Electric Power Systems Research 214 (2023) 108898

Contents lists available at ScienceDirect

Electric Power Systems Research

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

Robust linear parameter varying frequency control for islanded hybrid AC/
DC microgrids
Abbas Aff a, Mohsen Simab a, *, Mehdi Nafar a, Alireza Mirzaee b
a
Department of Electrical Engineering, Marvdasht Branch, Islamic Azad University, Marvdasht, Iran
b
Department of Electrical Engineering, Dariun Branch, Islamic Azad University, Shiraz, Iran

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

Keywords: The applicability of a robust linear parameter varying (LPV) control for frequency fluctuation damping in an
Frequency regulation islanded hybrid microgrid (IHMG) system (DEG/WTG /PV/FC/ESSs) is investigated in this paper. Due to the
Hybrid AC/DC model erratic behavior of the majority of the renewable energy sources (RESs) and the varying time nature of loads,
Islanded microgrid
frequency deviation from the nominal value is inevitable, especially in the islanded mode. Load variation, solar
Linear matrix inequality (LMI)
Linear parameter varying (LPV) model
irradiation, and wind power disturbance, as well as system parametric uncertainties, can significantly disturb the
Parametric uncertainties system frequency operation. In this paper, the wind speed and rotated speed of the nonlinear wind turbine are
Renewable resources highlighted as scheduling parameters, and the nonlinearity of the wind turbine is hidden by the LPV approach. In
Robust control the proposed method, a dynamic output feedback controller is obtained based on the power balance equation, by
Stochastic disturbance solving three LMIs to minimize the upper bound of the L2 -Norm of the uncertain IHMG in an algorithmic
approach. For better evaluation, the ultimate LPV control approach is compared with two other methods on the
understudy IHMG model. The result represents a noticeable advantage of the proposed robust method compared
to other methods in terms of frequency stability based on time-domain and norms characteristics in presence of
the simultaneous disturbances and parametric uncertainties.

efficient, and reliable energy for the consumers. The hybrid AC/DC of­
1. Introduction fers both AC microgrid advantages such as utilizing the present AC grid
infrastructure and DC microgrid benefits such as improving the effi­
In the last two decades, the growing penetration of microgrids into ciency, size, and energy cost, because of fewer power conversion stages
traditional power systems promises to play an important role in future [8,9].
power generation applications. Increasing employment of various For the satisfactory stable operation of IHMG, developing improved
renewable energy sources (RESs) brings plentiful, sustainable, reliable, control strategies is vital to frequency regulation. The variable and un­
inexpensive power to home and industrial consumers. Integrating certain nature of RESs and loads have led to a significant dynamic
renewable energy sources (RESs) like biomass, photovoltaic solar arrays, characteristic variation in the system. One of the most substantial issues
and wind farms, geothermal with various energy storage systems as with IHMGs is wind power variations which depend on wind speed, can
primary frequency control [1, 2], and a reserve diesel engine generator cause severe deviations in the frequency of the IHMG, and can even lead
as secondary frequency control can play an effective role to meet the to a breakout [10]. Therefore, efficient control algorithms for handling
energy necessity of the connected loads in an islanded hybrid microgrid these uncertainties, disturbances, and dynamic variations are required.
(IHMG) [3,4]. In the next subsection, the research necessity and motivation are
Microgrids are power distribution architectures comprising RESs represented.
that can be operated in a managed, controlled way either grid-connected
or disconnected mode, which is called the islanded mode [5,6]. There 1.1. Research motivation
are three kinds of microgrids considering their structures and voltage
dynamics: DC, AC, and hybrid AC/DC microgrids that are composed of For having a stable frequency operation in a microgrid system, power
DC microgrids and AC microgrids [7]. The integration of these resources consumption and production must be matched persistently and instan­
can be performed through a single bus system that must offer stable, taneously as much as possible [11]. Several factors influence power

* Corresponding author.
E-mail address: msimab@miau.ac.ir (M. Simab).

https://doi.org/10.1016/j.epsr.2022.108898
Received 18 July 2022; Received in revised form 10 October 2022; Accepted 11 October 2022
Available online 28 October 2022
0378-7796/© 2022 Elsevier B.V. All rights reserved.
A. Aff et al. Electric Power Systems Research 214 (2023) 108898

Nomenclature Bs The damping of the transmission

β Blade pitch angle
IHMG Islanded hybrid microgrid V nom Rotational speed at maximum power for base wind speed
ESS Energy storage system RES Renewable energy sources
FC Fuel cell SEMS Superconducting magnetic energy storage
PV Photovoltaic UC Ultra-capacitors
WTG Wind turbine generator LMI Linear Matrix Inequality
MT Microturbine GA Genetic algorithm
L2 2-Norm LFT Linear fractional transformation
Fρ Parameter variation set
The IHMG parameters R DG speed regulation
TG Governor time constant T BESS Time constant of Flywheel energy storage system
δ Ramp rate limits T FESS Time constant of Battery energy storage system
D Damping coefficient K PV Time constant of solar photovoltaic
M Inertia constant K AE Gain of the aqua electrolyzer
T FC Fuel cell constant Ka The ratio of net power to PWTG
T AE Gain of the aqua electrolyzer K BESS Gain of Battery energy storage system
K FC Gain of the fuel cell K FESS Gain of Flywheel energy storage system
T DEG Diesel generator time constant Bs The damping of the transmission
The wind turbine’s model parameters Bg Intrinsic generator damping
Ks The stiffness of the transmission Jg Generator inertia
Jg Generator inertia R The length of the rotor blades
Jr Rotor inertia Pnom The nominal mechanical output power
ρ the air mass density ci s Performance coefficient of the turbine
kr,V Rotor torque - wind speed gain τ Delay time
Pw base The base value of wind speed Pe base The base power of the electrical generator

balancing in microgrids. Disturbances in a microgrid can cause fre­ secondary frequency control in IHMGs. In [20] a conventional PI/PID
quency deviations from the desired value and reduce the quality of the controller is tuned by Firefly Algorithm (FA) and Cuckoo Search Algo­
power supply. For an acceptable power system operation, the microgrid rithm (CSA) to regulate frequency in IHMG with a hybrid electric
frequency should remain close to its desired value and any fluctuations vehicle. In [21], WTG’s impact on the frequency deviation of the IHMG
must be damped quickly. Therefore, in every hybrid microgrid (HMG), is studied. A quasi-oppositional harmony search (QOHS) optimization
the principal control objective is disturbance rejection which leads to method is employed to tune the SMES and PID controller gains. Also,
frequency stability. It is notable that in the islanded mode, unlike the Fuzzy logic and numerous meta-heuristic methods as optimization
grid-connected mode, the power imbalance is not solved by the up­ procedures are more relevant in the literature. In [22], a fuzzy logic grey
stream power grid and it might cause serious frequency instability and wolf optimization (GWO)-PSO-based PI controller is represented. In
power quality problem [12,13]. The IHMG operation essentially suffers [23] an optimal PI control strategy is designed based on the constrained
from load-demand power imbalance and its frequency may undergo population extremal optimization. (Algorithm 1)
abrupt oscillations due to the small-time constants and low inertia of the The robust control strategies due to formulating uncertainties in the
employed RESs [14]. Therefore, a robust frequency control design is controller synthesis procedure, deal with the issue systematically, and
challenging due to the variable dynamics of RESs and their parametric therefore have a higher capability to manage uncertainties in a wide
uncertainties. range, unlike classical control procedures that can solely attain a satis­
factory tradeoff between nominal and robust performance. The sto­
chastic nature of the load, solar radiation, and wind speed profiles, make
1.2. Literature review the use of robust control procedures, preferable to other methods. Based
on the research contribution, in the following, the preceding robust
Numerous studies have been conducted on handling uncertainties to control studies in IHMGs are highlighted.
regulate frequency and power in HMGs. One of the most applicable In, some recent research on robust control applications for various
traditional frequency regulation methods is dedicated to the islanded microgrid systems is presented. In this paper, only load
proportional-integral-derivative (PID) control approaches and the spe­ disturbance is considered an irregular wind turbine, and solar irradia­
cial P, PI, and PD control cases that are directly or indirectly combined tion disturbances are neglected in the design procedure. A decentralized
by other methods. the basic frequency fluctuation minimization problem robust frequency control approach for an islanded MG that comprises
using a PI controller in an IHMG that included several energy storage several RESs with the same number of robust controllers for each RES is
systems such as SMES, BESS, FESS, and UC is studied in [15]. studied in [24]. μ-based robust control strategies for secondary fre­
Among applied nonlinear control approaches in IHMGs, sliding quency regulation in islanded microgrids are presented in [25–27].
mode control is more prominent although the feedback linearization A robust H∞ control approach design in the presence of parametric
method is also studied [16,17]. The optimized sliding mode load fre­ uncertainties and load disturbances for a hybrid microgrid is performed
quency controller is designed for the diesel generator system to regulate and compared with a standard PID controller in [28]. In another
system frequency through the secondary frequency control loop [18]. In research, for adjusting the initial frequency in stand-alone microgrids, a
[19] the nonlinear sliding surface is introduced to guarantee the IHMG robust H∞ multivariable control design is suggested [29].
state’s convergence to the origin in the finite time that is based on In [30], a robust μ-synthesis controller based on Linear fractional
fractional power terms in a surface. transformation (LFT) in an AC MG is employed. Robust mixed H2 /H∞
Intelligent control techniques are employed to deal with the

2
A. Aff et al. Electric Power Systems Research 214 (2023) 108898

Fig. 1. The frequency control problem in the IHMG.

control is applied on IHMG to performance and robustness enhancement is characterized as scheduling parameters for the LPV synthesis
in the presence of various uncertainties [31–33]. A mixed H2 procedure.
/H∞ controller based on static output feedback is studied in [31]. A
multi-objective energy sharing-frequency regulation H2 /H∞ control 1.4. The main contribution and novelty
approach for interconnected AC microgrids is investigated [34]. In [32],
a robust controller based on the coefficient diagram method is proposed. To overcome the shortcomings of the aforementioned published
An application of H∞ PID controller for regulating frequency in an IHMG works, this paper addresses a robust LPV control approach. As a new
is studied [33]. The authors of [35], propose a robust H∞ frequency contribution to earlier studies, it explores the use of an LPV controller
control for a hybrid AC/DC microgrid as a master/slave framework. based on a polytopic IHMG model to achieve minimum frequency de­
An LMI-based H∞ control method to regulate the DC bus voltage viation. The substantial contributions can be briefly outlined as follows:
fluctuation for an islanded MG is presented in [36]. To handle un­
certainties and nonlinearities, a class of nonlinear stochastic differential 1 A polytopic model is proposed for the IHMG system composed of
equations is defined in which parameter uncertainties appear in the DEG, PVs, WTs, FC, FESS, BESS, and loads to design the proposed
system and controller coefficients. Unlike [36], our proposed approach LPV controller. With such an LPV model, the uncertainty and sto­
considers nonlinearities differently. The key point is hiding the non­ chasticity of PVs, WTs, and loads can be considered mathematically.
linearities via the LPV approach to reach an infinite set of linear systems 2 The robust control methods reported in the preceding published
in a convex hull. Another different aspect of the suggested approach is papers related to IHMG frequency control have made use of H∞ , H2 /
that the upper bounds of external and internal disturbances errors are H∞ or μ-synthesis like [30–33] ignore nonlinearities in wind tur­
formulated in the form of LMIs. bines, while the unique character of the proposed method is model
It is notable that in the vast majority of the aforementioned research, nonlinearities considerations as scheduling parameters in the design
IHMG dynamics are considered linear and with linear robust methods procedure which makes it suitable and efficient for operating in a
try to solve frequency regulation issues. Due to considering nonlinear broad operating range.
terms in the dynamic model of IHMG such as wind speed, in addition to 3 Compared with robust H∞ controller and optimal PID controller, the
the parameters considered in the aforementioned robust methods, the proposed LPV controller attains a smaller objective function
proposed LPV approach, provides a more effective control synthesis magnitude, faster convergence, and more efficient control perfor­
structure for IHMG systems. It suggests a less complex dynamic output mance under several load-disturbance scenarios.
feedback controller with lower order than the IHMG model and hence it 4 Due to the dynamic nature of the suggested output feedback
is possible for practical implementation. controller, it can handle disturbances and parameter uncertainties
persistently.
1.3. The necessity of research
The next section is dedicated to clarifying the frequency control
Isolated microgrids have been broadly established particularly for problem in a general form.
distant and sensitive areas consisting of several low inertia RESs with
nonlinear and complex dynamics. The IHMG is isolated from the power 1.5. The robust LPV frequency control approach at a glance
grid and feeds its loads utilizing its dispersed generations. One of the
challenging issues in this mode is the discussion of frequency regulation. The main goal of this research is to diminish or reduction of the
Due to the small grid size, parameter variations, and the weather- disturbance’s effects on the IHMG frequency deviation based on a robust
dependent nature of RESs like wind turbines and photovoltaic arrays LPV control approach as it is shown in Fig. 1.
in standalone operation, the IHMG frequency might lead to abrupt In this structure, there are three inputs. One control input (ΔuDEG ) as
fluctuations. The robust LPV control approach is one of the most effi­ a secondary control and two input disturbances caused by solar irradi­
cient candidates to tackle IHMGs [14]. ation for photovoltaic arrays (ΔuPV ) and wind speed for wind turbine
Three distinguished disturbances that are common in HMGs are wind generators (ΔuWTG ), and one output disturbance as the load variation
turbine power, photovoltaic solar power, and load variations. Further­ (ΔPL ). Exploiting a robust control approach is one of the main ap­
more, the IHMG model has a simplified approximation of the real one proaches to handling these disturbances with intermittent natures.
and mostly its parameters are not deterministic. As a result, the necessity
of a satisfactory robust control approach is required. In the literature, 1.6. Structure
there is a lack of systematic robust control strategies for IHMGs to tackle
these issues effectively. In this paper, for better handling the wind tur­ After the introduction section, the rest of the study is presented in the
bine disturbance, the wind speed and rotated speed of the wind turbine subsequent sections. The nonlinear model of the IHMG dynamics and its

3
A. Aff et al. Electric Power Systems Research 214 (2023) 108898

Fig. 2. The general scheme of an IHMG.

equivalent LPV model are presented in Section 2. In Section 3, the electric loads are connected through a common bus bar [37]. As seen,
closed-loop structure of the proposed algorithm with the controller the DEG, PV, WT, FC, BESS, and FESS sources are connected to the bus
design procedure is stated along with the performance indices. In Sec­ bar utilizing AC/DC or DC/AC converters. The load power PL and
tion 4, the proposed LPV control scheme assessment and the simulation- generated power PS equitation is depicted in Fig. 2 and is determined by
based performance investigation under several disturbance patterns and the (1).
parameter uncertainties scenarios are declared. Finally, the current
∑
n ∑
m
shortcomings and concluding remarks for future studies are addressed in PL = P S = PWTGi + PPVi − PAE + PDEG + PFC +PFESS +PBESS , for t
Section 5. i=1 i=1

= ti
2. The investigated IHMG and its LPV state space representation (1)

2.1. The IHMG model description where, PWTGi and PPVi are the output power from ith wind turbine
generator, and ith photovoltaic array, respectively. PAE is the aqua
The concept of an IHMG induces a combination of numerous RESs electrolyzer power. PDEG and PFC are the output power of diesel gener­
using a coordinated control framework to regulate the entire system ators, and fuel cells. Also PFESS , and PBESS are the exchanged power of
output frequency under different operating points and uncertainties. FESS, and BESS, respectively. It is notable that ti is the ith definite time
Fig. 2 represents the schematic diagram of an IHMG investigated in this for the WTGi s or PVi s to connect to the system.
research. The aforementioned IHMG system offers a nonlinear and complex
RESs in the IHMG include wind turbine (WT), photo voltaic (PV) dynamic with several uncertainties. Since, usually most parts of the
solar cells, diesel engine generator (DEG), Fuel Cell (FC), Battery Energy system, change slightly during normal operation time, the linearization
Storage System (BESS), Flywheel Energy Storage System (FESS), and technique can be employed to characterize the IHMG dynamics around

4
A. Aff et al. Electric Power Systems Research 214 (2023) 108898

Fig. 3. The detailed model of an islanded HMG.

the normal operating condition.

Table 1
The IHMG model is composed of several subsystems, which are
Br (θ) and kr,V (θ) values for four vertexes of the LPV model [39].
connected to the single bus bar as shown in Fig. 3. The model has the
control input (ΔuDEG ) that its task is the frequency regulation as a sec­ Vertexes: θ = (V, ω) Br (θ) kr,V (θ)
ondary control level by the DEG. The IHMG microgrid parameters are θ1 = (3, 1.36) 2.4472 × 10 4
1.7052 × 104
listed in Appendix [38]. θ2 = (3, 2.25) 0 0
Due to RESs and load intermittent variation nature, fixed-parameters θ3 = (12, 1.36) − 4.6453 × 105 3.0030 × 104
(12, 2.25) 1.5438 × 105 1.2584 × 105
control synthesis may not have satisfactory performance for real sys­ θ4 =

tems. Hence, the IHMG model should be enhanced by considering

parametric uncertainties. Before defining the IHMG LPV representation,
in the next section, the LPV model for a wind turbine generator is
illustrated.

2.2. The wind turbine generator LPV model

In the low wind speed operating region, the pitch angle is constant
(β=0), and there are two variables in the scheduling parameter θ =
[ VW ω ]T where VW is the wind speed and ω is the turbine rotated
speed. The state variables of the wind turbine are defined as x =
[ ]T
θs ωr ωg where θs is the torsion angle, ωr is the rotor speed, and
ωg is the generator speed. The control input is u = ωz where ωz is the
zero-torque speed. Therefore, the wind turbine LPV model is defined as
(2).
⎡ ⎤ Fig. 4. The scheduling parameter box for the wind turbine model.
0 1 − 1
⎢ Ks Bs + Br (θ) Bs ⎥ [ ]T
⎢− − ⎥
⎢
A(θ) = ⎢ Jr Jr Jr ⎥ k (θ)
⎥, Bv (θ) = 0 r,VJr 0 , 1.36 rad
s
≤ ω ≤ 2.25 rad
s
, the range of wind speed is 3 ms ≤ VW ≤ 12 ms. The
⎢ ⎥
⎣ Ks Bs Bg + Bs ⎦ parameters of the turbine model are given in Appendix [39].
Jg Jg
−
Jg To calculate Br (θ) and kr,V (θ) Table 1 is presented.
Thus, in the next subsection, a comprehensive LPV model is illus­
[ ]T trated for the IHMG considering RESs and loads, and the wind turbine
(2)
Bg
B(θ) = 0 0 Jg , Ct = [ Ks Bs − Bs ], C = [ 0 0 1 ], D = 0. power nonlinearity.

The range of turbine rotated speed in the low-speed field is

5
A. Aff et al. Electric Power Systems Research 214 (2023) 108898

Fig. 5. The closed-loop structure of the proposed approach.

2.3. The LPV state-space representation for the investigated IHMG ⎛ ⎞ ⎡ ⎤

⎡ ⎤ 0
KG 0 0 0 ⎢ ⎥
⎜ ⎟
Due to the disturbances, nonlinearities, and uncertainties, LPV ⎢ ⎥ ⎜ ⎟ ⎢
⎢ 0 ⎥
⎥
⎢ TG ⎥ ⎜ 0 0 0 ⎟
frameworks are successful approaches for modeling power systems for ⎢
⎢ 0 ⎥
⎥ ⎜
⎜
⎟
⎟
⎢
⎢ 1 ⎥
⎥
⎢ ⎥
control synthesis purposes. The significant feature of LPV control syn­ ⎢
⎢
⎥
⎥
⎜ 0
⎜
0 0 ⎟
⎟ ⎢ ⎥
⎢ M ⎥
thesis is the inherent handling of the parametric uncertainties through ⎢ 0 ⎥
⎢ ⎥
⎜
⎜ 0 0 0 ⎟
⎟ ⎢ ⎥
⎢ KBESS ⎥
defining scheduling parameters in LPV model structures [40]. ⎢
⎢ 0 ⎥
⎥ ⎜
⎜
⎟
⎟ ⎢
⎢
⎥
⎥
⎢ ⎥ ⎜ 0 0 0 ⎟ T
LPV model of the IHMG system can be reached by defining θ(t) ⎢ ⎥ ⎜ ⎟ ⎢ BESS ⎥
B u = ⎢ 0 ⎥ , Bi = ⎜ ⎟, Bo = ⎢ ⎥
⎢ KFESS ⎥, (3)
= [ VW ω ]T . The lower and upper bounds are [ 3 1.36 ]T ≤ θ(t) = ⎢
⎢
⎥
⎥
⎜ KPV
⎜
0 0 ⎟
⎟ ⎢ ⎥
⎢ 0 ⎥ ⎜ ⎟ ⎢ T ⎥
[ θ1 θ2 ]T ≤ [ 12 2.25 ]T . the parameter box based on θ(t) is depicted in ⎢ ⎥ ⎜ 0 0 0 ⎟ ⎢ FESS ⎥
⎢ ⎥
⎢ ⎥ ⎜ ⎟ ⎢ 0 ⎥
Fig. 4. ⎢ 0 ⎥
⎢ ⎥
⎜
⎜ 0 kr,V (θ) ⎟
⎟ ⎢ ⎥
0 ⎟ ⎢ 0 ⎥
The resultant LPV state-space model of the IHMG is defined as fol­ ⎢
⎢
⎥
⎥
⎜
⎜ Jr ⎟ ⎢ ⎥
⎢ 0 ⎥ ⎢ ⎥
lows: ⎣ ⎦
⎜
⎜
⎟
⎟
Bg ⎠
⎢ 0 ⎥
⎢ ⎥
⎝ 0 0 ⎣ ⎦
[ ] [ ][ ] 0
ẋ A Bu Bi Bo x Jg 0
Gρ : = ,
y C2 0 I 0 u
[ ]T
Where x = ΔXG ΔPDEG Δf ΔPBESS ΔPFESS ΔPPV Δθs Δωr Δωg ,
and u = [ ΔuDEG di do ], that control input ΔuDEG is the first element
⎛ ⎞
1 KG of u. di denotes uncontrollable input disturbances [ ΔuPV ΔVW Δωz ]T ,
⎜− T 0 − 0 0 0 0 0 0
⎜ G RTG ⎟
⎟ and do represents a stochastic load variation w = ΔPL with normal
⎜ ⎟
⎜ KDG 1 ⎟ probability distribution function. Therefore, three matrices Bu , Bi and Bo
⎜ 0 0 0 0 0 0 0 ⎟
⎜T − TDG ⎟ are related to control signal ΔuDEG , input disturbance di =
⎜ DG ⎟
⎜ ⎟
⎜ 1 D 1 1 ⎟ [ ΔuPV ΔuWTG ]T , and output disturbance do = ΔPL , respectively.
⎜ 0 − 0 0 0 0 ⎟
⎜
⎜ M M M M ⎟
⎟ The next section shows the robust LPV approach based on the L2
⎜
⎜ 1
⎟
⎟ norm as the IHMG frequency regulation solution.
⎜ 0 0 0 − 0 0 0 0 0 ⎟
⎜ TBESS ⎟
⎜ ⎟
⎜
A= ⎜ 1 ⎟
⎟ 3. The L2 norm robust control approach
⎜ 0 0 0 0 − 0 0 0 0 ⎟
⎜ TFESS ⎟
⎜
⎜
⎟
⎟ The frequency control objective needs to be formulated in terms of
⎜ 0 1 ⎟
⎜ 0 0 0 0 − 0 0 0 ⎟ input/output disturbances and the control signal. H∞ control technique
⎜ TPV ⎟
⎜
⎜ 0
⎟
⎟ is mainly valuable for this disturbance rejection problem by shaping the
⎜ 0 0 0 0 0 0 1 0 ⎟ closed-loop frequency response to meet the desired design requirements.
⎜ ⎟
⎜
⎜ 0 0 0 0 0 0 −
Ks Br (θ)+Bs
−
Bs ⎟
⎟ the proposed control strategy is a straightforward H∞ control extension
⎜
⎜ Jr Jr Jr ⎟ ⎟ to LPV systems and a branch of L2 -norm-based approaches, which is a
⎜ ⎟
⎝ Ks Bs Bs +Bg ⎠ suitable approach for handling uncertainties in the IHMG system. In
0 0 0 0 0 0 −
Jg Jg Jg general form, the L2 -norm of a system from input d to output e is the
maximum gain of Hρ or the largest amplification of L2 input signals over
all frequencies, admissible parameter trajectories, and input/output
directions. ‖ Hρ ‖∞ is characterized as follows:

‖ e ‖2
‖ Hρ ‖∞ = sup
‖ d ‖2 (4)
d ∈ L2 , ρ ∈ Fρ

6
A. Aff et al. Electric Power Systems Research 214 (2023) 108898

Fig. 6. The proposed control algorithm procedure.

where, ‖ . ‖2 is the 2-norm index, e and d belong to the L2-norm space, [

Ac (θ) Bc (θ)
]
and Fρ is an admissible parameter set. Gc : (7)
Cc (θ) Dc (θ)
The IHMG model in the form of Gρ is specified as (5) which eu and er
represent input and output disturbance errors. These error variables are The Generalized control loop for the frequency control problem is
related to the control signal u = UDEG , desired microgrid frequency (fr ), demonstrated in Fig. 5.
and the measured frequency value (fmeas ) as eu = Wu .u, er = (fmeas − fr + The ultimate control law minimizes the objective function J. The Gc
do ).Wr . It is notable that Wu and Wr are design parameters for weighting controller matrices can be attained by solving a convex optimization
the control effort and the frequency fluctuations from the desired value problem to reach the feasible values of X, Y, and γ based on Theorem 1
on the frequency regulation issue. The weights for the current design [42,43].
are: Wu = I and Wr = 0.2I. Theorem 1. Given the HMG system in (1) and the performance level
⎡ ⎤ ⎡ ⎤⎡ ⎤ γ > 0, the output feedback control matrices are achievable, if and only if
ẋ A(θ) Bi (θ) Bo Bu x there exist positive definite matrices X(θ) ∈ C1 (RS , Sn×n ), Y(θ) ∈ C1 (RS ,
⎢ eu ⎥ ⎢ C11 D11 D12 0 ⎥ ⎢ ⎥
Gρ : {⎢ ⎥=⎢ ⎥⎢ di ⎥ (5) Sn×n ), such that:
⎣ er ⎦ ⎣ C12 D21 D22 I ⎦⎣ do ⎦
⎡ ⎤
y C2 0 I 0 u T T 1
φ
⎢ X (θ) − B B
u u X(θ)C 11 ([ B i (θ) B o ] − B u [ D 21 D 22 ]) ⎥〈
γ
The structure (5) can be attained by scaling techniques and feed­ ⎢
⎢
⎥
⎥
⎢ ⎥ 0 φX
forward looping [41,42]. ⎢ − I
1
[ D D ] ⎥
⎢ ∗ 11 12 ⎥
The objective function J is attained by induced L2-norm from d to e ⎣ γ ⎦
and it is characterized by (8). ∗ ∗ − I
⎧ T ⎫
⎨1 ∫ ( ⎬ = X(A(θ) − B2 C12 )T + (A(θ) − B2 C12 )X(θ) (8)
)
J=E eTu eu + eTr er dt (6) ⎡
⎩T ⎭ 1 ([ T ] [ ]) ⎤
T T T
0
⎢ φY (θ) − C2 C2 Yρ Bi (θ) C11 C12 − C2T DT12 D22 ⎥〈
⎢ γ ⎥
⎢ ⎥
where E(.) is the expected value of the signal. ⎢ 1[ T ] ⎥ 0 φY (θ)
⎢ DT21 ⎥
The final control objective is to minimize both the input disturbances ⎢ ∗ − I
γ
D11 ⎥
⎣ ⎦
di and the noise effect of do on the frequency error and ensuring the
∗ − I
stability of the closed-loop system. ∗
T
The robust LPV controller representation of Gc is illustrated as (8) = Y(θ)(A(θ) − Bo C2 ) + (A(θ) − Bo C2 ) Y(θ).
which Ak , Bk , Ck , Dk are the controller matrices. (9)

7
A. Aff et al. Electric Power Systems Research 214 (2023) 108898

Fig. 7. The HMG frequency output in presence of the wind power change profile.

⎡ ⎤
1 variables to obtain their magnitudes, then by inserting X(θ) and Y(θ) in
⎢ X(θ) I ⎥
LMI (10) and perturbing the X(θ) for satisfying that, the value of γ is
Q(θ) = ⎣ γ ⎦≥0 (10)
Y(θ) calculated. By inserting the γ value in LMIs (8) and (9), and achieving
∗
these two matrix variables again, the next γ value is calculated, and this
If LMI’s in (8)-(10) are satisfied, feasible solutions of Xθ and Yθ is loop procedure will be stopped when the optimal upper bound γ is ob­
derived. Then, the robust L2 induced controller Gc matrices are char­ tained. After calculating the ultimate values of X(θ), Y(θ), γ, with the use
[ ]
acterized in the form of (11): Ak (θ) Bk (θ)
of Theorem 1, the L2 induced controller matrices are
Ck (θ) Dk (θ)
Ac (θ) = A + B2 F(θ) + Q− 1 YL(θ)C2 − γ− 2 Q− 1 M,
achieved.
Bc (θ) = Q− 1 YL(θ),
(11) In summary, this algorithm proceeds as follows.
Cc (θ) = F(θ);
Dc (θ) = Ω. In the next section, the proposed controller performance on the
investigated IHMG system through several scenarios is studied in terms
Therefore, the controller matrices are obtained by Theorem 1, and of time domain and norms characteristics.
the control design procedure is completed.
According to Theorem 1, the whole controller design approach is 4. Simulation results
presented in Fig. 6.
To apply Theorem 1, the dynamic system equations must be con­ To investigate the proposed controller performance, simulation
verted to a standard form (1). In the second step, LMIs (8) and (9) are studies under several scenarios are employed by comparing two control
solved considering the positive definite matrix X(θ) and Y(θ) as decision methods. In these scenarios, the proposed control approach is compared

8
A. Aff et al. Electric Power Systems Research 214 (2023) 108898

Fig. 8. The HMG frequency output in presence of the photovoltaic power change profile.

with a H∞ robust control [35], and an optimal PID control tuned by the the H∞ robust control.
simplex search method [44] in terms of frequency fluctuation responses
and robustness. The IHMG system which is depicted in Fig. 5, is the base
of our simulations. The simulation time, Tsim = 120 s is considered and 4.2. Only photovoltaic power fluctuations in the investigated IHMG
the closed loop system is examined in the face of changing the wind
speed for the wind turbine system, changing the irradiance for the PV Fig. 8 demonstrates the time-domain response for only PV power
arrays, and changing the DC load. disturbance scenarios. The simulation results under the mentioned
multi-step PV power changes are analyzed.
This figure represents the IHMG frequency deviation response
4.1. Only wind power fluctuations in the investigated IHMG behavior for the proposed approach is smoother, faster response with
low oscillations.
In the first scenario, based on the scheduling parameter limitations,
the wind power pattern is derived for the multi-step wind speed pattern.
In Fig. 7a, the wind power pattern which is obtained from the wind 4.3. Only load deviation pattern in the investigated IHMG
speed profile is depicted. In Fig. 7b, controllers’ performances are
compared in terms of frequency deviation. the superiority of the LPV In this scenario, multi-level DC load change is employed to the IHMG
controller over the H∞ robust control and the optimal PID control in as a disturbance. The multi-step load pattern is shown in Fig. 9-a, while
terms of maximum overshoot, settling time, and steady-state error are the frequency aberration responds to the LPV control, the H∞ robust
demonstrated. It is notable that the H∞ robust control is a slightly control and the optimal PID controls are depicted in Fig. 9b.
smaller maximum overshoot than the optimal PID control, but in terms According to Fig. 9, the suggested LPV controller is better perfor­
of relative stability, the optimal PID control has better performance than mance with a more satisfactory transient response.

9
A. Aff et al. Electric Power Systems Research 214 (2023) 108898

Fig. 9. The HMG frequency output in presence of the load deviation profile.

4.4. All three types of disturbances in the investigated IHMG 4.5. Robustness against system parameter uncertainties with disturbances

In this scenario, simultaneous multi-step disturbances in wind speed, In this section, to evaluate the robust capability of the proposed
solar power, and DC load are characterized. The IHMG frequency output controller in practical situations, the parameter uncertainties as well as
in presence of the all three types of disturbances is depicted in Fig. 10. the disturbances variations are studied simultaneously. Considering
It’s obvious from Fig. 10 that by applying the LPV controller to the Table 4, three scenarios are addressed based on
IHMG, the peak magnitude of frequency deviation reduces in the least R, D, TDEG , Tg , M, TBESS , and TFESS changes in the IHMG model.
amount of time as compared to the magnitude and time when the H∞ To achieve this goal, it is assumed that IHMG system parameters are
robust and the optimal PID controllers are employed. defined according to Scenario R3 and all disturbances are occurred ac­
In the following, the characteristics of the time domain and norm cording to Scenario 4.
specifications for the three control approaches are compared in the form The frequency fluctuation response is presented for the three con­
of Tables 2 and 3 respectively. trollers in Fig. 11. Furthermore, the time domain specifications and
In the next subsection, the controller’s robust performance in face of norms of the frequency fluctuation of the proposed LPV control method
three parametric uncertainty scenarios solely and with all disturbances are addressed in Table 5 which reveals the superiority of the suggested
occurring scenario (Scenario 4) is analyzed. controller robustness against external disturbances and parameter un­
certainties simultaneously.
In the ultimate section, conclusions will be demonstrated.

10
A. Aff et al. Electric Power Systems Research 214 (2023) 108898

Fig. 10. The IHMG frequency output in presence of the all three types of disturbances.

Table 2 Table 4
Time domain characteristics comparison for different considered controllers. The parameter changes of the IHMG model.
Scenarios Control approaches MP (%) tp (sec) ts (sec) ess Parameter Scenario R1 (%) Scenario R2 (%) Scenario R3 (%)
(10− 6 )
R +30 +35 +60
Scenario 1 Optimal PID controller 0.0442 59.49 94.55 3.39 D -22 -40 -70
The H∞ controller 0.0483 59.65 86.19 0.0016 TDEG +25 +50 +58
The proposed controller 0.0311 59.31 81.89 0 Tg -20 -45 -75
Scenario 2 Optimal PID controller 0.0047 2.06 156.2 6.46 M +30 +55 +60
The H∞ controller 0.0043 2.02 101.4 0.0003 TBESS -25 -50 -75
The proposed controller 0.0022 1.25 97.32 0 TFESS +30 +40 +65
Scenario 3 Optimal PID controller 0.0063 50.05 97.2 27.6
The H∞ controller 0.0068 50.03 94.7 0.0032
The proposed controller 0.0044 50.01 91.4 0 uncertainties.
Scenario 4 Optimal PID controller 0.0442 59.64 99.05 9.833 A mathematical LPV model for the nonlinear wind turbine is
The H∞ controller 0.0483 59.62 93.27 3.353
The proposed controller 0.0312 59.61 92.14 0
extracted based on wind speed and turbine rotated speed as scheduling
parameters and the state-space representation of the whole IHMG sys­
tem is derived. The robust LPV synthesis algorithm is exploited to
eliminate frequency fluctuations for different uncertain conditions. For
Table 3 this purpose, the dynamic output feedback controller matrices are
Norms of the frequency deviation comparison for different considered
calculated according to an LPV design procedure. The performance
controllers.
assessment of the proposed controller can be observed from the re­
Scenarios Control approaches ‖ Δf ‖2 ‖ Δf ‖∞ sponses of frequency deviation after closing the secondary control loop.
Scenario 1 Optimal PID controller 0.1894 0.0249 Accordingly, For controller performance evaluation, four different sce­
The H∞ controller 0.2148 0.0270 narios are considered and the results are compared to a H∞ control and
The proposed controller 0.1368 0.0174 an optimal PID control method in terms of time domain and norms
Scenario 2 Optimal PID controller 0.0254 0.0024
The H∞ controller 0.0190 0.0022
characteristics. The simulation results reveal the superiority of the
The proposed controller 0.0055 0.0011 proposed controller performance improvement in all evaluated sce­
Scenario 3 Optimal PID controller 0.0302 0.0039 narios (MP = 0.0312 %, ts = 97.32 s, ess = 0, ‖ Δf‖2 = 0.1376 and ‖
The H∞ controller 0.0338 0.0043 Δf‖∞ = 0.0174) in the worst-case conditions.
The proposed controller 0.0220 0.0028
Further, sensitivity analysis is carried out by varying the system
Scenario 4 Optimal PID controller 0.1937 0.0251
The H∞ controller 0.2196 0.0271 parameters in several ranges (between -75% and +65% relative to the
The proposed controller 0.1376 0.0174 nominal values) and it is found that the designed LPV controller per­
formance is satisfactorily under wide changes in the system parameter
uncertainties. In terms of the ITAE (integral time absolute error) ITAE <
5. Conclusion and future directions 1 that about 85% smaller than the other controllers’ ITAE.
In summary, the highlights of the study are as follows.
Besides various intermittent disturbances, hybrid microgrid (DEG/
WTG/PV/FC/ESSs) parameters are continuously changing and the • A polytopic LPV model is presented for the IHMG system composed
closed-loop system performance may be degraded, seriously. Unlike the of DEG, PVs, WTs, AE, FC, FESS, BESS, and DC loads. With such an
vast majority of classic and intelligent controllers that have no guarantee LPV model, the uncertainty and stochasticity of PVs, WTs, and loads
to provide an acceptable performance over a broad range of operating can be considered in the controller design procedure.
conditions, a robust LPV control approach is presented for frequency
regulation in the presence of several disturbances and parameter

11
A. Aff et al. Electric Power Systems Research 214 (2023) 108898

Fig. 11. The IHMG frequency output in Scenario 4 and Scenario R3.

Table 5 Table 6
the time domain specifications and norms against external disturbances and IHGM parameter values were used in the simulation studies [38,39].
parameter uncertainties simultaneously. Parameter Value Parameter Value Parameter Value
Scenarios Control approaches ‖ Δf ‖2 ‖ Δf ‖∞ ITAE TG 2.000 s Ka 0.6 TPV 1.8 s
Scenario 4-R1 Optimal PID controller 0.2494 0.0390 6.589 TDEG 1.000 s KBESS -1/100 KDEG 1/300
The H∞ controller 0.4744 0.0377 4.602 R 3.000 KFESS -1/300 KPV 1
The proposed controller 0.1835 0.0286 1.034 δ 0.010 KFC 1/100 KAE 1/500
Scenario 4-R2 Optimal PID controller 0.2274 0.0341 6.585 D 0.012 Jg 59 kg ⋅m2 Ka 0.6
The H∞ controller 0.2984 0.0320 2.953 M 0.200 Bg 15.993 Nms/ KBESS -1/100
The proposed controller 0.1637 0.0241 1.026 rad
Scenario 4-R3 Optimal PID controller 0.1997 0.0251 6.576 TBESS 0.100 s Jr 830000 KFESS -1/300
The H∞ controller 0.2221 0.0251 2.594 TFESS 0.100 s ρ 1.225 kg/m3 KFC 1/100
The proposed controller 0.1376 0.0174 1.007 TFC 4.000 s R 30.3 m Pelec base 1.66 ×
106
TAE 0.5 s kr,V 1.8 s Pwind base 0.73 (pu)
• The proposed secondary frequency control approach suggests a TPV 1.8 s Pnom 1.5 × 106 (W) Vnom 1.2 (pu)
KDEG 1/300 500 μs 0
practical dynamic output feedback controller with lower order than τ β

the IHMG model.

• The robust control methods reported in the state of the art, neglect Declaration of competing interest
nonlinearities in wind turbines, while the proposed method can
handle these nonlinearities in the LPV control design procedure to None.
operate efficiently in a wide operating
Data Availability
In this study, the nonlinearity of the wind turbine generator in the
form of an LPV framework is considered; The nonlinearities of the other Data will be made available on request.
IHMG elements, power convertors, and power losses may be considered
in future studies. In the presented paper, the primary/local control
strategy is not exploited. A combination of primary control by ESSs and Appendix
secondary control by DEG can be led to a more stable operation. We plan
to extend the application of the proposed LPV control approach to The values of parameters used in the IHMG system are listed in
voltage and reactive power control in hybrid AC/DC microgrids. Also, Table 6.
hardware in loop and real-time implementation on the IHMG test bed is
remarkable in future directions.

CRediT authorship contribution statement

Algorithm 1
Implementation of the proposed controller.
Abbas Aff: Conceptualization, Resources, Data curation, Writing –
Design procedure
original draft, Methodology, Software. Mohsen Simab: Methodology,
1 Obtaining the LPV state-space model (3) from the IHMG dynamics, depicted in
Supervision, Data curation, Project administration. Mehdi Nafar: Fig. 3.
Investigation. Alireza Mirzaee: Formal analysis, Software, Writing – 2. Determine the design parameters Wu and Wr for weighting the control effort and the
review & editing. frequency fluctuations.
3. Use Theorem 1 to compute the LPV controller matrices based on the LPV model
given in (3).
4. Applying the control law to the closed loop structure depicted in Fig. 5.

12
A. Aff et al. Electric Power Systems Research 214 (2023) 108898

References [22] A. Annamraju, S. Nandiraju, Robust frequency control in an autonomous

microgrid: a two-stage adaptive fuzzy approach, Electr. Power Components Syst.
46 (1) (2018) 83–94.
[1] S. Rajamand, Load frequency control and dynamic response improvement using
[23] K. Lu, W. Zhou, G. Zeng, Y. Zheng, Constrained population extremal optimization-
energy storage and modeling of uncertainty in renewable distributed generators,
based robust load frequency control of multi-area interconnected power system,
J. Energy Storage 37 (2021), 102467.
Int. J. Electr. Power Energy Syst. 105 (2019) 249–271.
[2] U. Datta, A. Kalam, J. Shi, The relevance of large-scale battery energy storage (BES)
[24] A.H. Etemadi, E.J. Davison, R. Iravani, A decentralized robust control strategy for
application in providing primary frequency control with increased wind energy
multi-DER microgrids—Part I: Fundamental concepts, IEEE Trans. Power Delivery
penetration, J. Energy Storage 23 (2019) 9–18.
27 (4) (2012) 1843–1853.
[3] M.S. Javed, T. Ma, J. Jurasz, M.Y. Amin, Solar and wind power generation systems
[25] S. Som, S. De, S. Chakrabarti, S.R. Sahoo, A. Ghosh, A Robust Controller for Battery
with pumped hydro storage: Review and future perspectives, Renewable Energy
Energy Storage System of an Islanded AC Microgrid, IEEE Trans. Ind. Inf. (2021).
148 (2020) 176–192.
[26] S.C. Sahoo, A. Latif, S. Naidu, S. Patel, R. Kumar, D.C. Das, Demand Side
[4] D. Ribó-Pérez, P. Bastida-Molina, T. Gómez-Navarro, E. Hurtado-Pérez, Hybrid
Management-Based Load Frequency Control of Islanded Microgrid Using Direct
assessment for a hybrid microgrid: A novel methodology to critically analyse
Load Control. Modeling, Simulation and Optimization, Springer, 2021,
generation technologies for hybrid microgrids, Renewable Energy 157 (2020)
pp. 503–514.
874–887.
[27] K. Chen, M. Baran, Robust Controller for Community Microgrids for Stability
[5] A. Rahman, L.C. Saikia, N. Sinha, Automatic generation control of an
Improvement in Islanded Mode, IEEE Trans. Power Syst. (2022).
interconnected two-area hybrid thermal system considering dish-stirling solar
[28] P.S. Prasad, A.M. Parimi, L. Renuka, Control of hybrid AC/DC microgrids.
thermal and wind turbine system, Renewable Energy 105 (2017) 41–54.
Microgrids, Elsevier, 2022, pp. 191–225.
[6] J. Uche, A. Muzás, L. Acevedo, S. Usón, A. Martínez, A. Bayod, Experimental tests
[29] Q.L. Lam, A.I. Bratcu, D. Riu, C. Boudinet, A. Labonne, M. Thomas, Primary
to validate the simulation model of a Domestic Trigeneration Scheme with hybrid
frequency H∞ control in stand-alone microgrids with storage units: A robustness
RESs and Desalting Techniques, Renew. Energy 155 (2020) 407–419.
analysis confirmed by real-time experiments, Int. J. Electr. Power Energy Syst. 115
[7] P. Mansoorhoseini, B. Mozafari, S. Mohammadi, Islanded AC/DC microgrids
(2020), 105507.
supervisory control: a novel stochastic optimization approach, Electr. Power Syst.
[30] P.S. Prasad, A.M. Parimi, Robust μ-Synthesis Controller Design and Analysis for
Research 209 (2022), 108028.
Load Frequency Control in an AC Microgrid System, IFAC-PapersOnLine 55 (1)
[8] J. Li, P. Liu, Z. Li, Optimal design and techno-economic analysis of a solar-wind-
(2022) 547–554.
biomass off-grid hybrid power system for remote rural electrification: A case study
[31] M. Khan, H. Sun, Y. Xiang, D. Shi, Electric vehicles participation in load frequency
of west China, Energy 208 (2020), 118387.
control based on mixed H2/H∞, Int. J. Electr. Power Energy Syst. 125 (2021),
[9] C. Ardalan, V. Vahidinasab, A. Safdarian, M. Shafie-khah, J.P. Catalão, Pave the
106420.
way for sustainable smart homes: A reliable hybrid AC/DC electricity
[32] H. Ali, et al., A new frequency control strategy in an islanded microgrid using
infrastructure, Electr. Power Syst. Res. 210 (2022), 108128.
virtual inertia control-based coefficient diagram method, IEEE Access 7 (2019)
[10] M. Ranjan, R. Shankar, A literature survey on load frequency control considering
16979–16990.
renewable energy integration in power system: Recent trends and future prospects,
[33] A.D. Rosaline, U. Somarajan, Structured H-Infinity controller for an uncertain
J. Energy Storage 45 (2022), 103717.
deregulated power system, IEEE Trans. Ind. Appl. 55 (1) (2018) 892–906.
[11] M.A. Jirdehi, V.S. Tabar, S. Ghassemzadeh, S. Tohidi, Different aspects of
[34] H. Hua, Y. Qin, Z. He, L. Li, J. Cao, Energy sharing and frequency regulation in
microgrid management: A comprehensive review, J. Energy Storage 30 (2020),
energy network via mixed H 2/H∞ control with Markovian jump, CSEE Journal of
101457.
Power and Energy Systems 7 (6) (2020) 1302–1311.
[12] R. Carli, M. Dotoli, Decentralized control for residential energy management of a
[35] A.H. Tayebi, R. Sharifi, A.H. Salemi, F. Faghihi, Presentation of an H∞ based
smart users ʼ microgrid with renewable energy exchange, IEEE/CAA J. Automatica
frequency control for islanding provisional microgrid consisting of hybrid AC/DC
Sinica 6 (3) (2019) 641–656.
microgrid, Int. Trans. Electr. Energy Syst. 31 (10) (2021) e12715.
[13] J. Heidary, M. Gheisarnejad, H. Rastegar, M.H. Khooban, Survey on microgrids
[36] H. Hua, J. Cao, G. Yang, G. Ren, Voltage control for uncertain stochastic nonlinear
frequency regulation: Modeling and control systems, Electr. Power Syst. Res. 213
system with application to energy Internet: Non-fragile robust H∞ approach,
(2022), 108719.
J. Math. Anal. Appl. 463 (1) (2018) 93–110.
[14] H. Bevrani, H. Golpîra, A.R. Messina, N. Hatziargyriou, F. Milano, T. Ise, Power
[37] H. Bevrani, M.R. Feizi, S. Ataee, Robust Frequency Control in an Islanded
system frequency control: An updated review of current solutions and new
Microgrid: H∞ and μ -Synthesis Approaches, IEEE Trans. Smart Grid 7 (2) (2015)
challenges, Electr. Power Syst. Res. 194 (2021), 107114.
706–717.
[15] P.K. Ray, S.R. Mohanty, N. Kishor, Proportional–integral controller based small-
[38] F.A. Inthamoussou, F.D. Bianchi, H. De Battista, R.J. Mantz, LPV wind turbine
signal analysis of hybrid distributed generation systems, Energy Convers. Manage.
control with anti-windup features covering the complete wind speed range, IEEE
52 (4) (2011) 1943–1954.
Trans. Energy Convers. 29 (1) (2014) 259–266.
[16] M. Zolfaghari, G.B. Gharehpetian, Decentralized power exchange control methods
[39] A.M. Shawqran, A. El-Marhomy, M.A. Attia, A.Y. Abdelaziz, H.H. Alhelou,
among subsystems in future power network. Decentralized Frameworks for Future
Enhancement of wind energy conversion system performance using adaptive
Power Systems, Elsevier, 2022, pp. 345–367.
fractional order PI blade angle controller, Heliyon 7 (10) (2021) e08239.
[17] N.S. Boghrabadi, M.R.J. Motlagh, A.M. Sayidi, Chaos suppression in the uncertain
[40] H. Atoui, O. Sename, V. Milanes, J.-J. Martinez-Molina, Toward switching/
Van Der Pol oscillator, in: 2008 International Conference on Computational
interpolating LPV control: a review, Ann. Rev. Control (2022).
Intelligence for Modelling Control & Automation, IEEE, 2008, pp. 208–213.
[41] A. Mirzaee, M. Dehghani, M. Mohammadi, Robust LPV control design for blood
[18] M. Shouran, F. Anayi, M. Packianather, Design of sliding mode control optimised
glucose regulation considering daily life factors, Biomed. Signal Process. Control
by the Bees algorithm for LFC in the Great Britain power system, Mater. Today:
57 (2020), 101830.
Proc. 52 (2022) 937–943.
[42] F. Wu, X.H. Yang, A. Packard, G. Becker, Induced L2-norm control for LPV systems
[19] A. Bagheri, A. Jabbari, S. Mobayen, An intelligent ABC-based terminal sliding
with bounded parameter variation rates, Int. J. Robust Nonlinear Control 6 (9-10)
mode controller for load-frequency control of islanded micro-grids, Sustain. Cities
(1996) 983–998.
Soc. 64 (2021), 102544.
[43] J. Cheng, M. Wu, F. Wu, C. Lu, X. Chen, W. Cao, Modeling and control of drill-
[20] A. Latif, A. Pramanik, D.C. Das, I. Hussain, S. Ranjan, Plug in hybrid vehicle-wind-
string system with stick-slip vibrations using LPV technique, IEEE Trans. Control
diesel autonomous hybrid power system: frequency control using FA and CSA
Syst. Technol. 29 (2) (2020) 718–730.
optimized controller, Int. J. Syst. Assurance Eng. Manag. 9 (5) (2018) 1147–1158.
[44] L. Fu, L. Zhou, J. Liang, M. Lin, PID parameters self-tuning based on simplex
[21] S. Ganguly, C.K. Shiva, V. Mukherjee, Frequency stabilization of isolated and grid
method, in: 2018 Chinese Control And Decision Conference (CCDC), IEEE, 2018,
connected hybrid power system models, J. Energy Storage 19 (2018) 145–159.
pp. 4769–4774.

13