Journal of Energy Storage 19 (2018) 145–159

Contents lists available at ScienceDirect

Journal of Energy Storage

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

Frequency stabilization of isolated and grid connected hybrid power system T

models
⁎
Somnath Gangulya, , Chandan Kumar Shivab, V. Mukherjeec
a
Department of Electrical Engineering, Bankura Unnayani Institute of Engineering, Bankura, West Bengal, India
b
Department of Electrical and Electronics Engineering, SR Engineering College, Ananthsagar, Hasanparthy, Warangal, Telangana, India
c
Department of Electrical Engineering, Indian Institute Technology (Indian School of Mines), Dhanbad, Jharkhand, India

A R T I C LE I N FO A B S T R A C T

Keywords: This paper investigates the load frequency control (LFC) of two different configurations of power system model.
Diesel engine generator The first one is worked as an isolated hybrid power system (HPS) (IHPS) while the second one is a grid tied HPS
Grid connected hybrid power system (GHPS). In IHPS, the power generating units are diesel engine generator (DEG), wind turbine generator (WTG)
Isolated hybrid power system and solar thermal power generation (STPG). It is considered in such a way that the generation persists
Optimization
throughout the day with uninterrupted power supply. The energy storage device used in this work is super-
Superconducting magnetic energy storage
Wind turbine generator
conducting magnetic energy storage (SMES). The three scenarios considered in isolated mode are (a) HPS with
one DEG, one WTG and one STPG, (b) HPS with one DEG and one STPG and (c) HPS with one DEG and one WTG.
As concerned to GHPS, a combination of WTG and steam thermal power plant is considered. A conventional
proportional-integral-derivative (PID) controller has been used as supplementary control mechanism. The PID
controller gains and the tunable parameters of the SMES are tuned by the use of quasi-oppositional harmony
search (QOHS) algorithm. Initially, integral of time absolute error based objective function is minimized and,
subsequently, three performance indices such as integral of absolute error, integral of square error and integral
of time square error are calculated to test the designed efficiency of the proposed QOHS based PID controller.
The transient responses of systems pertaining to step and random changes in the load demand are analyzed. The
simulated results obtained by the QOHS algorithm reveals that it may be imposed to boost LFC performance of
the power system.

1. Introduction characteristic as well as fast development in the designed technology.

1.1. General
The demand of electrical energy is continuously increasing in phy-
sical world due to high population growth and industrial development.
At present, the high percentage of electrical energy requirement is met
from fossil fuel generating units. However, due to the gradual depletion
of fossil fuel and hike in fuel cost, requirement of load demand may not
be fulfilled by conventional steam thermal power station (STPG). In this
context, renewable energy resource (RES) presents a promising future
as it provides clean energy and eliminates dependency on fossil fuels
[1]. The electricity and heat generation from solar radiation, wind
energy, ocean energy, hydro power, biomass, geothermal resources,
biofuels and hydrogen are derived from the clean RESs. Among these
RESs, wind energy and solar photovoltaic (PV) generating units have
gained popularity due to their inexhaustible environment friendly
Also, it has a positive implication on financial as well as social growth
of the country [2]. Along with the conventional resources, both the
reliability as well as the power quality of the generated electricity may
be improved, if these RESs are added.
1.2. Literature survey
The increasing concerns on environmental problems as well as the
shortage and rising costs of fossil fuels have exhibited interest towards
the integration of RESs to the power systems. The power generated by
renewable sources is irregular and cannot be easily predicted. This af-
fects the scheduled frequency regulation. To tackle these problems, the
concept of distributed generation (also termed as small isolated power
system with or without connection to the grid) may be implemented to
support and regulate the system frequency at rural application, large
commercial areas and process industries. In small isolated power

⁎
Corresponding author.
E-mail addresses: somnath_aec82@yahoo.co.in (S. Ganguly), chandankumarshiva@gmail.com (C.K. Shiva), vivek_agamani@yahoo.com (V. Mukherjee).

https://doi.org/10.1016/j.est.2018.07.014
Received 12 May 2018; Received in revised form 24 July 2018; Accepted 24 July 2018
Available online 31 July 2018
2352-152X/ © 2018 Elsevier Ltd. All rights reserved.
S. Ganguly et al.
system, electrical power is supplied by the diesel generators.
Additionally, wind and solar radiation based generation have gained
increased attention as green energy in most of the isolated power sys-
tems [1]. Due to sudden change either in load or in wind speed, the
system frequency and, hence, the output power fluctuate. The fluctu-
ating frequency and power seriously affects the stability, control
strategy and security of the locally connected grid [3].
A standalone hybrid power system (HPS) consisting of wind turbine
generator (WTG), diesel engine generator (DEG), fuel cell and aqua
electrolyzer has been used in [4]. In this work, the effect of designed
system on the load stabilization has been considered to ensure quality
power supply. The development of hybrid PV and fuel cell generation
system for stand-alone application has been presented in [5]. The role
of this system is the production of electric energy without interruption
in isolated places. The modeling, control and power management of
hybrid PV fuel cell and battery bank system supplying electric vehicle
have been under taken in [6]. It consists of a PV generator, a proton
exchange membrane fuel cell and a battery bank supplying an electric
vehicle of 3 kW. In another work (see [7]), modeling of a hybrid PV,
wind and fuel cells based power system has been presented. The
mathematical modeling topology and its power management with
battery bank system are significant contributions of this work. In [8],
the modeling and the control of HPS are presented by Aissou et al. It
comprises of wind and PV sources with battery storage supplying a load
via an inverter. Tamalouzt et al., in [9], have highlighted the modeling
and simulation of a micro-grid based renewable power system. It
comprises of wind turbine, double fed induction generator, PV gen-
erator, fuel cell and battery bank.
The generation from RES is stochastic in nature and depends upon
the weather condition at any instant of time. This might results in si-
tuation where the electrical load is higher than the generation. If there
is surplus power available from these RESs, then the energy storage
devices (ESDs) may store the excess power for a short period of time
and releases them later on to the grid when the load demand is higher
than the generation. The most popularly used ESDs are flywheel energy
storage system, battery energy storage system, compressed air energy
storage system, super capacitor and superconducting magnetic energy
storage (SMES) unit. These are being considered to store the surplus
energy and deliver the same at the time of peak load demand [10–12].
Among the mentioned storage devices, SMES may be considered to be
the efficient for some of its advantages. The SMES gives very high en-
ergy storage efficiency, typically 97% for very large systems and com-
paratively less for small application [13]. This is the device which
stores the electrical energy with no loss by a magnetic field of coil
comprising of superconducting wire. This storage device is capable of
discharging high power within a fraction of cycle to compensate the
sudden drop in the line power. Such strategic injection of brief bursts of
power plays a crucial role in maintaining the grid reliability. This
makes SMES more environmentally beneficial and viable as compared
to the batteries [14].
A number of evolutionary computational intelligence based tech-
niques have been developed as well as employed for HPS application.
The genetic algorithm (GA) has been applied for the designing of hybrid
solar-wind system with battery bank as energy storage (see [15]). To
minimize the deviation in frequency, integration of different energy
resources along with the energy storage elements has been carried out
in [16]. In this work, proportional-integral (PI) controller has been
addressed in order to achieve improvements in the deviation of fre-
quency profiles. The tuning of a PI controller using particle swarm
optimization (PSO) algorithm for simulation studies in autonomous
hybrid energy generation/energy storage system has been dealt with in
[17].
In order to improve the utilizing efficiency of renewable energy, a
146
Journal of Energy Storage 19 (2018) 145–159
combined PSO and simulated annealing has been used in the literature
for hybrid energy storage station as wind-solar generation system (refer
[18]). The artificial bee swarm optimization algorithm has been used to
optimize the size of HPS model which consists of solar-wind-fuel cell.
The simulated work of [19] showed that the designed HPS system is a
cost effective one. The hybrid generation systems consisting of WTGs,
solar thermal power generation (STPG), PV, DEGs, fuel cells, battery
energy storage system, flywheel, ultra capacitors and aqua electrolyzer
have been considered for simulation studies in [20]. In this work, GA is
used for optimization of controller gains of the hybrid systems. Frac-
tional order fuzzy control of hybrid power system with renewable
generation using chaotic PSO has been studied in [21].
The present paper deals with the study and analysis of an isolated
and grid connected HPS models. The power generating units like WTG,
DEG and STPG form the studied IHPS models. The grid connected HPS
(GHPS) is formed by using one WTG and one STPS. Two proportional-
integral-derivative (PID) controllers are used in isolated HPS (IHPS) in
which one is to control the blade pitch angle of the WTG while the other
is to control the speed governor of the DEG. Two more controllers are
used in GHPS model to control the blade pitch angle of the WTG and to
control the speed governor of the thermal turbine for obtaining better
dynamic performance.
1.3. Motivation of present work
Despite of a lot of progress in the field of metaheuristic approaches
employed for the optimization problem, there are still possibilities of
some more avenues to improve further their searching capabilities and
convergence characteristics. In the year 2000, harmony search algo-
rithm (HSA) came into picture (see [22]). In [22], this algorithm was
employed to solve the optimization problem of water distribution.
Afterwards, this algorithm has found a number of successful applica-
tions in different aspects of engineering problems including industry,
construction design, information technology, medical science, bench-
mark function optimization, power system and control system domain
(refer [23–25]). A new concept based on opposition based learning
(OBL) has been introduced in [26] to enhance the exploration potential
of HSA. The work introduced in [27] has showed a new type of popu-
lation initialization concept and named it as quasi-oppositional based
learning (QOBL) concept. It uses random number population with its
quasi-opposite number aiming to (a) enhance performance, (b) trounce
the premature convergence, (c) improve the diversity of the solution
and (d) accelerate the convergence speed of the basic HSA. Due to the
increased exploring capability of QOBL, quasi-opposition harmony
search (QOHS) algorithm has been used for load frequency control
(LFC) problem in [28].
With reference to the above discussions, the main motivation of this
paper evolves the utilization of QOHS algorithm to tune the controller
parameters for frequency stabilization in different scenarios of the de-
ployed IHPS models and GHPS model subjected to load disturbances.
1.4. Contribution of the present work
The main contributions of this paper are as follows:
(a) The modeling of IHPS embrace of WTG, DEG and STPG is done for
handling rural electricity crisis.
(b) The modeling of GHPS comprising of one RES (mainly WTG) and
one non RES (popularly STPS) is done for uninterrupted power
supply.
(c) For reliable power quality and satisfying the power demand of the
local area, SMES has been used in case of IHPS. The benefits of
SMES are explored in this case.
S. Ganguly et al. Journal of Energy Storage 19 (2018) 145–159

(d) A PID controller is added to get an improved power quality and a
scheduled frequency.
(e) A QOHS algorithm is investigated to tune the controller gains and
the SMES parameters.
(f) Simulation results pertaining to IHPS and GHPS are presented and
analyzed.
supports its candidature for use as most reliable standby emergency
generators in supplying continuous and high quality power.
Fig. 1 shows the dynamic model of DEG with simplified dynamics of
speed governor and diesel engine turbine [30]. The transfer function
equations are presented in (3)–(5), in order,
1
ΔPGD = ⎛ ⎜
⎞ ΔPGT ⎟

⎝ 1 + sTD4 ⎠ (3)
1.5. Layout of the paper
KD (1 + sTD1 )
This paper covers the issues concerning the frequency regulation ΔPGT = ⎛ ⎜ ⎟
⎞ ΔPG
⎝ (1 + sTD2 )(1 + sTD3 ) ⎠ (4)
perspective of both isolated and grid connected HPS models subjected
to load perturbation in the system. After the introduction section, the
1
rest of the study is presented in the following sequences. In the next ΔPG = ΔPCD − ⎛ ⎞ Δf
⎜ ⎟

section of ‘System configuration and characteristics’, basic concept on ⎝ RD ⎠ (5)

DEG, WTG, STPG, SMES and PID controller are described. In the
‘Mathematical problem formulation’, the designed objective function is
stated along with the performance indices and the transient perfor-
mance parameters. A brief description of the applied algorithm is de-
picted in the section ‘Adopted optimization technique’. Simulation
based performance investigation and their analysis are done in the
‘Result and analysis’ section. Finally, the concluding remarks are sum-
marized in section ‘Conclusion’.
2. System configuration and characteristics
2.1. Designed IHPS model
In the modeling part of the present work, the studied IHPS is
composed of a DEG, a WTG and a STPG. The diagrams of the different
configurations of the studied IHPS system under three scenarios are
shown in Fig. 1(a)–(c). Table A1 of Appendix Section may be referred
for the chosen parametric values of the studied IHPS and GHPS models.
In these figures, a SMES block is added as an ESD. The system model
considered is of a 150 kW WTG and a 150 kW DEG [29]. In the IHPS
model, the total power balance equation is given by (1)
ΔPIHPS = ΔPGD + ΔPSTPG + ΔPGW − ΔPL − ΔPSMES (1)
where ΔPG is the control signal for speed governor (in p.u.), ΔPGT is the
control signal of diesel generator (in p.u.), ΔPCD is the control signal of
diesel governor (in p.u.), Δf is the frequency deviation, KD is the gov-
ernor gain, RD is the regulation of speed governor, Hz/p.u., TD1, TD2 and
TD3 are the time constants of speed governor (in sec) and TD4 is the time
constant of the turbine (in sec).
2.3.2. WTG: basic concept
The WTG system converts the kinetic energy of moving air into
mechanical to electrical energy. The power output of aerodynamic
power turbine (Pt ) may be written in (6)
1
Pt = CP (λ, β ) ρπR2v 3
2 (6)
where ρ is the air density, R is the blade radius, β is the pitch angle and
v is the linear wind speed and CP is the power coefficient which is a
function of tip speed ratio and blade pitch angle. Fig. 1 shows a dy-
namic model for WTG. The transfer function equations are presented in
(7)–(11). The fluid coupling block transfers the speed which is pro-
portional to the difference between the turbine and generator fre-
quency. In response to this, the change in generated power is stated in
(7) [31]

where ΔPIHPS is the total output deviation of IHPS model, ΔPGD is the ΔPGW = KIG [ΔFT − Δf ] (7)
deviation of output power of DEG, ΔPGW is the deviation in output where KIG is the fluid coupling gain and ΔFT is the speed deviation of
power of WTG, ΔPSTPG is the deviation of output power of STPG, ΔPSMES the WTG.
is the deviation of output power of the SMES and ΔPL is the change in The speed of wind turbine induction generator is shown in (8)
load demand.
1
ΔFT = ⎛ ⎜
⎞ [KTP ΔFT − ΔPGW + KPC ΔX3 + ΔPIW ]
⎟

2.2. Designed GHPS model ⎝ 1 + sTW ⎠ (8)

The GHPS model under study is consists of a WTG and a STPS and
the same is shown in Fig. 2. The power balance equation for this model
is given by (2).
ΔPGHPS = ΔPGW + ΔPM − ΔPL (2)
In (2), ΔPGHPS is the total output deviation of the GHPS model, ΔPGW
is the deviation in output power of the WTG and ΔPM is the deviation in
output power of the steam turbine.
2.3. Description of the power generating units
2.3.1. DEG: basic concept
As concern to IHPS, DEG plays an impotent role while the other
power generating units (PGUs) (such as WTG, STPG etc.) are dependent
on the weather conditions. DEG is the only power generation unit that
works autonomously. When operation, maintenance and control for the
non-conventional PGUs is done in HPS, DEG is the only PGU which
where KPC is the gain of blade characteristic, ΔPIW is the input of wind
power deviation (in p.u.) and ΔX3 is the output of data fit pitch re-
sponse. A data fit pitch response block acts as a simple lag compensator
which is matched with the gain characteristics of the model. The output
of data fit pitch response may be presented in (9).
KP3 ⎤
ΔX3 = ΔX2 ⎡
⎢
⎣ 1 + sTP3 ⎥
⎦ (9)
In (9), ΔX2 is the output of the hydraulic pitch actuator, KP3 is the
gain of the data fit pitch response block and Tp3 is its time constant.
The hydraulic pitch actuator controls pitch angle of the turbine. The
output of hydraulic pitch actuator may be expressed by (10).
KP 2 ⎤
ΔX2 = ΔX1 ⎡
⎣ 1 + sTP 2 ⎥
⎢ ⎦ (10)
In (10), ΔX1 is the output of the pitch controller, KP2 is the gain of
the hydraulic pitch actuator, TP2 is the time constant of hydraulic pitch

147
S. Ganguly et al. Journal of Energy Storage 19 (2018) 145–159

Fig. 1. Block diagram of aggregated studied IHPS pertaining to studied (a) Scenario-I, (b) Scenario-II and (c) Scenario-III.

148
S. Ganguly et al. Journal of Energy Storage 19 (2018) 145–159

Fig. 2. Block diagram of aggregated studied GHPS model.

Fig. 3. Block diagram of aggregated SMES model.

actuator (in sec). The pitch angle of the WTG is to select very finely to
get the maximum output power of the turbine. The pitch angle block
output may be expressed by (11)

KP1 (1 + s TP1) ⎤
ΔX1 = ΔPCW ⎡
⎣ 1+s ⎦ (11)

where ΔPCW is the control signal of pitch control block (in p.u.), KP1 is
the gain of the pitch control block, TP1 is the time constant of the pitch Fig. 4. Structure of PID controller.
control block (in sec). The electrical parameters of wind turbine are
stated in Appendix Section (refer Tables A2–A4)

2.3.3. STPG: basic concept

1 ⎤ ⎡ 1 + sT1 ⎤ ⎡ 1 + sT3 ⎤ Δf KF K SMES
STPG has emerged as a new technology and its acceptance by power ΔPSMES = ⎡
⎢
⎣ 1 + sTSMES ⎥
⎦⎢⎣ 1 + sT2 ⎥
⎦⎢⎣ 1 + sT4 ⎥
⎦ (13)
utilities is increasing day by day. It has same working principle as that
of a conventional steam thermal power plant with only difference lies in
In (13), ΔPSMES is the incremental change in output power of SMES;
the process of production of steam.
TSMES is time constant (in sec); KF , K SMES are the gains of the control
The transfer function model (GSTPG (s )) of the STPG unit may be
loop of SMES; T1, T2 , T3 , T4 are the time constants (in sec) with phase
presented in (12) [32].
compensation to provide appropriate phase lead characteristics to
compensate for phase lag between input and output signals.
KT ⎞ ⎛ K S ⎞ ΔPSTPG
GSTPG (s ) = ⎛⎜ ⎟⎜ = ⎟

⎝ 1 + sTT ⎠⎝ 1 + sTS⎠ ΔPSOL (12)

In (12), KT and K S are the gains of STPG, TT and TS are the time
constant of STPG (in sec). The ΔPSTPG and ΔPSOL are the deviations of
output and input powers of STPG.
2.4. Description on energy storage device
2.4.1. SMES: basic concept
SMES is a device which can store electrical power from the grid in
the form of magnetic field of a coil. The magnetic field of a coil is made
up of superconducting wire which is nearer to zero loss of energy [33].
The structure of SMES as a frequency stabilizer is shown in Fig. 3. The
incremental change in power of SMES (ΔPSMES ) is expressed in (13).
2.5. Description on supplementary control mechanism
2.5.1. PID controller: basic concept
In this work, PID controller is used to stabilize the load frequency
and improve the dynamic responses of the system. The proper design of
PID controller requires determination of optimal values of its three
parameters viz. proportional (KP ), integral (Ki ) and derivative (K d )
gains connected in parallel (refer Fig. 4). The controlling action of PID
controller may be define in (14)
t
u (t ) = Kp e (t ) + Ki ∫ e (t ) dt + Kd dtd e (t )
0 (14)
where u (t ) and e (t ) are actuating signal and error signal, respectively.

149
S. Ganguly et al. Journal of Energy Storage 19 (2018) 145–159

The error inputs to the PID controller for IHPS are given in (15) and The expression of FOD is stated in (19), where ts is the total simulation
(16), respectively. time (in sec).
e1 (t ) = Δf (15) ts
FOD = ITAE = ∫ t Δf dt
e2 (t ) = ΔPGW − ΔPGW (max) = ΔPgw (16) 0 (19)
The error input to the PID controller for GHPS are presented in (17)
and (18), in order.
3.2. Measure of performances
e (t ) = Δf (17)
In addition to the chosen FOD, the other three measures of perfor-
e4 (t ) = ΔPGW − ΔPGW (max) = ΔPgw (18) mance adopted for the design are integral of absolute error (IAE), ISE
and ITSE. The expression for these three performance indices are
started in (20)–(22), in order [35].
3. Mathematical problem formulation
ts

At attempt made in this paper is to design the QOHS based PID IAE = ∫ Δf dt
controllers and SMES parameters for the studied test systems. Here, the 0 (20)
main emphasis is to minimize the frequency deviation of studied IHPS
ts 2
and GHPS models as swiftly as possible.
ISE = ∫ Δf dt
0 (21)
3.1. Designed objective function
ts 2
In the present work, integral of time multiplied absolute error ITSE = ∫ Δf t dt
(ITAE) is chosen as an objective function for the optimization task. It 0 (22)
may be also called as figure of demerit (FOD). The ITAE weights error
which exists after a long time much more heavily as compared to start
of the response. The time multiplication term penalizes the error more 3.3. Parameter constraints of the optimization problem
at the later stages than at the beginning. The other point is ITAE based
tuned system settles much more quickly than integral of square error In the present optimization task, the tunable parameter constrains
(ISE) and integral of time multiplied squared error (ITSE) criteria. Since are the PID controller gains and SMES adjustable parameters. These are
the absolute error is included in the ITAE criterion, the maximum confined to a particular limit range (under the superscripts min and
percentage of overshoot is also minimized. It gives better performance max) during the optimization process.
as compared to other integral based performance criteria (refer [34]). The PID controller gains are restricted as started in (23).

Fig. 5. Profiles of studied load perturbation for IHPS system (a) 1.0 p.u. SLP and (b) random SLP.

Table 1
QOHS based optimized PID controller and SMES parameters for IHPS.
Studied Controller-I Controller-II
scenarios
KP1 KI1 KD1 KP2 KI2 KD2

Scenario-I 0.8599 0.9997 0.0062 0.9301 0.0045 0.5311

Scenario-II 0.9971 1.0000 0.0787 – – –
Scenario-III 0.9124 0.9976 0.0349 0.9996 0.0011 0.6519
SMES T1 T2 T3 T4 K SMES TSMES
Scenario-I 1.7336 1.9924 1.0739 1.9992 2.000 1.9485
Scenario-II 0.1084 0.9992 0.2602 0.0250 0.2505 1.0000
Scenario-III 0.0460 0.9998 0.0048 0.8207 0.0249 0.9671

For entry “-” means not applicable.

150
S. Ganguly et al. Journal of Energy Storage 19 (2018) 145–159

KPmin ≤ KP ≤ KPmax ⎫ time (tr ), settling time (ts ), peak value (Mp) , peak time (tp ) and steady
⎪ state error (eSS ). It may be noted that the value of measure of perfor-
KImin ≤ KI ≤ KImax
⎬ mances and transient parameter specifications are calculated at the end
KDmin ≤ KD ≤ KDmax ⎪
⎭ (23) of the developed program.
The SMES parameters are limited as stated in (24).
4. Adopted optimization technique
T1min ≤ T1 ≤ T ⎫
T2 ≤ T2 ≤ T2max
min ⎪
⎪ In this work, the QOHS algorithm has been applied to the load
T3min ≤ T3 ≤ T3max ⎪ frequency stabilization problem of both isolated and grid connected
T4min ≤ T4 ≤ T4max ⎬ power systems. In the QOHS algorithm, the concept of QOBL is added to
⎪ the parent HSA. In line with the work reported by Banerjee et al. [36],
min max
KSMES ≤ K SMES ≤ KSMES ⎪
min max ⎪
the pseudo code of basic HSA may be shown in Algorithm 1. The work
TSMES ≤ TSMES ≤ TSMES ⎭ (24) carried out on OBL showed that the projection of quasi-opposition
concept (in this work, symbolized as quasi-oppositional based learning
In (23) and (24), the superscripts min and max represent the
(QOBL)) is more likely to be closer to the solution as compared to OBL
minimum and the maximum values of the respective variables.
[37–39] that enhances both the convergence characteristics towards
the optimal solution and quality of the solution. To explore this ad-
3.4. Transient parameter specifications vantage offered by QOBL, it is being applied in this work. The pseudo
code for the calculation of quasi-opposite number and quasi-opposite
The main objective in this study is to optimize the tuning para- point are given in Algorithms 2 and 3, respectively. The pseudo code of
meters of the adjustable gains so that the power system dynamic re- the QOHS algorithm, compatible with the present LFC problem, is
sponses are neither too fast nor too slow. In the present prospect, the presented in Algorithm 4 [40].
dynamic response of the system is studied, mainly, in the form of rise

151
S. Ganguly et al.
The applications of QOHS algorithm may be noted in the design of
the scaling factor based fuzzy classical controller of an IHPS model in
[31]. The same may be seen for the optimization of the controller
parameters of three-area thermal power stations in [40].
152
Journal of Energy Storage 19 (2018) 145–159
5. Results and analysis
The IHPS and the GHPS models comprise of different combinations
of energy resources and storage unit are simulated and analyzed under
various operating conditions. For the time domain analysis, the simu-
lation work is carried out for two different types of load disturbances
S. Ganguly et al. Journal of Energy Storage 19 (2018) 145–159

Table 2 Table 4
QOHS based optimized PID controller parameters for GHPS. Value of different performance indices for IHPS configuration with 1.0 p.u. SLP.
Studied model Controller-III Controller-IV Models Performance index Only model Model + SMES

KP1 KI1 KD1 KP2 KI2 KD2 Scenario I ITAE 40.7314 9.5719
ITSE 9.9153 1.6522
Only STPS 0.9983 0.4609 0.4957 – – – IAE 5.1291 2.0229
WTG and STPS 4.9898 3.1345 0.1503 4.2508 3.3838 0.1401 ISE 2.0883 0.6593
Scenario II ITAE 48.0738 7.1718
For entry “-” means not applicable. ITSE 44.2020 4.4760
IAE 10.1542 2.3281
ISE 17.3400 3.4022
(refer Fig. 5). In a more concise way, a comparison is being reported for
Scenario III ITAE 59.7409 16.060
different configurations of the isolated power systems to show the ITSE 22.7409 3.8517
tuning efficiency of the QOHS algorithm. IAE 7.6030 3.1067
It has already been shown that due to the presence of unreliable ISE 4.3324 1.3046
renewable energy generation components like wind and solar radiated
power, a continuous variation in grid frequency is observed throughout
the system [41]. In effect of this, the power quality needs to be kept Table 5
within limits. To sort out this problem, a controller is introduced in the Dynamic performance analysis of GHPS configuration with 0.01 p.u. SLP.
loop which sends a signal to the energy storage systems to absorb/re- Type of model MP (%) tr (sec) ts (sec) tp (sec) eSS (10−5)
lease additional/deficit power from/into the grid. The controller also
sends a command to the generating unit to release high burst of power Only STPS 0.0322 0.0040 13.2232 2.0027 3.486
WTG + STPS 0.0058 0.8191 4.0750 1.0393 1.142
into the grid to meet short term load demands [41]. The parameters of
PID controllers and SMES unit are tuned for the supplementary control
action in the studied test systems using QOHS technique. The adopted
parameters of the QOHS algorithm are defined in Table A5 of Appendix Table 6
Section. Tables 1 and 2 represent the QOHS optimized PID controller Value of different performance indices for GHPS configuration with 0.01 p.u.
parameters and SMES parameters for Case-I and Case-II, respectively. SLP.
Tables 3 and 4 show the transient details and the obtained performance Performance index Only STPS WTG + STPS
indices, in that order, for all the studied scenarios of the IHPS system.
The same is shown, in order, in Tables 5 and 6 for the GHPS. The results ITAE 0.1521 0.0229
ITSE 0.0025 0.0001
of interest are bold faced in the respective tables. The investigations
IAE 0.0512 0.0052
considered in Case I and Case II are described below. ISE 0.0012 0.0003

5.1. Case I. Basic details

with and without SMES unit.

In this case, the system considered is in isolated mode where power

system units are composed of different generating units. The system
5.1.1. Performance investigation of case I
time domain behavior is studied under different load conditions. The
The time domain analysis of IHPS is done in this case. For this
IHPS models are run subjected to 1.0 p.u. step load perturbation (SLP)
model, 0.01 p.u. is applied as base wind power whereas 0.2 p.u. is
for 100 s as shown in Fig. 5(a). The same for random SLP for 200 s is
applied at the solar power system.
revealed in Fig. 5(b). The random load input is increasing initially up to
80 s after wards it is in decreasing mode up to 160 s, then again in-
creasing up to 200 s. Each increment and decrement is taking place
5.1.2. Scenario I: HPS with one DEG, one WTG and one STPG
after every 20 s of time duration with 0.4 p.u. change in load. In the
The block diagram representation of the model studied in this sce-
designed isolated HPS, the considered three scenarios are as follows.
nario is shown in Fig. 1(a). At present, it is studied under two different
Scenario I: HPS with one DEG, one WTG and one STPG.
conditions. The first one is at under full load and the second one is
Scenario II: HPS with one DEG and one STPG.
under random load variation. The QOHS algorithm is run for 100
Scenario III: HPS with one DEG and one WTG.
iterations. The obtained results are described below.
The considered three scenarios are operated under the influence of

Table 3
Performance analysis of IHPS configuration.
Scenarios SLP Type of MP (%) tr (sec) ts (sec) tp (sec) eSS
model

Scenario I 1.0 p.u. Only model 0.5821 0.0045 50.9047 2.5393 0.00058
Model + SMES 0.5428 0.0011 14.1062 1.6935 0.00054
Random Only model 0.5995 0.0813 199.729 182.00 0.00539
Model + SMES 0.5347 0.0069 195.263 180.81 0.00023
Scenario II 1.0 p.u. Only model 4.0762 0.0042 16.3396 1.2255 0.02356
Model + SMES 1.8890 0.0218 7.66090 1.0432 0.00330
Random Only model 3.6874 0.1022 196.405 180.24 0.02306
Model + SMES 1.5075 0.0627 180.859 180.05 0.01417
Scenario III 1.0 p.u. Only model 0.7901 0.0028 50.9221 3.2815 0.00012
Model + SMES 0.6781 0.0061 14.7300 1.9207 0.00010
Random Only model 0.5990 0.0805 199.716 182.00 0.01594
Model + SMES 0.0546 63.133 194.493 180.91 0.00368

153
S. Ganguly et al. Journal of Energy Storage 19 (2018) 145–159

Fig. 6. Dynamic response profile pertaining to Scenario-I with 1.0 p.u. SLP (a) Δf without SMES, (b) Δf with SMES and (c) QOHS based convergence profile of FOD.

Fig. 7. Dynamic response profiles corresponding to Scenario-I with random SLP (a) Δf without SMES and (b) Δf with SMES.

Fig. 8. Dynamic response profiles pertaining to Scenario-II with 1.0 p.u. SLP (a) Δf without SMES, (b) Δf with SMES and (c) QOHS based convergence profile of FOD.

154
S. Ganguly et al.
Fig. 9. Dynamic response profiles corresponding to Scenario-II with random SLP (a) Δf without SMES and (b) Δf with SMES.
5.1.3. Simulation results with 1.0 p.u. load demand
This section discusses the frequency domain analysis of HPS model
consisting of WTG, DEG and STPG. The simulation results of this test
model are shown in Fig. 6. The subplots of this figure i.e. Fig. 6(a) and
(b) are the frequency deviation profile of the model obtained without
SMES and with SMES. When both the curves are compared, it is re-
vealed that the scheduled frequency is reached within 14.1062 s when
SMES is applied to the system. In the same system, the stable condition
is reached in 50.9047 s when SMES is not applied. Thereby, the small
signal stability of the system is improved by introducing the SMES to
the system. It may also be noted that the oscillations are attenuated and
the system has reached its steady state condition. Therefore, QOHS
based PID controller with designed SMES enhances both the system
stability and damping characteristics.
Fig. 6(c) shows the convergence mobility of the FOD when SMES is
applied to the system. Convergence movement of FOD value, as ob-
tained by the proposed QOHS algorithm, shows promising convergence
characteristic. It is seen that the FOD value of only model is 40.73 and
that of the model with SMES is 9.5719 (refer Table 4). This shows that
the FOD value is less for the model with SMES. Low FOD value always
offers better optimized gain and better tuned efficiency of the design
controller.
Fig. 10. Dynamic response profile pertaining to Scenario-III configurations with 1.0 p.u. SLP (a) Δf without SMES, (b) Δf with SMES and (c) QOHS based con-
vergence profile of FOD.
155
Journal of Energy Storage 19 (2018) 145–159
5.1.4. Simulation results under random load variation
A very new kind of load profile having unequal magnitude at dif-
ferent instant of time is studied. The same is applied to the studied IHPS
model for the LFC performance analysis. This load profile is adopted for
better understanding of the performance of the designed QOHS based
PID controller. Fig. 7(a) shows the frequency deviation profile of IHPS
(refer Fig. 1(a)) without SMES while Fig. 7(b) portrays the same with
SMES. It is evident from both of these figures that when SMES is applied
to the system, the system is much more stable and the oscillations are
also less as compared to the system without SMES unit.
5.1.5. Scenario II: HPS with one DEG and one STPG
In this scenario, one DEG and one STPG are taken into consideration
as test system (refer Fig. 1(b)).
5.1.6. Simulation results with 1.0 p.u. load demand
The dynamic response of this system subjected to 1.0 p.u. load
perturbation has been shown in Fig. 8. The sub-plots of this figure show
the frequency deviation plot of the system without and with SMES as
well as convergence profile of FOD. Same as Scenario-I, it may be ob-
served that the system has better dynamic response with SMES than the
system without SMES. The system without SMES has reached the steady
state at a little bit more time than that of the same system with SMES
(expected to be around 8.6787 s, refer Table 3). Moreover, when the
S. Ganguly et al.
Fig. 11. Dynamic response profiles corresponding to Scenario-III configurations with random SLP (a) Δf without SMES and (b) Δf with SMES.
system is connected with SMES, the change in maximum undershoots
for frequency deviation plot is 0.0648 p.u. less.
The value of FOD pertaining to iteration cycle is also recorded to get
the convergence profile of QOHS algorithm as applied in the system
with SMES (refer Fig. 8(c)). The FOD of only model is 48.073 and that
of model with SMES are 7.1718 (refer Table 4). This shows that the
model with SMES has better frequency regulation capability.
5.1.7. Simulation results under random load variation
The system frequency deviation of this model without SMES and
with SMES is depicted in sub-plots of Fig. 9(a) and (b), in order. While
comparing both the plots altogether, it may be observed that both the
magnitude of peak overshoot and undershoot are suppressed with SMES
as compared to without SMES. Therefore, the QOHS based designed
controller achieves better damping even for random load variation.
5.1.8. Scenario III: HPS with one DEG and one WTG
The third scenario discusses the simulation analysis with the above
described controllers in hybrid WTG, DEG based power system (refer
Fig.1(c)). In this scenario, 0.01 p.u. SLP is applied in the system base
wind power and 1.0 p.u. SLP and random load variation is applied to
the system.
5.1.9. Simulation results with 1.0 p.u. load demand
The load demand of the hybrid WTG and DEG isolated system are
Fig. 12. (a) SLP of 0.01 p.u. load disturbance applied to GHPS, (b) dynamic response profile of GPHS configuration, (c) QOHS based convergence profile of FOD for
STPS and (d) QOHS based convergence profile of FOD for WTG and STPS.
156
Journal of Energy Storage 19 (2018) 145–159
considered as ΔPL = 1 p.u. SLP. The frequency deviation profiles are
presented in Fig. 10(a) and (b). When the system is without SMES, the
settling time is found to be 50.9221 s whereas the settling time is de-
creased to 14.730 s when the SMES is applied to the system. Comparing
the two settling times, it is clear that the system with SMES has reached
the stability state much earlier. The convergence of the FOD value with
SMES is presented in Fig. 10(c). The FOD values of only model and
model with SMES are 59.7409 and 16.0604, respectively (refer
Table 4). Thereby, it is showed that the system with SMES has better
convergence sketch than the system without SMES.
5.1.10. Simulation results under random variation
The system frequency deviation plot is provided in Fig. 11(a) and
(b). Here, QOHS algorithm has been used in tuning controller-I, con-
troller-II and SMES gain values for achieving better damping for fre-
quency deviation profile. Similar to Scenario-I and Scenario-II, the same
inferences may be made. The system with SMES is much more stable
than the system without SMES.
5.2. Case II. Basic detail
The second case is the combination of one WTG and one STPS. In
the GHPS model, SLP of 1% is taken into account for 100 s.
S. Ganguly et al. Journal of Energy Storage 19 (2018) 145–159

5.2.1. Performance investigation pertaining to case II 6. Conclusion

The second case deals with GHPS model with one STPS and one
WTG. In the studied model, a SLP of 0.01 p.u. is applied to the system.
The profile of applied SLP is shown in Fig. 12(a). The system response
of this model is depicted in Fig. 12(b). In this figure, the comparative
time domain analysis of frequency deviation plot is done for the system
with only STPS and with the system consists of WTG and STPS. The
dynamic response of this system, employing WTG and STPS, shows that
Δf is driven back to zero with fewer oscillations as referred to the
system with only STPS unit. From Table 5 it would be fascinating to
note that the settling time for the frequency deviation with STPS as well
as with WTG and STPS models are 13.2232sec and 4.0750 s, respec-
tively.
The convergence plot of the proposed QOHS algorithm for the
model with STPG and the model with WTG and STPS are presented in
Fig. 12(c) and (d), respectively. As tabulated in Table 6, the FOD values
of only STPS and WTG with STPS model are 0.1521 and 0.0229, re-
spectively. This shows that the designed controller using QOHS per-
formed well even in grid connected HPS model. It may also be noted
that the WTG may be implemented for the frequency regulation of the
power system.
The impacts of renewable energy sources for the load frequency
stabilization of the power system have been studied in this article. For
this work, a critical investigation is performed for load frequency sta-
bilization of the HPS models in terms of both isolated and grid con-
nected modes having degrees of uncertainties. In the isolated mode, a
SMES unit is connected to the IHPS to reduce the power oscillations.
The QOHS algorithm has been effectively utilized as a proposed control
strategy to design the gain parameters of PID controller and the tunable
parameters of the SMES which helps to bring the system to its normal
operating condition following load perturbation.
From the simulation results of the IHPS models presented in this
work, it is clear that the frequency stabilization is going to be better
when SMES is connected to the system. Moreover, when performance
index is evaluated, it is found that the system error is reduced with the
SMES connected system. The QOHS based convergence profile of FOD
is also plotted for IHPS model. The dynamic response of the system in
grid connected mode has been also studied. It is observed that the
system response gets improved with the addition of WTG to the STPS as
talk about the system with only STPS.

Appendix A

Nominal system parameters of the studied IHPS and GHPS models are presented in Table A1. Electrical parameters of the wind turbine are
presented in Tables A2–A4. Table A5 may be referred for the parameters of the QOHS algorithm.

Table A1
Nominal system parameters of the studied IHPS and GHPS models.
Model Parameters

Wind unit KP1 = 1.25 KP2 = 1.00 KP3 = 1.40 KTP = 0.0033 KIG = 0.9969
KPC = 0.080 TP1 = 0.60 sec TP2 = 0.041 sec TP3 = 1.0 sec

Diesel unit KD = 0.3333 RD = 3.0Hz / pu TD1 = 1.00 sec TD2 = 2.00 sec TD3 = 0.025 sec
TD4 = 3.0 sec
Thermal units Tg = 0.8 sec Tt = 0.3 sec RD = 2.4 Hz/p.u KP = 120 TP = 14.4 sec
SMES KF = 50
Solar unit K S = 1.8 KT = 1 TS = 1.8 sec TT = 0.3 sec

Table A2
Characteristics for rotor model of the simulated wind
turbine [42].
Rotor Characteristic Value

Minimal rotor speed 9 rpm

Rated rotor speed 18 rpm
Rotor diameter 60 m
Rotor swept area 2827 m2
Rated wind speed 14 m/s
Inertia constant 0.72 s

Table A3
Characteristics of the simulated shaft.
Shaft characteristic Variable Value

Rated power Pshaft 1.5 MW

Rated speed ωn 18 rpm
Turbine damping Dturb 14 × 106 Nms/rad
Rotor inertia Jturb 6.1 × 106 Kg mm
Shaft stiffness Ks 83 × 106 Nm/rad

157
S. Ganguly et al. Journal of Energy Storage 19 (2018) 145–159

Table A4
Characteristics of the simulated pitch angle controller.
Characteristic Variable Value

Blade angle controller gain ka 100°/p.u.

Lead time constant Tr 5
Speed reference ωref 1.25 p.u.
Servo time constant T 0.5 sec
Closing rate of change limit Rate Closing −15°/sec
Opening rate of change limit Rate closing 15°/sec
Min. Blade angle Bmin 0°
Max. Blade angle Bmax 70°

Table A5
Parameters of the QOHS algorithm.
Parameters Values Parameters Values

Harmony memory size 60 Harmony consideration rate 0.9

Minimum pitch adjusting 0.45 Maximum pitch adjusting 0.98
rate rate
Minimum bandwidth 0.00005 Maximum bandwidth 50
Jumping rate 0.8 number of iteration 100

References
[1] R. Ali, T.H. Mohamed, Y.S. Qudaih, Y. Mitani, A new load frequency control ap-
proach in an isolated small power systems using coefficient diagram method, Electr.
Power Energy Syst. 56 (2014) 110–116.
[2] H. Holttinen, R. Hirvonen, Power system requirements for wind power, in:
T. Ackermann (Ed.), Wind Power in Power Systems, John Wiley & Sons Ltd,
2005Chapter 8.
[3] M. Gouveia Eduardo, A. Matos Manuel, Evaluating operational risk in a power
system with a large amount of wind power, Electr. Power Syst. Res. 79 (2009)
734–739.
[4] Senjyu Tomonobu, et al., A hybrid power system using alternative energy facilities
in isolated island, IEEE Trans Energy Conver. 20 (2) (2005) 406–414.
[5] D. Rekioua, S. Bensmail, N. Bettar, Development of hybrid photovoltaic-fuel cell
system for stand-alone application, Int. J. Hydrogen Energy 39 (3) (2014)
1604–1611.
[6] Z. Mokrani, D. Rekioua, T. Rekioua, Modeling, control and power management of
hybrid photovoltaic fuel cells with battery bank supplying electric vehicle, Int. J.
Hydrogen Energy 39 (27) (2014) 15178–15187.
[7] N. Mezzai, D. Rekioua, T. Rekioua, A. Mohammedi, K. Idjdarane, S. Bacha,
Modeling of hybrid photovoltaic/wind/fuel cells power system, Int. J. Hydrogen
Energy 39 (27) (2014) 15158–15168.
[8] S. Aissou, D. Rekioua, N. Mezzai, T. Rekioua, S. Bacha, Modeling and control of
hybrid photovoltaic wind power system with battery storage, Energy Convers.
Manag. 89 (2015) 615–625.
[9] S. Tamalouzt, N. Benyahia, T. Rekioua, D. Rekioua, R. Abdessemed, Performances
analysis of WT-DFIG with PV and fuel cell hybrid power sources system associated
with hydrogen storage hybrid energy system, Int. J. Hydrogen Energy 41 (45)
(2016) 21006–21021.
[10] L. Wang, D.J. Lee, W.J. Lee, Z. Chen, Analysis of a novel autonomous marine hybrid
power generation/energy storage system with a high-voltage direct current link, J.
Power Sources 185 (2008) 1284–1292.
[11] W. Li, G. Joos, Comparison of energy storage system technologies and configura-
tions in a wind Farm, IEEE Power Electronics Specialist’s Conference, PESC, (2007),
pp. 1280–1285.
[12] I. Ngamroo, Robust frequency control of wind-diesel hybrid power system using
superconducting magnetic energy storage, Int. J. Emerg. Electr. Power Syst. 10 (2)
(2009).
[13] N. Koshijuka, F. Ishikawa, H. Nasu, Progress of superconducting bearing technol-
ogies for flywheel energy storage system, Physica C 386 (2003) 444–450.
[14] http://www.superpower-inc.com/content/superconducting-magnetic-energy-
storage-smes#.
[15] H. Yang, Z. Wei, L. Chengzhi, Optimal design and techno-economic analysis of a
hybrid solar-wind power generation system, Appl. Energy 86 (2009) 163–169.
[16] P.K. Ray, S.R. Mohanty, N. Kishor, Proportional-integral controller based small-
signal analysis of hybrid distributed generation systems, Energy Convers. Manag. 52
(4) (2011) 1943–1954.
[17] D.C. Das, A. Roy, N. Sinha, PSO based frequency controller for wind-solar-diesel
hybrid energy generation/energy storage system, Proc. ICEAS (2011) 458–463.
[18] T. Zhou, W. Sun, Optimization of battery super-capacitor hybrid energy storage
station in wind/solar generation system, IEEE Tran. Sustain. Energy (5) (2014)
408–415.
[19] A. Maleki, A. Askarzadeh, Artificial bee swarm optimization for optimum sizing of a
stand-alone PV/WT/FC hybrid system considering LPSP concept, Sol. Energy 107
(2014) 227–235.
[20] D.C. Das, A.K. Roy, N. Sinha, GA based frequency controller for solar thermal–-
diesel–wind hybrid energy generation/energy storage system, Electr. Power Energy
Syst. 43 (1) (2012) 262–279.
[21] I. Pan, S. Das, Fractional order fuzzy control of hybrid power system with renewable
generation using chaotic PSO, ISA Trans. 62 (2016) 19–29.
[22] Z.W. Geem, J.H. Kim, G.V. Loganathan, A new heuristic optimization algorithm:
harmony search, search, Simulation 76 (2) (2001) 60–68.
[23] K.S. Lee, Z.W. Geem, A new structural optimization method based on the harmony
search algorithm, Comput. Struct. 82 (9–10) (2014) 781–798.
[24] K.S. Lee, Z.W. Geem, A new meta-heuristic algorithm for continuous engineering
optimization, harmony search theory and practice, Comput. Methods Appl. Mech.
Eng. 194 (36–38) (2005) 3902–3933.
[25] D. Manjarresa, I. Landa-Torresa, S. Gil-Lopeza, et al., A survey on applications of the
harmony search algorithm, Eng. Appl. Artif. Intel. 26 (8) (2013) 1818–1831.
[26] H.R. Tizhoosh, Opposition-based learning: a new scheme for machine intelligence,
Int. Conf. Comput. Intell. Model. Control Autom. 1 (2005) 695–701.
[27] S. Rahnamayan, H.R. Tizhoosh, M. Salama, Quasi-oppositional differential evolu-
tion in evolutionary computation, IEEE Congress, (2007), pp. 2229–2236.
[28] T. Mahto, V. Mukherjee, A novel quasi-oppositional harmony search algorithm and
fuzzy logic controller for frequency stabilization of an isolated hybrid power
system, Electr Power Energy Syst. 66 (1) (2015) 247–261.
[29] T.S. Bhatti, A.A.F. Al-Ademi, N.K. Bansal, Load frequency control of isolated wind
diesel hybrid power systems, Energy Convers. Manag. 38 (9) (1997) 829–837.
[30] B. Sedaghat, A. Jalilvand, R. Noroozian, Design of a multilevel control strategy for
integration of stand-alone wind/diesel system, Int. J. Electr. Power Energy Syst. 35
(1) (2012) 123–137.
[31] S. Ganguly, T. Mahto, V. Mukherjee, Integrated frequency and power control of an
isolated hybrid power system considering scaling factor based fuzzy controller,
Swarm Evol. Comput. 32 (2017) 184–201.
[32] D.C. Das, A.K. Roy, N. Sinha, GA based frequency controller for solar thermal-
diesel-wind hybrid energy generation/energy storage system, Electr. Power Energy
Syst. 43 (1) (2012) 262–279.
[33] B. Praghnesh, R. Ranjit, S.P. Ghosal, Load frequency stabilization by coordinated
control of thyristor controlled phase shifters and superconducting magnetic energy
storage for three types of interconnected two-area power system, Int. J. Elec. Power
Energ. Syst. 32 (2010) 1111–1124.
[34] R.K. Sahu, S. Panda, S. Padhan, Optimal gravitational search algorithm for auto-
matic generation control of interconnected power system, Ain Shams Eng. J. 5 (3)
(2014) 721–733.
[35] A. Banerjee, V. Mukherjee, S.P. Ghoshal, Modeling and seeker optimization based

158
S. Ganguly et al.
simulation for intelligent reactive power control of an isolated hybrid power
system, Swarm Evol. Comput. 13 (2013) 85–100.
[36] A. Banerjee, V. Mukherjee, S.P. Ghoshal, An opposition-based harmony search al-
gorithm for engineering optimization problems, Ain Shams Eng. J. 5 (1) (2014)
85–101.
[37] C.K. Shiva, V. Mukherjee, A novel quasi-oppositional harmony search algorithm for
automatic generation control of power system, Appl. Soft Comput. 35 (2015)
749–765.
[38] H.R. Tizhoosh, Opposition based learning: a new scheme for machine intelligence,
Proc. Int. Conf. Comput. Intell. Model. Control Autom. 1 (25) (2005) 695–701.
159
Journal of Energy Storage 19 (2018) 145–159
[39] S. Rahnamayan, H.R. Tizhoosh, M.M.A. Salama, Opposition versus randomness in
soft computing techniques, Appl. Soft Comput. 8 (2) (2008) 906–918.
[40] C.K. Shiva, G. Shankar, V. Mukherjee, Automatic generation control of power
system using a novel quasi-oppositional harmony search algorithm, Electr. Power
and Energy Syst. 73 (2015) 787–804.
[41] I. Pan, S. Das, Fractional order fuzzy control of hybrid power system with renewable
generation using chaotic PSO, ISA Trans. 62 (2016) 19–29.
[42] F.M. González-Longatt, P. Wall, V. Terzija, A simplified model for dynamic behavior
of permanent magnet synchronous generator for direct drive wind turbines, June,
Power Tech, 2011 IEEE Trondheim, IEEE, 2011, pp. 1–7.