Anovel hyprid AVR-LFC based on sliding mode

scheme for coupled multi-area power systems

Gafary Mahmoud Touria El Mezyani
Institute of Electronic Engineering College of Engineering, University of North Florida
Bilbies, Egypt 1-UNF drive, Jacksonville, FL
[email protected] [email protected]

Abstract— The significance of the AVR and the LFC distributed generation sources [1]. This complexity raises
systems in electrical systems cannot be overstated, as they significant issues of network performance, stability and
are essential for their operation, stability, and safety. These security [2]. A critical challenge is how generation facilities
control mechanisms are designed using both numerical and respond to fluctuations in active loads, where frequency
analytical techniques. Numerical methods, frequently harmonics propagate into multi-regional generation networks,
based on heuristics, vary in quality and optimization creating significant safety and stability issues. In addition, bus
methods, focusing on a single objective criterion that voltage deviations from set levels cause power losses that
balances various performance aspects like steady-state and affect the efficiency and stability of power generation units [3].
transient responses. This research proposes a (NSTW- The primary goal of Automatic Voltage Regulation (AVR) is
SMC) that accommodates a multi-area coupling power to deliver an optimized and consistent voltage reference
system (MACPS) which consists of multiple zones. This tracking function to reduce reactive losses. The topic of
innovative approach is employed to ascertain the control of regulating output voltage has been extensively explored in
an AVR-LFC system connected in a two-area electrical numerous research studies [4]. The Load Frequency Control
network, ensuring swift and stable regulation of AC (LFC) system manages fluctuations in the grid's primary
generator bus voltage and frequency under dynamic frequency in response to changes in active loads [ [5]
disturbances. Furthermore, a (FOSMC) is applied for the In a different context, the methods used to design AVR-
VI compensation system (VICS) to ensure the frequency LFC schemes often involve single-objective cost functions,
stability due to the lack of inertia of renewable energy which can result in conflicting optimization criteria. For
sources. This method surpasses traditional heuristics such instance, a shorter rise time might cause a higher percentage
as artificial bee colony, differential evolution, and particle overshoot when minimizing integrated squared tracking errors.
swarm optimization, as well as analytical approaches like Consequently, robust techniques are essential to enhance the
internal model control systems. efficiency of combined AVR-LFC systems based on multiple
optimization goals, addressing the nonlinearity and
Index Terms-- STW-SMC, MACPS, AVR-LFC, FOSMC. VICS uncertainties of power system networks [6].
Integrated AVR-LFC systems are crucial for ensuring the
Tab.1. The symbols and acronyms
reliability, security, and performance of interconnected power
Parameters Description
generation systems [7]. This study investigates a coupled
DER Distributed energy resources AVR-LFC system in which a linearized LFC loop is expanded
AVR Automatic Voltage Regulation. to incorporate an excitation system, and a damper winding is
LFC Load Frequency Control added to the AVR loop as presented in . In this context, a novel
AC Alternatting Current super twisting sliding mode control (NSTW-SMC) has been
STW Super twisting developed to determine the control effort of a combined AVR-
NSTW-SMC Novel super twisting sliding mode control LFC system for significant synchronous generators within a
VI Virtual inertia two-area power grid. This method demonstrated improved
SMC sliding mode control performance in terms of peak deviations, as confirmed by
MACPS multi-area coupling power system
Feed-back observer based SMC
research indicating that under dynamic disturbances, specific
FFOSC
VICS Virtual inertia compensation system interactions occur between the two AVR-LFC loops . This is
because AVR loops directly influence the power generation
voltage magnitude, as reported in . A combined AVR-LFC
I. INTRODUCTION
system utilizing SMC control is presented in [8], [9] examining
The structure of the electricity network is becoming
the mutual effects of the AVR and LFC loops, including the
increasingly complex and non-linear, especially in terms of
excitation system for a single-area power system.

1
 A hybrid controller based on super twisting sliding mode It can also improve the convergence rate and robustness of
control is proposed in [10] to study the interactions between the system, making it a promising control strategy for various
the LFC and AVR loops for a single-area power system using applications.Simulation results have shown that the proposed
a Simulink model. Practical Implementation of a Hydro Power NSTW-SMC control system assures better dynamic
Unit Active Power Regulation Based on an MPC Algorithm is performance and illustrates a significant reduction in
introduced in [11], outperforming a traditional Proportional- chattering compared to traditional control methods [23].
Integral (PI) controller in multiple performance criteria. The The STSMC has also been applied to SRM control strategy,
FGSPI approach showed better performance in terms of where it has been shown to eliminate the chattering
convergence stability, ease of implementation, and phenomenon, particularly for electric vehicle applications. In
computational efficiency. conclusion, the STSMC is a robust and efficient control
In [12], an AVR LFC system for a single-area power technique that can achieve adaptive tracking performance and
grid.method is utilized to ascertain the SMC control . This minimize chattering, making it a promising control strategy for
research has demonstrated high efficiency in minimizing various applications, including speed regulation of PMSM,
settling time, overshoot percentage, and oscillations. electro-pneumatic actuators, and unmanned aerial vehicles
Additional studies addressing the coupling between AVR-LFC (UAVs) [24].
regulation schemes are introduced in [13]. These advantages have been demonstrated in various studies,
Although there is a wide range of control designs in including [25], which compares the performance of STSMC
literatures, the unpredictability of power generation unit with other robust controls, such as PI and SM controllers, and
dynamics has affected their performance and robustness. For concludes that STSMC performances are better. [26] proposes
example, Heuristics such as nonlinear threshold acceptance an improved second-order STSMC strategy to enhance system
algorithm (NLTA), genetic algorithm (GA), bacterial feed stability and robustness. [27] uses STSMC to improve the
optimization (BFO), and particle swarm optimization (PSO) performances and robustness of a switched reluctance machine
have been proposed to optimize the PID control gain of AVR for an electric vehicle. [28] highlights the advantages of
control systems in [14]. STSMC, including its ability to adjust the control gains to
Alternative methods are presented, which include a teach- optimize the system's response to changing
learn controller, a fractional PID controller, and an H∞ control conditions. [29] discusses the windup effect of the super-
method with μ-analysis are introduced in [15]. twisting algorithm in the case of a saturated control input.
Recent research investigations have focused on customizing The paper is organized as follows: Section II introduces the
and developing innovative control mechanisms that can dynamic model of a combined AVR-LFC scheme for a multi-
withstand high levels of uncertainty in dynamic parameters, area power system and outlines the optimization problem's
ultimately achieving superior performance [16] [17] objectives. Section III presents the optimization problem's
Super-twisting sliding mode control (STSMC) is a modern solution using a single-objective cost function with the NLTA
control technique known for its capability to achieve adaptive algorithm. Section IV provides insights into the solutions of
tracking performance and minimize chattering [18]. It is an multi-objective optimization problems and the structure of the
extension of the traditional sliding mode control (SMC) MOSA approach. Section V analyzes the performance of the
method, which incorporates an integrator to smooth out abrupt MOSA mechanism, and Section VI concludes the paper with
signals, making it suitable for various systems, including speed some final remarks.
regulation of permanent magnet synchronous motors (PMSM),
electro-pneumatic actuators, and unmanned aerial vehicles II. MULTI-ZONE AVR-LFC POWER SYSTEM
(UAVs) [19]..The STSMC algorithm consists of a continuous
state-dependent term and a discontinuous term that switches A. The LFC system model
sign based on the system's state position [20]. The continuous
term ensures the system's trajectory spirals towards the origin, A power system that includes load frequency control (LFC)
while the discontinuous term guarantees robustness and consists of a rotating mass, a controller, a turbine, and load
stability by moving the system towards the target trajectory. requirements, as shown in Figure 1. The main purpose of the
The super-twisting algorithm in the STSMC is designed to LFC structure is to distribute dynamic load fluctuations
minimize the effects of uncertainties and disturbances, equally among different generators, maintain consistent
ensuring stability, fast convergence, and robustness towards frequency usage values and regulate route replacement
the intended trajectory [21]. It also avoids chattering and schedules. Frequency deviations in each area 𝒪 are represented
ensures smooth control action by adding a non-linear term that by changes in the generator rotor angle ∆𝛿(𝒪) which are
smoothes transitions between various control actions.The expressed and amplified to form the corresponding real power
STSMC has several advantages over other control techniques. correction 𝑃𝑇𝑟 . This correction signal tells the power unit to
It is simple to implement and can reduce chattering while adjust the torque. The LFC dynamics is described using three
maintaining the same performance and robustness as standard differential equations that show the relationship
traditional SMC methods [22]. between the controller, turbine and generator/load units. These

2
equations for a single-region electrical system 𝒪 are expressed χT = [ΔF ΔPgov ΔPTr ΔPδ ΔPω ΔPsol ]𝑇
as follows [19], [30]. (10)
Δ𝑃𝑇𝑟 − Δ𝑃𝑙 = 2 𝐻 (Δ𝐹). 𝑠 + 𝐷(Δ𝐹) (1) u = [ΔPω∗ ΔPsol∗ ΔPI ΔP𝑅 ]𝑇
Where, Δ𝑃𝑇𝑟 is the power change of the generator, ΔPℓ is the
𝑢̅d = [ΔPvi ]𝑇 (11)
load power change, 𝐹 represents the frequency of the power
system, ΔF is the deviation of system frequency, 𝐻 is the The state space-model (3)-(10) can be presented as follows:
system inertia, and 𝐷 is the system damping parameter. The
governor's dynamic behavior, represented as ∆𝑃𝑔𝑜𝑣(𝒪) , is 𝜒̇
𝐿𝐹𝐶
influenced by the LFC control signal 𝒰(𝒪) and a negative −𝛽 1 1 1 1
feedback loop with a gain of 1/𝑅, which a ctuates the 0
2𝐻 2𝐻 2𝐻 2𝐻 2𝐻
governor's valve openings as depicted in Figure 1. The
1 1
governor's actuation function is dependent on the LFC control − − 0 0 0 0
gains and the speed regulation term 𝑅, with the coupled load 𝑅𝒯𝑔𝑜𝑣 𝒯𝑔𝑜𝑣
frequency deviations serving as inputs. The LFC control signal
= 1 1 𝜒
𝐿𝐹𝐶
𝒰(𝒪) , which is the output signal of the NSTW-SMC controller, 0 − 0 0 0
𝒯𝑇𝑟 𝒯𝑇𝑟
is responsible for regulating the frequency variation ∆𝐹(𝒪) 1 1 0 0
caused by load adjustments within each power system area 𝒪, 0 0 − 1 0
0 0 𝒯𝛿 𝒯𝛿 −
as well as controlling the other coupled frequency deviations
𝒯𝜔 1
∆𝐹𝒴 for ∀𝒴 ∈ 𝑁𝒪 . This signal plays a crucial role in 0 0 0 0 −
[ 0 0 0 𝒯𝑠𝑜𝑙 ]
maintaining the stability and reliability of the power system by
adjusting the generator output power to match the load
+
demand. The frequency for Area 𝒪 with penetration of DERs,
VICS, and primary and secondary controllers shown in Figure
1, can be defined as follows:
̄ 1 −1
0 − 2𝐻 0
1
0 2𝐻 −1
2𝐻. 𝑠 𝛥𝐹 = 𝛥𝑃𝑇𝑟 + 𝛥𝑃𝜔 + 𝛥𝑃𝑠𝑜𝑙 + 𝛥𝑃𝑣𝑖 − 𝛥𝑃ℓ − 𝐷𝛥𝐹 (2) 2𝐻
0 0 0 0 0 𝒯𝑔𝑜𝑣
𝑠. 𝒯𝑇𝑟 𝛥𝑃𝑇𝑟 = −𝛥𝑃𝑇𝑟 + 𝛥𝑃𝑔𝑜𝑣 (3) 0 0 0 0 0
𝜐+ 0 𝑢̄ 𝑑 (12)
0 0 0 0 0 0
1
0 0 0 −1 0 0
𝑠. 𝒯𝑔𝑜𝑣 𝛥𝑃𝑔𝑜𝑣 = 𝑢̄ 𝑑 − 𝛥𝐹 − 𝛥𝑃𝑔𝑜𝑣 (4) 1
𝑅 0 0 2 𝒯w − 2𝒯 ] [ 0 ]
[0 0 𝑠𝑜𝑙
𝐿𝐹𝐶
𝒰(𝒪) = 𝑢̅̇𝑑 (5)
Consequently, the second-order sliding mode control is
𝑠. 𝒯𝜔 𝛥𝑃𝜔 = 𝛥𝑃𝜔∗ − 𝛥𝑃𝜔 (6) suitable to ensure the system is on an established sliding
function.
𝑠. 𝒯𝑠𝑜𝑙 𝛥𝑃𝑠𝑜𝑙 = 𝛥𝑃𝑠𝑜𝑙 ∗ − 𝛥𝑃𝑠𝑜𝑙 (7) On the further point of view, A multi-zone AVR-LFC
electrical system consists of multiple zones with their own
𝛥𝑃ℓ = 𝑃𝐼 + 𝑃𝑅 (8) automatic voltage regulator and load frequency control
mechanisms. These systems are essential for regulating
𝐿𝐹𝐶
Where 𝑢̄ 𝑑 is the control input, 𝒰(𝒪) is the new switching voltage and frequency, especially with the increasing use of
control signal produced by the proposed controller, 𝛥𝑃𝜔 is the renewable energy sources and the complexty of electrical
energy established by the wind system, 𝛥𝑃𝑤∗ is the initial systems [31]. The dynamic structure of the combined AVR-
influence change of the wind power system, 𝛥𝑃𝑠𝑜𝑙∗ is the initial LFC system for a two-layer multi-region power system. The
power variation of photovoltaic systems, 𝛥𝐸𝑠𝑜𝑙 is the developed lower layer is responsible for regulating the frequency
power from the solar system, 𝑃𝐼 is the commercial/industrial deviations caused by active load disturbances and controlling
load demand, 𝑃𝑅 is the residential load demand, 𝒯𝑖 is the inertia the output voltages of the generation units using LFC and AVR
time constant of the element i, and 𝛥𝑃𝑣𝑖 is the energy of the VI systems [32]. The second layer reflects the dynamic energy
system. Assume an enlarged state factor ΔPδ ∈ Rn in (9), and exchange between different regions of the power system or the
let's consider the state variables χT , and the input vectors 𝑢̅d dynamics of highway linesynchronization between regions.
This structure is discussed in detail and the dynamic energy
s. 𝒯δ Δ𝑃δ = ΔPTr − Δ𝑃δ (9) exchange between regions is emphasized.

3
 LFC System

Control
AREA1
∆𝐹(𝒪)

Loads
Primary
Commercia ∆𝑃
𝐼 ∆𝑃𝑅 Residenti
l/Industrial 𝚺
Loads
+ + al Loads
𝑩𝓞
Droop 𝟏
CONTROLLER 𝑹
𝐿𝐹𝐶 ∆𝑃𝑙 Inertia & Damping
𝒰(𝒪)
Limiter -
𝐿𝐹𝐶
𝒦𝑆𝑀𝐶 + ∆𝑃𝑇𝑟
𝚺 න⬚ 1 ∆𝐹(𝒪)
𝚺 Governor Turbin 𝚺 +𝚺
∆𝑃𝑔𝑜𝑣 e + 2𝐻. 𝑠 + 𝐷
STW-SMC 𝑢𝑑
̅̅̅ ∆𝑃𝑣𝑖
∆𝑃𝑤 + Virtua
FOSC
∆𝑃𝑇𝑖𝑒 ∆𝑃𝑤∗ Wind l Inertia

∆𝑃𝑠𝑜𝑙∗ PV 𝚺 𝒦1 1
𝑠
∆𝑃𝑠𝑜𝑙

∆𝑃𝑠𝑜𝑙∗ PV
∆𝑃𝑤∗ 𝒦2
Win ∆𝑃𝑤
d
∆𝑃𝑣𝑖 𝒦4 𝒦3
𝒜𝒱ℛ 𝒱𝑎(𝒪) 𝒱𝑔(𝒪)
𝒰(𝒪) 𝒱𝒪
+

𝒱𝒪∗ 𝑢𝑑
̅̅̅ 𝒦𝑒 𝒦𝓃
𝒜𝒱ℛ 𝒦𝑎
𝒦𝑆𝑀𝐶 𝒦5
𝚺 න⬚ 𝚺 1 + 𝒯𝑎 1 + 𝒯𝑒
𝚺 𝚺+ 1 + 𝒯𝓃
+ 𝒱𝑒(𝒪) - Generator Field
Exciter ∆𝑃𝑙
Amplifier
STW-SMC
Droop
Commercia
AREA2

l/Industrial ∆𝑃𝐼 + ∆𝑃𝑅

+ Residenti
Loads
Primary 𝚺 al Loads
Loads
Control

AVR System

Figure 1The block diagram of a coupled AVR-LFC scheme for area 𝓞.

PHYSICAL CONNECTION

4
VIRTUAL CONNECTION
 B. AVR System model solve. This coupling scheme reflects the
interdependencies between the voltage and frequency
The goal of an Automatic Voltage Regulator (AVR) variables in each area, and the dynamical behavior of
system is to minimize reactive power losses caused by the governor, which is influenced by the LFC control
1
discrepancies between the desired voltages and the signal and a negative feedback loop with a gain of .
𝑅
exciter terminal voltage 𝒱𝑒(𝒪) in a specific area 𝒪. The LFC control signal, which is the output of a SMC,
When there are fluctuations in the reactive power load regulates the frequency variation and coupled
of a generator, the terminal voltage 𝒱𝑒(𝒪) changes, frequency deviations within each power system area,
which is detected using a single-phase potential as well as the other coupled frequency deviations in
transformer (voltage 𝒱𝒪 ). This measured voltage is the system. The dynamical behavior of the governor
compared to a desired reference voltage 𝒱𝒪∗ . This is and the LFC control signal, along with the
achieved by adjusting the exciter field to maintain interdependencies between the voltage and frequency
desired voltage levels and minimize reactive power variables, contribute to the increased dynamical
loss due to differences between actual and desired challenges in the multi-objective optimization
voltage. The SMC controller compares the sensed problem.
voltage 𝒱𝒪 with the desired reference voltage 𝒱𝒪∗ and
amplifies the error signal to adjust the exciter field and D. The coupling AVR-LFC scheme
adjust the exciter pole voltage, resulting in oscillations within a multi-area power system.
in the generator field. current and induced emf. This
process is depicted in Figure 1. Interaction in a multi-regional power system
1
𝒱̇𝑔(𝒪) = (𝒦𝑛(𝒪) 𝒱𝑔(𝒪) + 𝒱𝑒(𝒪) ) (13) involves dynamic exchanges between various
𝒯𝑛(𝒪)
1 components and regions, such as power transmission
𝒱̇𝑒(𝒪) = (𝒦ℯ(𝒪) 𝒱𝒶(𝒪) + 𝒱𝒶(𝒪) ) (14)
𝒯ℯ(𝒪) and regulation. These interactions impact the system's
1
𝒱̇𝒶(𝒪) = 𝒜𝒱ℛ
(𝒦𝒶(𝒪) 𝒰(𝒪) + 𝒱𝒶(𝒪) ) (15) stability and reliability, necessitating the study of
𝒯𝒶(𝒪)
1 complex power systems to develop effective control
𝒱̇𝒪 = (𝒦𝒮(𝒪) 𝒱ℊ(𝒪) + 𝒱𝒪 ) (16)
𝒯𝒮(𝒪) strategies. In a multi-zone power system, interactions
Therefore, the purpose of the SMC (Sliding include controller dynamics, LFC control signals, and
Mode Control) in the AVR system is to adjust the AVR-LFC switching patterns, affecting frequency
generator output voltage 𝒱ℊ(𝒪) to the desired voltage deviations and synchronization among regions. Active
level 𝒱𝒪∗ for each region. load fluctuations in one region can create interference
C. Compound LFC-AVR Scheme for others, leading to unwanted harmonics and energy
exchange. Considering these interactions and energy
The AVR and LFC systems are weakly coupled,
exchanges is crucial for planning and operating multi-
allowing for separate control designs for voltage and
region power systems effectively.
frequency variables in each area 𝒪. However, a
The dynamic expression of the tie-line power
coupled AVR-LFC scheme exists for each area, as
exchange in each region 𝒪 is described as.
shown in Figure 1. In this scheme, the AVR system
regulates the output voltage𝒱𝒪 for each area 𝒪 to the 𝒩
𝛥𝒫𝑡𝑖𝑒(𝒪) = ∑𝒦=1,
𝒪
𝒞(𝒪𝒦) (∆𝐹(𝒪) − ∆𝐹(𝒦) ) (17)
desired levels. The relationship between the small 𝒦≠0
changes in the stator emf on the real electric power 𝒦2 ,
Where 𝒩𝒪 denotes to the quantity of generation
the small adjustments in rotor-angle on the terminal
regions in the vicinity of zone 𝒪. The synchronization
voltage 𝒦3 ,, the small changes in the stator emf on the
system in a multizone power system involves dynamic
rotor angle 𝒦4 , and the small variations in rotor angle
equations representing electrical exchanges between
on the stator emf 𝒦5 , is depicted in this coupling
regions. The system forms a graph with weights based
scheme. The weak coupling enables separate control
on synchronization coefficients 𝒞(𝒪𝒦) , acting as an
designs, but the coupling scheme highlights the
interdependencies between the voltage and frequency agreement mechanism. Frequency offsets represent
variables in each area. The inclusion of the coupling regions, and synchronization speed is influenced by
AVR-LFC scheme for each area in the multi-objective coefficients determining connections' strength.
optimization problem is expected to increase its Coefficients impact system dynamics, potentially
dynamical complexity, making it more challenging to affecting stability and reliability. Studying these

5
factors is crucial for understanding complex power 2. Determine 𝐿 such that: 𝑑𝑒𝑡( 𝜆𝐼 − (𝐴 − 𝐿𝐶)) = 0
system behavior and developing control strategies. has roots in the left half-plane and places the poles
to achieve acceptable observer performance.
III. CONROLLER DESIGN PROCEDURES Oobserver poles in the complex plane are enabled
if the system is fully observable. Connect the
A. Secondary phase control scheme observer to the full-state feedback law using 𝑢 =
The sliding surface is considered as follows:
−𝐾𝑥̑ (𝑡) where 𝑥̑ (𝑡) is the estimate of 𝑥(𝑡) The
𝛿(𝑡) = 𝛼𝛸(𝑡) ∈ 𝑅𝑛 (18)
With 𝛼 = 𝑑𝑖𝑎𝑔{𝛼1 , 𝛼2, . . . . . . . . . . . , 𝛼𝑛 } ∈ 𝑅𝑛×𝑛 . In the whole compensator function can be given
following, the time representation of the sliding surface consistently with input 𝑢(𝑡) and output 𝑦(𝑡) as
𝛿(𝑡) will be 𝛿. The STW-based sliding mode controller (25):
(STW-SMC) can be defined as: 00 = −𝛥𝑃 + 𝛥𝑃𝑔𝑜𝑣 (25)
𝐿𝐹𝐶
𝒰(𝒪) = 𝜎̄1 ||𝛿||1/2 𝑠𝑖𝑔𝑛(𝛿) − 𝛯̄,
𝛯̄ = −𝜎̄2 𝑠𝑖𝑔𝑛(𝛿). (19) IV. STABILITY ANALYSIS
Let's consider, 𝜑̄ 1 = 𝛿 and 𝜑̄ 2 = 𝛿̇ parameter used in
the design steps of the controller with:
𝜑̄ 1 = 𝜑̄ 2
𝜑̄̇ 2 = 𝜓̄ + 𝜍̄𝑢̄𝑠𝑡
𝑢̅̇𝑑 = 𝑢̄ 𝑠𝑡 (20)
Assumption i: Assume that 𝜓̄ and 𝜍̄ have the
bounds defined in (17) and (18).
where 𝜓̃̄ > 0 and 𝜍̄̃ > 0 are constants. [19]. The
proposed virtual inertia control scheme is represented
as shown in Figure 2. Figure 2. FFOSC virtual inertia control block diagram.
|𝜓̄| ≤ 𝜓̃̄ (21) When the system in (12) touches the sliding surface
|𝜍̄| ≤ 𝜍̄̃ (22) in a finite time 𝑡𝑟𝑒 ≥ 0, then.
0 = Δ𝑃Tr + Δ𝑃ω + Δ𝑃sol + Δ𝑃vi − ΔPℓ − 0 (23)
1 𝛥𝑃𝑔𝑜𝑣 = −𝜎̄2−1 (𝛼1 𝛥𝐹 + 𝛼3 𝛥𝑃𝑇𝑟 + 𝛼4 𝛥𝑃𝛿 +
0 = ū d − ∗ 0 − Δ𝑃gov (24)
R 𝛼5 𝛥𝑃𝜔 + 𝛼6 𝛥𝑃𝑠𝑜𝑙 ) (26)
Assumption ii Assuming a parameterΔEℓ , control Based on (23) and the model dynamics equations in
effort ū d and state (ΔF= 0, ΔPTr ΔPgov , ΔPω , ΔPsol , (12), the equivalent system on the sliding phase is
ΔPδ ) that fulfills (23), and (24). given as follows:

Where 𝜎̄1 ,𝜎̄2 are the control parameters, which can 𝛥𝐹̇ = 𝐶𝛥𝐹
−1
(𝛥𝑃𝑇𝑟 + 𝛥𝑃𝜔 + 𝛥𝑃𝑠𝑜𝑙 + 𝛥𝑃𝑣𝑖 − 𝛥𝑃ℓ −
be defined as in [18]. 𝐻𝛥𝑃𝑠𝑜𝑙 (27)
B. Stabilizing controller for virtual inertia 𝛤𝑇𝑟 𝛥𝑃̇𝑇𝑟 = −𝛼2−1 𝛼1 𝛥𝐹 + (𝛼2−1 𝛼3 − 1)𝛥𝑃𝑇𝑟 +
Due to the absence of inertia of the system, an 𝛼2 𝛼4 𝛥𝑃𝛿 + 𝛼2−1 𝛼5 𝛥𝑃𝜔 +
−1
FFOSC is applied to ensure the frequency stability. In 𝛼2−1 𝛼6 𝛥𝑃𝑠𝑜𝑙 (28)
general, battery systems (BS), and inverters, with
appropriate control, are the main components used in 𝛤𝛿 𝛥𝑃𝛿̇ = 𝛥𝑃𝑇𝑟 − 𝛥𝑃𝛿 (29)
the virtual inertia control concept where 𝚫𝐅, 𝚫𝐏𝐓𝐫 ,
𝚫𝐏𝛚 , 𝚫𝐏𝐬𝐨𝐥 , 𝚫𝐏𝐯𝐢 , 𝚫𝐏𝓵 , 𝚫𝐏𝛅 , and 𝐮̄ 𝐝 are the upper 𝛿=0 (30)
bound standards of 𝚫𝐅, 𝚫𝐏𝐓𝐫 , 𝚫𝐏𝛚 , 𝚫𝑷, 𝚫𝐏𝐯𝐢 , 𝚫𝐏𝓵 ,
𝚫𝑷𝛅 and 𝐮̄ 𝐝 respectively where 𝐶𝛥𝐹 = 1/2𝐷
Assumption iii (Selection of sliding function)
Theorem 1. Under assumptions (i-iii), the sliding
Assume 𝜶𝟏 , 𝜶𝟐 , . . . . . . , 𝜶𝒏 are positive constants.
function (18), and the proposed (STW-SMC), the
The proposed control system for VI can be defined
model dynamics (12) will asymptotically downward to
as in Figure 2:
the sliding function 𝛿 = 0. 𝛿 = 0 with zero steady-
1. Define 𝐾such that 𝑑𝑒𝑡( 𝜆𝐼 − (𝐴 − 𝐵𝐾)) = 0 has
state frequency error (i.e.,𝛥𝐹 = 0).
roots in the left half-plane and places the poles to
meet the control design conditions. The capability Proof 1: The following function is considered for
to identify the poles arbitrarily in the complex the system dynamics (23-26).
plane is guaranteed if the system is controllable.
6
 1 plants have an installed capacity of 7.5MW. However,
𝑓(𝛥𝐹, 𝛥𝑃𝑔𝑜𝑣 , 𝛥𝑃𝑇𝑟 , 𝛥𝑃𝛿 , 𝛥𝑃𝜔 , 𝛥𝑃𝑠𝑜𝑙 ) = 𝛥𝐹 𝑇 (2𝐻)𝛥𝐹 − 1
2 the renewable energy sources lack inertia and export
1 1
+ 𝛥𝑃𝑇𝑟 𝛼1 𝛼2 𝛤𝑇𝑟 𝛥𝑃 + 𝛥𝑃𝛿𝑇 𝛼1−1 (𝛼2 + 𝛼3 )𝛤𝛿 𝛥𝑃𝛿
𝑇 −1
controllers, which can significantly reduce the
2 2
system's stability.
(31)
Battery Energy Storage System (BESS)
The renewable-based DERs (𝛥𝑃𝜔 and𝛥𝑃𝑠𝑜𝑙 ) and the To maintain the stability and reliability of the
load (𝛥𝑃𝑅 and𝛥𝐸𝑖𝑛𝑑 ) doesn't contribute to the electrical system in the face of dynamic disturbances,
frequency deviation regulation. These DERs and loads a battery energy storage system (BESS) with a rated
are considered as additional disturbances to the power of 4.5MW is integrated into the multi-area
dynamics of the system. Therefore, the Bregman coupling power system (MACPS). This BESS serves
distance approach can be used to prove the system's as a crucial component in ensuring the swift and stable
stability. It's necessary to demonstrate the existence of regulation of AC generator bus voltage and frequency,
asymptotic system stability if and only if the derived particularly in the presence of renewable energy
of the Bregman distance formula is negative. The sources that lack inertia.The BESS is designed to
Bregman distance accompanying with 𝛿 is distinct as mitigate the risks associated with thermal runaway and
[33] we should demonstrate that the derivative of the cell failures, which are critical safety concerns in
Bregman distance equality (31) fulfills 𝛢̇𝐵𝑟𝑔 ≤ 0 energy storage systems. By incorporating advanced
along with the system dynamics (24-27) solutions. safety features and robust control mechanisms, the
BESS enhances the overall safety and reliability of the
𝜕𝛢𝐵𝑟𝑔 𝜕𝛢𝐵𝑟𝑔 electrical system, enabling it to withstand dynamic
𝛢𝐵𝑟𝑔 = | (𝛥𝐹 − 0) + | (𝜕𝑃𝑇𝑟 −
𝜕𝛥𝐹 𝑦=𝑦̄ 𝜕𝑃𝑇𝑟 𝑦=𝑦̄
disturbances and maintain stable operation across
𝜕𝛢𝐵𝑟𝑔
𝜕𝑃𝑇𝑟 ) + | (𝜕𝑃 − 𝜕𝑃𝛿 ) (32) multiple zones within the MACPS.
𝜕𝑃𝛿 𝑦=𝑦̄
System Load
𝛢̇𝐵𝑟𝑔 = 𝛥𝐹 𝑇 𝑉̄ 𝛥𝐹 − (𝛥𝑃𝑇𝑟 − 𝛥𝑃𝛿 )𝑇 (−𝛼1−1 (𝛼2 + The plant base rated is 15MW. The STW-SMC
𝛼3 ))(𝛥𝑃𝑇𝑟 − 𝛥𝑃𝛿 ) (33) elements are σ̄_1=3 and σ̄_2=6, and the FFOSC
controller parameters are k=[2 2] and l=[59.5 29.5] T.
Where, 𝑉̄ = [𝛼
⃡ − 2𝐻 0.5𝛼⃡] With 𝛼
⃡ > 2𝐻 → The system load, as depicted in Figure .3, exhibits
0.5𝛼
⃡ 1 significant fluctuations over the entire operating
𝑉̄ > 0 → 𝛢̇𝐵𝑟𝑔 ≤ 0. period of 1000 seconds. This variation includes both
internal load perturbations and external uncertainties
𝛥𝐹 = 0 can be proved if the Bregman function (32) that affect the system's performance. According to the
has a local lowest at (𝛥𝐹 = 0, 𝛥𝑃𝑇𝑟 , 𝛥𝑃𝛿 ) [19-22]. information provided, the highest value of the system
Then, it is adequate to check that (28) is convex at the load is approximately 0.6 per unit (pu), which occurs
states(𝛥𝐹 = 0, 𝛥𝑃𝑇𝑟 , 𝛥𝑃𝛿 ). Since the Hessian matrix at the 800-second mark during the 1000-second.
𝐻𝑚 (𝛥𝐹 = 0, 𝛥𝑃𝑇𝑟 , 𝛥𝑃𝛿 ) assessed at (0, 𝛥𝑃𝑇𝑟 , 𝛥𝑃𝛿 ).
Further analysis illustrates that:

𝑈 0
𝐻𝑚 (𝛢𝐵𝑟𝑔 ) = [ ] (34)
0 𝑈̄
where 𝑈 > 0 and 𝑈̄ > 0. Hence 𝐻𝑚 (𝛢𝐵𝑟𝑔 ) > 0,
which concludes proof 1. In the same way we can
𝐴𝑉𝑅
design and analyse the AVR controller 𝒰(𝒪) for area
𝒪.

V. SIMULATION RESULTS
Thermal Power Plant and Renewable Energy
Sources
The thermal power plant has a fixed capacity of
15MW, generated from a conventional synchronous
machine. This power is distributed to 10MW industrial
plants and 5MW residential loads. The wind plant has Figure 3. The load profile.
an installed capacity of 8.5MW, and the solar sub-
7
 The analysis reveals that the system load undergoes This comparison highlights the effectiveness of the
significant fluctuations and disturbances throughout proposed method in maintaining frequency stability
the 1000-second operating period, with a peak value and minimizing oscillations in the electrical system.
of approximately 0.6pu occurring at the 800-second
mark. These variations, encompassing both internal
load perturbations and external uncertainties, can pose
substantial challenges for maintaining the stability and
reliability of the electrical system.

Figure 4 illustrates the power generated by the

Distributed Energy Resources (DERs). This power is
considered as an additional perturbation to the system,
with the following contributions:

 Solar power: ΔP_sol≈0.35pu from t=100 seconds to

t=800 seconds.

 Wind power: ΔP_win≈0.45pu from t=350 seconds Figure 4. The nominal output power of DERs.
to t=1000 seconds. These perturbations can further
impact the stability and reliability of the electrical
system, particularly in the presence of the renewable
energy sources with limited inertia.

Figure 5 compares the battery power dissipation

between the improved simulated method, the FFOSC
technique, and the suggested methodology. The graph
reveals the following key points:

 The battery power consumption is significantly

lower for the suggested control method
compared to the other approaches.

 The peak battery power dissipation levels are: Figure 5. The battery energy consumption.

 Suggested control method: ΔP_vi≈0.15

 FFOSC control: ΔP_vi≈0.25

 Improved virtual control method:

ΔP_vi≈0.35

The comparison in Figure 5 demonstrates the

superiority of the suggested control method in
reducing battery power dissipation. This efficiency is
critical for the optimal performance and extended
lifespan of the battery energy storage system (BESS)
integrated into the electrical system.

Figure 6 compares the frequency deviation

performance of various control methodologies. The
results indicate that the proposed method: Figure 6. The frequency deviation comparison.

 Exhibits the highest level of robustness Figure 7 compares the tie-line power exchange
between coupled areas under the proposed control
 Displays the smallest overshoot levels during method and an optimized-parameter PID control
transient stages method from a reference source.The key points are:

8
 Despite the presence of disturbances and analytical methods like internal model control
unsymmetrical load distribution between the two systems.
areas, the proposed control method maintains the tie-
line power exchange between the coupled areas at very VII. REFERENCES
small values. This indicates that the proposed control
method is able to ensure a stable operation throughout
the entire power system, even in the face of
[1] Z. Yang, F. Yang, H. Min, H.Tian, W.Hu,
challenging conditions such as disturbances and
J.Liu, "Review on optimal planning of new
unbalanced load distribution. The superior
power systems with distributed generations and
performance of the proposed control method in
electric vehicles," Energy Reports, vol. 9, no.
regulating the tie-line power exchange highlights its
2352-4847, pp. 501-509, 2023.
effectiveness in maintaining the overall stability and
reliability of the interconnected power system..
[2] P. E. Muñoz, R. J. Mantz and S. A. González,
"Control-based Fault Current Limiter for
Minimizing Impact of Distributed Generation
Units on Protection Systems," Journal of
Modern Power Systems and Clean Energy, vol.
11, no. 2, pp. 643-650, March 2023.

[3] S. A. Adegoke , Y. Sun, "Power system

optimization approach to mitigate," Cogent
Engineering, vol. 10, no. 1, p. 2153416, 2023.

[4] Y. Liu et al.,, "A Multimode Wide Output

Range High-Voltage Power Supply for
Magnetrons," IEEE Transactions on Industrial
Electronics, vol. 70, no. 11, pp. 11153-11162,
Nov. 2023.
Figure 7. Tie-Line power exchange between coupled
areas. [5] C. Li et al, "Continuous Under-Frequency
Load Shedding Scheme for Power System
VI. CONCLUSION Adaptive Frequency Control," IEEE
The Automatic Voltage Regulator (AVR) and Load Transactions on Power Systems, vol. 35, no. 2,
Frequency Controller (LFC) systems are crucial pp. 950-961, March 2020.
components of electrical systems, ensuring their stable
operation and safety. These control mechanisms are [6] N. Nahas, M. Abouheaf, A. Sharaf and W.
designed using a combination of numerical and Gueaieb, "A Self-Adjusting Adaptive AVR-
analytical techniques. This research proposes a novel LFC Scheme for Synchronous Generators,"
approach, the Non-Standard Twin-Widening Sliding IEEE Transactions on Power Systems, vol. 34,
Mode Controller (NSTW-SMC), to control an AVR- no. 6, pp. 5073-5075, Nov. 2019.
LFC system in a two-area electrical network. This
innovative method is designed to regulate the AC [7] F. Yang et al., "Fractional-Order Sliding
generator bus voltage and frequency quickly and Mode Load Frequency Control and Stability
stably under dynamic disturbances in a multi-area Analysis for Interconnected Power Systems
coupling power system (MACPS) consisting of With Time-Varying Delay," IEEE
multiple zones. Additionally, the research employs a Transactions on Power Systems, vol. 39, no. 1,
Frequency-Oriented Sliding Mode Controller pp. 1006-1018, Jan. 2024.
(FOSMC) for the Voltage Inverter Compensation
System (VICS) to maintain frequency stability in the [8] J. Gui, Z. Sun, Y. Wen, D. Tao and J. Ye, "A
presence of renewable energy sources, which lack Review on Generative Adversarial Networks:
inertia. This approach outperforms traditional Algorithms, Theory, and Applications," IEEE
heuristics such as artificial bee colony, differential Transactions on Knowledge and Data
evolution, and particle swarm optimization, as well as

9
 Engineering, vol. 35, no. 4, pp. 3313-3332, [16] J. Fei, Z. Wang, X. Liang, Z. Feng and Y.
April 2023. Xue, "Fractional Sliding-Mode Control for
Microgyroscope Based on Multilayer
[9] F. Yang, D. Huang, D. Li, S. Lin, S. M. Recurrent Fuzzy Neural Network," IEEE
Muyeen and H. Zhai, "Data-Driven Load Transactions on Fuzzy Systems, vol. 30, no. 6,
Frequency Control Based on Multi-Agent pp. 1712-1721, June 2022.
Reinforcement Learning With Attention
Mechanism," IEEE Transactions on Power [17] J. Wang and X. Xing, "Identification of
Systems, vol. 33, no. 6, pp. 5560-5569, Nov. Continuous-Time Dynamic Systems With
2023. Uncertainties Measured by Fuzzy Sets Subject
to Model Structure Errors," IEEE Transactions
[10] A. Nurettin and N. İnanç, "High-Performance on Fuzzy Systems, vol. 32, no. 5, pp. 3293-3300,
Induction Motor Speed Control Using a Robust May 2024.
Hybrid Controller With a Supertwisting Sliding
Mode Load Disturbance Observer," IEEE [18] G. Mahmoud, Y. Chen, L. Zhang and M. Li,
Transactions on Industrial Electronics, vol. 70, "FHOSM Tolerant Control for Networked
no. 8, pp. 7743-7752, Aug. 2023. Control," Int. J. Control Autom. Syst, vol. 19,
no. 9, pp. 3049-3061, 2021.
[11] M. Beus and H. Pandžić, "Practical
Implementation of a Hydro Power Unit Active [19] G. Mahmoud, Y. Chen, L. Zhang and M. Li,
Power Regulation Based on an MPC "Sliding Mode Based Nonlinear Load
Algorithm," IEEE Transactions on Energy Frequency Control for Interconnected Coupling
Conversion, vol. 37, no. 1, pp. 243-253, March Power Network," International Journal of
2022. Control, Automation and Systems, vol. 20, no.
11, pp. 3731-3739, 2022.
[12] Gupta, A., Chauhan, A., & Khanna, R.,
"Design of AVR and ALFC for single area [20] G. Mahmoud, Y. Chen, L. Zhang, M. Li,
power system including damping control," in "Finite-time Fault Tolerant Control for
Recent Advances in Engineering and Uncertain-Robot Mechanism Using Adaptive-
Computational Sciences (RAECS), Chandigarh, Gain SMC," in 6th International Conference on
India, 2014. Robotics and Automation Engineering
(ICRAE), Guangzhou, China, 2021.
[13] E. Rakhshani, K. Rouzbehi and S. Sadeh, "A
New Combined Model for Simulation of [21] G. Mahmoud, T. El Mezyani, "Resilient
Mutual Effects between LFC and AVR Loops," Interconnected Power-Grid Control Using
in Asia-Pacific Power and Energy Engineering Sliding Mode Approach," in 2024 IEEE Texas
Conference, Wuhan, China, 2009. Power and Energy Conference (TPEC), Texas,
2024.
[14] Y. -S. Cheng, H. -Y. Pai and Y. -H. Liu, "PID
parameter tuning for flyback converter with [22] L. Xu, Y. Chen, M. Li, L. Zhang, G.
synchronous rectification using particle swarm Mahmoud, "Extended observer‐based hybrid
optimization," in Joint 12th International tracking control strategy for networked system
Conference on Soft Computing and Intelligent with FDI attacks," Asian Journal of Control,
Systems and 23rd International Symposium on vol. 25, no. 4, pp. 3092-3104, 2023.
Advanced Intelligent Systems (SCIS&ISIS), Ise,
Japan, 2022,, 2022. [23] G. Mahmoud, Y. Chen, L. Zhang, M. Li,
"Sliding Mode Control for Networked
[15] M. Xiao, B. Tao, W. X. Zheng and G. Jiang, Markovian Jump Systems with Disturbances,"
"Fractional-Order PID Controller Synthesis for in China Automation Congress (CAC), Beijing,
Bifurcation of Fractional-Order Small-World China, 2021.
Networks," IEEE Transactions on Systems,
Man, and Cybernetics: Systems, vol. 51, no. 7,
pp. 4334-4346, July 2021.

10
[24] L. Zhang, G. Mahmoud, M. Li, Y. Chen, [32] C. H. N. S. Kalyan and G. S. Rao, "Impact of
"Robust Sliding Mode Controller with communication time delays on combined LFC
Suppressor for DC Motor System with and AVR of a multi-area hybrid system with
Disturbances," in Chinese Automation IPFC-RFBs coordinated control strategy,"
Congress (CAC), Hangzhou, China, 2019. Protection and Control of Modern Power
Systems, vol. 6, no. 1, pp. 1-20, Jan 2021.
[25] Y. Chen, L. Xu, G. Mahmoud, "Sliding mode
control for direct current motor system with [33] S. Trip, M. Cucuzzella, C. De Persis, A. van
multi-channels and external disturbances," der, "Non-linear sliding mode based load
Journal of Systems and Control Engineering, frequency control for power," IEEE
vol. 236, no. 4, pp. 742-753, 2022. Transactions on Control Systems Technology,
vol. 27, no. 5, pp. 1893-1906, Sept. 2019.
[26] J. Song, Z. Wang, Y. Niu, J. Hu and D. Yue,,
"First-and Second-Order Sliding Mode Control
Design for Networked 2-D Systems Under
Round-Robin Protocol," IEEE Transactions on
Cybernetics, pp. 1 - 14, January 2024.

[27] L. Feng, X. Sun, D. Guo, M. Yao and K. Diao,

"Advanced Torque Sharing Function Strategy
With Sliding Mode Control for Switched
Reluctance Motors," IEEE Transactions on
Transportation Electrification, vol. 10, no. 1,
pp. 2302-2311, June 2023.

[28] G. A. Ghazi, E. A. Al-Ammar, H. M.

Hasanien, W. Ko, "Circle Search Algorithm-
Based Super Twisting Sliding Mode Control
for MPPT of Different Commercial PV
Modules," IEEE Access, vol. 12, pp. 33109 -
33128, March 2024.

[29] X. Yang, X. Xiong, Z. Zou and Y. Lou,

"Semi-Implicit Euler Digital Implementation of
Conditioned Super-Twisting Algorithm With
Actuation Saturation," IEEE Transactions on
Industrial Electronics, vol. 70, no. 8, pp. 8388-
8397, Aug. 2023.

[30] H. Bevrani, M. R. Feizi, and S. Ataee,

"Robust frequency control in an islanded
microgrid: H1 and msynthesis approaches,"
IEEE Trans. Smart Grid, vol. 7, no. 2, p. 706–
717, Mar. 2016.

[31] P. K. Pathak, A. K. Yadav, S. Padmanaban

and I. Kamwa, "Fractional Cascade LFC for
Distributed Energy Sources via Advanced
Optimization Technique Under High
Renewable Shares," IEEE Access, vol. 10, pp.
92828-92842, Aug. 2022.

11