Electrical Engineering

https://doi.org/10.1007/s00202-024-02551-0

ORIGINAL PAPER

Design of cascade P-P-FOPID controller based on marine predators

algorithm for load frequency control of electric power systems
Jahanzeab Hussain1 · Runmin Zou1 · Samina Akhtar1 · Khalid A. Abouda1

Received: 4 January 2024 / Accepted: 17 June 2024

© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024

Abstract
As power network complexity increases, efficient control of frequency deviations and tie-line power fluctuations becomes
crucial. Load frequency control (LFC) addresses these issues by managing generation-consumption mismatches. This paper
introduces a unique P-P-FOPID cascade controller for LFC, combining a proportional controller with a fractional-order PID.
The marine predators algorithm (MPA), known for its benefits like being parameter-less, derivative-free, user-friendly, flexible,
and simple, is employed to optimize the controller’s parameters using the integral time absolute error criterion. Tested on three
power systems, the proposed controller outperforms single-structure FOPID controller based on MPA and recent approaches in
reducing ITAE, settling time, and frequency and tie-line power deviations. In comparative analysis, the proposed MPA-tuned
P-P-FOPID controller achieves a 60% and 11% decrease in ITAE compared to the second-best DSA-optimized FOPI-FOPD
controller for Systems 1 and 2, respectively, and a 25% decrease compared to the second-best MPA-tuned FOPID controller
for System 3. Furthermore, the controller’s robustness against parametric uncertainties is confirmed through variations in
fundamental parameters.

Keywords Load frequency control · Interconnected power systems · Fractional-order PID controller · Cascade controller ·
Marine predators algorithm

List of symbols 1 Introduction

f Frequency deviation In order to guarantee a reliable power supply to con-

Ts Settling time sumers, maintaining a balance between power generation
Ptie Tie-line power deviation and consumption is necessary. In modern energy manage-
ACEi Area control error ment systems, one crucial issue is regulating the actual power
Pref Incremental change in controller output output of generation units in response to fluctuations in sys-
Kp Proportional gain of controller tem frequency and power exchange between interconnected
Kd Derivative gain of controller power units. Any unfavorable disparities in power produc-
Ki Integral gain of controller tion and load demands resulting from unforeseen disruptions
λ Order of integration or load changes may cause system frequency and tie-line
μ Order of derivation power flow to diverge from their intended thresholds. The
control mechanism designed to rectify these inconsistencies
and keep the system within specified limits is known as load
frequency control (LFC). LFC is a widely studied subject by
researchers, who have proposed various control schemes to
keep system frequency and tie-line power within acceptable
limits. A critical literature review on LFC is available in [1].
B Jahanzeab Hussain The literature reveals that various controller designs based on
[email protected] modern control theory have been reported to improve load
frequency control (LFC). These include fuzzy logic theory [2,
1 School of Automation, Central South University, Changsha
3], H-infinity control [4], artificial neural networks (ANNs)
410017, Hunan, China

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Electrical Engineering

[5], adaptive neuro-fuzzy inference systems (ANFIS) [6, 7], resilience, robustness, and flexibility when compared to sin-
and adaptive integral second-order sliding mode control [8]. gle controllers. A cascade controller has a higher number of
Although, these strategies improve the system’s steady-state tuning parameters than a single controller. However, with
performance to some extent, they pose challenges in terms this extended set of tuning knobs within its structure, it
of design complexity and high computational costs [9]. offers the potential of producing better results. Consequently,
Conventional controllers like integral (I), Proportional- cascade controllers have been used to improve the per-
Integral (PI), Proportional-Integral-Derivative (PID), and formance of interconnected power systems by effectively
other similar controllers were widely used in LFC design dealing with load frequency control (LFC) issues. Various
due to their simple structure, low development requirements, cascade strategies have been proposed for LFC in different
and several other advantages [10]. However, in modern inter- power systems, including an improved GWO algorithm-
connected power systems, the complicated network structure based cascade PI–PD controller [23], hybrid stochastic frac-
and complex module expressions make empirical parameter tal search (SFS) and local unimodal sampling (LUS)-based
tuning methods ineffective, especially when using a con- PID controller [24], hybrid stochastic fractal search (hSFS)
ventional controller-based LFC scheme [11]. Furthermore, and pattern search (PS)-based PI–PD cascade controller
using conventional PI/PID controller-based LFC design and [25], DSA-optimized (1 + PD)-PID cascade controller [26],
empirical parameter tuning methods may reduce the system’s DSA technique-based fractional-order PI–fractional-order
robustness and performance when dealing with parametric PD cascade controller [27], novel chaos game optimization-
uncertainties and highly variable loads. In response to these tuned fractional-order PID fractional-order PI controller
challenges, researchers have developed various parameter [28], weighted geometric center method-tuned PI–PD con-
tuning and optimization techniques to enhance conventional troller [29], Jaya algorithm-tuned fractional-order fuzzy PID-
controller-based load frequency control (LFC). These opti- integral double derivative (CFOFPID-IDD) controller [30],
mization techniques include genetic algorithms (GA) [12], and black widow optimization algorithm (BWOA)-tuned
Improved stochastic fractal search algorithm [13], Whale PIDF-(1 + I) cascade controller [31]. These cascade con-
optimization algorithm [14], big bang big crunch optimiza- trollers exhibit promising control results in LFC. However,
tion [15], gray wolf optimizer [16], sea horse optimizer [17], the authors believe that there is still room for improve-
Modified Harris Hawks Optimization [18], and so on. ment in the outcomes of these works, particularly concerning
In recent years, researchers have made various modifica- time domain performance metrics such as undershoot, over-
tions to the PID controller, and a new type of PID controller shoot, settling time, and system performance metrics like
known as the fractional-order PID (FOPID) controller has ITAE. Thus, this study introduces a novel approach combin-
been introduced in load frequency control (LFC) applica- ing a proportional (P) controller with a fractional-order PID
tions. Fractional-order PID (FOPID) controllers offer more (FOPID) in a cascade setup, which has not been explored in
computational power compared to conventional methods. LFC so far. The parameters of this controller are tuned using
This is due to their additional design parameters, namely a robust optimizer known as the marine predators algorithm,
λ and μ, alongside the standard Kp, Ki, and Kd gain val- which has demonstrated superior performance compared to
ues. These additional parameters broaden the search space many other algorithms [32].
for controller gain values, leading to improved performance. The marine predator algorithm (MPA) is a well-known
Various studies have investigated the effectiveness of FOPID meta-heuristic optimization algorithm developed and intro-
controllers in addressing LFC issues in interconnected power duced by Faramazi and his colleagues in 2020 [33]. MPA has
systems. For instance, in [19], an LFC strategy employ- superior attributes such as being derivative-free, parameter-
ing FO controllers was introduced for interconnected power free, user-friendly, adaptable, straightforward, sound, and
systems. Furthermore, [20] proposed a FOPID structure for comprehensive. Furthermore, it is capable of producing bet-
LFC in interconnected power systems with renewable energy ter outcomes than many other algorithms [32]. Therefore,
sources. The parameters of this FOPID controller were tuned MPA has found diverse applications in various fields, includ-
using the jellyfish search optimizer (JSO) technique. Addi- ing COVID-19 image processing [34], the estimation of
tionally, [21] described an adaptive FO-Fuzzy-PID controller photovoltaic (PV) solar cell parameters [35], and the opti-
optimized with the TLBO algorithm. The robustness of mization of power generation in various photovoltaic solar
FOPID controllers was further investigated in [22] when elec- cells (PSCs) [36]. Additionally, MPA has recently been
tric vehicles were introduced into microgrid systems. The applied to estimate controller gains in load frequency con-
parameters in this case were fine-tuned using the slap swarm trol (LFC) of electric power systems [37, 38]. The marine
algorithm. predator algorithm’s (MPA) extraordinary achievements in
Nowadays, the concept of cascade control is gaining the past are a blatant demonstration of its effectiveness.
popularity due to its improved performance in terms of

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Electrical Engineering

Collectively, these findings provide strong motivation for of participation factors for all contributing generators within
developing a novel cascade controller known as the (P-P- a control area must be equal to one. In Fig. 1, R1 , R2 , and R3
FOPID) cascade controller. This controller makes use of are regulatory parameters; K T , K H , and K G are participa-
the remarkable capability of the MPA technique for pre- tion factors; and UT , U H , and UG are the control outputs of
cise adjustment of its parameters, such as proportional gain thermal, hydro, and gas-generating units, respectively. Sim-
(K p ), integral gain (K i ), derivative gain (K d ), differentiator ilarly, Tgs is the time constant of the speed governor of the
order (μ), and integral order (λ). The objective of this adjust- thermal unit, Tt is the time constant of the steam turbine,
ment is to minimize the integral time absolute error (ITAE) and Tr and K r are the reheat time constant and reheat gain,
criterion. The MPA-tuned cascade controller (P-P-FOPID) respectively, Tr h is the transient droop time constant of hydro
uses two error signals, namely the area control error (ACE) turbine speed governor, Tr s is the rest time, Tgh is the main
and the area frequency ( f ), to mitigate disturbances. It dis- servo time constant of speed governor of the hydro turbine,
plays superior performance compared to modern controllers Tw is the nominal starting time of water in the penstock; X c
described in the literature. Finally, the significance of this and Yc are lead time and leg time constants of speed governor
work can be summarized as follows: of gas turbine, C g is valve positioner of gas turbine, bg is the
constant of the valve positioner, Tcd , T f , and Tcr are com-
A simple yet efficient cascade (P-P-FOPID) controller is con- pressor discharge volume-time constant, fuel time constant
structed to address the LFC issue. and combustion reaction time delay of gas turbine, respec-
The robustness of the MPA-tuned (P-P-FOPID) controller is tively, T ps and K ps are the time constant and gain of power
investigated on three different power systems. system in sec,  f is the deviation in frequency and P D is
Finally, a thorough comparative analysis is conducted to the incremental load change.
demonstrate the superior performance of the proposed (P-
P-FOPID) controller over existing works in the literature. 2.2 Two-area multi-source power system

The rest of the paper are organized as follows: Sect. 2 pro- To assess the performance of the proposed controller in more
vides an exposition of the power systems being investigated. realistic conditions, we subjected it to a test in a two-area
Section 3 describes the architecture and the optimization multi-source system as shown in Fig. 2. This system contains
procedure for the proposed controller. An overview of the two identical multi-source single-area power systems inter-
MPA technique is presented in Sect. 4. Section 5 provides connected through an AC–DC tie-line to share the load during
a comparative analysis, demonstrating the performance of transient state. Therefore, the explanations provided for pre-
the suggested controller compared to existing approaches. vious systems apply equally to this system. The optimization
Finally, Sect. 6 presents the summary of this work. process becomes straightforward and economical as a result
of the similarity between the two regions, which obviates
the need for precise calibration of six controllers per gener-
2 Power systems under investigation ating unit. This particular power plant is referred to as Test
system-2. The Appendix provides the nominal parameters
The efficiency and robustness of the proposed cascade (P- of system-1, system-2, and system-3, which were obtained
P-FOPID) controller have been verified through extensive from the previous paper.
testing on various power systems. These systems include
a single-area multi-source power system, a two-area multi- 2.3 Three-area hydrothermal power system
source power system, and a three-area hydrothermal power
system. Finally, the ability of the proposed controller has been tested
by increasing the number of areas in the power system.
2.1 multi-source single-area power system The power system under examination, named Test system-
3, comprises three equal-sized generating zones. Thermal
Figure 1 represents the transfer function model of a multi- reheat turbine units are found in the first and second gen-
source single-area power system. In this study, this particular erating zones, whereas hydro-generator units are found in
power plant is referred to as Test system-1. This system is the third zone. In all three areas, the proposed (P-P-FOPID)
made up of a single isolated region and three independently cascade controller has been employed. However, due to the
controlled generating units. Along with a hydropower and similarity of the first two thermal areas, the constants of both
gas power unit, the first unit is a thermal power unit with controllers for these areas are set to be identical and different
a reheat turbine. Each generating unit of this power system from those in the third area.
has its regulatory parameters and participation factors that The power production rate in a power system is subject
determine their contribution to overall generation. The sum to physical limitations imposed by system dynamics and

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Electrical Engineering

Fig. 1 Transfer function model of the single-area multi-source system with a cascade controller

Fig. 2 A transfer function model of a two-area six-unit power system

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Electrical Engineering

Fig. 3 A model of three areas hydrothermal power system

mechanics. The generation rate constraint (GRC) is used to applications and research. Significant advantages of FOPID
estimate the rate of generation for all areas for this purpose. controllers include their flexible design, quick response,
For a hydroelectric plant, the standard range of GRC has val- robust stability in a variety of operating conditions, and
ues of 270%/min for boosting generation and 360%/min for exceptional resilience to disturbances. As a result, many
decreasing generation. On the other hand, the GRC range is researchers have investigated several FOPID controller appli-
3%/min for thermal plants (Fig. 3). Therefore, for the nth cations such as heat flow processes [39], trajectory control
thermal system, GRC is given by: [40], twin-rotor helicopter [41], and LFC of diverse electric
power systems. The FOPID that is referred to as P I λ D μ con-
|PGi (s)| ≤ ±0.0005 p.u M W /s troller consists of an integrator with a fractional-order λ and
a differentiator with an order μ, yielding the fractional-order
GRC set for the hydro area is transfer function given as follows:

|PG (s)| ≤ 0.045 p.u M W /s For raising Ki

G FOPID (s)  K p + + Kd sμ (1)
sλ
|PG (s)| ≤ 0.06 p.u M W /s For Lowering
where K p ,K i , and K d are the gain parameters of the P, I, and
D controller, respectively, whereas μ and λ are the order of
differentiation and integration. When λ and μ both are equal
3 The proposed controller to one, the FOPID controller behaves as a traditional PID
controller. Similarly, if λ  0 and μ  1, it acts as a P D μ
3.1 Fractional-order PID (FOPID) controller controller, and if λ  1 and μ  0, it acts as a P I λ controller.
A popular approach for generating fractional-order transfer
The fractional-order PID (FOPID) controller, with its unique functions in practical applications and computer simulation is
integer order and extra two degrees of freedom, is widely rec- to compute fractional-order transfer functions using integer-
ognized as a remarkably efficient controller in both industrial based functions for transfer. The Oustaloup filter given by

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Electrical Engineering

Fig. 4 Structure of proposed

(P-P-FOPID) cascade controller

Eq. 2 is commonly utilized to approximate the fractional potential to surpass the performance of existing controllers.
derivative or integral s α by a rational transfer function within The structure of the (P-P-FOPID) cascade controller is illus-
a chosen frequency range [ω h , ωl ], where ωh is the higher and trated in Fig. 4.
ωl is the lower frequency limit of the approximate rotational As shown in Fig. 4, this controller uses two signals, area
transfer function, as described as follows [42]: frequency deviation  f and area control error (ACE). The
ACE of two-area power systems can be given as follows:

N
S + ωkz
α
G f (s)  s  K p (2) ACE  −B f + Ptie (3)
K −N
S + ωk

p where B is the frequency response of the parameter,  f

where K represents the gain of the filter, while ωkz and ωk frequency deviation, and Ptie is the tie-line power deviation.
denote the zeros and poles of the G f (s), and N is the order of The transfer function in the s-domain for the (P-P-FOPID)
approximation, respectively. The performance of this approx- cascade controller can be expressed as follows:
imation is significantly influenced by the selected values of
k and N . The value of K should be adjusted such that at a
G p1  K p1 (4)
frequency of 1 rad/s, the gain of the approximation equals
one. Similarly, the factor N also requires precise adjustment,
as smaller values of N simplify the approximation and make G p2  K p2 (5)
it easier to implement, but they also lead to decreased per-
formance characterized by prominent ripples in both phase Ki
and magnitude response. On the other hand, higher values G FOPID  K p3 + + Kd sμ (6)
sλ
of N could be able to solve the problem of high ripples, but
at the expense of making the approximation more complex where K p1 , K p2 , K p3 , K d , K i , λ, and μ are seven unknown
and making it more challenging to implement in real life. In parameters of the cascade controller that need to be opti-
this research, we have selected the order of approximation to mized. This proposed controller has the advantage of having
N  5 and the fitting frequency range within [10–3 , 103 ]. seven unknown parameters, which allows us to fine-tune
its parameters to increase system performance. This study
3.2 (P-P-FOPID) cascade controller structure adopted an MPA optimization approach to fine-tune these
parameters, aiming to enhance the overall performance of
LFC is a critical issue in power systems, and this paper the system. ACE multiplied by K p1 to produce u and then
introduces a novel approach in the form of a (P-P-FOPID) we subtract  f × K p2 from u forming the reference signal
cascade controller to advance LFC performance in power sys- for fractional-order PID (FOPID) controller, and finally we
tems. This combines the merits of a classical proportional obtained control signal Pref from the output of P-P-FOPID
controller and a fractional-order PID controller to create a controller for power system, which can be mathematically
novel cascade controller. The cascade controllers in the lit- given as follows:
erature have obtained favorable results, but they still need
Ki
to develop a classic controller to enhance LFC performance. Pref  ({AC E × K p1 } − { f × K p2 }) × K p3 + + Kd sμ
After extensive experimentation with various controller com- sλ
binations, we finally discovered this (P-P-FOPID) cascade (7)
controller that is simple and easy to implement and has the

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Electrical Engineering

3.3 Optimization of P-P-FOPID cascade controller where min and max refer to the minimum and maximum P-
P-FOPID cascade controller gains, respectively. To ensure
3.3.1 Objective function the stability of power systems and the fruitful optimization
process, the minimum and maximum range for K p1, p2, p3, i, d
It is very important to identify the optimal values of controller gains are chosen −2 to 2 for test system 1 and test system
parameters, as the performance of a well-designed controller 2, whereas the range is of respective gains for system 3 is
can significantly deteriorate if its parameters are not appro- between 0 and 3. However, the range for μ and λ is chosen
priately tuned. Thus, to get the optimal value of parameters, between 0 and 1.5 for all power systems [27].
we need an optimization technique and an objective func-
tion. It is essential to minimize the objective function to the
greatest extent possible through the utilization of optimiza-
tion algorithms to achieve the best solution. Various objective 4 Marin predators algorithm
functions exist, including integral time absolute error (ITAE),
integral time absolute error (ITAE), integral absolute error The marine predators algorithm (MPA) is a swarm-based
(IAE), and integral squared error (ISE). In this work, we meta-heuristic optimization approach that simulates the for-
chose ITAE as an objective function for the optimization task. aging tactics of marine predators in search of prey [33].
ITAE has been used in many published works because of its Two well-known patterns of motions that MPs display when
faster response, less oscillation, and less settling time. Equa- searching for food are Levy and Brownian motion. Brow-
tion 8, 9, and 10 represent the expression of ITAE for single- nian motion is better suited for exploration, whereas the
area, two-area, and three-area power systems, respectively, Lévy motion tends toward efficient exploitation. The com-
plete MPA optimization technique is detailed as follows:
tsim
JITAE  | f |.t.dt (8)
0 4.1 Initialization
tsim
MPA involves an initialization step where it is necessary to
JITAE  | f 1 +  f 2 + Ptie |.t.dt (9)
equally distribute all populations across the search space as
0 given in Eq. (12).
tsim
JITAE  | f 1 +  f 2 +  f 3 + Ptie1 + Ptie2 + Ptie3 |.t.dt
0 X 0  u + rand(l − u) (12)
(10)

where tsim is simulation time its value should be long enough

where l and u are minimum and maximum boundaries,
for the system to reach a stable state.
whereas rand is a random vector in the range of 0 and 1.
Based on ‘survival of the fittest’ an elite matrix is created
3.3.2 Problem constraints −
→
that includes fittest solutions X I among the marine preda-
tors. The Elite matrix can be expressed as:
The problem constraints are the limits of controller param-
eters that must be guaranteed and not violated while
attempting to solve the optimization problem. Therefore, the ⎡ ⎤
X 1,I 1 X 1,I 2 · · · X 1,I d
optimization problem can be expressed as follows: ⎢ XI XI
Minimize JITAE subject to: ⎢ 2, 1 2, 2 · · · X 2,I d ⎥
⎥
⎢ . .. ⎥
⎢ . .. .. ⎥
⎧ min ⎢ . . . . ⎥
⎢ . .. ⎥ (13)
⎪
⎪ K p1 ≤ K p1 ≤ K max ⎢ . .. .. ⎥
⎪
⎪
p1
⎢ . . . . ⎥
⎪
⎪ K p2 ≤ K p2 ≤ K p2
min max ⎢ ⎥
⎪
⎪ ⎢ .. .. .. .. ⎥
⎪
⎪ ⎣ . . . . ⎦
⎪
⎪
⎨ K p3 ≤ K p3 ≤ K p3
min max
⎪ I
X n, X I · · · X n,
I
1 n, 2 d
K imin ≤ K i ≤ K imax (11)
⎪
⎪
⎪
⎪
⎪
⎪ K dmin ≤ K d ≤ K dmax
⎪
⎪ Because both predator and prey are searching for food,
⎪
⎪ λmin ≤ λ ≤ λmax
⎪
⎪ therefore both are search agents. Given that the Elite matrix
⎪
⎩ min
μ ≤ μ ≤ μmax will be modified if a new best predator is discovered, the

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Electrical Engineering

updated matrix is referred to as the prey matrix and is pro- −−→ −−→ −−→
Preyi  Preyi + P. R × Stepi (18)
vided as:
⎡ ⎤ The second half of the population
X 1, 1 X 1, 2 · · · X 1, d
⎢X ⎥
⎢ 2, 1 X 2, 2 · · · X 2, d ⎥ −−→ −→ −→ −−→ −−→ 
⎢ . .. ⎥ Stepi  R B × R B × Elitei − Preyi , i  n 2 . . . . . . . . . .n (19)
⎢ . .. .. ⎥
⎢ . . . . ⎥
⎢ . .. ⎥ (14)
⎢ . .. .. ⎥ −−→ −−→ −−→
⎢ . . . . ⎥ Preyi  Elitei + P.C F × Stepi (20)
⎢ ⎥
⎢ .. .. .. .. ⎥
⎣ . . . . ⎦ −
→
X n, 1 X n, 2 · · · X n, d where R L is a vector of random values generated by Lévy
distribution, and CF is a step size controlling parameter that
where d represents the number of problem variables and n represents predator motion. The formula of CF is given as
is the number of prey. During each phase, the two matrices  
mentioned above are updated.   2×Iter
Iter Max_Iter
CF  1 − (21)
Max_Iter
4.2 Optimization scenarios
Phase 3: In the last phase of optimization, when the preda-
The entire MPA optimization process can be divided into tor velocity is higher than the prey. This phase is mostly
three main phases based on the velocity ratio and simulating associated with high exploration capability and during this
pattern of predators and prey. The details of three phases are phase, the most effective strategy for predators is the Lévy
described as follows: movement. The equations for phase 3, where a low velocity
Phase 1: In the early stage of optimization, when the ratio is maintained, are as follows:
velocity of prey is higher than the predator, then the best strat-
egy of predators for this stage is to not change their position 2
WhileIter > Max_Iter
and remain stationary. The mathematical model of phase 1, 3
where a high-velocity ratio is maintained are as follows: −−→ − → − → −−→ −−→
Stepi  R L × R L × Elitei − Preyi , i  1, . . . . . . ., n
1
WhileIter < Max_Iter (22)
3
−−→ −−→ −−→
−−→ − → −−→ − → −−→ Preyi  Elitei + P.C F × Stepi (23)
Stepi  R B × Elitei − R B × Preyi , i  1, 2, 3, .......n
(15) To prevent getting trapped in local minima, MPA has intro-
−−→ −−→ −−→ duced a concept known as “fish aggregating effects (FAEs)”.
Preyi  Preyi + P. R × Stepi (16) FAEs represent a situation in which the predator primar-
−
→ ily focuses on hunting around the prey vortex and uses the
where R B is a vector that has random values generated by remaining time to do lengthy leaps in search of the prey. The
Gaussian distribution (GD) to identify the Brownian motion. mathematical expression of FAE is denoted as follows:
Phase 2: In the intermediate part of optimization, when
⎧ −−→  
the prey and predator move at the same velocity, this stage
−−→ ⎨ Preyi + C F X min + R × (X max − X min ) × U
−−−→ −−−→
i f r ≤ ρ
encompasses both exploration and exploitation. Therefore, Preyi  −−→
⎩ Prey + [ρ × (1 − r ) + r ] Prey − Prey i f r > ρ
half of the population is allocated for exploration and the i r1 r2

rest of the population is allocated for exploitation. Notably, (24)

the prey takes on the responsibility of exploitation, while the
predators take on the responsibility of exploration. The equa- where X min and X max denote the minimum and maximum
−−→ −−→
tions for phase 2 involve both exploration and exploitation: dimensional vectors of each population, Preyr 1 and Preyr 2
represent the random positions produced by the prey matrix,
−
→
1 2 U is a binary vector with values of 0 and 1, and ρ is the
WhileIter < Max_Iter < Max_Iter
3 3 trigger coefficient of FAE. The flowchart of MPA is shown
in Fig. 5.
For the first half of the population
For implementation of the MPA, a population size of 30
−−→ − → −−→ − → −−→  and 100 iterations are specified. Each test system undergoes
Stepi  R L × Elitei − R L × Preyi , i  1, 2, 3, .......n 2 an optimization procedure 20 times, and the final result is
(17) determined by selecting the best results obtained. Figure 6

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Electrical Engineering

Fig. 5 Flowchart of MPA

algorithm

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Electrical Engineering

settling time, objective function, and optimal gain values for

each controller unit. The controller gains, settling time and
ITAE values for the multi-source single-area power system
(PS) are presented in Table 1. This table highlights that com-
pared to various controllers such as hSFS–LUS: PID (ITAE 
0.0223), MGWO: PID (ITAE  0.0292), DSA: FOPD–FOPI
(ITAE  0.0070), and MPA: FOPID (ITAE  0.0108), the
MPA-adjusted (P-P-FOPID) cascade controller outperforms
them, exhibiting the lowest ITAE objective function with a
value of 0.00281. Additionally, it is evident from Table 1
that the proposed controller significantly enhances the set-
tling time of f, improving it by 84.78%, 57.95%, 46.87%,
and 36.702%, respectively, compared to hSFS-LUS: PID,
MGWO: PID, DSA: FOPD-FOPI, and MPA: FOPID con-
Fig. 6 Convergence curve of MPA algorithm for system-1 troller.
Figure 7 depicts the dynamic response of the proposed
controller and other methods at 1% SLP. It is clear from
displays the convergence curve of the MPA for the selected the figure that the proposed controller notably reduces both
objective function. frequency and tie-line power deviations.

5 Simulation results and discussions

5.1.2 Scenario 2: sensitivity analysis
In order to evaluate the effectiveness and superiority of the
proposed P-P-FOPID controller, we investigate three differ- Sensitivity analysis was conducted to assess the robustness
ent interconnected power system networks in this research. of the suggested MPA-optimized P-P-FOPID controller. The
The transfer-function models of these systems are con- stability of the system might be at risk if the proposed control
structed using MATLAB/SIMULINK environment, and the scheme fails to adapt to fluctuations in system parameters.
marine predators algorithm (MPA) is coded in an m-file. The parameter values of the power system time constant (TP )
The ITAE criterion is employed to optimize the parame- and speed regulator (R) were intentionally drifted by ± 30%
ters of the controllers. For each test system, two independent from their nominal values, while considering a 1% SLP, and
scenarios are developed to analyze the performance of the then compared to their nominal parameter response. Figure 8
proposed controller. In first scenario, the system response illustrates the results achieved by adjusting the system param-
is investigated for a step load perturbation (SLP) and the eters within a range of ± 30%. Examination of Fig. 8 reveals
performance of proposed controller is compared with the that the system response closely aligns with nominal values
MPA-tuned FOPID controller and well-established existing for numerous parameters. This demonstrates the robust per-
works, and in scenario 2, robustness test analysis is conducted formance of the intended MPA-tuned P-P-FOPID controller
by changing the system parameters. across a ± 30% variation in power system parameters. More-
over, the optimal values of the proposed controller eliminate
the necessity of re-adjustment the controller for a wide range
5.1 Single-area multi-source power system (Test
of parameters at nominal capacity with nominal parameters.
system 1)

5.1.1 Scenario 1: load disturbance analysis

5.2 Two-area multi-source power system (Test
To evaluate the effectiveness and efficacy of the proposed system 2)
(P-P-FOPID) cascade controller for load frequency control
(LFC) studies, we performed a performance comparison test Furthermore, to analyze the performance of our proposed (P-
against MPA-tuned FOPID controller and the recently pub- P-FOPID) controller, we have expanded our investigation to
lished studies such as [23, 24, 27], concentrating on a 1% include a more realistic two-area multi-source power system,
SLP condition in area 1. The power system model shown where a total of six generating units, with three units in each
in Fig. 1 consists of three different power plants, each of area, interconnected via an AC-DC tie line, as shown in Fig. 2.
which is linked to a (P-P-FOPID) cascade controller. The Two independent scenarios were studied for this test system
MPA optimizer is used to analyze the performance of the as well.

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Electrical Engineering

Table 1 Optimized controller parameters and Ts/ITAE results for test system-1

Algorithms MGWO: PIPD [23] hSFS-LUS: PID [24] DSA: FOPI-FOPD [27] MPA: FOPID MPA: P-P-FOPID

Controller Parameters Unit 1: Thermal Unit 1: Thermal Unit 1: Thermal Unit 1: Thermal Unit 1: Thermal
Kp1  − 0.2828 Kp  3.9988 Kp1  − 2.000 Kp  1.9910 Kp1  − 1.9815
Ki  − 1.7898 Ki  3.9996 Kp2  − 2.000 Ki  1.9999 Kp2  − 1.8679
Kd  − 1.9978 Kd  2.2747 Ki  − 2.000 Kd  1.9172 Kp3  − 1.9705
Kp2  0.0084 Unit 2: Hydro Kd  − 2.000 λ  0.5946 Ki  − 2.000
Unit 2: Hydro Kp  3.9984 λ  0.8883 μ  0.1246 Kd  − 1.0952
Kp1  0.1610 Ki  1.5289 μ  0.5116 Unit 2: Hydro λ  0.4192
Ki  -0.5022 Kd  1.2681 Unit 2: Hydro Kp  − 0.5227 μ  0.9019
Kd  − 0.4581 Unit 3: Gas Kp1  − 1.2377 Ki  1.9997 Unit 2: Hydro
Kp2  − 0.6917 Kp  3.9972 Kp2  1.3189 Kd  − 1.3676 Kp1  − 1.7355
Unit 3: Gas Ki  3.9956 Ki  0.3657 λ  0.8920 Kp2  2.000
Kp1  0.0115 Kd  0.3447 Kd  0.6020 μ  0.3923 Kp3  − 1.9097
Ki  − 1.9978 λ  0.6953 Unit 3: Gas Ki  0.4479
Kd  − 0.1134 μ  0.8743 Kp  1.7873 Kd  1.7354
Kp2  − 1.9978 Unit 3: Gas Ki  1.9999 λ  1.1987
Kp1  − 2.000 Kd  0.6119 μ  1.1997
Kp2  − 1.3423 λ  0.9962 Unit 3: Gas
Ki  2.000 μ  1.2422 Kp1  − 1.9996
Kd  1.2254 Kp2  − 1.9999
λ  0.7158 Kp3  − 1.8002
μ  0.5336 Ki  − 1.9999
Kd  1.9999
λ  1.1705
μ  0.5639
Ts (s) 2% band f 2.83 7.82 2.24 1.88 1.19
ITEA 0.0292 0.0223 0.0070 0.0108 0.00281

Fig. 7 Output results of test system-1 after a 1% SLP

5.2.1 Scenario 1: load disturbance analysis recently published MGWO-based PID controller [23], hSFS-
PS-based PID controller [25], and DSA-tuned FOPI-FOPD
In this scenario, a SLP of 2% is applied in area-1 at time controller [27], to highlight the competency of our proposed
0. Table 2 lists the associated system results and the con- controller under identical conditions. The statistics in Table 2
troller settings that have been adjusted using the MPA provide strong support for the (P-P-FOPID) controller’s
method of minimization of Minimum Integral Time Absolute superiority, as it achieves the lowest ITEA value (0.0411)
Error (ITAE). Like in the previous exercise, the final results when compared to MGWO: PID (ITAE  0.9197), hSF-
are compared with MPA-tuned FOPID controller and the S–PS: PID (ITAE  0.3818), MPA: FOPID (ITAE  0.1375),
and the DSA: FOPI–FOPD (ITAE  0.461). Moreover,

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Electrical Engineering

Fig. 8 Frequency deviation response of test system-1 for 1% SLP with system parameters variation (a) Frequency response for ± 30% variation in
TP (b) Frequency response for ± 30% variation in R

Table 2 illustrates that the settling times for area frequency 5.3 Three-area hydrothermal power system (Test
and tie-line power are notably enhanced with the proposed system 3)
MPA-optimized P-P-FOPID controller. The improvements
in settling times are as follows: for  f 1 : 89.48%, 85.61%, To realize the full potential of the suggested controller, it is
78.98%, and 8.84%, for  f 2 : 82.79%, 74.74%, 55.31%, and essential to demonstrate its robustness in various power sys-
23.82%, and for Ptie : 75.76%, 64.82%, and 32.82% respec- tems. To achieve this objective, we expanded the scope of our
tively compared to MGWO: PID, hSFS–PS: PID, MPA: study to include a three-area hydrothermal power system. The
FOPID, and DSA: FOPI–FOPD controllers. transfer function model for a three-area hydropower system
Figure 9 displays the transient response of test system 2 for is shown in Fig. 3, and Appendix C provides the nominal sys-
a 2% SLP in region 1. From the results shown in Fig. 9, it is tem parameters, which were obtained from references [27],
clear that the proposed controller has significantly minimized and [28]. Similar to previous test systems, two scenarios are
the overshoot/undershoot and settling time of  f 1 ,  f 2 , and investigated.
Ptie .
5.3.1 Scenario 1: load disturbance analysis
5.2.2 Scenario 2: Sensitivity analysis
In this case, a step load perturbation (SLP) of 1% is uniformly
In this scenario, a sensitivity analysis was conducted to exam- applied in each area, and the same approach as described
ine the variability in the dynamic performance of test system for previous test systems is employed to find the optimal
2 when subjected to a specific change in a small portion of controller gains. Table. 3 illustrates the comparison between
its fundamental parameters. The objective of this analysis the optimized gains and the outcomes with other meth-
was to evaluate the resilience of the proposed controller’s ods, including MPA-optimized FOPID, ISFS-optimized PID
performance in response to changes in system parameters. [13], DSA-based FOPI–FOPID [27], and CGO-optimized
Using the same controller parameters employed in scenario FOPID–FOPI [28]. Analyzing the data in Table 3 reveals
1, different metrics for parameters such as power system time that, in comparison to other techniques, the MPA-tuned (P-
constant (TP ), speed regulator (R), and tie-line coefficient P-FOPID) cascade controller yields a lower value of Integral
(T12 ) were changed by up to ± 30% to gauge the sensitiv- Time Absolute Error (ITAE  55.24). Moreover, the pro-
ity of various parameters at their nominal values. Figure 10 posed controller stands out as the best performer, efficiently
illustrates the dynamic response of the proposed controller to restoring system frequencies and tie-line powers to stable
frequency deviations in area-1. The results depicted in Fig. 10 values with the shortest settling time.
affirm the robustness of the proposed MPA-tuned P-P-FOPID Notably, compared to alternative controllers like MPA-
cascade controller, showcasing its ability to withstand varia- tuned FOPID, CGO-tuned FOPID–FOPI, DSA-tuned
tions in system parameters. FOPI–FOPD, and ISFS-tuned PID controller, the proposed
controller showcases significant improvements. Specifically,
it achieves a 37% improvement in settling time of frequen-
cies compared to MPA-tuned FOPID, 12% compared to
CGO-tuned FOPID–FOPI, 41% compared to DSA-tuned

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Electrical Engineering

Table 2 Optimized controller parameters and Ts/ITAE results for test system 2

Algorithms MGWO: PID [23] hSFS-PS: PID [25] DSA: FOPI-FOPD [27] MPA: FOPID MPA: P-P-FOPID

Controller Unit 1: Thermal Unit 1: Thermal Unit 1: Thermal Unit 1: Thermal Unit 1: Thermal
Parameters Kp  1.7502 Kp  − 1.7074 Kp1  − 2.000 Kp  2.0000 Kp1  − 2.000
Ki  − 0.0087 Ki  − 1.9589 Kp2  − 2.000 Ki  1.9999 Kp2  − 1.9953
Kd  0.7499 Kd  − 1.3934 Ki  − 2.000 Kd  1.9999 Kp3  − 1.9986
Unit 2: Hydro Unit 2: Hydro Kd  − 2.000 λ  0.5928 Ki  − 2.000
Kp  0.3110 Kp  − 0.7453 λ  0.5040 μ  0.0301 Kd  − 1.9461
Ki  0.3102 Ki  0.1375 μ  0.0100 Unit 2: Hydro λ  0.7382
Kd  0.0034 Kd  − 0.9896 Unit 2: Hydro Kp  − 1.7043 μ  2 × 10–7
Unit 3: Gas Unit 3: Gas Kp1  1.7433 Ki  1.9927 Unit 2: Hydro
Kp  0.0091 Kp  − 1.8253 Kp2  − 2.000 Kd  1.8251 Kp1  − 2.000
Ki  1.2409 Ki  − 1.6813 Ki  − 2.000 λ  0.9139 Kp2  0.05599
Kd  0.6901 Kd  − 0.1628 Kd  − 1.3626 μ  0.7547 Kp3  − 2.000
λ  0.9506 Unit 3: Gas Ki  − 2.000
μ  0.0877 Kp  1.9999 Kd  1.9960
Unit 3: Gas Ki  1.9999 λ  0.4522
Kp1  − 2.000 Kd  1.9999 μ  1.0070
Kp2  − 1.7815 λ  1.00722 Unit 3: Gas
Ki  − 2.000 μ  1.4999 Kp1  − 2.000
Kd  − 2.000 Kp2  − 2.000
λ  0.8692 Kp3  − 1.4574
μ  1.2000 Ki  − 2.000
Kd  − 1.6723
λ  1.2000
μ  1.0721
Ts (s) 2% band
f 1 15.69 11.47 1.81 7.85 1.65
f 2 20.05 13.66 4.51 7.72 3.45
Ptie 17.82 12.28 3.98 6.43 4.32
ITAE 0.9197 0.3818 0.0461 0.1375 0.0411

FOPI–FOPD, and 50.5% compared to ISFS-tuned PID con- from the figure that despite substantial variations in power
troller. Similarly, regarding settling time of tie-line powers, system parameters, the system output remains within accept-
the proposed controller outperforms with a 35% improve- able limits without significant differences. This leads to the
ment compared to MPA-tuned FOPID, 39% compared to conclusion that the proposed controller approach effectively
CGO-tuned FOPID–FOPI, 30% compared to DSA-tuned ensures robust and stable control.
FOPI–FOPD, and 29% compared to ISFS-tuned PID con-
troller. Furthermore, the comparative dynamic reaction to a
1% SLP at t  0 is shown in Fig. 11. The results indicate that
the proposed MPA-tuned P-P-FOPID controller has signifi- 6 Conclusion
cantly improved the system dynamic’s response by reducing
overshoot, undershoot, and settling time. This paper introduces an efficient cascade controller named
(P-P-FOPID) to address the LFC issue in various single-
and multi-area power systems. A modern optimizer, MPA, is
5.3.2 Scenario 2: sensitivity analysis employed to find the optimal values of controller parameters.
To test the capability of the proposed controller, three differ-
In order to evaluate the robustness of the proposed control ent typical test systems—a multi-source single-area power
system and confirm its effectiveness, we changed some key system, a two-area multi-source power system, and a three-
parameters of test system 3, including the power system time area hydrothermal power system—are investigated. Each test
constant
  (TP ), speed regulator (R), and tie-line coefficient system was subjected to two scenarios: step load perturba-
Ti j , within a range of ± 15%, while keeping the optimal tion and variation in system parameters. The performance
values of the controller gain unchanged. Figure 12 depicts of the proposed approach was then evaluated by comparing
the frequency deviation in area-1 resulting from adjustments its results with those obtained using the MPA-tuned FOPID
to system settings within the range of ± 15%. It is evident controller and other recent significant contributions in the

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Electrical Engineering

Fig. 9 Output results of test system-2 after a 2% SLP in area 1 (a) Frequency deviation in area 1 (b) Frequency deviation in area 2 (c) Tie-line
power deviation

Fig. 10 Frequency deviation response of area-1 of test system-2 for 2% SLP in area-1 with system parameters variation (a) Response of F1 for ±
30% variation in TP (b) Response of F1 for ± 30% variation in R (c) Response of F1 for ± 30% variation in T12

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Electrical Engineering

Table 3 Optimized controller parameters and Ts/ITAE results for test system 3

Algorithms ISFS: PID [13] DSA: FOPI–FOPD [27] CGO: FOPID–FOPI [28] MPA: FOPID MPA: P-P-FOPID

Controller Area 1: Thermal Area 1: Thermal Area 1: Thermal Area 1: Thermal Unit 1: Thermal
parameters Kp  1.1539 × Kp1  0.0316 Kp1  1.9248 Kp  0.6163 Kp1  0.00092
10–12 Kp2  0.3346 Kp2  1.4875 Ki  0.2199 Kp2  3.000
Ki  0.0577 Ki  0.3424 Ki1  0.3518 Kd  2.4972 Kp3  1.1797
Kd  0.1606 Kd  0.0271 Ki2  1e − 12 λ  1.0000 Ki  0.2765
Area 2: Thermal λ  1.0000 Kd  1.8881 μ  1.1619 Kd  2.8859
Kp  1.1539 × μ  0.6354 λ1  1.0000 Area 2: Thermal λ  1.0000
10–12 Area 2: Thermal λ2  0.0930 Kp  0.6163 μ  1.1500
Ki  0.0577 Kp1  0.0316 μ  1.000 Ki  0.2199 Unit 2: Thermal
Kd  0.1606 Kp2  0.3346 Area 2: Thermal Kd  2.4972 Kp1  0.00092
Area 3: Hydro Ki  0.3424 Kp1  0.9875 λ  1.0000 Kp2  3.000
Kp  1.1539 × Kd  0.0271 Kp2  1.7e − 7 μ  1.1619 Kp3  1.1797
10–12 λ  1.0000 Ki1  0.2773 Area 3: Hydro Ki  0.27655
Ki  0.0577 μ  0.6354 Ki2  1.3016 Kp  0.0556 Kd  2.8859
Kd  0.1606 Area 3: Hydro Kd  1.9500 Ki  0.00914 λ  1.0000
Kp1  0.0001 λ1  1.0000 Kd  0.0891 μ  1.2500
Kp2  0.0001 λ2  1e − 5 λ  1.4734 Unit 3: Hydro
Ki  0.2249 μ  1.0000 μ  1.1648 Kp1  0.0386
Kd  0.0396 Area 3: Hydro Kp2  0.0037
λ  1.4999 Kp1  0.0020 Kp3  2.6608
μ  0.0100 Kp2  0.5422 Ki  0.3002
Ki1  0.0305 Kd  3
Ki2  0.4657 λ  1.3400
Kd  0.1605 μ  1.0031
λ1  1.0000
λ2  0.1376
μ  1.0000
Ts (s) 2% band
f 1 107.6 90.3 60.59 85.2 53.25
f 2 107.6 90.3 60.59 85.2 53.25
f 3 107.6 90.0 58.71 84.5 53.00
Ptie12 65.28 68.7 77.43 72.8 46.56
Ptie13 65.28 68.7 77.43 72.8 46.56
Ptie23 94.83 74.7 62.47 83.6 66.24
ITAE n/a 147.56 82.67 73.94 55.24

field. The carefully gathered results have demonstrated the Appendix

superior performance of the proposed approach. This supe-
riority is evident in terms of achieving a minimum value of A. Parameters of single-area three-unit system [25,
ITAE and enhancing the settling time while reducing under- 38] (Test System 1)
shoot and overshoot in the responses related to area frequency
deviation and tie-line power. R  2.4Hz/pu, f  60 Hz, Tgs  0.08 s, Tt  0.3 s, Kr
In the future, the performance of the proposed controller  0.3, Tr  10 s, Tgh  0.2 s, Trh  28.75 s, Trs  5 s,
can be improved by employing the latest optimizers. More- Tw  1 s, bg  0.05 s, cg  1, Yc  1 s, Xc  0.6 s, Tf
over, the potential of the proposed controller can be evaluated  0.23 s, Tcr  0.01 s, Tcd  0.2 s, KT  0.543478 pu,
with various complicated practical systems in future works. KH  0.326084 pu, KG  0.130438 pu, Kps  68.9566
Hz/pu MW, Tps  11.49 s.

B. Parameters of two-area six units system [23, 26,

27](Test System 2)

f  60 Hz, R  2.4Hz/pu, B  0.4312 pu MW/Hz, Tgs

 0.08 s, Tt  0.3 s, Kr  0.3, Tr  10 s, Tgh  0.2 s,
Trh  28.75 s, Trs  5 s, Tw  1 s, bg  0.05 s, cg 

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Electrical Engineering

Fig. 11 Output results of test system-3 after a 1% SLP in all three areas (a) Frequency deviation in Area 1 (b) Frequency deviation in Area 2
(c) Frequency deviation in Area 3 (d) Tie-line power deviation in Area 1 (e) Tie-line power deviation in Area 2 (f) Tie-line power deviation in Area
3

1, Yc  1 s, Xc  0.6 s, Tf  0.23 s, Tcr  0.01 s, Tcd C. Parameters of Three areas hydro thermal power
 0.2 s, KT  0.543478 pu, KH  0.326084 pu, KG  system [27, 28] (Test System 3)
0.130438 pu, Kps  68.9566 Hz/pu MW, Tps  11.49
s, T12  0.0433. B  0.425 p.u. MW/Hz; R  2.4 Hz/p.u. MW; Tg 
0.08 s; Tr  10 s, Tt  0.3 s; Tw  1.0 s; Tr  5 s; KPS
 120 Hz/p.u. MW; TPS  20 s; T12  T23  T31 
0.086 p.u.

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Electrical Engineering

Fig. 12 Frequency and tie-line power deviation response of area-1 of R (d) Response of Ptie12 for ± 15% variation in R (e) Response of F1
test system-3 for 1% SLP in all areas with system parameters variation for ± 15% variation in Ti j (f) Response of Ptie12 for ± 15% variation
(a) Response of F1 for ± 15% variation in TP (b) Response of Ptie12 in Ti j
for ± 15% variation in TP (c) Response of F1 for ± 15% variation in

Author contributions Jahanzeab Hussain contributed to investigation, References

analysis, validation, and writing—original draft. Zou Runmin pro-
vided supervision. Samina Akhtar prepared figures. Khalid A. Abouda 1. Ram Babu N, Bhagat SK, Saikia LC, Chiranjeevi T, Devarapalli
involved in writing—review & editing. R, García Márquez FP (2023) A comprehensive review of recent
strategies on automatic generation control/load frequency control
Funding No funding was received for conducting this research work. in power systems. Archiv Comput Methods Eng 30:543–572
2. Sabahi K, Ghaemi S, Pezeshki S (2014) Application of type-2 fuzzy
Data availability All data supporting the findings of this study are avail- logic system for load frequency control using feedback error learn-
able within the paper. ing approaches. Appl Soft Comput 21:1–11
3. Shafei MAR, Ibrahim DK, Bahaa M (2022) Application of PSO
Declarations tuned fuzzy logic controller for LFC of two-area power system with
redox flow battery and PV solar park. Ain Shams Eng J 13:101710
4. Rosaline AD, Somarajan U (2019) Structured H-infinity controller
Conflict of interest There is no conflict of interest.
for an uncertain deregulated power system. IEEE Trans Ind Appl
55:892–906
5. Saikia LC, Mishra S, Sinha N, Nanda J (2011) Automatic genera-
tion control of a multi area hydrothermal system using reinforced

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Electrical Engineering

learning neural network controller. Int J Electr Power Energy Syst 23. Padhy S, Panda S, Mahapatra S (2017) A modified GWO tech-
33:1101–1108 nique based cascade PI-PD controller for AGC of power systems
6. Fathy A, Kassem AM (2019) Antlion optimizer-ANFIS load in presence of Plug in Electric Vehicles. Eng Sci Technol Int J
frequency control for multi-interconnected plants comprising pho- 20:427–442
tovoltaic and wind turbine. ISA Trans 87:282–296 24. Sivalingam R, Chinnamuthu S, Dash SS (2017) A hybrid stochastic
7. Selvaraju RK, Somaskandan G (2017) ACS algorithm tuned fractal search and local unimodal sampling based multistage PDF
ANFIS-based controller for LFC in deregulated environment. J plus (1+PI) controller for automatic generation control of power
Appl Res Technol 15:152–166 systems. J Franklin Inst 354:4762–4783
8. Tran A-T, Le Ngoc Minh B, Tran PT, Huynh VV, Phan V-D, Pham 25. Padhy S, Panda S (2017) A hybrid stochastic fractal search and pat-
V-T, Nguyen TM (2021) Adaptive integral second-order sliding tern search technique based cascade PI-PD controller for automatic
mode control design for load frequency control of large-scale power generation control of multi-source power systems in presence of
system with communication delays. Complexity 2021:5564184 plug in electric vehicles. CAAI Trans Intell Technol 2:12–25
9. Cui M, Zhao Y, Cao P, Tang Y, Lu Y (2024) Load frequency control 26. Çelik E, Öztürk N, Arya Y, Ocak C (2021) (1 + PD)-PID cascade
of interconnected hydrothermal power system based on FOPI + controller design for performance betterment of load frequency
FOPD controller. Int J Dyn Control 12:1073–1085 control in diverse electric power systems. Neural Comput Appl
10. Daraz A, Malik SA, Mokhlis H, Haq IU, Laghari GF, Mansor 33:15433–15456
NN (2020) Fitness dependent optimizer-based automatic gener- 27. Çelik E (2021) Design of new fractional order PI–fractional order
ation control of multi-source interconnected power system with PD cascade controller through dragonfly search algorithm for
non-linearities. IEEE Access 8:100989–101003 advanced load frequency control of power systems. Soft Comput
11. Wang P, Chen X, Zhang Y, Zhang L, Huang Y (2024) Fractional- 25:1193–1217
order load frequency control of an interconnected power system 28. Barakat M (2022) Novel chaos game optimization tuned-fractional-
with a hydrogen energy-storage unit. Fractal and Fractional 8:126 order PID fractional-order PI controller for load-frequency control
12. Zhou X, Gao H, Zhao B, Zhao L (2018) A GA-based parameters of interconnected power systems. Prot Control Mod Power Syst
tuning method for an ADRC controller of ISP for aerial remote 7:1–20
sensing applications. ISA Trans 81:318–328 29. Güler Y, Kaya I (2023) Load frequency control of single-area power
13. Çelik E (2020) Improved stochastic fractal search algorithm and system with PI–PD controller design for performance improve-
modified cost function for automatic generation control of intercon- ment. J Electr Eng Technol 18:2633–2648
nected electric power systems. Eng Appl Artif Intell 88:103407 30. Singh K, Arya Y (2023) Tidal turbine support in microgrid fre-
14. Guha D, Roy PK, Banerjee S (2020) Whale optimization algorithm quency regulation through novel cascade Fuzzy-FOPID droop in
applied to load frequency control of a mixed power system consid- de-loaded region. ISA Trans 133:218–232
ering nonlinearities and PLL dynamics. Energy Syst 11:699–728 31. Pathak PK, Yadav AK, Shastri A, Alvi PA (2023) BWOA assisted
15. Mbuli N, Ngaha WS (2022) A survey of big bang big crunch opti- PIDF-(1+I) controller for intelligent load frequency management
misation in power systems. Renew Sustain Energy Rev 155:111848 of standalone micro-grid. ISA Trans 132:387–401
16. Ibrahim MH, Peng AS, Dani MN, Khalil A, Law KH, Yunus S, 32. Al-Betar MA, Awadallah MA, Makhadmeh SN, Alyasseri ZAA,
Rahman MI, Au TW (2023) A Novel computation of delay margin Al-Naymat G, Mirjalili S (2023) Marine predators algorithm: a
based on grey wolf optimisation for a load frequency control of review. Archiv Comput Methods Eng 30:3405–3435
two-area-network power systems. Energies 16:2860 33. Faramarzi A, Heidarinejad M, Mirjalili S, Gandomi AH (2020)
17. C. Andic, S. Ozumcan, M. Varan, and A. Ozturk, (2023) A novel Marine predators algorithm: a nature-inspired metaheuristic.
sea horse optimizer based load frequency controller for two-area Expert Syst Appl 152:113377
power system with pv and thermal units, in: preprints, Preprints 34. Sahlol AT, Yousri D, Ewees AA, Al-qaness MAA, Damasevicius
(eds) R, Elaziz MA (2020) COVID-19 image classification using deep
18. Sahoo G, Sahu RK, Panda S, Samal NR, Arya Y (2023) Modified features and fractional-order marine predators algorithm. Scientific
harris hawks optimization-based fractional-order fuzzy PID con- Rep 10:15364
troller for frequency regulation of multi-micro-grid. Arabian J Sci 35. Soliman MA, Hasanien HM, Alkuhayli A (2020) Marine predators
Eng 48:14381–14405 algorithm for parameters identification of triple-diode photovoltaic
19. Morsali J, Zare K, Tarafdar Hagh M (2017) Applying frac- models. IEEE Access 8:155832–155842
tional order PID to design TCSC-based damping controller in 36. Vankadara SK, Chatterjee S, Balachandran PK, Mihet-Popa L
coordination with automatic generation control of interconnected (2022) Marine predator algorithm (MPA)-based MPPT technique
multi-source power system. Eng Sci Technol Int J 20:1–17 for solar PV systems under partial shading conditions. Energies
20. Daraz A, Malik SA, Basit A, Aslam S, Zhang G (2023) Modified 15:6172
FOPID controller for frequency regulation of a hybrid intercon- 37. Padhy S, Sahu PR, Panda S, Padmanaban S, Guerrero JM, Khan B
nected system of conventional and renewable energy sources. (2022) Marine predator algorithm based PD-(1+PI) controller for
Fractal and Fractional 7:89 frequency regulation in multi-microgrid system. IET Renew Power
21. Ismail MM, Bendary AF (2018) Load frequency control for multi Gener 16:2136–2151
area smart grid based on advanced control techniques. Alexandria 38. Halmous A, Oubbati Y, Lahdeb M, Arif S (2023) Design a new
Eng J 57:4021–4032 cascade controller PD-P-PID optimized by marine predators algo-
22. Babaei F, Lashkari ZB, Safari A, Farrokhifar M, Salehi J (2020) rithm for load frequency control. Soft Comput 27:9551–9564
Salp swarm algorithm-based fractional-order PID controller for 39. Al-Saggaf U, Mehedi I, Bettayeb M, Mansouri R (2016) Fractional-
LFC systems in the presence of delayed EV aggregators. IET Electr order controller design for a heat flow process. Proc Inst Mech Eng
Syst Trans 10:259–267 Part I J Syst Control Eng 230:680–691

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Electrical Engineering

40. Mousavi Y, Alfi A (2015) A memetic algorithm applied to trajec- Publisher’s Note Springer Nature remains neutral with regard to juris-
tory control by tuning of Fractional Order proportional-integral- dictional claims in published maps and institutional affiliations.
derivative controllers. Appl Soft Comput 36:599–617
41. Azarmi R, Tavakoli-Kakhki M, Sedigh AK, Fatehi A (2015) Ana- Springer Nature or its licensor (e.g. a society or other partner) holds
lytical design of fractional order PID controllers based on the exclusive rights to this article under a publishing agreement with the
fractional set-point weighted structure: case study in twin rotor author(s) or other rightsholder(s); author self-archiving of the accepted
helicopter. Mechatronics 31:222–233 manuscript version of this article is solely governed by the terms of such
42. Oustaloup A, Levron F, Mathieu B, Nanot FM (2000) Frequency- publishing agreement and applicable law.
band complex noninteger differentiator: characterization and syn-
thesis. IEEE Trans Circuits Syst I Fundam Theory Appl 47:25–39

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:

1. use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
2. use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at

[email protected]