simulink and it has been porated with SMES and Thyristor controlled phase shifter. And some
observed that it provides improved settling time as well as overshoot it non-linearities such as GRC, DB has also employed in the system. With
comparison of other techniques. large renewable penetration proposed method is gone through sensi­
Jenna et al. [78] proposed an adaptive control strategy to regulate tivity analysis by changing the load conditions and it has been seen that
frequency fluctuation and power deviation, while improving the reli­ gains of proposed controller need not to be change under such condi­
ability and performance of microgrid network. As the sources incorpo­ tions. It is evident that due to low inertia support , intermittent nature of
rated in microgrid is stochastic in nature, they have suffered from lack of renewables enlarge the change of frequency deviation under islanded
inertia which in turn requires a reliable controller to achieve the load mode of operation. Alteration rate of frequency is maximum in islanded
generation balance. Non linearities such as GRC, communication delay , mode as compared to conventional system. So, it is require to in co-
load perturbation are certain issues reside in microgrid. In this litera­ operate an effective solution which compensate the shortage of ancil­
ture, author integrates small hydro power plant (HPP) in the MG thereby lary services required for frequency regulation.
enhancing the functioning of the system. Selfish Herd optimization In [85] author proposed a novel supplementary control strategy for
(SHO) and fuzzy based adaptive SHO (FASHO) is employed to fine tune DER based islanded microgrid system. Due to these converter based
the controller gain thereby increasing its response on any system vari­ sources requirement of fast frequency regulation arises. Some of these
ation. Fuzzy logic control decides the optimizing parameters of SHO sources like flywheel energy storage system (FESS), solar etc. is taken as
which in turn improves the efficacy of algorithm [79]. Centralized potential source of power for active power regulation in LFC scheme.
sliding mode control based controller is employed whose parameter are Author developed a more efficient dynamic model in comparison of
tuned using FASHO and its performance is compared with other PID, previous model of energy storage system. Proposed fuzzy droop
FOPID etc. controller. Performance analysis shows the superiority of controller is optimized using artificial bee colony optimization method.
FASHO integration in the microgrid. Nayak et al. [80] proposed a High energy density, good power exchange are some of the benefits of
Grasshopper Optimization Approach (GOA) based PDF plus (1+ PI) FESS which will be utilized for frequency excursions under load dis­
controller and this controller is employed for AGC in FACTS containing turbances. Similarly, wind system are capable to make use of the dy­
power systems. In this paper author considered three thermal units namic energy saved in rotational mass whereas fuzzy droop enhances
having different GRC rates along with FACTS devices. Author introduces frequency regulation in comparison with traditional droop of wind
a metamorphic PDF plus (1+PI) controller keeping filter with PD and system. Several control strategies based on DERs has been proposed and
(1+ PI ) controllers. For better efficiency transient response of PI, PID, proper co-ordination of these sources provides reasonable solutions
and PIDF controllers are equated. Performance analysis of GOA based under several operating scenarios. Propose approach is tested under
controller is tested under OPAL RT environment. Controller gains are MATLAB environment and it has been shown that solar , wind, flywheel ,

10
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

DEG system has the ability to diminish the frequency deviation under gain of developed controller is tuned using ABC method and is tested
various operating scenario even when small signal dynamic model is under MATLAB environment. ABC tuned TSMC controller gives
incorporated in the standalone microgrid. magnificent performance in microgrid for AGC by giving excellent
Paper [86] proposed a novel distributed secondary control technique damping efficiency, low steady state error etc. Propose controller damp
for AGC and AVR which maintains the load generation balance while out frequency oscillation under disturbance and uncertainty conditions
reducing the generation cost of power in microgrids. As DGs suffers from and proves its robustness and sturdiness for the LFC.
lack of inertial support, and to obtain the frequency voltage recovery of Author [92] proposed a secondary frequency control also called
such system [87], author employs continuous time distributed control to rapid start operation by using the concept of several machine learning
initiate a control variable for the cause. On the other hand it employs a algorithms. ML algorithms are firmed on neural networks, clustering
distributed economical increment also in secondary control to reduce techniques, decision trees and helps in starting several rapid start units
the economic dispatch problem (EDP) by neglecting the issue of speed to perform LFC in the system and its classification is shown in Fig. 15.
inconsistency in communication between secondary and tertiary level This paper basically focused on rapid start units which means system
control. This method is employed in both on-grid and synchronizing will continue to sustain the operation close to the nominal frequency
mode for voltage frequency recovery and EDP. As it is known that low range for maintaining better efficiency.
proportion of X/R results in sturdy linking of real power and phantom ML based algorithms are used to anticipate that the condition of AGC
power and author is primarily focusing on voltage frequency recovery is fulfilled and rapid start (RS) unit is connected, as the initiation of RS
under the influence of inductive line impedance. Proposed method is unit is one of the complex issue due to its operation in real time. This
tested while switching between isolated mode to on-grid mode and work is classified into two parts (i) proposing multiple machine learning
initializing the plug and play operation. Author considered an AC based algorithm for AGC and (ii) selection of these algorithm in multi
microgrid consisting five DGs for voltage frequency regulation and EDP objective decision problems. A future research objective is found as it
in secondary level control. Two level algorithm is employed in the should to be employed for economical reliability of the Spanish system.
microgrid and it provide helpful response in optimal economical oper­ In addition to that a machine learning process is shown in Fig. 16.
ation by reducing the communication fault condition between second­ In [93] author focuses on the issues of AGC in isolated two area
ary and tertiary control. Proposed system is tested and result shows numerous sources based microgrid when there is inconstancy and vague
improved stability under the communication link failure. in generation and load demand. Author proposed a microgrid based two
Paper [88] proposed an Integrated FOPI and TID controller for fre­ area model having multiple energy sources and energy storage equip­
quency excursion in hybrid microgrid system. Suggested controller ment for LFC. Dynamic characteristic of discrete time Markov based
make use of the properties of both the controller stated above and method is used for the formulation of solar, wind and load variations.
provides effective and robust AGC in modelled system. Proposed Frequency controller is proposed whose parameter are tuned using
controller is optimized by a novel metaheuristic algorithm named as simulated annealing-based Q learning (SAQ) algorithm [94]. Q learning
wild horse optimization(WHO). Author considered renewable such as method is used to optimize controller parameters without falling into
solar PV and wind integrated conventional power system along with local optima. Dynamic response of the controller for two area microgrid
electric vehicle employed in each area for study. Suggested controller is tested under MATLAB environment. Author basically works with the
performance is verified under various operating scenarios such as load secondary frequency control where small variation in frequency should
perturbation, renewable fluctuation , sensitivity analysis etc. reliability be diminished in multiple sources based microgrid system. Response of
of suggested controller is assessed by applying wave oscillation in area 2 suggested controller is tested and analyzed with other conventional Q
. Suggested controller is corelated with other traditional controller learning based controller and outcomes shows the supremacy of SAQ
optimized using WHO and result proves the dominance of developed controller as it map with the CPS economic evaluation index. CPS index
method over other in respect of better settling time, oscillation ampli­ of interconnected system is the measure of quality of power system
tude etc. Inclusion of electric vehicle also provides advantage such as control considering environmental index as well as economic assessment
improving the frequency regulation together with minimization of index for long drawn accomplishment of AGC. CPS is used to assess the
power flow in tie line. Time domain analysis has been implemented on execution of SAQ learning controller and given as :
proposed controller under Simulink framework. An intelligent inte­
CPS1 = (2 − CF) ∗ 100% (14)
grated standalone microgrid control is proposed in [89]. Microgrid
system comprises of solar PV , wind turbine generator , DEG as ( )
ACE∗ Δf
distributed source of power as well as battery and flywheel based energy CF = AVG (15)
− 10Bi ∗ ∈21
storage system. For extracting maximum power from solar PV system an 10min

improved incremental conductance based MPPT method is used where √̅̅̅̅̅̅̅̅̅̅̅̅̅̅

CPS2 : AVG(ACE)10min < 16.5∈10 Bi ∗ Bs (16)
as fuzzy wing power generation is used as an MPPT for wind system.
Prime novelty of the work is controlling power flow in microgrid thereby Where,
improving the load frequency control in the system. For maintaining
optimal flow between generation, load, and energy storage an intelligent ∈1 = control target value of frequency deviation in 1 min.
fuzzy based adaptive control is developed. Fuzzy based PID controller is ∈10 = control target value of frequency variation in 10 min
employed to obtain wind and DEG droop control . Fuzzy observer is Bs = total frequency deviation coefficient used as evaluation index
developed to direct the flow of power among the load and source. Pro­
posed system is tested under IEEE 39 bus system and result shows that
the suggested approach is 90% more efficient than the primary MPPT
method as it has high tracking efficiency and fast MPP tracking speed.
In [90,91] an artificial bee colony(ABC) optimization inspired ter­
minal SMC (TSMC) is developed to overcome the frequency oscillations
in stand alone system occurred due to the penetration of renewables in
the microgrid system. It is a finite time control method This control
strategy is used to design a non linear sliding surface to make sure the
system state should be converge to origin in the definite time by using
fractional power terms. Due to the distinct stability of the control it
ensures, convergence of state variable along with higher precision. The Fig. 15. Classification of ML.

11
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

System performance is tested under MATLAB environment and it has

shown that TDAC approach gives exceptional learning and exploratory
stability as well as improved dynamic performance of the designed
system.
Electric vehicle (EV) is one of the emerging area of research nowa­
days due to their support to grid stability. Vehicle to grid technologies
are drawing keen interest of researchers due to their benefits of con­
Fig. 16. Process of ML [92]. trolling and planning of large power network. As these vehicles are
employed with large number of high rating batteries and bidirectional
A novel multiagent control strategy is proposed in article [95,96] for capabilities of converters put together it for better frequency support. In
AGC primarily focusing on the deep reinforcement learning (RL). DL is [99] author integrates EV to the grid as a controllable load along with
subset in ML which represents the operation of human brain to perceive other generation units so as to provide frequency support, when dis­
interdependence and patterns by filtering data based on defined logical turbances occurs in multi area power system delineated in Fig. 19.
anatomy and its basic process is shown in Fig. 17. Deep learning Author employs tilt ID (TID) controller for EV and controller gain is
methods provides accurate output response as they depict supervised tuned using artificial bee colony algorithm. For better performance
data as an input. This concept mimics the functioning of biological analysis AGC parameters such GRC, DB etc. are taken into consideration.
neural network and compares the new data with unspecified data to System is tested for three condition (i) step load change (ii) sensitivity
obtain the decision. Concept of deep learning and machine learning is analysis of coordinated control and (iii) frequency and tie line power
achieving great interest for active power regulation schemes. deviations. Suggested method is tested under MATLAB environment and
As deep learning utilizes large aggregate of hidden layers in NN it has been shown that controller has improved reliability for 50%
which defies the primary neural network and restrain only few numbers change in GRC of turbine and synchronizing coefficient independently.
of hidden layers. Fig. 18 shows the performance comparison of deep In [100] author focused on electric vehicle based frequency regulation.
learning from other AI techniques. In order to progress an efficient AGC Due to the operational benefits of EV such as fast response and its sup­
scheme, RL based control strategy has been developed for grids having port to the grid, control strategies are proposed which withdraw high
high renewable penetration. So, in expansion of this a multi agent deep frequency signal from ACE and ARR (active regulation requirement)
RL algorithm is proposed in this article and it sinks the phenomenon of using digital filter.
Double Deep Q Neural (DDQN) algorithm and action discovery strategy. Regulation capacity of EV is evaluated on the basis of maximum
This technique possesses self-exploration and optimization ability. For vehicle to grid power and its present power and it will be consign the
real time precise control, action set of DDQN does not map with the regulation task among EVs. Simulation result shows that efficiency of
condition specified for AGC as they are artificially fixed. Whereas this proposed approach as it reduces frequency deviation and tie line power .
issue was overcome by the action discovery (AD) strategy as it will After reviewing all these literature , benefits and limitation of employed
interrupt the boundaries of action set and optimize it in the real time techniques are concluded in Table 3 given below :
whereas enhancing the dynamic efficiency of LFC system. Simulation is
performed on Simulink platform and result shows that deep learning- 7. Discussion
based method gives fast convergence speed under high penetration of
renewable sources. After the thorough study of AGC schemes, controller implemented
In [97,98] author focused on neural network based AGC in microgrid for frequency regulation and optimization methods for tuning controller
system. As it is evident that conventional AGC methods are not more parameter, it has been observed that all these robust control techniques
promising with the evolving distributed generation system due to their lacks on several parameters such as large settling time, low controller
stochastic nature. As a result large disturbance from source as well as efficiency, slow operation, unreliable operation etc. [101]. These issues
load side causing issue in power grid. So, a DL firmed three network dual are more severe when DGs comes into picture as there is large pene­
delay actor critic (TDAC) approach is developed for LFC to overcome the tration of renewable which causes stochasticity in generation also
issues caused by large penetration of renewable and non linear loads. continuously varying load in the existing system influence system sta­
This control strategy make use of several neural network to assess the bility severely. Some methods like SMC, MPC suffers from high cost
system value and accommodate its action strategies. Modified actor optimization, absence of real time optimization etc. Several AI based
critic approach as well as incentive heuristic mechanism is wielded to controllers are implemented and enhances the system response but
maximize the exploration efficiency , system dominance performance certain performance parameters still lags these techniques to completely
and standard of AGC. TDAC based approach take care the large distur­ dominate the existing conventional droop control scheme. As AI based
bance caused by renewables to the power grid. This method merges techniques are metaheuristic techniques and mimics the environmental
DDQN and modified actor critic strategy to enhance the learning and scenarios for controlling purpose, so to find the finest optimal solution it
decision making ability. As stated DDQN is a type of deep reinforced requires more convergence time which lags the process and delays the
learning technique entrenched on iteration function which is a fusion of regulation. Some adaptive and coordinated control methods are also
neural network and Q learning. To achieve optimal coordinated control, developed by some researchers for better LFC due to their good per­
a novel iterative method which minimizes the optimization bias is used. formance output but these methods are not applicable for large and
multi area interconnected system. Some authors implemented EV based
control schemes for active power regulation in microgrid system. EVs
are integrated with the grid and their support is tested under various
operating condition. After the study it has been observed that controllers
are proposed which charges and discharges the battery of EV under
different loading conditions but there is a gap for proposing an adaptive
controller which performs three function (i) charging when load de­
mand is less (ii) discharging during high load condition (iii) Employed
battery acts as reserve for lightning the EV charging station. For such
purpose properties of several controllers are combined and their benefits
are taken into consideration. Lithium ion batteries are used in EVs which
Fig. 17. Deep learning Process [95]. are expensive but has good life span, compact, better energy density as

12
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

Fig. 18. Comparison between deep learning and other Neural Network [96].

Fig. 20. Authors in [106] proposed a hybrid Demand Response (DR)

control method by introducing communication latency and result shows
it improves the stability margin and proposed method is validated using
OPAL -RT based real time simulator following some experimental
analysis. It has been observed that demand response based frequency
regulation in traditional generation system is performed but in order to
improve system performance, existing literature lags the concept of
demand response regulating frequency stability in DG incorporated
hybrid system. Some researchers [107,108] involves Renewables in
frequency control but the effect of sporadic behavior of RE’s ,sudden
load change while considering communication latency and physical
constraints still needs to be explored. After thorough study, it has been
observed that area of demand side management for frequency regulation
in presence of DG is still unexplored. Therefore, suitable algorithm and
techniques can be developed to implement demand side management
activities for frequency regulations in DG incorporated Hybrid power
system.
Fig. 19. EV incorporated Grid for AGC [100]. In [109], author discussed recent work done in the area of smart grid
and cyber security issues is one of the prominent issue occurred in the
compared to lead acid batteries, nickel batteries etc. system. This paper also deals with the new controller like FACTS
Paper also briefs about Newly emerging area of research in ML and controller (Distributed Power Flow, DPFC) and Energy storage devices
DL based optimization methods [102,103]. As both the methods are based strategies developed for restructured power system. This paper
subset of AI techniques but possess several advantages over conven­ also presented frequency regulation in 5 and 6 area of multi inter­
tional approaches such as it can generate automation feature, improved connected power system. Similarly, [110] presented challenges like
learning capabilities, coordination with unstructured data, cost inten­ stability, low virtual inertia of hybrid microgrid in optimal regulation of
sive, scalable etc. These methods are different from other AI methods as voltage and frequency. Author proposed a bio energy model for hybrid
in AI methods, data is given as input and after some traditional pro­ microgrid system by incorporating renewable sources as well as energy
gramming, output is obtained. Whereas in case of ML, data is given as storage system. This model helps in replacing the costly diesel engine
input along with some desired output at initial stage and after data generator by waste to energy concept and provides stability and effec­
training, latest data in the form of program is generated and training tive power regulation. In [111] author proposed a novel Green
algorithm will finally produce the desired output. Along with this, in Leaf-hopper Flame optimization (GLFO) algorithm for synchronized
deep learning data is given as input and after training the data (using voltage and frequency regulation in hybrid MG system. An isolated
deep learning algorithm) desired output is achieved. Efficiency and microgrid is modelled to produce power using various energy sources
effectiveness of Machine learning is depend upon both input data and like solar, wind , agricultural waste etc. along with bio-hydrogen based
performance of training algorithm. DL and ML methods required large energy storage system. Stochastic behavior of these sources leads to
number of data sets to produce a trained output and these benefits certain issues like optimal power flow, reliable power supply etc. A PID
should be further employed for cyber security attacks on microgrids controller is tuned using novel GLFO approach with Integral-Square of
where the attackers injects faulty data or false data to disorient the Weighted Absolute Error (ISWAE) based objective function. This
system objective. Existing cyber security solutions slow down the power controller helps in minimizing the power fluctuations in the system.
system operation and they took some time to recover from pre attack Proposed controller is tested under MATLAB environment and shows the
point [104,105]. A schematic block diagram has been developed which superiority over other conventional optimized controllers. Author [112]
shows the implementation of ML in the system for LFC is shown in developed a novel proportional fractional integral derivative filter plus

13
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

Table 3 double derivative ((PI)^λ (DFD)^2) controller optimized using Memetic

several control methods with their benefits and limitations. Salp swarm algorithm for low voltage AC microgrid to improve the
Control Techniques Benefits Limitations stability and robustness of the system, under the variations of operating
parameters. Concept of virtual power and virtual impedance is utilized
Classical linear Elementary design and easy Long settling time, large
control i.e. P,PI,PD, to implement frequency fluctuations, to introduce an improved droop controller which provides improved
PID [8,11,16] poor results over wide accuracy in comparison to other distributed controller during load
operating range perturbation, activating controller parameters as well as reconfiguring
Deterministic Provides optimal control No real time testing only the structure. Proposed controller is tested and validated using real time
technique such as gains. simulation study has done.
Newton- Raphson
simulator test bench using OPAL-RT 4510 for practical implementation
[5]. in the power system. In [113], a control strategy has been developed to
Metaheuristic Preferable dynamic optimization is not easy in improve the power quality in residential Solar PV system. A modified
techniques like behavior, lower controlling actual time, large settling power control based strategy is implemented in grid integrated auton­
PSO,ACO etc. [9, order. time.
omous MG having multiple solar PV units. Proposed strategy provide
19,20]
AI methods like Better dynamic response, For better efficiency, smooth automatic transition from grid tied mode to off grid mode and
ANFIS+PSO, real time optimization controller requires large vice versa while multiple PV units are operating in parallel. During grid
ANFIS+ ANN [48, training sets tied mode, power obtained from solar PV is used to fed the load demand
60,72] whereas the remaining power is supplied to the grid at unity power
Hybrid control Optimization techniques are Simulation study shows
methods like useful for large that it is not useful for
factor while working under current control mode. During grid failure,
GA+SA, PSO+CRO interconnected systems online optimization, the current control mode of operation in grid tide mode is switched to
[37,55] having various non - voltage control mode in autonomous mode of MG to maintain contin­
linearities. uous power supply to the system. During grid restoration, solar PV
BBO [51,57] Improved dynamic response Hit and trial error method
inverter units at PCC resynchronizes the voltage at PCC with the utility
in terms of oscillation is used to find parameters.
magnitude, peak variation voltage and switched to grid tied mode. Result shows smooth transition
etc. takes place during grid failure and gris restoration with out any
Adaptive control [78, Good performance under Not practical for large area communication setup while maintaining optimal power flow in the
89] limited condition system. network. This method minimizes the requirement of large communica­
Fuzzy , neuro fuzzy Good performance in small Impractical tuning of
tion setup , thereby reducing complexity as well as cost of the system.
control [29,52] area system parameters
QOHS-PID [44,50] Beneficial optimization for Trial and error method for Proposed method tested under simulation environment and shows that it
several non linearity present B and R parameter works efficiently during load fluctuations, attenuates harmonics,
multi area system determination. maintains the total harmonic distortion as per IEEE standard 519.
Sliding mode control Low computational cost, Execution is impractical
Author [114] proposed a control strategy for secondary frequency
[78,90,91] less sensitive for parameter
variations regulation during cyber attacks in AC microgrids. A cyber resilient
MPC and DMPC [62, Stability performance in Costly on line optimization control strategy is developed to provide end to end security against
65] closed loop increases cyber attacks for micro AGC. For this purpose, data driven based
Mixed control [44] Improved stability, better Need of LMI and SFS. microgrid modeling is done which utilizes the rank deficient properties
experience of error tracking
of inverter based renewable sources. This modeling provides end to end
H∞ [45] Effective in multi area Acquire higher order
system control cyber security during false data injection and developed for micro AGC.
Analyzer control Productive results in Min. frequency regulation Scheme does not requires frequency computation and can be imple­
[61] disturbance rejection against load uncertainty mented during load as well power fluctuations. Control scheme is tested
Electric vehicle based Bidirectional capabilities of Increased cost due to Li-ion
for two MG system and result shows that it minimizes the deviation
[99,100] converter Batteries
Machine Learning Trained data will improve —
under various operating conditions along with fast detection of cyber
and Deep Learning the system performance, attacks. A future scope is provided to test the system for large microgrids
[92,95,96] wide range application, as it requires communication links and centralized controller which is at
learning from large data set high risk of cyber attacks and leads to system black out. A novel
distributed event-triggered hierarchical control scheme is developed in
[115] to improve the multi time scale frequency response in solar PV

Fig. 20. ML based load frequency control scheme.

14
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

systems. This method involves primary as well secondary control loops 2. Natured inspired optimization techniques provides inherent solution
of AGC. For primary control, solar PV inverters are operating under to our problems. In order to achieve a stable solution, some new data
regulatory mode responding in sub- seconds using automatic de-loading driven and model based approaches should be proposed which is not
strategy whereas for secondary control, PV invertors fives response to­ critically explored for microgrid incorporated hybrid power system.
wards reference instruction given by high level AGC control in seconds 3. Small signal stability analysis is not performed in multi area based
of time duration. Primary control is easy to employed as energy storage, hybrid power system.
irradiance sensors is not required along with its stable operation. Pro­ 4. Present literature lacks in showing optimal planning of energy stor­
posed scheme is independent of initial operating conditions and age systems for enhancing the stability of MG.
designed to reduce the cost of communication setup among PV invertors. 5. Fault diagnosis approach for AGC is not shown in the present
Proposed system is tested on IEEE − 13 bus MG system having multiple literature.
PV units and results shows that it will refines the frequency regulation 6. Cyber attacks and fault data injections are some crucial issues in
while minimizing the requirement of energy storage systems. In [116] a hybrid systems. A fast decentralized control approach needs to be
PI-PD cascade controller is employed to reduce the frequency fluctua­ developed to enhance the capability of AGC loops.
tions in electric vehicle integrated RE based microgrid. A test system is
developed by incorporating biogas, biodiesel, solar etc. sources and it is 8. Conclusion
evident that these sources are inherent in nature so optimal power
balancing is one of the difficult task in a system. So to improve the power This work is predominantly emphasizes on the opportunities in the
flow, energy storage and electric vehicle is introduced in the existing area of AGC. Prime focus of this article is to review and compare mul­
system. Optimal gain of cascaded controller is obtained from tiple control approaches available in the required area. Since several
magneto-tactic bacteria optimization method using integral square error control methods has been developed to overcome the issues occurred
as an objective function. Proposed controller is tested under MATLAB- due to the load perturbation and power generation deficit. In order to
SIMULINK environment and output shows the superiority of controller deal with the non linearities like GDB, GRC, DB etc. available in the
over other classical controller during several operating conditions. system, several control strategies has been developed. This literature
Similarly, [117] proposed a Salp Swarm Algorithm(SSA) based FOPID also shows that a research gap is available in the area of resilient
controller in grid integrated virtual power plant(VPP). Developed VPP controller optimized using deep learning and machine learning tech­
system consist several power generating entities like solar PV, bio diesel niques for LFC in microgrids against cyber security attacks. Further
engine generator, plug in hybrid EV etc. interconnected to conventional demand response based frequency regulation in RE integrated hybrid
power system. Centralized control method is embraced to attain fre­ power system considering operational constraints is one of the unex­
quency regulation in VPP system. Optimal gain of proposed controller is plored domain in the field of power system. In addition, advanced
obtained from Salp optimization method by using ISE and ITAE as an methods in the moder power system, such as fractional order-based
objective function. ISE and ITAE provides minimum settling time as well controllers, fuzzy aided approaches, cascaded controllers, electric ve­
as reduced overshoots. Proposed controller is compared with other PID, hicles, demand response, local frequency estimator, and resilience, are
PI controllers optimized using GOA, ALO, PSO optimization techniques thoroughly examined. In addition, recently published works on the AGC
and comparative analysis proves the efficacy of controller for frequency of existing power system utilizing both conventional and renewable
regulation as well as for tie line power flow. Proposed controller is energy sources are presented. Finally, the comprehensive review of this
investigated under different operating conditions such as load change, article covers all technical control aspects and current trends in the field
irregular power generation and output shows improved system perfor­ of AGC-based modern power systems and identifies the future research
mance. Author [118] proposed a distributed hybrid scheme for fre­ trajectory for mitigating frequency fluctuations in future power grids.
quency regulation in an islanded AC microgrid. Primary and secondary
control loops are utilized to obtain large signal stability of a non linear Declaration of Competing Interest
network. Control strategy includes nonlinearities like switching and
affine in inverters and minimizing dwell time obstruction in the signal. The authors declare that they have no known competing financial
Consensus strategy exerted secondary control loop improves battery life interests or personal relationships that could have appeared to influence
by analyzing state of charge (SOC) of batteries, robustness of the system the work reported in this paper.
during plug and play operation or any failure in communication link.
Similarly primary control loop controls the DC-AC power conversion of References
battery during discharging mode by utilizing dynamic hybrid system
model. Singular perturbation theory is applied to achieve large signal [1] O. Elgerd, et al., Optimum megawatt-frequency control of diverse-area electric
energy systems, IEEE Trans. Power Appl. Syst. (1970) 556–563. PAS-89.
stability analysis by selecting the gains of controller properly. Proposed [2] P. Kundur, Power System Stability and Control, Tata Mc-Graw Hill, Inc, 1994.
strategy ensures MG scalability as well as improves the health of energy [3] R. Verma, S.K. Gawre, N.P. Patidar, An extensive study on optimization and
storage [119]. A distinctive strategy is developed in [120,121] to control techniques for power quality improvement, in: Proceedings of the IEEE
2nd International Conference on Electrical Power and Energy Systems (ICEPES),
maintain the frequency fluctuation in a three area interconnected and 2021, pp. 1–6.
independent hybrid system. Suggested topology comprises of solar [4] R. Sharma, M. Karimi, Addressing abrupt PV disturbances, and mitigating net
chimney, solar gas turbine, ESS, SMES etc. and integration of these units load profile’s ramp and peak demands, using distributed storage devices,
Energies 13 (2020) 1024.
in three area hybrid system is one of the recent work done by the au­
[5] H.A. Hassan, et al., Challenges and opportunities of load frequency control in
thors. A cascaded PI-TID controller is used for frequency regulation of conventional, modern and future smart power systems: a comprehensive review,
hybrid microgrid and controller parameters are optimized using chaotic Energies 11 (2018) 2497, 2018.
[6] M. Gomez-Gonzalez, et al., Optimal sizing and power schedule in PV household-
butterfly optimization algorithm. Control strategy shows better results
prosumers for improving PV self-consumption and providing frequency
over different algorithms and controllers during sensitivity analysis and containment reserve, Energy 191 (2020), 116554.
variation in operating conditions. Proposed control schemes is tested [7] A. Hirsch, Y. Parag, J. Guerrero, Microgrids: a review of technologies, key drivers,
while considering real time scenarios. and outstanding issues, Renew. Sustain. Energy Rev. 90 (July 2018) 402–411.
[8] Xilin, et al., Research on automatic generation control with wind power
After critical review of these literature it has been found that participation based on predictive optimal 2-degree-of-freedom PID strategy for
following research gap are available in the field of AGC : multi-area interconnected power system, Energies 11 (2018) 3325.
[9] M. Prateek, et al., Taylor series based protection starting element for STATCOM
compensated transmission line, Electr. Power Syst. Res. 204 (2022) 107700.
1. Role of demand side management in renewable based hybrid power [10] R. Shankar, K. Ashiwani, R. Utkarsh, C. Kalyan, Fruit fly algorithm-based
system is not explored thoroughly. automatic generation control of multiarea interconnected power system with

15
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

FACTS and AC/DC links in deregulated power environment, Int. Trans. Electr. [40] S. Singh, R. Verma, A. Shakya, Frequency regulation of microgrid connected
Energy Syst. 29 (2018) 2690–2715. hybrid power system with SMES, Technol. Econ. Smart Grids Sustain. Energy 2
[11] R. Arun, N.P. Patidar, Reliability assessment using probabilistic modelling of (13) (2017) 1–13.
pumped storage hydro plant with PV-Wind based standalone microgrid, Int. J. [41] X. Lin, et al., Coordinated Control Strategies For SMES-Battery Hybrid Energy
Electr. Power Energy Syst. 106 (2019) 17–32. Storage Systems, IEEE, 2017. Volume 5.
[12] T. Diambomba Hyacinthe, Y. Sun, Load frequency controllers considering [42] F. Katiraei, et al., Micro-grid autonomous operation during and subsequent to
renewable energy integration in power system, Energy Rep. 5 (2019) 436–454. islanding process, IEEE Trans. Power Deliv. 20 (1) (2005) 248–257.
[13] IEEE Standard for the specification of microgrid controllers, IEEE Std. 2030. 7, [43] N. John, K. Ramesh, An overview of load frequency control strategies: a literature
2018. survey, Int. J. Eng. 1 (10) (2012).
[14] IEC TS 62898-3-1:2020 Microgrids - Part 3-1: technical requirements - protection [44] H. Bevrani, Y. Mitani, K. Tsuji, Bilateral based robust load frequency control,
and dynamic control and real implementations, 2020. Energy Convers. Manag. 46 (7–8) (2005) 1129–1146, vol.
[15] A. Sirisha, K. Radha Rani, S. Naick, Automatic generation control of three area [45] H. Shayeghi, H. Shayanfar, Application of ANN technique based on μ-synthesis to
power system with artificial intelligent controllers, Int. J. Adv. Res. Electr. load frequency control of interconnected power system, Int. J. Electr. Power
Electron. Instrum. Eng. 6 (8) (2017), 6576-6591. Energy Syst. 28 (7) (2006) 503–511.
[16] A. Rathore, N.P. Patidar, Interval type-2 fuzzy logic based determination of [46] I. Pan, S. Das, Fractional order AGC for distributed energy resources using robust
battery charging efficiency applied to standalone hybrid power system, J. Intell. optimization, IEEE Trans. Smart Grid 7 (5) (2015) 2175–2186.
Fuzzy Syst. 35 (2018) 1–12. [47] T.S. Bhatti, et al., Load frequency control of isolated wind diesel hybrid power
[17] W. Zhang, K. Fang, Controlling active power of wind farms to participate in load systems, Energy Convers. Manag. 38 (9) (1997) 829–837. VolPp.
frequency control of power systems, IET Gener. Transm. Distrib. 11 (9) (2017) [48] P.K. Pandey, Automatic generation control of multi-area power system with wind
3194–3203. power generation, Int. J. Sci. Eng. Technol. 6 (2) (2014).
[18] C. Tan, et al., Multi-area automatic generation control scheme considering [49] Y. Xu, et al., Dynamic gain-tuning control (DGTC) approach for AGC with effects
frequency quality in Southwest China grid: challenges and solutions, IEEE Access of wind power, IEEE Trans. Power Syst. 31 (5) (2016) 3339–3948.
8 (2020) 199813–199828. [50] Basit, et al., Real-time impact of power balancing on power system operation with
[19] W. Long, An exploration-enhanced grey wolf optimizer to solve high-dimensional large scale integration of wind power, J. Mod. Power Syst. Clean Energy 5 (2)
numerical optimization, Eng. Appl. Artif. Intell. 68 (2018) 63–80. (2017) 202–210.
[20] Dipayan, et al., Load frequency control of interconnected power system using [51] A. Rahman, L. Saikia, N. Sinha, Automatic generation control of an
grey wolf optimization, Swarm Evol. Comput. 97–115 (2016). interconnected two-area hybrid thermal system considering dish-stirling solar
[21] A. Daraz, M. Abdullah, H. Altaf, Review of automatic generation control for thermal and wind turbine system, J. Renew. Energy 105 (2017) 41–54.
multi-source interconnected power system under deregulated environment, in: [52] Y. Arya, N. Kumar, BFOA-scaled fractional order fuzzy PID controller applied to
Proceedings of the 4rth International Conference on Power Generation System AGC of multi-area multi-source electric power generating systems, Swarm Evol.
and Renewable Energy Technology (PGSRET), 2018. Comput. 32 (2017) 202–218.
[22] G. Dipayan, R. Provas, B. Subrata, Load frequency control of large scale power [53] P. Ram, A.N. Jha, Automatic generation control of interconnected hydro-thermal
system using quasi-oppositional grey wolf optimization algorithm, Eng. Sci. system in deregulated environment considering generation rate constraints, in:
Technol. Int. J. 19 (2016) 1693–1713. Proceedings of the International Conference on Industrial Electronics, Control
[23] O. Singh, I. Nasiruddin, Hybrid evolutionary algorithm based fuzzy logic and Robotics, 2010, pp. 148–159.
controller for automatic generation control of power systems with governor dead [54] Y. Zou, et al., Constrained predictive control synthesis for quantized systems with
band non-linearity, Cogent Eng. 3 (1) (2016) 1–16. mark ovian data loss, Automatica 55 (2017) 217–225, volume.CCA.
[24] S. Bhongade, R. Verma, Grey wolf optimized PI controller for hybrid power [55] S.K. Pandey, et al., A literature survey on load frequency control for conventional
system using SMES, in: Proceedings of the Intelligent Computing Techniques for and distribution generation power systems, Renew. Sustain. Energy Rev. 25
Smart Energy Systems, Lecture Notes in Electrical Engineering 607, Springer, (2013) 318–334, volume.
2019. [56] V. Donde, M.A. Pai, I.A. Hiskens, Simulation and optimization in an AGC system
[25] Y. Mi, et al., Decentralized sliding mode load frequency control for multi-area after deregulation, IEEE Trans. Power Syst. 16 (3) (2001) 481–489, volume.
power systems, IEEE Trans. Power Syst. 28 (4) (2013) 4301–4309. [57] L. Rahman, et al., Load frequency control of a hydro-thermal system under
[26] M. Prathibha, M. Bhavani, Automatic generation control in restructured power deregulated environment using biogeography-based optimized three degree-of-
system with wind integrated system, in: Proceedings of the International freedom integral-derivative controller, IET Gener. Transm. Distrib. 9 (15) (2015)
Conference on Innovations in Engineering and Technology (ICIET’14), March 2284–2293, volume.
2014. Vol. 3, Spcl.. Issue 3. [58] T.K. Panigrahi, et al., Novel approach to Automatic Generation Control with
[27] S. Singh, et al., Operation and control of a hybrid wind-diesel-battery energy various Non-linearities using 2- degree-of-freedom PID controller, Energy
system connected to micro-grid, in: Proceedings of the International Conference Procedia 138 (2017) 464–469.
on Robotics, Automation, Control and Embedded Systems, 2014. [59] K.C. Sekha, K. Vaisakh, Application of VSC/SMC controllers for automatic
[28] S. Singh, et al., Frequency regulation of isolated hybrid Wind/Diesel, power generation control in deregulated environment, in: Proceedings of the 7th
generation with fuel cell system, in: Proceedings of the Lecture Notes in Electrical International Conference on Intelligent Systems and Control (ISCO), 2013,
Engineering, Springer, 2015, pp. 853–862, 326. pp. 210–218.
[29] T. Chaiyatham et al. “Design of optimal fuzzy logic-PID controller using bee [60] A. Pappachen, A.P. Fathima, NERC’s control performance standards based load
colony optimization for frequency control in an isolated wind-diesel system” IEEE frequency controller for a multi area deregulated power system with ANFIS
T&D Asia 2009. approach, Ain Shams Eng. J. 9 (4) (2017) 2399–2414.
[30] A. Shayanfar, H. Shayeghi, A. Younesi, Optimal PID controller design using Krill [61] P.K. Hota, B. Mohanty, Automatic generation control of multi source power
Herd algorithm for frequency stabilizing in an isolated wind-diesel system, in: generation under deregulated environment, Int. J. Electr. Power Energy Syst. 75
Proceedings of the International Conference on Artificial Intelligence (ICAI), (2016) 205–214.
2015. [62] T.H. Mohamed, et al., Decentralized model predictive based load frequency
[31] P.C. Sekhar, S. Mishra, R. Sharma, Data analytics based neuro-fuzzy controller for controlin an interconnected power system, Energy Convers. Manag. 52 (2) (2011)
diesel-photovoltaic hybrid AC microgrid, IET Gener. Transm. Distrib. 9 (2) (2015) 1208–1214.
193–207. Iss. [63] M. Ma, et al., Distributed model predictive load frequency control of the multi-
[32] S. Padhy, et al., A modified GWO technique based cascade PI-PD controller for area power system after deregulation, IEEE Trans. Ind. Electron. 64 (6) (May
AGC of power systems in presence of Plug in electric vehicles, Eng. Sci. Technol. 2016) 5129–5139.
Int. J. 20 (2) (2017) 427–442. [64] S. Sajadian, et al., Model predictive-based maximum power point tracking for
[33] I. Miranda, et al., A Holistic approach to the integration of battery energy storage grid-tied photovoltaic applications using a Z-source inverter, IEEE Trans. Power
systems in island electric grids with high wind penetration, IEEE Trans. Sustain. Electron. 31 (11) (2016) 7611–7620.
Energy 7 (2) (2015) 775–785. [65] F. Oldewurtel, et al., Use of model predictive control and weather forecasts for
[34] N. Dattatraya, et al., Performance study of roof top wind solar microgrid system energy efficient building climate control, Energy Build. 45 (2012) 15–27.
in isolated mode of operation, in: Proceedings of the IEEE International [66] C.K. Shiva, V. Mukherje, Design and analysis of multi-source multi-area
Conference on Power Electronics, Drives and Energy Systems (PEDES), 2014. deregulated power system for automatic generation control using quasi-
[35] V. Lazarov, et al., Hybrid power systems with renewable energy sources – types, oppositional harmony search algorithm, Electr. Power Energy Syst. 80 (2016)
structures, trends for research and development, in: Proceedings of the Eleventh 382–395.
International Conference оn Electrical Machines, Drives аnd Power Systems, [67] S. Mishra, et al., Improving the frequency response of a microgrid using an
2005. auxiliary signal in the AGC loop, IFAC-Papers Online 48-30 (2015) 300–305.
[36] J. Shen, et al., Controllable load management approaches in smart grids, Energies [68] A. Jahangir, et al., Control topology of hybrid energy storage system for AC-DC
8 (10) (2015) 11187–11202. microgrid, in: Proceedings of the IEEE International Conference on Power
[37] D. Maiti, A. Acharya, M. Chakraborty, A. Konar, R. Janarthana, Tuning PID and Electronics, Drives and Energy Systems (PEDES), 2018.
PIλDδ Controllers using the Integral Time Absolute Error Criterion, IEEE, 2008. [69] D. Rerkpreedapong, A. Feliache, Decentralized load frequency control for load
[38] A. Khamis, et al., Voltage and frequency control of microgrid systems with following services, in: Proceedings of the IEEE Winter Meeting Power Engineering
demand response, Int. J. Simul. Syst. Sci. Technol. 17 (41) (2017) 25.1–25.6. Society 2, 2002, pp. 1252–1257.
[39] S. Singh, et al., Operation and Control of a Hybrid Wind-Diesel-Battery Energy [70] D. Rerkpreedapong, et al., Robust load frequency control using genetic algorithms
System Connected to Micro-Grid, IEEE, 2013. and linear matrix inequalities, IEEE Trans. Power Syst. 18 (2) (2003) 855–861.
Volume.

16
R. Verma et al. Electric Power Systems Research 226 (2024) 109945

[71] A. Demiroren, H.L. Zeynelgil, GA application to optimization of AGC in three area [97] L. Xi, et al., Automatic generation control based on multiple neural networks with
power system after deregulation, Electr. Power Energy Syst. 29 (3) (2007) actor-critic strategy, IEEE Trans. Neural Netw. Learn. Syst. 32 (6) (2021)
230–240, volume. 2483–2493.
[72] H. Hosseini, A.H. Etemadi, Adaptive neuro-fuzzy inference system based [98] D. Li, et al., Neural network controller design for a class of nonlinear delayed
automatic generation control, Electr. Power Syst. Res. 78 (7) (2008) 1230–1239, systems with time-varying fullstate constraints, IEEE Trans. Neural Netw. Learn.
volume. Syst. 30 (9) (2019) 2625–2636. Sep.
[73] S. Debbarma, et al., AGC of a multi-area thermal system under deregulated [99] A. Oshnoei, Automatic generation control incorporating electric vehicles, Electr.
environment using a non-integer controller, Electr. Power Syst. Res. (2013) Power Compon. Syst. 0 (0) (2019) 1–13.
175–183. [100] H. Liu, et al., Real-time vehicle-to-grid control for frequency regulation with high
[74] H.P. Li, Y. Shi, Distributed model predictive control of constrained nonlinear frequency regulating signal, Prot. Control Mod. Power Syst. 3 (2018) 13.
systems with communication delays, Syst. Control Lett. 62 (10) (2013) 819–826, [101] K. Jagatheesan, et al., Design of a proportional-integral-derivative controller for
volume. an automatic generation control of multi-area power thermal systems using firefly
[75] X. Zhang, et al., Bi-objective optimization of real-time AGC dispatch in A algorithm, IEEE/CAA J. Autom. Sin. 6 (2) (2019) 503–515.
performance-based frequency regulation market, CSEE J. Power Energy Syst. 21 [102] S. Swarup, Automatic generation control with competing GENCOs-a
(2020) 1–16. reinforcement learning based approach, in: Proceedings of the 20th National
[76] M. Ali, et al., Design of cascaded PI-fractional order PID controller for improving Power Systems Conference, NPSC, 2018.
the frequency response of hybrid microgrid system using gorilla troops optimizer, [103] I.H. Sarker, Machine learning: algorithms, real-world applications and research
IEEE Access 9 (2021) 150715–150732. directions, SN Comput. Sci. 2 (2021) 160.
[77] O. Singh, I. Nasiruddin, Optimal AGC regulator for multi-area interconnected [104] S. Sahoo, et al., Resilient synchronization strategy for AC microgrids under cyber
power systems with parallel AC/DC links, Cogent Eng. 3 (1) (2018), 1209272. attacks, IEEE Trans. Power Electron. 36 (1) (2020) 73–77.
[78] N.K. Jena, et al., Fuzzy adaptive selfish herd optimization based optimal sliding [105] S. Sahoo, et al., Distributed screening of hijacking attacks in DC microgrids, IEEE
mode controller for frequency stability enhancement of a microgrid, Eng. Sci. Trans. Power Electron. 35 (7) (2020) 7574–7582.
Technol. Int. J. 33 (2022) 101071. [106] A. Saxena, R. Shankar, S.K. Parida, R. Kumar, Demand response based optimally
[79] N.R. Babu, L.C. Saikia, AGC of a multiarea system incorporating accurate HVDC enhanced linear active disturbance rejection controller for frequency regulation
and precise wind turbine systems, Int. Trans. Electr. Energy Syst. 30 (2019) in smart grid environment, IEEE Trans. Ind. Appl. 58 (4) (2022) 4337–4349.
e12277. [107] S. Bhuyan, S.H. Nee Dey, S. Paul, Role of demand side management in automatic
[80] P. Chandra Nayak, et al., Grasshopper optimization algorithm optimized load frequency control, in: Proceedings of the International Conference on
multistage controller for automatic generation control of a power system with Emerging Frontiers in Electrical and Electronic Technologies (ICEFEET), Patna,
FACTS devices, Prot. Control Mod. Power Syst. 6 (8) (2021) 1–15. India, 2020.
[81] R.B. Naladi, L.C. Saikia, Load frequency control of a multi-area system [108] C. Albea, C. Bordons, M.A. Ridao, Robust hybrid control for demand side
incorporating realistic high-voltage direct current and dish-Stirling solar thermal management in islanded microgrids, IEEE Trans. Smart Grid 12 (6) (2021)
system models under deregulated scenario, IET Renew. Power Gener 15 (5) 4865–4875.
(2021) 1116–1132. [109] K. Peddakapu, M.R. Mohamed, P. Srinivasarao, Y. Arya, P.K. Leung, D.J.
[82] Y. Li, et al., Primary frequency control scheme for a fixed-speed dish-Stirling K. Kishore, State-of-the-art review on modern and future developments of AGC/
solar–thermal power plant, IEEE Trans. Power Syst. 33 (2) (2019) 2184–2194. LFC of conventional and renewable energy-based power systems, Renew. Energy
[83] T. Washima, et al., Deregulated AGC of multi-area system incorporating dish- Focus 43 (2022) 146–171.
Stirling solar thermal and geothermal power plants using fractional order cascade [110] B. Amar Kumar, S. Jaiswal, D.C. Das, Recent trends and development in hybrid
controller, Electr. Power Energy Syst. 101 (2018) 60–74. microgrid: a review on energy resource planning and control, Int. J. Sustain.
[84] D. Khamari, R.K. Sahu, T.S. Gorripotu, S. Panda, Automatic generation control of Energy 41 (4) (2022) 308–322. VOL.NO.
power system in deregulated environment using hybrid TLBO and pattern search [111] S.C. Sahoo, A.K. Barik, D.C. Das, Synchronized voltage-frequency regulation in
technique, Ain Shams Eng. J. 11 (3) (2020) 553–573. sustainable microgrid using novel Green Leaf-hopper Flame optimization,
[85] A. Abazari, et al., Coordination strategies of distributed energy resources Sustain. Energy Technol. Assess. 52 (2022), 102349. Elsevier.
including FESS, DEG, FC and WTG in load frequency control (LFC) scheme of [112] M.K. Behera, L.C. Saikia, Combined voltage and frequency control for diverse
hybrid isolated micro-grid, Int. J. Electr. Power Energy Syst. 109 (2019) 535–547. standalone microgrid networks using flexible IDC with novel FOC: a real-time
[86] M. Dong, L. Li, et al., A distributed secondary control algorithm for automatic validation, IETE J. Res. 69 (7) (2021) 1–26.
generation control considering EDP and automatic voltage control in an AC [113] Y. Singh, B. Singh, S. Mishra, Control strategy for multiple residential solar pv
microgrid, Energies 11 (2018) 932, 2018. system in distribution network with improved power quality, IEEE Trans. Ind.
[87] P.K. Pathak, A.K. Yadav, Design of battery charging circuit through intelligent Appl. 59 (3) (2023) 3686–3699.
MPPT using SPV system, Sol. Energy 178 (2019) 79–89. [114] T. Huang, D. Wu, M. Ili´c, Cyber-resilient automatic generation control for systems
[88] M. Ali, et al., Frequency regulation of hybrid multi-area power system using wild of AC microgrids, IEEE Trans. Smart Grid 1 (2023) 1949–3061.
horse optimizer based new combined fuzzy fractional-order PI and TID [115] Z. Li, Z. Cheng, J. Si, S. Xu, Distributed event-triggered hierarchical control of PV
controllers, AEJ Alex. Eng. J. 61 (2022) 12187–12210. inverters to provide multi-time scale frequency response for AC microgrid, IEEE
[89] M. Ramesh, et al., Intelligent adaptive LFC via power flow management of Trans. Power Syst. 38 (2) (2023) 1529–1542.
integrated stand alone micro-grid system, ISA Trans. 112 (2021) 234–250. [116] S.R. Bhupathiraju, M. Sundararajan, K.R. Pedakota, D.P. Chinta, Frequency
[90] A. Bagheri, S. Jabbari, Mobayen, An intelligent ABC-based terminal sliding mode control of electric vehicles integrated isolated renewable microgrid using a
controller for load-frequency control of islanded micro-grids, Sustain. Cities Soc. magnetotactic bacteria optimized cascade controller, IETE J. Res. 22 (2023) 1–10.
64 (2021), 102544. [117] R. Abishek, D.C. Das, A.K. Srivastava, A. Latif, Frequency control strategy for grid-
[91] A.T. Tran, et al., Load frequency regulator in interconnected power system using tied virtual power plant using SSA-tuned fractional order PID controller, IETE J.
second-order sliding mode control combined with state estimator, Energies 14 (4) Res. 22 (2023) 1–17.
(2021) 863. [118] M. Camila Merchán-Riveros, C. Albea, Hybrid dynamical control for discharging
[92] I. Saboya, et al., Machine learning based algorithms to dispatch multiple rapid- rate consensus in AC-bus microgrids, IEEE Control Syst. Lett. 7 (2023)
start units in AGC of power systems, Electr. Power Energy Syst. 115 (2020), 2323–2328.
105412. [119] W. Feng, C. Xuebing, C. Shuyu, S.C. Beng, C. Zhiqiang, T.K. Jet, D. Mahinda
[93] C. Feng, et al., Automatic generation control of isolated two-area microgrid with Vilathgamuwa, MPC based dynamic voltage regulation using grid-side BESPS
multiple resources based on learning optimization, in: Proceedings of the IEEE with the consideration of communication delay, IEEE Trans. Energy Convers. 38
9th Data Driven Control and Learning Systems Conference, 2020, pp. 20–22. (2) (2023) 838–848.
[94] B. Kiumarsi, et al., Optimal and autonomous control using reinforcement [120] M. Bhuyan, D.C. Das, A.K. Barik, Chaotic butterfly optimization algorithm based
learning: a survey, IEEE Trans. Neural Netw. Learn. Syst. 29 (6) (Jun. 2018) cascaded PI-TID controller for frequency control in three area hybrid microgrid
2042–2062. system, Optim. Control Appl. Methods 44 (5) (2023) 2595–2619.
[95] J. Pan, et al., Multisource transfer double DQN based on actor learning, IEEE [121] W. Jiao, J. Chen, Q. Wu, C. Li, B. Zhou, S. Huang, Distributed coordinated voltage
Trans. Neural Netw. Learn. Syst. 29 (6) (2018) 2227–2238. control for distribution networks with DG and OLTC based on MPC and gradient
[96] L. Xi, Lu Yu, Y. Xu, S. Wang, Xi Chen, A novel multi-agent DDQN-AD method- projection, IEEE Trans. Power Syst. 37 (1) (2022) 680–690.
based distributed strategy for automatic generation control of integrated energy
systems, IEEE Trans. Sustain. Energy 11 (4) (2020) 2417–2426.

17