Journal of Electrical Engineering ISSN 1582-4594 - Vol 21 No 3 (2021)

A NOVEL PARTICLE SWARM OPTIMIZATION ALGORITHM FOR MAXIMUM

POWER POINT TRACKING BASED PHOTOVOLTAIC SYSTEM
E.MALARVIZHI1, J.KAMALA2
Assistant Professor, Department of Electronics and Communication Engineering,
St.Joseph’s College of Engineering1, Chennai-119, College of Engineering Guindy2,
Annauniversity, Chennai-25.TamilNadu, India.
[email protected], [email protected].
A.SIVASUBRAMANIAN3
Professor, School of Electronics, VIT University, Chennai-127, TamilNadu, India.
[email protected].

Abstract: In recent years, solar energy is effectively 1. Introduction

utilized as an alternate energy source for generating These days increased demand for electricity
electricity. Maximum Power Point Tracking(MPPT) consumption and negative effects on the environment
is applied to the photovoltaic (PV) system to extract are the major issues [1]. These are due to depleting of
maximum power (MP). Numerous algorithms are fossil fuels at a faster rate. To overcome these issues
developed and implemented to extract the MP under continuous research is being conducted to seek
varying environmental conditions. One such efficient alternate sources of energy for power
algorithm is the Particle Swarm Optimization generation [2]. Therefore attention is being given to
algorithm(PSO).This article introduces a novel PSO the utilization of Renewable Energy Sources (RES).
algorithm with Cauchy distribution to track MP from RES are solar, hydropower, biomass, wind,
the PV system. It is designed to overcome the geothermal energy etc. RES are economical and they
drawback of slow convergence rate of conventional also reduce the environmental issues caused by power
PSO. Parameters required for conventional PSO are generation from non-renewable energy sources such
inertial weight, acceleration coefficients, and a as fossil fuels. Because of the merits of RES, these are
number of particles. In case of Cauchy PSO(CPSO), utilized as best alternate for generation of electrical
tuning parameter is the number of particles. In order energy. Nowadays, the generated electricity from
to increase the convergence speed, Cauchy various sources is distributed via the centralized grid.
distribution is used instead of normal distribution Various resources are referred to as Distributed
function to generate the random numbers. The Energy Resources (DER)[3].
advantage of this algorithm is that it provides the Among the RES solar energy and wind energy
global best-optimized result at a faster are currently widely deployed. Compared to solar
convergence speed. It has the ability to track the MP energy, wind energy has some limitations. The
in extreme climatic conditions with varying loads. rotatary motion of wind turbine is used to produce
The proposed method outperforms the standard PSO electricity. Following are some of the important
and some of the existing methods in terms of quick practical difficulties. Wind turbines are very large
convergence and also has a simplified structure. machines. It occupies larger area and also wind
turbine blades make more noise. Another important
Keywords: Photovoltaic(PV) system, Maximum fact is that the wind mills are installed near shores, and
Power Point Tracking (MPPT), Particle Swarm on high mountain. Site selection also plays a crucial
Optimization (PSO) Algorithm, Cauchy PSO role in several aspects. Financial returns from the
(CPSO). project, ease of construction, maintenance and overall
safety are the factors that decide the overall design and

24
 Journal of Electrical Engineering ISSN 1582-4594 - Vol 21 No 3 (2021)

implementation of wind power generation magnitude is zero. On either side of MPP, its
systems[4].Therefore solar energy has become widely magnitude has negative and positive increments
essential to compensate the energy demand. Its [9-11]. It is developed to increase the convergence
environmental friendly nature, abundance presence on time of tracker. HC approach uses the PV array output
the earth’s surface, it being totally free from pollution, current. Change in current is measured before and
low operational cost and also renewable and after a change in duty cycle. Therefore this method
sustainable characteristics are the merits of solar perturbs the duty cycle of the converter.
energy [5]. Although the operational cost is low for Traditional MPPT techniques have two major
the solar energy , the cost of solar modules are high difficulties. Under PSC it tracks only the first local
.Still the grid connected PV power generation systems MPP.It is unable to track the global MPP and they also
are being commercialized in many countries because suffer from steady state oscillations [12-14].This
of its long term benefits. Further more generous means after capturing MPP, the operating point
financial schemes for example the feed in tariff and oscillates back and forth around MPP. This limits
subsidized policies are introduced in various countries maximum power generation.
resulting in rapid growth of the industry [5].Recently Fuzzy based controllers and neural network are
the power generation systems designed using another type of controllers. The fuzzy controller uses
integration of solar and wind energy. The reason is to an ad-hoc technique to handle nonlinearity problems.
generate interminable power depend upon the Its execution is complex with a process of
environmental conditions [6]. fuzziffication, rule based storage, inference and
Conversion of photon energy to electrical energy defuzziffication operations. Implementation requires
is called Photovoltaic (PV) effect. PV cells are the more knowledge for both user and designer [15].
smallest element used to produce power. Integration Neural Network is an Artificial Intelligence (AI)
of large number of PV cells produces PV module, based technique and its implementation process is
large number of series and parallel connected modules much complicated. It uses a large amount of data, for
produce PV arrays. Nowadays PV arrays are tuning. Because of the varying weather conditions,
constructed pyramidal like shapes used to upgrade the real time implementation needs intelligence
radiation intensity falling on solar panels.[7]. To processors for better performance [15].
extract the MP solar panels are to be operated at Evolutionary Algorithms (EA) such as PSO ,
Maximum Power Point (MPP). Because of the Genetic Algorithm (GA), Ant Colony Algorithm
dynamic environmental conditions such as (ACO), Evolutionary Programming (EP), Simulated
temperature and irradiance, PV characteristic varies Annealing (SA) etc. are introduced to improve the
nonlinearly. It is difficult to locate the MPP. Another performance of MPPT under dynamic atmospheric
common issue is Partial Shading Condition (PSC) due conditions. In SA, setting of control parameters is a
to surrounding buildings, trees, passing clouds in the difficult task and convergence speed is slow when
sky etc. Under PSC the PV curve exhibits multiple applied to a real system. Though the GA techniques
MPPs with one global and several local peaks [8]. To have been employed successfully to solve complex
overcome these issues MPPT controllers are optimization problems, its efficiency is degraded due
integrated with DC-DC converters. to its premature convergence. EP seems to be good
Conventional MPPT algorithms are Perturb and choice to solve multimodal optimization problem. But
Observe (P&O), Incremental conductance (INC) and it takes long simulation time to obtain the optimum
Hill climbing (HC). These are very common MPPT solution for this kind of problems [16]. The word
techniques. P&O compares PV array voltage (or optimization broadly refers to locating either the
current) to the constant reference voltage (or current) maxima or minima of real functions. It uses a
at the MPP under specific atmospheric condition. The systematic approach to choose real or an integer
difference of voltage (or current) is used to decide the variable from within the allowed set is called search
direction of tracking. Simply this method perturbs the space.
operating voltage (or current)[9,10]. In INC output Among these PSO shows outstanding
conductance is taken for reference. At MPP its performance in attaining optimized solutions. The

25
 Journal of Electrical Engineering ISSN 1582-4594 - Vol 21 No 3 (2021)

most suitable and attractive approach to obtain MP  Modifying the position and velocity update
from PV array is the PSO algorithm. Merits of PSO rules, outside the equations, and so on
are its ability to track the MPP location quickly and [22,23].
accurately irrespective of the climatic conditions. It Changes in velocity formula is implemented in
eliminates the steady state oscillations and produces improved PSO–MPPT algorithm. Vanxay
more stable output unlike the conventional methods Phimmasone et al.[24] introduced the fourth term
[17]. It doesn’t require the gradient of the objective which is called as repulsive term. The repulsive term
function and uses only the values of the objective k k k k 3
is given by c3r3 (cent  Si ) /( cent  Si  d ) .
function. It eliminates complex computational
procedures. It also has a simple structure with fast Where, c3 r3 term is same as those used in
computational ability. Characteristics of Conventional
conventional PSO. Where, cent is the center of all
PSO are, it converges quickly and has good search
particles. The term‘d’ is small constant added to avoid
accuracy. Convergence is defined when the particles
k k
stop at the desired operating point to provide global divide by zero error for the values of ( cent  Si ).
optimal result [18, 19]. In practice, the algorithm stuck Qing Zhang et al. [25] presented a modified PSO
at local maxima is called premature convergence and with both mutation and crossover functions to
it requires larger search space. increase the convergence speed and to avoid the
To tackle the issue of slow convergence of PSO trapping of PSO into local maxima. Authors
and Gaussian PSO, CPSO is proposed and introduced Cauchy random number in velocity and
implemented. It is an improved version of Gaussian position update expressions.
PSO [GPSO] [20] with enhanced convergence speed. Venkatesh Kumar et al.[16] introduced the
Other advantages are reduction in the number of Cauchy mutation in the position update equation of
tuning parameters, speedy convergence rate and faster PSO to optimize the results of economic dispatch
output settling time. problems.
Remainder of this article is organized as follows. Hui wang et al.[26] proposed a hybrid PSO with
Section 2 explains the related work to CPSO Method. Cauchy mutation. The inertial weight is modified with
Section 3 presents the concepts in conventional PSO. the help of Cauchy random number. It is used to
Functional blocks of the proposed method, enhance the position of best particle to achieve faster
configuration and implementation of CPSO convergence.
algorithm are discussed in section 4. In section 5 The difference between the above mentioned
functional modules are validated through simulation methods and the proposed method is faster
results. Section 6 presents the merits of the proposed convergence is achieved using minimum number of
method as compared with other techniques and tuning parameters.
concluding remark on future scope of this work.
3. Conventional PSO
2. Related work PSO is a population based optimization
Several solutions are dealt in the literature to algorithm. It is one of the metaheuristic methods
overcome the PSO convergence issue. These are as and it follows the random search techniques.
follows: Tuning parameters required for the execution of
 Introducing variable parameters the conventional PSO algorithm are acceleration
 Use of different swarm topologies coefficients (c1,c2), inertial weight(  ) and
 Changing the velocity update formula with maximum velocity as given by equations(1) and
the enhancements applied in the inertial (2) [17,27]. Equation (1) is the velocity update
weight [21]. equation and (2) is the positions update equation.
 Modifying the particle position and velocity
through the mutation operator. v i ( k  1)  vi ( k )  c1 r1 (( p best ,i  x i ( k ) 

c 2 r 2 ( g best  xi ( k ))

26
 Journal of Electrical Engineering ISSN 1582-4594 - Vol 21 No 3 (2021)

(1) the optimization precision.

Important property of Cauchy distribution is that it
xi ( k  1)  xi ( k )  vi ( k  1)
(2) has heavier tail [28]. This means its probability
Where r1, r2 are uniformly distributed random distributions are not exponentially bounded. That is,
numbers, pbest and gbest are particles individual and its distribution is greater than exponential distribution.
global best positions respectively from the target Some of the other properties of this distribution is it
position and k is number of iterations. has no expected value and other moments also. Its
median and mode are equal. Probability Density
4. Proposed Cauchy PSO method Function (PDF) of Cauchy distribution is given in
equation (5) [29].
Existing methods have complex computational 1
structure. Proposed method has less complex f ( x; x0 ,  ) 
 2
computational structure with faster convergence.  x  xo 
 1    
In the proposed method the velocity update     
(5)
formula is modified. Velocity equation is constructed
x0 - Location parameter specifies the location of the
by without inertial weight, maximum velocity and
acceleration coefficients (cognitive and social peak of distribution.  - Scale parameter or
parameters) to achieve the convergence. Gaussian and probable error. Its value is half the width of PDF at
Cauchy probability distributions are very effective for the half maximum height. Amplitude of Cauchy PDF
updating the velocity equation. 1
Cauchy PSO is also stochastic, population based is located at x  x0 .  =1, x0 =0, the Standard
swarm intelligence evolutionary algorithm. Equations 
(3) and (4) present the velocity and position updates of Cauchy distribution with PDF[30] ,
the proposed PSO. 1
f ( x;0,1) 
2
 (1 x ) (6)
vi (k  1)  vi (k )  r1 (( p best,i
 xi ( k )  r 2 ( g
best
 xi (k ))
In this work , the random numbers are
characterized by standard Cauchy distribution. It is a
(3) special case of the student’s t-distribution. Student’s t-
distribution is looks almost identical to normal
xi ( k  1)  xi ( k )  vi ( k  1) distribution but it possess heavier tail like Cauchy
(4) distribution [31].
Here r1 and r2 are the Cauchy distributed random PDF of student’s t- distribution is given in
numbers. Cauchy random numbers possesses the equation (7).
long jump ability. It enhances the velocity of the
particles. Also this property helps the particles to get  n 1
 
best position [3]. Another useful property is that, it is f ( x) 
 2    < x< ;
capable of providing huge search space for the  n 
 n  x 
2 2
particles. This is due to its tail decay very slowly. n     1
This yields better result for the search.  2  n 
The proposed method and the conventional PSO (7)
are similar except that the initialization of the tuning Where, ‘n’ is called degrees of freedom. If
parameters. Importance of tuning parameters in PSO Student’s t-distribution has degree of freedom as 1 its
is that, it greatly influences the algorithm PDF is same as that of the standard Cauchy
performance. It is often referred to as exploration and distribution. It is presented in equation (8).
exploitation tradeoff. Exploration is the ability to
1
check the test space and provide global optimal f ( x)  2
  < x <  ; for n=1. (8)
results. Exploitation is the ability to concentrate on  (1  x )

27
 Journal of Electrical Engineering ISSN 1582-4594 - Vol 21 No 3 (2021)

Now it is seen that the Equations (8) and (5) are for the PV array by adjusting the duty cycle of the
the same. buck boost (BB) converter.
Figure 1 shows the flowchart of the proposed Start
method. The implementation starts with three
different duty cycles. These are referred to as the
particles, di, (i=1, 2, 3). Selection of initial duty cycles Initial Position Of The
Particle di= d1,d2,d3
are based on the MP output of the system. First duty
cycle value (d1) is chosen such that the duty cycle
Set di to
value required generating 80% of the MP of the
converter
system. The subsequent duty cycles (d2, d3) are 65%
and 50% of MP outputs respectively [13]. In the first
Measure Vi
iteration these are considered to be the personal best and Ii, Pi=Vi
(Pbest) values of the algorithm. For each duty cycle, the
current and voltage are sensed and the corresponding
power is calculated. This power is the objective
function of the system [32]. Its mathematical Is ΔP > Is
relationship is given in equation (9). It is compared for threshold maximum
different duty cycles. Depending on the received Y generation?
power, the Pbest value is modified. It is presented in the
flow chart. N
Velocity V=0
k k 1
P  d i  > P  d i 
Generation Is first Pbest
(9)
    K=1 generation ?
Y
k k 1
P  d i  -New power and P  d i  -Old power
N
    Calculate
Is Pi>Ppre, vi ?
position of N
each particle di
The Gbest value is computed based on the Y
maximum output power. The value of Pbest Pbest i=dpre ,v i
corresponds to the maximum output power is taken as Pbesst i =di
Gbest. The duty cycle corresponding to the MP is
applied to the power converter for each cycle.
Find Gbest , Update
Efficiency of the system is computed using position and velocity of
equation (10). Where, Po is the output power and each particle
Increase generation
Pin is the input power of the system.
P
Tracking efficiency:   o  100 % (10) Fig.1.Flow chart of CPSO algorithm [33].
Pin
Load
4.1 Block diagram of the proposed system DC/DC Converter
PV
PV system consists of PV array, controller
unit and load called Power Conditioning System [5].
Figure 2 shows the block diagram of the proposed
V& I Sensing P WM-Generator
system. Buck-Boost converter is designed to transfer
the regulated DC output to the load. The output
voltage and current of PV array are the inputs of the CPSO- MPPT Algorithm and Duty Cycle
digital controller to perform iterative tracking for the Testing
MPP. The controller decides the operating voltage
Fig.2. Block diagram of MPPT based PV system [11]

28
 Journal of Electrical Engineering ISSN 1582-4594 - Vol 21 No 3 (2021)

A. System Configuration and Implementation Table 1 Parameters of PVpanel

using CPSO Peak Power(W) PMPP 42.7
Figure 3 shows the MATLAB/SIMULINK
Peak power voltage(V) VMMP 21.7
simulation model for the PV system used in this
Peak power current(A) IMMP 1.97
discussion. As per the block diagram the simulation
model consists of PV panel, converter and the MPPT Open circuit voltage(V) Voc 26.4
control circuitry. PV array consists of series Short circuit current(A) Isc 2.1
connection of 44 solar cells. Each cell has open circuit Number of series cells Ns 44
voltage, Voc = 0.6V and short circuit current Isc = 2.1
A. The sampling period (or time interval) for sensing Table 2 shows the specifications of the BB
voltage and current from the PV panel is 0.01 seconds converter. It is designed with the help of BB
to obtain the input power. Input power is the objective converter design equations [34].
function. The next important block is a CPSO function
block. Inputs are panel voltage and current. Based on Table 2 Specifications of BB Converter.
the power product, the algorithm computes duty
cycle for the converter to obtain the MP. Third section Input Inductor 5mH
is converter section and the system is implemented Output capacitor 25µF
with a BB converter.
Table 1 shows the specifications of solar panel Output ripple voltage 0.05 V
at 25°C, temperature and insolation of 1000 W/m2. Output ripple current 0.1A
Switching frequency 25kHz

Converter Section

Fig.3.Simulation circuit for CPSO based MPPT system

5. Results and Discussion  Output settling time and

In order to evaluate the effectiveness of the  Convergence time of particles.
algorithm and to validate the results of the proposed Apart from that the system response is analyzed for,
method, two different existing algorithms  Varying load conditions
“conventional PSO and Deterministic PSO  Varying irradiance conditions.
(DPSO)[17]” are implemented and analyzed. The
following parameters of the proposed method are
compared,

29
 Journal of Electrical Engineering ISSN 1582-4594 - Vol 21 No 3 (2021)

5.1 Output settling time DPSO. Therefore from the results it is seen that the
proposed method has the fastest settling time.
The output settling time of conventional PSO
is presented in Figure 4 .The settling time is
around 1.2s. Figure 5 shows output of DPSO.
Output settling time is nearly 0.35s and Figure 6 is
output wave form of CPSO. It shows the output
settling time as 0.25s.

Fig. 7. Comparison chart for output setting time

5.2 Particles converging time of PSOs
Figure 8 shows the duty cycle exploration of
standard PSO for 0.2s. All the particles are converged
around 0.65s. Figure 9 is for DPSO algorithm.
Fig. 4. PSO_Output
Particles converged at nearly 0.2s. Figure 10 is duty
cycle exploration of CPSO. Convergence of particles
takes just 0.12s.Among these algorithms DPSO
technique doesn’t use any random numbers to avoid
randomness of PSO. It is designed such that the
particles follow deterministic behavior to speed up the
operation. Though the proposed method uses the
random numbers it still converges earlier than DPSO.

Fig. 5.DPSO_Output

Time (s)
Fig. 8. Duty Cycle exploration of PSO

Fig. 6. CPSO_Output
Figure 7 is the comparison chart between the
output settling time of all the different implemented
versions of PSO discussed in this work. Ratio
between the output settling time of PSO ,DPSO and
CPSO are analyzed. From the figure CPSO is 4.8 Time (s)

times faster than PSO and 1.4 times faster than Fig. 9.Duty cycle exploration of DPSO

30
 Journal of Electrical Engineering ISSN 1582-4594 - Vol 21 No 3 (2021)

Time,(s)

Fig. 10.CPSO duty cycle exploration

Figure13.Load resistor and CPSO P_out for high range of loads
Figure 11 shows the converging time of all
Figure 14 shows the load versus duty cycle (duty
different PSOs. From this chart it is seen that the in
ratio) curve for the proposed method. For the load
CPSO algorithm particles converged quickly. It is fluctuations, the algorithm gradually expand the duty
5.41 times faster than PSO and 1.66 times faster than cycle to track the maximum panel power.
DPSO.

Fig. 14. load resistance Vs Duty cycle of CPSO

Fig.11. Comparison chart for particles converging time
5.4 Respose to Irradiance changes
5.3 Response to load resistance variations Figure 15 is the plot between irradiance versus
Figure 12 and 13 are plot between load resistance power. As we know these two quantities are directly
and output power of the proposed system. The output propotional with each other, for lower irradiance( Irs)
power of the BB converter using CPSO algorithm is the power output is low and is increases linerly.
recorded for various loads. It provides maximum
output power irrespective of load variations.Output
power fluctuates approximately by only 1W from
minimum load value to maximum load.

Ir (W/m2)
Fig.15.IrradianceVs P_out of CPSO
Figure 16. is a plot between duty cycle versus
Fig. 12.Load resistor and CPSO P_out for low range of loads power for the proposed method.This duty ratio

31
 Journal of Electrical Engineering ISSN 1582-4594 - Vol 21 No 3 (2021)

variations are obtained for various load resistances of hybrid MPPT systems. The hybrid system uses more
the system. Due to the load variations duty ratio is than one renewable energy sources, combination of
expanded with the help of the proposed PSO wind energy and solar energy. Furthermore it is
method.The system provides the almost constant useful for supplying uninterrupted power to the
output power with very slight variations. commercial buildings and industries.
References
1. Yi-Hwa Liu ,Shyh-Ching Huang, Jia-wei
Huang,Wen Cheng- Liang,:AParticle Swarm
Optimization-Based Maximum Power Point
Tracking Algorithm for PV Systems
Operating Under Partially Shaded
Conditions. In:IEEE TRANSACTIONS N
ENERGY CONVERSION, VOL. 27, NO. 4,
(2012), pp. 1027-1035.
Fig.16.Duty cycle Vs P_out of CPSO
Simulation results illustrate the proposed 2. Moacyr Aureliano Gomes de Brito, Luigi
method outperforms the existing methods in terms of Galotto,: Evaluation of the Main MPPT
convergence speed and output settling time. Merits of Techniques for Photovoltaic Applications, In:
CPSO are (i) It converges quickly than other PSOs IEEE Transactions on Industrial Electronics,
(ii) It requires only the number of particles and no Vol.60(2013),No.8, pp.1156-1167.
other tuning parameters.
3. Anna Rita Di Fazio, Mario Russo et al.,:
6. Conclusion Sensitivity-Based Model of Low Voltage
DistributionSystemswithDistributedEnergyR
In this article, a novel Cauchy PSO based MPPT esources.In:Energies,Vol.9(2016 ),No.10,80
technique is implemented for PV system. The 1:https://doi.org/10.3390/en91001.
proposed method is designed to overcome the
drawback seen in conventional PSO. The enhanced 4. Independent statistics & Analysis U.S.Energy
feature of the proposed Cauchy PSO is its less Information Association.
susceptibility of getting stuck at local optima. This is https://www.eia.gov/energyexplained..
achieved with the help of its long jump ability. It
provides rapid convergence rate with high 5. Kashif Ishaque, Zainal Salam,Muhammad
convergence precision. The positive trait of the Amjad, sad Mekhilef.:An Improved Particle
algorithm is its fast convergence. It is a good thing Swarm Optimization (PSO)–Based MPPT for
because it helps to find faster solutions. Hence it can PV With Reduced Steady-State Oscillation.
be applied for time critical applications. Though the In: IEEE TRANSACTIONS N
efficiency of all PSOs are same, CPSO’s major POWER ELECTRONICS, VOL. 27
advantage is its simplified control structure due to the (2012), NO. 8, pp. 3627-3638.
reduced number of tuning parameters. Therefore the 6. M.Talaat,Abdullaziz,S.Alsayarri,Mohamand
system is economical and less complex. Compared to A Essa, M.A.Yousef,: Investigation of
other algorithms it possess fast convergence rate. The
results proved that this algorithm performs better and transparent pyramidal covers effect to PV
extracts the maximum power irrespective of load power output using detected wireless sensors
fluctuations. It delivers a wide range of duty cycle incident radiation.In:Measurement,
output for a low and high range of loads. It provides Elsevier.Vol.136(2019),pp.775-785.
global maximum power even for uneven insolation.
In future, the proposed system can be used in the 7. M.Talaat, M.A.Farahat,M.H.Elkholy,:

32
Journal of Electrical Engineering ISSN 1582-4594 - Vol 21 No 3 (2021)

Renewable power integration: Experimental JOURNAL OF

and simulation study to investigate the ability PHOTOVOLTAICS,Vol.4(2014),No.2.pp.62
of integrating wave, solar and wind 6-633.
energies.In:Energy,Elsevier,Vol.170(2019
),pp.668-682. 14. QiangFu, Nan Tong,: A new Fuzzy Control
Method Based on PSO for Maximum Power
8. Duy Huynh,C, Tuong Nguyen,M, Matthew Point Tracking of photovoltaic System.In:
Dunnigan,W. Markus Mueller,A,: Global 2011 International Conference on Computer
MPPT of Solar PV Modules using a Dynamic Science and Network
PSO Algorithm under Partial Shading Technology,pp.1488-1491.
Conditions.In:IEEE Conference on Clean
Energy and Technology (CEAT), 2013, 15. Trishan Esram,
134-139. PatricL.Chapman..:Comparison of
PhotovoltaicArrayMaximumPower
9. MacIsaac, L, Knox, A,:Improved Maximum PointTracking Technique.In: IEEE
Power Point Tracking Algorithm for TRANSACTIONS ON ENERGY
Photovoltaic Systems,In: International CONVERSION, Vol.22(2007),No.2, pp.
Conference on Renewable Energies and 439-449.
Power Quality (ICREPQ’10), 2010.
16. Venkatesh Kumar, C, Ramesh Babu, M,:
10. Gomathy, S, Application of Cauchy Mutated Memetic
Saravanan,S,Thangavel,S,:Design and Particle Swarm Optimization Algorithm to
Implementation of Maximum Power Point Economic Dispatch Problem with Practical
tracking (MPPT) Algorithm for a Standalone Constraints,In:WSEAS,Transactions on
PV System, In:International Journal of Systems and Control, Vol.13(2018),pp.80-87.
Scientific & Engineering,,Vol.3(2012 ),No.3.
17. Kashif Ishaque, Zainal Salam.: A
Deterministic Particle Swarm Optimization
11. Saravanan Selvan.D,:Modeling and Maximum Power Point Tracker for
Simulation of Incremental Conductance Photovoltaic System n Under
MPPT Algorithm for Photovoltaic PartialShadingCondition.In:,IEEEANSACTI
Application. In: International Journal of ONS ON INDUSTRIAL ELECTRONICS,
Scientific Engineering and Technology, Vol. Vol. 60(2013), No. 8,pp.3195-3206.
No.2, (2013), No.7,PP.681-685.

12. Ramadan B.A.Koad, Ahmed. 18. Xueming Yang, Jinsha Yuan, Jiangye Yuan,
F.Zobaa.:Comparison between the Huina Mao.: A modified particle swarm
Conventional Methods and PSO Based optimizer with dynamic adaptation. In:
MPPT Algorithm for Photovoltaic Applied Mathematics and Computation, Vol.
System., In: International Journal of Electrical, 189,( 2007) PP. 1205–1213.
Computer, Energetic, Electronic and
Communication Engineering, Vol:8,(2014), 19. Po-Chen Cheng,Bo-Rei Peng ,Yi-Hua, Liu,
No:4, pp.691-696. Yu- Shan Cheng, Jia- Wei
Huang.:Optimization of a
13. K.L.Lian,J.H.Jhang,.I.S.Tian.:A Maximum Fuzzy-Logic-Control-Based MPPT Algorithm
Power Point Tracking Method Based on Using the Particle Swarm Optimization
Perturb-and-Observe Combined With Technique. In:Energies, Vol.8,( 2015),
Particle swarm optimization.In:IEEE pp.5338-5360.

33
Journal of Electrical Engineering ISSN 1582-4594 - Vol 21 No 3 (2021)

27. R.Pon Vengatesh, S.Edward

20. S. Komsiyah et.al.:Computational methods of Rajan.:evolutionary optimization approach
Gaussian Particle Swarm Optimization For finding GMPP Of a PV array system
(GPSO)and Lagrange Multiplier on under heterogeneous operating Condition. In:
economic dispatch issues (case study on the Journal of Electrical Engineering.
electrical system of Java-Bali IV area). ,Vol.17(2017 ),No.1,pp.1-8.
In:EPJ Web of Conferences, Vol. 68,( 2014),
No.00014. 28. Heavy Tailed Distribution & Light Tailed
Distribution: Definition & Examples.
21. M.Abdulkadir, A.H.M.Yatim, S.T.Yusuf.: An https://www.statisticshowto.datasciencecent
Improved PSO-Based MPPT Control .com/heavy-tailed-distribution.
Strategy for Photovoltaic Systems.In:
International Journal of Photoenergy, 29. QiWu,: Cauchy mutation for decision-making
Hindawi Publishing Corporation, Vol. variable of Gaussian particle swarm
2014.pp.1-12. optimization applied to parameters selection
of SVM. Expert Systems with Applications, ,
22. Vytautas Jancauskas et.al.:Empirical Study Elsevier,Vol.38 (2011),pp.4929–4934.
of Particle Swarm Optimization Mutation
Operators. In:Baltic J. Modern Computing, 30. Wang,H,Yong Liu,:A Hybrid Particle Swarm
Vol. 2,(2014) No.4, pp. 199–214. Algorithm with Cauchy Mutation, In: Swarm
intelligence Symposium, 2007,
ieeexplore.ieee.org.
23. Hao Gao, Wenbo Xu.: .A New Particle
Swarm Algorithm and Its Globally 31. Student t-Distribution ,
Convergent Modification. In: IEEE https://en.wikipedia.org/wiki/Student%27s_t-
TRANSACTIONS ON distribution.
SYSTEMS,MAN,ANDCYBERNETICS—P
ARTB: CYBERNETICS, VOL.41 (2011), 32. Seyed Mohsen Mirhassani, Sayedh Zahra
NO.5, pp.1334-1351. Mirhaberi Golroodbari..:An improved particle
swarm optimization based maximum power
24. Vanxay Phimmasone,: Evaluation of point tracking strategy with variable
Extracted Energy from PV with PSO-based sampling time. In: Electrical Power and
MPPT Against Various Types of Solar Energy Systems, ELSEVIER, Vol.64,(2015),
Irradiation Change,In:.International pp. 761-770.
Conference on Electrical, 2010. 33. Nattawat Jumpasri, Kittapas Pinsuntia,
Kaweepoj Woranetsuttikul.:Improved
25. Qing Zhang,Changhe Li,Yong Liu,: Fast Particle Swarm Optimization Algorithm
Multi-swarm Optimization with Cauchy using Average Madel on MPPT for Partial
Mutation and Crossover Operation. Shading in PV Array. In:Proceedings of the
Conference,In:2007,DOI International Electrical Engineering
10.1007/978-3-540-74581-5_38. Congress,(2014).
34. Abhinav Dogra et. al.:Design of Buck-Boost
Converter for Constant Voltage Applications
26. Hui Wang et. al.: A Hybrid Particle Swarm
and its Transient Response due to Parametric
Algorithm with Cauchy Mutation. In: Swarm
Variation of PI Controller.In: ternational
IntelligenceSymposium,(2007)
Journal of Innovative Research in Science ,
eeexplore.ieee.org.
Engineering and Technology.
Vol.3,(2014),No. 6.

34