Journal Pre-proof

Power System Stability Enhancement through Optimal PSS Design

Md. Rashidul Islam , Md. Samiul Azam , Md. Saber Hossen ,

Mohammad Saiful Islam , Muhammed Y. Worku ,
Mohammad Shoaib Shahriar , Md. Shafiullah

PII: S2772-6711(24)00315-2
DOI: https://doi.org/10.1016/j.prime.2024.100735
Reference: PRIME 100735

To appear in: e-Prime - Advances in Electrical Engineering, Electronics and Energy

Received date: 8 March 2024

Revised date: 4 July 2024
Accepted date: 14 August 2024

Please cite this article as: Md. Rashidul Islam , Md. Samiul Azam , Md. Saber Hossen ,
Mohammad Saiful Islam , Muhammed Y. Worku , Mohammad Shoaib Shahriar , Md. Shafiullah ,
Power System Stability Enhancement through Optimal PSS Design, e-Prime - Advances in Electrical
Engineering, Electronics and Energy (2024), doi: https://doi.org/10.1016/j.prime.2024.100735

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2024 Published by Elsevier Ltd.

This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
Highlights
• Formulation of mathematical modeling of power system networks.

• Deployment of efficient metaheuristic algorithms for PSS parameter optimization.

• Performance comparison with conventional PSS through various indicators.

 Power System Stability Enhancement
through Optimal PSS Design
1. Md. Rashidul Islam
Department of Electrical and Electronic Engineering
International Islamic University Chittagong
Chittagong 4318, Bangladesh
Email: [email protected]

2. Md. Samiul Azam*

Department of Electrical and Electronic Engineering
International Islamic University Chittagong
Chittagong 4318, Bangladesh
Email: [email protected]

3. Md. Saber Hossen

Department of Electrical and Electronic Engineering
International Islamic University Chittagong
Chittagong 4318, Bangladesh
Email: [email protected]

4. Mohammad Saiful Islam

Department of Electrical and Electronic Engineering
International Islamic University Chittagong
Chittagong 4318, Bangladesh
Email: [email protected]

5. Muhammed Y. Worku
Interdisciplinary Research Center for Renewable Energy and Power Systems
King Fahd University of Petroleum & Minerals
Dhahran 31261, Saudi Arabia
Email: [email protected]

6. Mohammad Shoaib Shahriar

Department of Electrical Engineering
University of Hafr Al-Batin
Hafr Al Batin 31991, Saudi Arabia
Email: [email protected]

7. Md. Shafiullah
Interdisciplinary Research Center for Renewable Energy and Power Systems
King Fahd University of Petroleum & Minerals
Dhahran 31261, Saudi Arabia
Email: [email protected]

Abstract- Low-frequency oscillations (LFO) are generated in interconnected electric networks because of the
weak tie lines between the parts of the networks. Such oscillations have been a significant challenge for engineers
that may lead to system instability if appropriate measures are not taken. This article proposes two new
metaheuristic algorithms, the jellyfish search algorithm (JSA) and the tunicate swarm algorithm (TSA), to design
the power system stabilizers (PSS) for single machine infinite bus (SMIB) and multimachine power system (MPS)
networks. The proposed method uses the JSA and TSA to dampen out the LFOs by modifying the vital parameters
of lead-lag type conventional PSS (CPSS). A damping ratio-based objective function improves both models'
system damping. These methods are tested on an SMIB system and two distinct MPS networks exposed to the
three-phase fault under various loading conditions. The effectiveness of the suggested approach has been studied
by comparing the system eigenvalues and minimum damping ratio (MDR) produced from JSA and TSA-tuned
PSS and the CPSS. In addition, time domain simulation results are compared, demonstrating that the new approach
outperforms the standard techniques, such as particle swarm optimization (PSO) and backtracking search
algorithm (BSA). In the SMIB system, MDR produced by JSA and TSA-based PSS is 2.55 times higher than that
of CPSS. Similarly, in comparison to PSO and BSA-based PSS, JSA, and TSA-tuned PSS offer more than 1.4
times greater MDR in the Two Area Four Machine network and up to 8.4 times larger MDR for the IEEE-39 Bus
network. Furthermore, the efficacy of the suggested JSA and TSA approaches' prediction capacity is validated by
satisfying the values of statistical performance metrics.

Keywords- Damping ratio; Dynamic stability; Electromechanical oscillations; Power system stabilizers; Jellyfish
search algorithm; Tunicate swarm algorithm.

1 Introduction
An electrical power system is a collection of electrical components which transmit, generate, and consume
electrical energy. It is a vast and intricate system with synchronous generators (SG) operating in parallel on the
generating side and various loads on the consumption side. The power system networks must expand
proportionately to accommodate the fast rise in energy consumption, which is not feasible owing to resource
restrictions and environmental issues. This circumstance compels most power systems to run at or near their
stability limitations. The interconnections of power systems through weak links result in weakly damped LFOs
with frequencies ranging from 0.1–3 Hz [15,25,34]. Without appropriate dampening, these LFOs continue to grow
in size, eventually bringing the system out of synchronization [25]. Because of the large-scale interconnected
power networks and the addition of renewable energy resources in the grids, maintaining stability in power
systems is becoming crucial. Power system stability is the system's capability to maintain a state of operating
equilibrium during normal circumstances and recover an acceptable equilibrium level after a fault. After any faults
in the power networks, if the oscillations cannot be mitigated within a short time, then the system becomes
unstable, which results in instability of the whole grid because of the cascading effect of the networks, and it may
cause total blackouts [27,47,50]. Engineers have spent several years doing various forms of study to overcome
this problem. Multiple controllers are used to deal with such issues [2]. An automatic voltage regulator (AVR) is
used with generator excitation to provide a way to improve the power system stability. Nevertheless, using large-
gain AVR in combination with SGs to keep a stable voltage also creates LFO by reducing rotor damping torque
[35]. As a result, PSS are extensively used to improve the stability of the power networks having LFOs
[10,28,39,42]

The PSS enhances system capacity using the excitation mechanism of a synchronous generator by providing an
extra control signal. Several methods have been employed to design the PSS for the last four decades. The classical
linear control theory was used to design most of the PSS employed in the power networks [19]. Linear control
theory is referred to describes a power system's linear model with a fixed configuration. PSS with
constant parameters is called conventional PSS since they can only perform at their best under specific operating
conditions and are ineffective when they vary substantially [30]. Due to its minimal construction, reliability, and
simplicity of implementation, conventional power system stabilizers (CPSS) are commonly utilized in power
companies. However, fluctuations in the operating circumstances of electrical networks might reduce PSS
performance. As the power networks are highly nonlinear, CPSS with constant parameters cannot deal with the
variations in operating conditions. Despite modern control techniques, including adaptive control, intelligent
control, optimal control, and variable structure control [3,6,51], the CPSS are still popular amongst power system
researchers and utilities for their easy-to-implementation features [26,29,46]. However, the CPSS parameters can
significantly impact the dynamic stability of the networks [26]. Therefore, they must be tuned as per the operating
conditions to enhance stability. The parameters are adjusted sequentially in conventional strategy, considering
one electromechanical mode at a time [12]. Nevertheless, the community does not accept the sequential techniques
as such selection at one mode might cause instability in the other modes [7]. In response, the gradient-based design
was introduced by the researchers [31], which can again get stuck to the local optima due to system nonlinearity
[1,21]. Ref. [13] and [18] demonstrated the development of the nonlinear PSS and excitation controller strategies
using nonlinear control theory.

In the last two decades, metaheuristic optimization methods have gained immense popularity, and many have been
used for PSS design. PSS for MPS networks is designed using the Backtracking Search Algorithm (BSA) [38]. It
has been tested on two different complex multimachine networks. However, no renewable energy sources are
considered in those networks. Particle swarm optimization (PSO) is used in this article to optimize the parameters
of PSS in the MPS network [49]. It has been tested on three different complicated networks and is simulated with
DigSilent Power Factory software, which is available for commercial usage. Nevertheless, online parameter tuning
is not considered in this work. This paper [9] uses the Whale optimization algorithm (WOA) for MPS-PSS design,
where WOA-based PSS performance is compared against PSO and Differential evolution (DE)-based PSS.
However, FACTS devices are not integrated with this system. Genetic algorithms (GA) [5], Bacterial Foraging
(BF) [17], Artificial bee colony (ABC) [32], Fuzzy gravitational search algorithm (FGSA) [14], etc, have also been
used for PSS design in MPS networks.

On the other hand, PSS for the SMIB network has also been designed using different optimization algorithms.
In[16], Multi-Verse Optimizer (MVO) algorithm-based PSS is designed for the SMIB network. Lead-Lag and PID-
based controllers are tuned using MVO in this paper. Although it outperforms GA, the performance could be
improved using UPFC. Article [23] presents a sine cosine algorithm (SCA) based PSS design where a lead leg
controller is used. It can outperform Evolutionary programming (EP) and Moth flame optimization (MFO)
techniques. However, this model uses a linearized model and does not consider the nonlinearity of the system.
Many other algorithms, such as the Firefly algorithm (FA) [36], GA [22], Honey bee mating optimization (HBMO)
[44] and so on are also used to optimize the parameters of PSS in the SMIB network. In some research, PSS has
been coordinated with other different control devices. In this article [43], PSS is coordinated with TCSC, and the
PSO Algorithm is used to tune the parameters of PSS and TCSC, which makes it a robust design. The limitations
are that no renewable sources are considered and are tested on a single network. This paper [41] represents a PSS
design with TCSC, and two different algorithms, i.e., Velocity update relaxation particle swarm optimization
(VURPSO) and GA, are used to optimize the parameters. However, the optimal placement of the controller is not
discussed, and the presence of distributed generators (DG) is not considered.

All these studies mostly considered either MPS or SMIB networks for performance evaluation. Considering this
limitation, this research utilizes both MPS and SMIB networks to analyze the performance of two proposed
metaheuristic algorithms. In the meta-heuristic optimization front, the artificial jellyfish search algorithm (JSA)
[39] and the tunicate swarm algorithm (TSA) [20] are novel bio-inspired swarm-based approaches. The algorithms
are easy to understand, implement, and code with several internal parameters. Due to their effectiveness and fast
convergence capability, they were already implemented for optimal power flow [11], distribution systems
automation [40], and parameter estimation of single-phase transformers [53]. To the authors' best knowledge, the
algorithms were not comprehensively explored to tune the PSS parameters in the single-machine infinite bus and
multimachine power networks. The major contributions of this article are as follows:
• Two efficient metaheuristic algorithms, JSA and TSA, are employed to tune the PSS parameters of the
SMIB and MPS networks.
• Performances are compared with conventional PSS through eigenvalue analysis and time domain plots
for performance parameters like angular frequency, rotor angle, control signal, field voltage response,
internal voltage response, and output power. In addition, the minimum damping ratios found from the
proposed models are compared with PSO and BSA algorithms.

This research article is organized in the following manner for the remaining sections: Section 2 discusses the
power system modeling and lead-leg controller design; Section 3 describes the optimization problem formulation
for the SMIB network and MPS network, and metaheuristic optimization algorithm (JSA and TSA) structures are
discussed with the flowchart in Section 4. Then, Section 5 provides the simulation results in detail, and the
conclusion is given in Section 6.

2 Power System Modeling

Power systems are intricate systems comprising numerous interrelated parts, including loads, transmission lines,
transformers, and generators. When modeling a power system, it is crucial to represent the behavior of these
elements precisely and how they interact in the real world. The practice of mathematically representing a power
system to examine its behavior and performance is known as power system modeling. The mathematical modeling
of the SMIB, MPS, and CPSS is provided in this section.

2.1 Single Machine Infinite Bus Network Modeling

The structure of the SMIB system is depicted in Fig. 1, in which an SG is connected to an infinite
bus via transmission lines (Network # 1). The infinite bus represents the remaining part of the power system,
which keeps a fixed voltage and frequency level.

Fig. 1 The SMIB network

The SG used in the SMIB network is modeled with fourth-order differential equations where DC excitation is
considered. The first-order equations of the proposed model are given below:

𝛿̇ = 𝜔𝑏 (𝜔 − 1) (1)

1
𝜔̇ = [𝑃𝑚 − 𝐷(𝜔 − 1) − 𝑃𝑒 ] (2)
2𝐻

1
𝐸̈𝑞 = [𝐸𝑓𝑑 − (𝑥𝑑 − 𝑥̇ 𝑑 )𝑖𝑑 − 𝐸̇𝑞 ] (3)
𝑇̇𝑑𝑜
In Eq. (1), (2), and (3), 𝑥̇ 𝑑 , 𝑥𝑑 , 𝐸𝑓𝑑 , Eq, δ, 𝜔, 𝜔𝑏 , D, Pe, and Pm are direct axis sub-transient reactance, direct axis
transient reactance, field voltage, generator's internal voltage, rotor angle, rotational speed of the generator,
synchronous speed, damping coefficient, electrical output power, and the mechanical input power, respectively.
𝑣𝑡 = 𝑣𝑑 + 𝑗𝑣𝑞 (4)

𝑣𝑑 = 𝑥𝑞′ 𝑖𝑞 (5)

𝑣𝑞 = 𝑥𝑑′ 𝑖𝑑 + 𝑒𝑞′ (6)

Here, 𝑥𝑞′ stands for the synchronous quadrature axis reactance. As shown in Eq. 4, the terminal voltage can be
determined through the direct and quadrature axis voltages. The d-q axis generator armature current is denoted by
𝑖𝑑 and 𝑖𝑞 , whereas the terminal voltages of the generator are represented by 𝑣𝑑 and 𝑣𝑞 . The following equations
represent 𝑣𝑑 and 𝑣𝑞 as Eqs. 5 and 6.
The generating power of the generator is expressed in quadrature, and the following are the quantities of the direct
axis:

Pe = 𝑣𝑑 𝑖𝑑 + 𝑣𝑞 𝑖𝑞 (7)

Six gain constants (K1-K6) represent the relationship between variables in a linear model for the SMIB system,
provided in [2]. The linearized formulas for the Fig. 1 system are shown below:

𝛥𝛿̇ = 𝜔𝑏 𝛥𝜔 (8)

𝑘1 𝐷 𝑘2
𝛥𝜔̇ = − 𝛥𝛿 − Δ𝜔 − Δ𝑒𝑞′ (9)
𝑀 𝑀 𝑀
𝑘4 𝑘4 1
𝛥𝑒𝑞′ = − ′ 𝛥𝛿 − ′ 𝛥𝑒𝑞′ + ′ 𝛥𝐸𝑓𝑑 (10)
𝑇𝑑𝑜 𝑘3 𝑇𝑑𝑜 𝑇𝑑𝑜

𝑘𝐴 𝑘𝑠 𝑘𝐴 𝑘6 1 𝑘𝐴
̇ = −
Δ𝐸𝑓𝑑 Δ𝛿 − Δ− 𝛥𝐸 + 𝑢 (11)
𝑇𝐴 𝑇𝐴 𝑇𝐴 𝑓𝑑 𝑇𝐴 𝑝𝑠𝑠

2.2 Multimachine power system modeling

Multimachine power system models the behavior of power systems involving multiple generators and loads
connected through transmission lines. These models are used to understand the dynamic behavior of the power
system, including the generation and transmission of electrical energy, as well as the stability and reliability of
the system. In multimachine power system modeling, each generator and load are represented by a separate
mathematical model, and the models are then combined to represent the entire system's behavior. This section
shows the mathematical modeling of a multimachine power system with two examples.

2.2.1 Synchronous Generator

The synchronous generator model affects voltage regulation, stability, transient performance, control system
design, and the generator’s interaction with the power system. Accurate modelling ensures that the excitation
system operates effectively and maintains system stability. The fourth-order model of the 𝑖 𝑡ℎ generator in an electric
network with 𝑛 number of machines is expressed mathematically as [54]:
𝛿𝑙̇ = 𝜔𝑏 (𝜔𝑖 − 1) (12)
1
𝜔̇ 𝑙 = (𝑃𝑚𝑖 − 𝑃𝑒𝑖 − 𝑃𝐷𝑖 ) (13)
𝑀
1 ′
𝑒̇𝑞 = [𝐸 − 𝑒𝑞𝑖 − (𝑥𝑑𝑖 − 𝑥𝑑𝑖 )𝑖𝑑𝑖 ]
𝑇𝑑𝑜 𝑓𝑑𝑖 (14)
1
𝐸𝑓̇ ⅆ = [𝐾 (𝑣 − 𝑣𝑡𝑖 + 𝑢𝑃𝑆𝑆𝑖 ) − 𝐸𝑓𝑑𝑖
𝑇𝐸𝑖 𝐸𝑖 𝑡𝑟𝑖 (15)
′ ′ ′
In the above equations, 𝑥𝑑 , 𝑥𝑑 , 𝐾𝐴 , 𝑒𝑞 , 𝑇𝑑𝑂 , 𝑢𝑃𝑆𝑆 , 𝜔, 𝑀, 𝑃𝑚 , 𝑃𝑒 and 𝐸𝑓𝑑 represents the direct quadrature axes
reactance, transient reactance, gain of the excitation system, internal generator voltage, open circuit field time
constant, control unit, angular frequency, inertia constant, input mechanical power, output electrical power, and
field excitation, respectively.

2.2.2 Power System Stabilizers and Location

Fig. 2 depicts the position and structure of a typical two-stage lead-lag CPSS. It is located within the control
system of each synchronous generator and integrated into the Automatic Voltage Regulator (AVR) system.
Reference voltage Vref, Terminal voltage Vt, and stabilizing signal 𝑈 from the PSS are fed into the AVR. AVR
adjusts the excitation system by comparing Vt with Vref and damping signal U. The PSS input is the variation in
the generator's angular frequency (𝜔𝑖 ) and the output is the variation in the control signal 𝑈𝑃𝑆𝑆𝑖 . 𝐾𝐶𝑖 represents
the gain of the controller. A washout/rest block with the time constant 𝑇𝑤 is attached to turn off PSSs during
steady-state operation. The limiter block controls the control signal's magnitude. Two lead-lag phase
compensators with time constants 𝑇1 , 𝑇2 , 𝑇3 , and 𝑇4 are sandwiched between the discharge filter and the limiter
block.

Fig. 2 Conventional lead-lag PSS structure with location.

2.2.3 Four-machine Two-Area Network

Fig. 3 illustrates a four-machine two-area system (Network # 2). The system described in the reference [24]
comprises eleven buses and is divided into two areas, with buses 7 and 9 having a weak link. This system has two
inductive loads, which are respectively connected to buses 7 and 9. In addition, the system contains two shunt
capacitors, which are connected to buses 7 and 9. The system consists of two areas, each with two identical
generators.

Fig. 3 Four-machine, two-area network.

2.2.4 IEEE 39-bus Test System Network

Fig. 4 demonstrates the structure of the IEEE 39-bus system (Network # 3: the New England Power System),
which comprises 12 tap-changing transformers, 19 loads, 10 generators, 39 buses, and 36 transmission lines [33].
 Fig. 4 IEEE 39-bus network.

3 Optimization Problem Formulation

This section presents the optimization problem formulation for enhancing the power system LFO damping with
metaheuristic optimization algorithms.

3.1 Optimization Problem Formulation for SMIB Network

To enhance the system damping of the SMIB network, maximizing the minimum damping ratio (MDR) (𝜁) can be
considered as the objective function (J) for the optimization problem:
𝐽 = max (minimum damping ratio ζ) (16)
The CPSS model is dependent on the following limits:
Maximize J
𝐾𝑐𝑚𝑖𝑛 ≤ 𝐾𝑐 ≤ 𝐾𝑐𝑚𝑎𝑥
𝑇1𝑚𝑖𝑛 ≤ 𝑇1 ≤ 𝑇1𝑚𝑎𝑥
𝑇2𝑚𝑖𝑛 ≤ 𝑇2 ≤ 𝑇2𝑚𝑎𝑥
Here, the lead-leg controller's time constants are T1 and T2, and Kc denotes the gain. The optimum value of the
controller elements must be determined to maximize the objective function J and maintain system stability during
power network transients. This research uses JSA and TSA to choose those optimal PSS parameters within the
abovementioned range.

3.2 Optimization Formulation for Multi-machine Networks

The problem is tuning the PSS parameters so that the damping performance of the system can be improved. The
goal is to maximize the objective function (J):

𝐽 = 𝑚𝑎𝑥 {𝜁𝑖 , 𝑖 = 1,2,3, … , 𝑛} (17)

Where, 𝜁𝑖 is 𝑖 𝑡ℎ machines MDR and 𝑛 is the number of machines. The problem of optimization can be stated in
the following way:
 Maximization of J
𝐾𝑐𝑖𝑚𝑖𝑛 ≤ 𝐾𝑐𝑖 ≤ 𝐾𝑐𝑖𝑚𝑎𝑥
𝑇1𝑖𝑚𝑖𝑛 ≤ 𝑇1𝑖 ≤ 𝑇1𝑖𝑚𝑎𝑥
𝑇3𝑖𝑚𝑖𝑛 ≤ 𝑇3𝑖 ≤ 𝑇3𝑖𝑚𝑎𝑥

The optimal values of the PSS parameters have to be obtained to maximize the objective function J and maintain
system stability. This study selects the PSS parameters from the range using JSA and TSA.

4 Metaheuristic Algorithms
Metaheuristic optimization algorithms are a class of algorithms used for solving complicated optimization
problems. These algorithms are named "metaheuristics" because they are high-level strategies that guide the
search for an optimal solution. Metaheuristic algorithms are often used when the exploration space is too large to
be searched exhaustively or when the optimization problem is too complex to be solved precisely [37]. This
research uses two well-known metaheuristic algorithms to solve the power system stability improvement problem.

4.1 Jellyfish Search Algorithm

The food-searching behaviour of the jellyfish was the main inspiration that led to the development of the JSA
technique in 2021 [8,52]. The significant steps of the algorithm are illustrated in Fig. 5. Three strategies guide the
JSA implementation (i) the jellyfishes migrate with the sea waves or within the swarm as per their time control
mechanism; (ii) they prioritize their search for the areas with a greater food supply over their current positions;
and (iii) their new positions impact the objective function.
During the implementation, the population is first initialized using the Logistic chaotic map of Eq. (18), where the
ocean current is represented by Eq. (19) [45]:
𝑋𝑖+1 = 𝜂𝑋𝑖 (1 − 𝑋𝑖 ), 0 ≤ 𝑋0 ≤ 1 (18)
𝑋𝑖 (𝑡 + 1) = 𝑋𝑖 (𝑡) + 𝑟𝑎𝑛𝑑(0,1) × (𝑋 ∗ − 𝛽 ×
(19)
𝑟𝑎𝑛𝑑(0,1) × 𝜇)
Their passive and active motions govern the movement of the jellyfish swarm as per Eq. (20) and (21), respectively.
𝑋𝑖 (𝑡 + 1) = 𝑋𝑖 (𝑡) + 𝛾 × 𝑟𝑎𝑛𝑑(0,1) × (𝑈𝑏 − 𝐿𝑏 ) (20)
⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗
𝑋𝑖 (𝑡 + 1) = 𝑋𝑖 (𝑡) + 𝑟𝑎𝑛𝑑(0,1) ×𝐷𝑖𝑟𝑒𝑐𝑡𝑖𝑜𝑛 (21)
Eq. (22) represents the direction of the jellyfish's active motion which is basically towards the best foodstuff.
𝑋𝑗 (𝑡) − 𝑋𝑖 (𝑡) 𝑖𝑓 𝑓(𝑋𝑖 ) ≥ 𝑓(𝑋𝑗 )
⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗
𝐷𝑖𝑟𝑒𝑐𝑡𝑖𝑜𝑛 = { (22)
𝑋𝑖 (𝑡) − 𝑋𝑗 (𝑡) 𝑖𝑓 𝑓(𝑋𝑖 ) < 𝑓(𝑋𝑗 )
The time control variable C(t) of Eq. (23) refers to the transition between passive and active motions in ocean
currents:
𝑡
𝐶(𝑡) = [(1 − ) × (2 × 𝑟𝑎𝑛𝑑(0,1) − 1)] (23)
𝑀𝑎𝑥𝑖𝑡𝑒𝑟
 Fig. 5 Flowchart of the JSA.

4.2 Tunicate Swarm Algorithm

The TSA optimization algorithm is a metaheuristic approach that draws inspiration from biological systems. It
utilizes principles observed in nature to solve complex problems through an iterative process of trial and error.
The suggested technique mimics the tunicate's swarm dynamics and jet propulsion during navigation and hunting
[4,20]. The flowchart of the TSA is shown in Fig. 6.
 Fig. 6 Flowchart of TSA.

Step 1: First, set up the TSA population 𝑃 𝑝

𝐹⃗ 𝑆 + 𝐴⃗ ⋅ 𝑃 𝐷, if 𝑟𝑎𝑛𝑑 ≥ 0.5
𝑃𝑝 (𝑥) = {
𝐹⃗ 𝑆 − 𝐴⃗ ⋅ 𝑃 𝐷, if 𝑟𝑎𝑛𝑑 < 0.5
Where ( 𝑥′) represents the tunicate's new position in relation to the current location of the food source 𝐹⃗ 𝑆.

Step 2: Use the equation below to determine the initial parameters and maximum iterations:
𝑃𝑜𝑠 = 𝑟𝑎𝑛𝑑(𝑆, 𝐷).∗ (𝑈𝑏 − 𝐿𝑏) + 𝐿𝑏. 𝑏 (24)
Here, Pos= 𝑃𝑜𝑠𝑖𝑡𝑖𝑜𝑛𝑠
 S= 𝑆𝑒𝑎𝑟𝑐ℎ𝐴⃗𝑔𝑒𝑛𝑡𝑠
D= 𝐷𝑖𝑚𝑒𝑛𝑡𝑖𝑜𝑛𝑠
Ub= 𝑈𝑝𝑝𝑒𝑟𝑏𝑜𝑢𝑛𝑑
Lb= 𝐿𝑜𝑤𝑒𝑟𝑏𝑜𝑢𝑛𝑑
Step 3: Then, evaluate every search agent's fitness value.

Step 4: After calculating the fitness value, the best search agent is evaluated within the specified search space.

Step 5: Use the equation below to update every search agent's positions:
𝑃𝑝 (𝑥) + 𝑃𝑝(𝑥 + 1)
𝑃𝑝(𝑥 + 1) = (25)
2 + 𝑐1

Step 6: Each updated search agent that surpasses the boundary conditions in a given search space will be
disqualified.

Step 7: Again, evaluate the updated fitness value of search agents if the optimal solution
𝑃𝑝(𝑥 + 1) gets updated better.

Step 8: If the desired result is achieved, stop the algorithm; else, repeat Steps 5-8.

Step 9: Then, the optimal result found so far is returned.

5 Results and discussions

The PSS performance results for SMIB and MPS networks are simulated and discussed in this section. The SMIB
network is simulated using JSA and TSA algorithms for four loading conditions. On the other hand, JSA and TSA
algorithms simulate two MPS networks: a two-area four-machine network and an IEEE 39-bus network. Both
cases compare PSS performance for JSA and TSA-optimized PSS.

5.1 Network 1: Single Machine Infinite Bus Network

In an SMIB network, transmission lines connect a single machine (SG) to an infinite bus. The infinite bus has a
constant voltage and frequency regardless of load variations. In this network, alternators are connected in parallel.

5.1.1. MDR and Eigenvalues analysis

Eigenvalue analysis can be used to estimate a system's stability. Four loading situations were simulated for the
SMIB with PSS. Table 1 provides four loading conditions for the SMIB network where Pe is real power, Qe is
reactive power and Vt is terminal voltage. Table 2 to Table 5 present the eigenvalues found for four loading
scenarios using JSA and TSA-optimized PSS and fixed-gain conventional PSS. The eigenvalues having an
imaginary part less than 0.01 are excluded. The eigenvalues found for loading condition 2 are plotted in the
complex plane, as shown in Fig. 7. The figures demonstrate that all eigenvalues have a negative real component,
which indicates that all the models can stabilize the network. However, the eigenvalue -0.8850 ± j4.0529 (λ₃)
found from the CPSS is much closer to the imaginary axis compared to -3.1253 ± j4.8149 (λ₂) and -3.1387 ±
j4.8360 (λ₂) found from the JSA- and TSA-tuned PSS, respectively, which suggests that the disturbance settles
more slowly than the optimal ones. Other loading conditions showed a similar response.

Moreover, the greater the MDR, the more capable the controller is in damping out the LFOs. MDRs obtained for
CPSS, JSA–tuned PSS, and TSA-tuned PSS under various loading situations are also shown in Table 2 to Table
5. For loading condition 2, the MDR found from JSA and TSA-tuned PSS is 0.5444, which is 2.5 times larger
than the MDR found from CPSS. This suggests that both JSA-tuned and TSA-tuned PSS can offer improved LFO
damping.
 (a)

(b) (c)

Fig. 7 (a) Eigenvalues of CPSS; (b) Eigenvalues of TSA-based PSS; (c) Eigenvalues of JSA-based PSS.

Table 1 Loading conditions.

Loading condition (LC) Pe (pu) Qe (pu) Vt (pu)
LC # 1 0.6 0.01 0.98
LC # 2 0.9545 0.2757 1.0311
LC # 3 0.4250 0.2095 1.0868
LC # 4 0.8751 0.1546 1.0486

Table 2 Comparison of eigenvalues for LC # 1.

Conventional PSS-based
JSA-optimized system TSA-optimized system
system
-18.2019 -17.4857 -17.4860
Eigenvalues
-5.1099 ± j7.2846 -3.1737 ± j5.1613 -3.1922 ± j5.1934
-0.8850 ± j4.0529 -3.1735 ± j5.1610 -3.1548 ± j5.1292
-0.3373 -0.3490 -0.3490
 MDR 0.2133 0.5238 0.5237
Kc 7.0900 26.6036 26.6027
T1 0.6851 0.2228 0.2228
T2 0.1000 0.1000 0.1000

Table 3 Comparison of eigenvalues for LC # 2.

Conventional PSS JSA-optimized PSS TSA-optimized PSS
-18.2019 -17.6710 -17.6711
Eigenvalues -5.1099 ± j7.2846 -3.1253 ± j4.8149 -3.1387 ± j4.8360
-0.8850 ± j4.0529 -3.1249 ± j4.8143 -3.1114 ± j4.7934
-0.3373 -0.3577 -0.3577
MDR 0.2133 0.5444 0.5444
Kc 7.0900 25.6979 25.6976
T1 0.6851 0.1910 0.1940
T2 0.1000 0.1000 0.1000

Table 4 Comparison of eigenvalues for LC # 3.

Conventional PSS JSA-optimized PSS TSA-optimized PSS

-18.2019 -17.2456 -17.2464
Eigenvalues
-5.1099 ± j7.2846 -3.2338 ± j5.1612 -3.2652 ± j5.2131
-0.8850 ± j4.0529 -3.2334 ± j5.1605 -3.2016 ± j5.1094
-0.3373 -0.3491 -0.3491
MDR 0.2133 0.5310 0.5308
Kc 7.0900 38.2083 38.2045
T1 0.6851 0.2157 0.2158
T2 0.1000 0.1000 0.1000

Table 5 Comparison of eigenvalues for LC # 4.

Conventional PSS JSA-optimized PSS TSA-optimized PSS
Eigenvalues
-18.2019 -17.7713 -17.7732
-5.1099 ± j7.2846 -3.1022 ± j5.2014 -3.1502 ± j5.2820
-0.8850 ± j4.0529 -3.1017 ± j5.2007 -3.0527 ± j5.1221
-0.3373 -0.3500 -0.3500
MDR 0.2133 0.5122 0.51200
Kc 7.0900 23.2297 23.2240
T1 0.6851 0.2229 0.2230
T2 0.1000 0.1000 0.1000

5.1.2. Time domain simulation results under disturbance.

In this part, time domain simulation is used to compare the effectiveness of a typical fixed gain PSS for loading
condition 2 to that of JSA and TSA-optimized PSS for damping LFOs in the presence of a fault. A 10% pulse
input of mechanical torque was provided at 1.0s for four cycles to imitate the disturbance, and the related system
responses were measured. Traditional PSS, JSA-optimized PSS, and TSA-optimized PSS can stabilize the power
system network following a mechanical disturbance. Fig. 8 shows the objective function variation curve with
respect to iterations for loading condition 2. It can be seen that JSA shows faster convergence compared to TSA,
and both algorithms give an optimal minimum damping ratio of 0.5444. Fig. 9 shows the angular frequency
response for the SMIB network after disturbance with CPSS and optimized PSS, where JSA- and TSA-optimized
 PSS stabilizes the angular frequency within 3.2s, whereas conventional PSS cannot stabilize within the simulated
period of 8s.

Fig. 8 Objective function change of JSA- and TSA-tuned PSS for SMIB (LC # 2).

Fig. 9 Angular frequency response of the SMIB system for Fig. 10 Rotor angle response of the SMIB system with
LC # 2. conventional PSS and optimized PSS for LC # 2.
Fig. 11 Control signal response of the SMIB system with Fig. 12 The field voltage response of the SMIB system with
conventional PSS and optimized PSS for LC # 2. conventional PSS and optimized PSS for LC # 2.

Fig. 13 The internal voltage response of the SMIB system

with conventional PSS and optimized PSS for LC # 2.

Fig. 10 indicates the response of rotor angle in the SMIB system, where optimized JSA and TSA take only 3.5s
to settle down, but the conventional approach needs more time to be stable. Fig. 11, Fig. 12 and Fig. 13 show the
control signal, field voltage, and internal voltage response, respectively, for the SMIB network using optimized
and unoptimized PSS. For all cases, JSA and TSA-optimized PSS provide much faster settling time than
conventional PSS. When the PSS settings are chosen randomly for CPSS, state responses take a long time to
stabilize. Each figure demonstrates that the system cannot attain the stable state within the 8s time duration
employed in the simulation when using a typical fixed gain PSS. A similar response can be found for all other
loading conditions.

5.2 Network 2: Four-machine Two-Area Network

The four-machine two-area network has two similar areas linked via transmission lines. There are two SGs and
one load in each area. Simulation results for the second network are discussed in this section.
5.2.1 MDR and Eigenvalues Analysis
The PSS-connected two-area four-machine system was simulated, and system stability was established by
analyzing its eigenvalues. JSA and TSA algorithms are used to optimize the parameters of the PSS. Eigenvalues
found for JSA, TSA-optimized PSS, and CPSS are shown in Table 6. The eigenvalues having an imaginary part
less than 0.01 are not included in the analysis. Fig. 14 represents the eigenvalues in the complex plane. Although
all eigenvalues have negative components, the complex eigenvalues of the CPSS are relatively close to the
imaginary (jx) axis, indicating that it requires more time to stabilize the fault than the TSA and JSA-tuned ones.

(a)

(b) (c)

Fig. 14 (a) Eigenvalues for CPSS; (b) Eigenvalues for JSA; (c) Eigenvalues for TSA

From Figure 14(a), -0.6436 ± j4.0814 () provides the smallest damping ratio, giving an MDR of 0.1558 for
CPSS. The eigenvalues produced by the JSA and TSA techniques are plotted in Figures 14(b) and 14(c). Except
for -24.8572 ± j5.6689 (), the remaining eigenvalues produce a damping ratio of 0.7549 for JSA-based PSS,
suggesting that no particular mode is excessively underdamped or overdamped, leading to a more predictable and
stable overall system behavior. Moreover, MDR obtained from JSA-tuned PSS is 4.85 times better than CPSS,
1.5 times greater than BSA-based PSS, and 2.3 times more than that of PSO-based PSS [38]. On the other hand,
-11.0639 ± j10.632 () provides an MDR of 0.721 for TSA-based PSS, which is 4.6 times better than CPSS, 1.4
times better than BSA-based PSS and 2.2 times better than PSO based PSS. MDR obtained for CPSS, PSO, BSA,
JSA, and TSA-optimized PSS are shown in Fig. 15.

Table 6 Comparison of eigenvalues

Conventional PSS JSA-optimized PSS TSA-optimized PSS

-38.2625 ± j8.2423 -17.7406 ± j15.4128 -20.0870 ± j17.2810
-46.6067 ± j0.0393 -24.8572 ± j5.6689 -25.7477 ± j7.785931
-2.4979 ± j8.7081 -13.1057 ± j11.3856 -11.0639 ± j10.632
-2.2925 ± j8.3509 -9.5042j ± j8.2567 -10.0470 ± j9.6190
-0.6436 ± j4.0814 -9.4273 ± j8.1903 -8.1956 ± j7.6097
0 -7.2032 ± j6.2581 -5.9384 ± j5.4016
- -3.6504 ± j3.1713 -5.1740 ± j4.6372
- -2.7589 ± j2.3969 -2.3761 ± j2.1127

Table 7 Optimized PSS parameters.

Gen No. JSA-tuned parameters TSA-tuned parameters
Kc T1 T3 Kc T1 T3
G1 18.8305 0.0209 0.0250 18.3778 0.0235 0.0200
G2 42.9914 0.0594 0.0796 39.8384 0.0504 0.0917
G3 30.7775 0.0906 0.1241 32.1290 0.1476 0.0693
G4 19.4093 0.0594 0.0860 17.4897 0.0772 0.0606

Fig. 15 The minimum damping ratio for different optimization techniques.

5.2.2 Time Domain Simulation Analysis

The JSA and TSA techniques are employed to carry out the optimization process for three parameters, 𝐾𝑐𝑖 ,
𝑇1𝑖 and 𝑇3𝑖 for a particular base case, while the 𝑇2𝑖 and 𝑇4𝑖 values were set to 0.0500. This MPS network was
previously tested with PSO and BSA techniques, and the damping ratios are given in [48]. Table 7 shows the
optimized PSS parameters for JSA and TSA-based methods.
 Bus 7 is subjected to a 3-ϕ fault for 0.1s, beginning at 0.5s. The whole process is simulated for 8 seconds. Fig.
16, Fig. 17, Fig. 18 and Fig. 19 show the variations in 𝐺3 ’s angular frequency, rotor angle, output power, and control
signal for the same fault using conventional, JSA, and TSA-optimized PSS. When the fault is applied at 0.5s, the
system becomes unstable, which is evident from the massive deviation in angular frequency (t) and rotor angle
(t). Following the angular frequency oscillation, output power Pe also changes significantly, causing low-
frequency oscillation in the transmission line. Fig. 16 and Fig. 17 show that JSA and TSA-based PSS can stabilize
the angular frequency within 2.3s and rotor angle within 3s, whereas CPSS takes more than 8s. As a result, output
power returns back to its normal value within 2.7s for JSA and TSA-based PSS, as shown in Fig. 18. Fig. 19 depicts
the control signal from different approaches where the signals from JSA and TSA-based PSS only last for 3.1s,
while conventional PSS takes much longer. It demonstrates that the JSA and TSA-based optimized PSS can
stabilize oscillations faster than CPSS. The JSA and TSA techniques have a substantially shorter settling time than
conventional PSS. Similar responses can be seen in other generators.

Fig. 16 Angular frequency response of 𝐺3 with Fig. 17 Rotor angle response of 𝐺3 with conventional, JSA-
conventional, JSA-optimized, and TSA-optimized PSS. optimized, and TSA-optimized PSS.

Fig. 18 Output power of 𝐺3 with conventional, JSA- Fig. 19 Control signal of 𝐺3 with conventional, JSA-
optimized, and TSA-optimized PSS. optimized, and TSA-optimized PSS
5.3 Network 3: IEEE 39-bus Network
The third network has 39 buses, 10 generators, and three areas connected through transmission lines. This section
discusses the simulation results for the IEEE 39-bus network.
5.3.1 MDR and Eigenvalues Analysis
The PSS-connected IEEE 39-bus network is simulated for ten generators using MATLAB software. The stability
of a system can be established by analyzing its eigenvalues. Table 8 compares the eigenvalues for JSA, TSA-
optimized PSS, and CPSS, including only the eigenvalues having an imaginary part greater than 0.01. These
eigenvalues are illustrated in complex planes in Fig. 20. As can be seen from Figure 20(a), two CPSS eigenvalues,
0.1946 ± j5.8220 () and 0.0851 ± j4.0141 (), are on the right side of the imaginary axis, resulting in a negative
MDR of -0.0334. This suggests that CPSS in the IEEE-39 bus network increases the LFO with time instead of
stabilizing. In contrast, all eigenvalues found from JSA and TSA-tuned PSS are on the left side of the imaginary
axis, as shown in Figures 20(b) and 20(c). Eigenvalue -2.2206 ± j9.7486 () provides an MDR of 0.221 for JSA-
based PSS, and -1.1902 ± j7.8225 () gives an MDR of 0.1503 for TSA-tuned PSS. This suggests that the
proposed method performed significantly better in dampening LFOs compared to CPSS. Moreover, the MDR
found from JSA and TSA-based PSS is compared with two famous algorithms, BSA and PSO, in Fig. 21. It
demonstrates that JSA provides 8.4 times and 7 times larger MDR compared to PSO and BSA-based PSS [38].
Likewise, TSA delivers 5.7 times and 4.8 times greater MDR compared to PSO and BSA-tuned PSS. These are
impressive improvements over already popular algorithms. From the eigenvalues and MDR analysis, both JSA
and TSA-optimized PSS show significant improvement over CPSS, PSO, and BSA-based PSS.

Table 8 Comparison of eigenvalues.

Conventional PSS JSA-optimized PSS TSA-optimized PSS
-0.1764 ± j9.7104 -29.3171 ± j0.6962 -4.0040 ± j24.7411
-0.0268 ± j9.6543 -3.1620 ± j13.7916 -3.2399 ± j16.1553
-0.2691 ± j9.3463 -5.4061 ± j11.3873 -20.0331 ± j0.1590
-0.0934 ± j8.1295 -2.9899 ± j13.0681 -1.9801 ± j13.0208
-0.2201 ± j8.0154 -2.8795 ± j12.4686 -1.6617 ± j10.6973
-0.0509 ± j7.1417 -2.8102 ± j12.1366 -1.4692 ± j9.5313
-0.0311 ± j6.4287 -2.6703 ± j11.6752 -1.3769 ± j8.6687
0.1946 ± j5.8220 -2.3576 ± j10.2175 -1.2864 ± j8.3366
0.0851 ± j4.0141 -2.2206 ± j9.7486 -1.1902 ± j7.8225
-5.7940 ± j1.7859 -1.9954 ± j8.7362 -1.1338 ± j7.4173
-6.7266 ± j0.6924 -3.2840 ± j4.5079 -6.2512 ± j1.1867
-1.1382 ± j0.14448 -4.7712 ± j2.6834 -0.6781 ± j3.5412
0 -6.2580 ± j0.3049 -1.3080 ± j2.9643
0 -6.1544 ± j0.1066 -3.0139 ± j2.0211
-0.6807 ± j2.9806 -1.1465 ± j0.3597
-1.8616 ± j2.4828 -0.6832 ± j0.0258
-1.3099 ± j0.1969
-0.7469 ± j0.0425
-0.5745 ± j0.0174

Table 9 Optimized PSS parameters.

JSA-tuned parameters TSA-tuned parameters
Gen No.
Kc T1 T3 Kc T1 T3
G2 17.85915 1.02731 0.97316 9.56016 0.45605 0.04899
G3 21.10907 0.95999 0.76108 29.6530 0.41534 1.02723
G4 31.67296 0.96694 0.77684 8.32946 0.52314 1.60714
G5 11.29446 0.15025 1.05456 19.9644 0.39170 0.11927
G6 10.49168 0.86949 1.31434 31.3544 0.15861 0.20616
G7 6.45810 0.50531 0.10901 5.55168 0.28128 1.01207
G8 16.19636 1.50054 1.04658 20.6615 1.10157 1.29015
G9 27.61241 0.37933 0.15423 30.6447 0.20029 0.41098
G10 57.23365 1.05829 0.77759 24.3617 0.94880 1.36906

(a)

(b) (c)

Fig. 20 (a) Eigenvalues for CPSS; (b) Eigenvalues for JSA; (c) Eigenvalues for TSA
 Fig. 21 The minimum damping ratio for different optimization techniques.

5.3.2 Time Domain Simulation Analysis

The JSA and TSA-based techniques are used to tune the PSS parameters. Bus 29 is subjected to a 3-𝜙 fault that
starts at 0.5s and lasts 0.1s for a particular base case. Response of the system is simulated for 8s after the fault.
Table 9 summarizes the JSA and TSA-optimized parameters. Fig. 22 shows the Angular frequency (t) of 𝐺3 when
the generator is equipped with CPSS, JSA, and TSA-based PSS. When the Angular frequency starts to oscillate
due to a fault at 0.5s, CPSS increases the oscillation with time as it provides a negative MDR for the IEEE-39 Bus
system. On the other hand, JSA stabilizes the angular frequency within 4.5s, and TSA stabilizes the signal within
5.1s. Rotor angle (t) follows the same trend as (t) = d/dt((t)) In Fig. 23, while JSA and TSA-tuned PSS damp-
out oscillation in the rotor angle of 𝐺3 within 5.2s, CPSS continues to grow it. Similarly, JSA and TSA-based PSS
only require 5.3s to fully stabilize the output power Pe, as shown in Fig. 24. Throughout this simulation, CPSS
produces an arbitrary control signal UPSS with a small magnitude, which is responsible for a gradual increase in
LFOs in G3. Unlike CPSS, JSA-based PSS generates control signals for 3.5s, and TSA-based PSS generates for 6s
to fully get the system back to its normal operating condition as shown in Fig. 25. From the time domain analysis,
it is clear that JSA and TSA provide significantly better damping of LFOs compared to CPSS. A similar response
can be seen for other generators, too.
 Fig. 23 Rotor angle response of 𝐺3 with conventional, JSA-
Fig. 22 Angular frequency response of 𝐺3 with
optimized, and TSA-optimized PSS.
conventional, JSA-optimized, and TSA-optimized PSS.

Fig. 24 Pe of 𝐺3 with conventional, JSA-optimized, and Fig. 25 Control signals of 𝐺3 with conventional, JSA-
TSA-optimized PSS. optimized, and TSA-optimized PSS.

6 Conclusions
The PSS is integrated with the synchronous generators of the power system networks for stability enhancement by
damping out the LFOs that occur due to various disturbances. This paper reported the two novel meta-heuristic
strategies (jellyfish search and tunicate swarm algorithms) for designing the PSS in single-machine infinite bus and
multimachine power networks (four-machine two-area and IEEE 39-bus networks). The robustness of the
employed strategies was demonstrated by converging the objective functions within a smaller number of iterations,
irrespective of the initial guess. Besides, the employed algorithms (JSA and TSA) achieved a better damping ratio
for the multimachine networks than the literature-reported algorithms (BSA and PSO). Moreover, both strategies
identified the optimal PSS parameters for the selected loading/operating conditions. Besides, time-domain
simulations (rotor angle, angular frequency, field voltage, internal voltage, and control signal) showed that the TSA
and JSA-tuned PSS can stabilize the systems much faster than the traditional fixed-gain PSS. Therefore, the JSA
and TSA-tuned PSS improved system stability against three-phase faults. However, the research can be further
extended in the future by integrating various FACTS devices and renewable energy sources into the system model.

Acknowledgment
The authors acknowledge the support provided by the Interdisciplinary Research Center for Renewable Energy and
Power Systems (IRC-REPS), Research Institute, King Fahd University of Petroleum and Minerals, through project
#INRE2315.

Funding
No funding was received for conducting this study.

Declarations

Conflict of interest: The authors declare that there are no conflicts of interest associated with this research study.

Ethical Approval: This research is the author’s original work and has not been published elsewhere. This paper is
not under consideration by another journal or conference proceedings.

References
[1] M.A. Abido, Robust design of multimachine power system stabilizers using simulated annealing, IEEE
Trans. Energy Convers. 15 (2000) 297–304.

[2] M.A. Abido, Power system stability enhancement using facts controllers: a review, Arab. J. Sci. Eng.
Sect. B Eng. 34 (2009).

[3] M.A. Abido, Y.L. Abdel-Magid, A hybrid neuro-fuzzy power system stabilizer for multimachine power
systems, IEEE Trans. Power Syst. 13 (1998) 1323–1330.

[4] A.H. Alharbi, S. Alkhalaf, Y. Asiri, S. Abdel-Khalek, R.F. Mansour, Automated Fruit Classification using
Enhanced Tunicate Swarm Algorithm with Fusion based Deep Learning, Comput. Electr. Eng. 108 (2023)
108657.

[5] H. Alkhatib, J. Duveau, Dynamic genetic algorithms for robust design of multimachine power system
stabilizers, Int. J. Electr. Power Energy Syst. 45 (2013) 242–251.

[6] Y. Cao, L. Jiang, S. Cheng, O.P. Malik, G.S. Hope, A nonlinear variable structure stabilizer for power
system stability, IEEE Trans. Energy Convers. 9 (1994) 489–495.

[7] C.L. Chen, Y.Y. Hsu, Coordinated Synthesis of Multimachine Power System Stabilizer Using an Efficient
Decentralized Modal Control (DMC) Algorithm, IEEE Trans. Power Syst. 2 (1987) 543–550.

[8] J.S. Chou, D.N. Truong, A novel metaheuristic optimizer inspired by behavior of jellyfish in ocean, Appl.
Math. Comput. 389 (2021) 125535.

[9] B. Dasu, M. Sivakumar, R. Srinivasarao, Interconnected multi-machine power system stabilizer design
using whale optimization algorithm, Prot. Control Mod. Power Syst. 4 (2019).

[10] A.M. El-Zonkoly, Optimal tuning of power systems stabilizers and AVR gains using particle swarm
optimization, Expert Syst. Appl. 31 (2006) 551–557.

[11] M. Farhat, S. Kamel, A.M. Atallah, B. Khan, Optimal Power Flow Solution Based on Jellyfish Search
Optimization Considering Uncertainty of Renewable Energy Sources, IEEE Access 9 (2021) 100911–
100933.

[12] R.J. Fleming, M.A. Mohan, K. Parvatisam, Selection of parameters of stabilizers in multimachine power
 systems, IEEE Trans. Power Appar. Syst. PAS-100 (1981) 2329–2333.

[13] G. Fusco, M. Russo, Nonlinear control design for excitation controller and power system stabilizer,
Control Eng. Pract. 19 (2011) 243–251.

[14] A. Ghasemi, H. Shayeghi, H. Alkhatib, Robust design of multimachine power system stabilizers using
fuzzy gravitational search algorithm, Int. J. Electr. Power Energy Syst. 51 (2013) 190–200.

[15] T. Guesmi, A. Farah, H.H. Abdallah, A. Ouali, Robust design of multimachine power system stabilizers
based on improved non-dominated sorting genetic algorithms, Electr. Eng. 2017 1003 100 (2017) 1351–
1363.

[16] A.M. Hasan, M.R. Islam, M. Shafiullah, M.B. Uddin, Multi-Verse Optimizer for LFO Damping in Single
Machine Infinite Bus Power Networks, 2022 Int. Conf. Adv. Electr. Electron. Eng. ICAEEE 2022 (2022).

[17] N.M.A. Ibrahim, B.E. Elnaghi, H.A. Ibrahim, H.E.A. Talaat, Performance Assessment of Bacterial
Foraging based Power System Stabilizer in Multi-Machine Power System, Int. J. Intell. Syst. Appl. 11
(2019) 43–53.

[18] Z. Jiang, Design of a nonlinear power system stabilizer using synergetic control theory, Electr. Power
Syst. Res. 79 (2009) 855–862.

[19] T.R. Jyothsna, K. Vaisakh, Multi-objective Evolutionary Programming Based Design of PSS, SVC, and
TCSC for Transient Stability Improvement, Proc. World Acad. Sci. Eng. Technol. 39 (2009) 859–863.

[20] S. Kaur, L.K. Awasthi, A.L. Sangal, G. Dhiman, Tunicate Swarm Algorithm: A new bio-inspired based
metaheuristic paradigm for global optimization, Eng. Appl. Artif. Intell. 90 (2020) 103541.

[21] A. Kazemi, M.R.J. Motlagh, A.H. Naghshbandy, Application of a new multi-variable feedback
linearization method for improvement of power systems transient stability, Int. J. Electr. Power Energy
Syst. 29 (2007) 322–328.

[22] A. Kharrazi, Tuning of power system stabilizer in Single Machine Infinite Bus (SMIB) using genetic
algorithm and Power Factory Modal Analysis, 2015 Australas. Univ. Power Eng. Conf. Challenges Futur.
Grids, AUPEC 2015 (2015).

[23] A.W. Khawaja, N.A.M. Kamari, M.A.A.M. Zainuri, Design of a Damping Controller Using the SCA
Optimization Technique for the Improvement of Small Signal Stability of a Single Machine Connected to
an Infinite Bus System, Energies 2021, Vol. 14, Page 2996 14 (2021) 2996.

[24] M. Klein, G.J. Rogers, P. Kundur, A fundamental study of inter-area oscillations in power systems, IEEE
Trans. Power Syst. 6 (1991) 914–921.

[25] P. Kundur, Power System Stability and Control, McGraw-Hill, Inc., Toronto, Ontario, 1994.

[26] P. Kundur, M. Klein, G.J. Rogers, M.S. Zywno, Application of power system stabilizers for enhancement
of overall system stability, IEEE Trans. Power Syst. 4 (1989) 614–626.

[27] L.L. Lai, H.T. Zhang, S. Mishra, D. Ramasubramanian, C.S. Lai, F.Y. Xu, Lessons learned from July
2012 Indian blackout, IET Conf. Publ. 2012 (2012).

[28] E. V. Larsen, D.A. Swann, Applying power system stabilizers Part I: General concepts, IEEE Trans.
Power Appar. Syst. PAS-100 (1981) 3017–3024.

[29] E. V. Larsen, D.A. Swann, Applying power system stabilizers, IEEE Trans. Power Appar. Syst. PAS-100
(1981) 3017–3046.

[30] A.A.P. Lerm, A.S. e Silva, PSS tuning via Hopf Bifurcation sensitivity analysis, in: 2003 IEEE Bol. Power
Tech Conf. Proceedings, IEEE, 2003: pp. 6-pp.

[31] V.A. Maslennikov, S.M. Ustinov, The optimization method for coordinated tuning of power system
regulators, in: Proc 12th Power Syst. Comput. Conf. PSCC, 1996: pp. 70–75.
[32] G. Naresh, M. Ramalinga Raju, M. Sai Krishna, Design and parameters optimization of multi-machine
power system stabilizers using artificial bee colony algorithm, 2012 Int. Conf. Adv. Power Convers.
Energy Technol. APCET 2012 (2012).

[33] P.W. Sauer and M.A. Pai, Power system dynamics and stability, Prentice Hall, New Jersey, 2006.

[34] M.I.H. Pathan, M.S. Shahriar, M.M. Rahman, M.S. Hossain, A. Nadia, M. Shafiullah, Comparative
analysis of machine learning approaches in enhancing power system stability, in: S. Padmanaban, P.
Sivaraman, C. Sharmeela, J.B. Holm-Nielsen (Eds.), Artif. Intell. Smart Power Syst., 1st ed., John Wiley
& Sons, Inc., 2023: pp. 159–180.

[35] D.K. Sambariya, R. Prasad, Design of PSS for SMIB system using robust fast output sampling feedback
technique, 7th Int. Conf. Intell. Syst. Control. ISCO 2013 (2013) 166–171.

[36] D.K. Sambariya, R. Prasad, D. Birla, Design and performance analysis of PID based controller for SMIB
power system using Firefly algorithm, 2015 2nd Int. Conf. Recent Adv. Eng. Comput. Sci. RAECS 2015
(2016).

[37] M. Shafiullah, M.A. Abido, A.H. Al-Mohammed, Metaheuristic optimization techniques, Power Syst.
Fault Diagnosis (2022) 27–68.

[38] M. Shafiullah, M.A. Abido, L.S. Coelho, Design of robust PSS in multimachine power systems using
backtracking search algorithm, 2015 18th Int. Conf. Intell. Syst. Appl. to Power Syst. ISAP 2015 (2015).

[39] M. Shafiullah, M.J. Rana, L.S. Coelho, M.A. Abido, A. Al-Subhi, Designing lead-lag PSS employing
backtracking search algorithm to improve power system damping, 2017 9th IEEE-GCC Conf. Exhib.
GCCCE 2017 (2018).

[40] A.M. Shaheen, A.M. Elsayed, A.R. Ginidi, E.E. Elattar, R.A. El-Sehiemy, Effective Automation of
Distribution Systems With Joint Integration of DGs/ SVCs Considering Reconfiguration Capability by
Jellyfish Search Algorithm, IEEE Access 9 (2021) 92053–92069.

[41] G. Shahgholian, A. Movahedi, J. Faiz, Coordinated Design of TCSC and PSS Controllers using VURPSO
and Genetic Algorithms for Multi-Machine Power System Stability, 13 (2015) 1–12.

[42] M.S. Shahriar, M. Shafiullah, M.I.H. Pathan, Y.A. Sha’aban, H.R.E.H. Bouchekara, M.A.M. Ramli, M.M.
Rahman, Stability improvement of the PSS-connected power system network with ensemble machine
learning tool, Energy Reports 8 (2022) 11122–11138.

[43] H. Shayeghi, A. Safari, H.A. Shayanfar, PSS and TCSC damping controller coordinated design using PSO
in multi-machine power system, Energy Convers. Manag. 51 (2010) 2930–2937.

[44] H. Shayeghi, H.A. Shayanfar, A. Akbarimajd, A. Ghasemi, PSS design for a single-machine power system
using honey bee mating optimization, in: Proc. Int. Conf. Artif. Intell. Las Vegas, USA, 2011: pp. 210–
216.

[45] Z. Sun, N. Wang, Y. Bi, D. Srinivasan, Parameter identification of PEMFC model based on hybrid
adaptive differential evolution algorithm, Energy 90 (2015) 1334–1341.

[46] C.T. Tse, S.K. Tso, Refinement of conventional pss design in multimachine system by modal analysis,
IEEE Trans. Power Syst. 8 (1993) 598–605.

[47] S. Ushkewar, S. Chaube, N. Komawar, N. Tirpude, M.S. Ansari, Controlled islanding scheme for power
system protection: Guidelines and approach: Case study: Proposed Bhopal islanding scheme, Proc. 3rd
IEEE Int. Conf. Adv. Electr. Electron. Information, Commun. Bio-Informatics, AEEICB 2017 (2017)
117–121.

[48] C. Ussawapusitgul, S. Kaitwanidvilai, Automatic design of robust PSS for multimachine power system
using PSO, IEEE Reg. 10 Annu. Int. Conf. Proceedings/TENCON (2011) 1065–1069.

[49] H. Verdejo, V. Pino, W. Kliemann, C. Becker, J. Delpiano, Implementation of particle swarm

optimization (PSO) algorithm for tuning of power system stabilizers in multimachine electric power
 systems, Energies 13 (2020).

[50] N.I.A. Wahab, A. Mohamed, A. Hussain, Fast transient stability assessment of large power system using
probabilistic neural network with feature reduction techniques, Expert Syst. Appl. 38 (2011) 11112–
11119.

[51] D. Xia, G.T. Heydt, Self-tuning controller for generator excitation control, IEEE Trans. Power Appar.
Syst. PAS-102 (1983) 1877–1885.

[52] B. Yang, M. Zhang, Z. Guo, P. Cao, J. Yang, G. He, J. Yang, R. Su, X. Huang, M. Zhu, H. Lu, D. Zhu,
Adaptive evolutionary jellyfish search algorithm based optimal photovoltaic array reconfiguration under
partial shading condition for maximum power extraction, Expert Syst. Appl. 215 (2023) 119325.

[53] H. Youssef, M.H. Hassan, S. Kamel, S.K. Elsayed, Parameter Estimation of Single Phase Transformer
Using Jellyfish Search Optimizer Algorithm, 2021 IEEE Int. Conf. Autom. Congr. Chil. Assoc. Autom.
Control (2021) 1–4.

[54] Y. Yu, Electric Power System Dynamics, Academic Press, New York, 1983.

Md. Rashidul Islam, Currently working as a faculty member at International Islamic

University Chittagong, Bangladesh. He received the M.Sc Engineering degree in Electrical
and Electronic Engineering from Japan. Prior to that, he completed his B.Sc Engineering
degree from Chittagong University of Engineering and Technology (CUET). His research
interest includes Power system operation and control, Integration of Renewable Energy with
microgrids, Optimization of complex energy systems, power electronics, and Electric vehicle
etc.
Md. Samiul Azam completed his B.Sc. in Electrical and Electronic Engineering from
International Islamic University Chittagong, Bangladesh, in 2022. He is currently working as
a Teaching Assistant at International Islamic University Chittagong. His research interests
include Power system stability improvement, grid integration of renewable energy resources,
power system control and stability, evolutionary algorithms, and machine learning techniques.

Md. Saber Hossen, Currently pursuing his M.Sc Engineering degree in Energy Technology
from Chittagong University of Engineering and Technology (CUET). He received his B.Sc
Engineering degree in Electrical and Electronic Engineering from International Islamic
University Chittagong. His research interest includes Power system operation and stability,
Control Lyapunov function, Stability of drones, power electronics, and Electric vehicle etc.

Mohammad Saiful Islam, Currently Studies at University of South Wales (USW), United
Kingdom. His MSc Subject is Renewable Energy and Sustainable Technology (Running). He
completed his B.Sc. Engineering degree from International Islamic University Chittagong
(IIUC), Bangladesh. His research interest includes Power system operation and control,
Integration of Renewable Energy with microgrids, Optimization of complex energy systems,
power electronics, and Electric vehicles, etc.
Muhammed Y. Worku (Member, IEEE) received the B.Sc. degree in electrical engineering
from Mekelle University, Ethiopia, in 2005, the M.Tech. degree in electrical engineering with
specialization in power electronics and electrical drives from the Indian Institute of Technology
Roorkee (IITR), India, in 2008, and the Ph.D. degree from the Electrical Engineering
Department, King Fahd University of Petroleum and Minerals (KFUPM), Saudi Arabia, in
2015., He is currently with the Interdisciplinary Research Center for Sustainable Energy
Systems (IRC-SES), Research Institute, KFUPM, as a Research Engineer II (Associate
Professor). He has published several papers in reputable journals and conferences. His research
interests include grid integration of renewable energy resources, energy storage applications to
power systems, microgrids, power electronics, and electrical drives.

Mohammad Shoaib Shahriar received the B.Sc. and M.Sc. degrees from the Islamic
University of Technology (IUT), Bangladesh, in 2009 and 2012, respectively, and the Ph.D.
degree in electrical power as a major and control systems as a minor from the EE Department,
King Fahd University of Petroleum & Minerals (KFUPM), Saudi Arabia, in 2018. He is
currently an Assistant Professor in the Electrical Engineering (EE) Department, University of
Hafr Al-Batin (UHB), Hafr Al-Batin, Saudi Arabia. He worked on monitoring, protection,
optimization, and stability of electrical power systems. His research interests include renewable
energy, drones and its application fields, and the use of machine-learning techniques in power
system applications.
MD SHAFIULLAH (Senior Member, IEEE) received the B.Sc. and M.Sc. degrees in electrical
and electronic engineering (EEE) from the Bangladesh University of Engineering and
Technology (BUET), Bangladesh, in 2009 and 2013, respectively, and the Ph.D. degree in
electrical power and energy systems from the King Fahd University of Petroleum and Minerals
(KFUPM), Saudi Arabia, in 2018. He was a Faculty Member with the Department of Electrical
and Electronic Engineering, International Islamic University of Chittagong (IIUC),
Bangladesh, from 2009 to 2013. He works as an Assistant Professor at the Control &
Instrumentation Department and a research affiliate at the Interdisciplinary Research Center
for Renewable Energy and Power Systems (IRC-REPS), KFUPM. His research interests
include grid fault diagnosis, grid integration of renewable energy resources, power quality
analysis, power system control and stability, evolutionary algorithms, and machine learning
techniques.

Declaration of interests

The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this paper.