World Academy of Science, Engineering and Technology

International Journal of Electrical and Computer Engineering

Vol:8, No:1, 2014

Selection of an Optimum Configuration of Solar PV

Array under Partial Shaded Condition Using Particle
Swarm Optimization
R. Ramaprabha

I. INTRODUCTION
Abstract—This paper presents an extraction of maximum energy
from Solar Photovoltaic Array (SPVA) under partial shaded
conditions by optimum selection of array size using Particle Swarm
S OLAR Photovoltaic array is formed by connecting number
of solar photovoltaic (SPV) modules in different
configurations to get a desired voltage and current levels. The
Optimization (PSO) technique. In this paper a detailed study on the
major problem in a larger SPVA/ building integrated PV
International Science Index, Electrical and Computer Engineering Vol:8, No:1, 2014 waset.org/Publication/9997238

output reduction of different SPVA configurations under partial

shaded conditions have been carried out. A generalized MATLAB (BIPV) arrays/ mobile SPV arrays is the reception of non-
M-code based software model has been used for any required array uniform insolation/ partial shade. In these cases, the
size, configuration, shading patterns and number of bypass diodes. occurrence of partial shading is frequent due to tree leaves
Comparative study has been carried out on different configurations falling over it, birds or bird litters on the array, shade of a
by testing several shading scenarios. While the number of shading
patterns and the rate of change are very low for stationary SPVA but neighboring construction etc. As each SPV module consists of
these may be quite large for SPVA mounted on a mobile platforms. number of series connected cells, all the cells are forced to
This paper presents the suitability of PSO technique to select carry the same current, even though a few cells under shade
optimum configuration for mobile arrays by calculating the global produce less photon current. The shaded cells may get reverse
peak (GP) of different configurations and to transfer maximum power biased, acting as loads, draining power from fully illuminated
to the load. cells. If the system is not properly protected, hot-spot problem
[1] can arise and in several cases, the system can be
Keywords—Global peak, Mobile PV arrays, Partial shading,
permanently damaged. In conventional SPV systems, this
optimization, PSO.
problem reduces the overall power generation to a larger level.
NOMENCLATURE Hence the SPVA installation cost is increased, because the
number of SPV modules must be increased, and as a result,
IPV - Solar module output current (A)
VPV - Solar module output voltage (V)
SPV power generation will be less attractive. This makes the
Iph - Photo current of the SPV module (A) study of partial shading of SPV modules a key issue. The
Diode reverse saturation current in the equivalent Voltage-Power characteristics of SPV module vary with solar
Ir -
circuit (µA) insolation and temperature. Researchers have analog and
Series resistance in the equivalent circuit of the mathematical models of SPV cells for varying environmental
Rse -
module(mΩ) conditions [2], [3]. The difference between all models is the
Parallel resistance in the equivalent circuit of the
Rsh -
module(Ω)
number of necessary parameters used in the computational.
n - Diode ideality factor There are several equations presented in literature to simulate
q - Electron charge ( =1.602×10-19 C ) the behavior of SPV cells. Typical SPV module consists of 36
k - Boltzman’s constant ( = 1.381×10-23 J/K) solar cells connected in series. SPV modules are tied together
T - Temperature (Kelvin) in different fashion to form an array with required voltage and
Vt - Thermal voltage (= nkT/q) current levels. The output power of a SPVA decreases
G - Irradiance level (at reference condition G=1000 W/m2)
α - Short circuit current temperature co-efficient
considerably, when voltage-current curves of solar cells are
β - Open circuit voltage temperature co-efficient not identical due to soiling, temperature variations, cell
Isc - Short circuit current of the module damaging and partial shading etc. [4]-[6]. In recent years, the
Voc - open circuit voltage of the module respectively impact of partial shadowing on the energy yield of SPV
Vm - Maximum power point voltage systems has been widely discussed [6]-[9]. Before trying to
Im - Maximum power point current eliminate or reduce mismatch effects, a detailed understanding
Pm - Maximum power of their performance is required. Hence it is convenient to
Additions subscripts indicate the parameters at
ref - carry out the simulation study with the help of a computer
reference conditions
model which properly allows the inclusion of mismatch
effects with high accuracy. In most of the studies [10]-[14],
the effect of partial shading in reducing the output power of
the SPVA has been discussed. But little attention has been
R. Ramaprabha is Associate Professor at the Department of EEE, SSN
College of Engineering, Rajiv Gandhi Road (OMR), Kalavakkam- 603 110, paid to the power dissipated by the shaded cells affecting the
Chennai, Tamilnadu, India (e-mail: ramaprabhasuresh@gmail.com).

International Scholarly and Scientific Research & Innovation 8(1) 2014 89 scholar.waset.org/1307-6892/9997238
 World Academy of Science, Engineering and Technology
International Journal of Electrical and Computer Engineering
Vol:8, No:1, 2014

array life and utilization of the array for the worst shaded case. with irradiation and temperature, respectively, and the other
The harmful effects in basic configurations and their parameters are determined by taking a reference operating
comparison have been discussed by [14]. Common use of by- condition [1], [8], [31], [34]. However, all of the circuit
pass diodes in antiparallel with the series-connected PV parameters depend on both irradiation and cell temperature
modules can partially improve the power reduction due to and the relationship between them is nonlinear and cannot be
partial shadow. In such cases a more complicated Maximum easily expressed by an analytical equation [3], [35]-[39]. In
Power Point Tracking (MPPT) algorithms capable to disregard addition, some differences can be seen in the equations that
local power maxima is required. Alternatively, the maximum describe the relationship between the parameters and
available DC power can be improved if the connection of the operational conditions [3], [37]. Consequently, every
SPV modules can be reconfigured such that panels with assumption forces the model to fall into error. For this reason
similar operating conditions are connected in the same series an assumption should be done carefully, especially in
string. Furthermore the parallel configuration should be simulation studies of SPV arrays under mismatch conditions
dominant under partial shaded conditions [14], [15]. However, and low irradiated PV modules [40]. Sharma et al. [39]
high output current at low voltage in parallel configuration showed that consideration of identical series and parallel
will have to be properly conditioned to the required level by resistances for illuminated and dark region of a PV module is
using suitable DC-DC converter. For the configuration types, not a valid assumption and enhancement of the resistances
International Science Index, Electrical and Computer Engineering Vol:8, No:1, 2014 waset.org/Publication/9997238

the generalized MATLAB programs have been developed must be considered in the analysis of partially shaded PV
which are capable of simulating any number of modules array. Dyk and Meyer [5] also showed that the effects of
connected in series or parallel and any type of shading parallel and series resistances on the PV module performance
patterns. The comparison study is made among the are significant. In this study, the dependence of all circuit
configurations with bypass diodes. For particular insolation parameters on module temperature and irradiance is included
and temperature different configurations will be dominant. It by improved model equations [17], [25], [41]. The effect of
is customary to select a proper size of SPV array [16], [17]. varying shunt resistance has been included by fitting a curve
Otherwise, a large change in SPV power because of insolation obtained from experimental results [26]. The model equations
variation caused by shading may lead to instability. Tracking presented in [17] by authors are used for modeling the SPV
the maximum power point is required in order to extract the system. As the importance of bypass diodes is well known, a
largest amount of power from a SPVA, regardless of weather bypass diode has been included as a part of every module in
or load conditions. Various MPPT methods have been the M-file code. This section considers that each module is
proposed and used to extract maximum power from SPVA connected with a bypass diode. To include the effect of bypass
under varying atmospheric conditions and partial shaded diode, negative voltages caused by shading is taken as diode
conditions [18]-[26]. forward drop (~0.7V) in m-file coding. The equations are for
This paper presents the reconfiguration of SPVA and single SPV module.
implementation of MPP tracking of a partially shaded SPVA
using PSO. The PSO [27]-[30] algorithm is one of the modern
evolutionary algorithms. This algorithm was first proposed by
[27]. PSO is a population-based search algorithm
characterized as conceptually simple, easy to implement and
computationally efficient. The authors aim to realize a power
tracking scheme that can find the GP to maximize the
generated power from the SPV source. It should be applicable
to large scale SPV system, resulting from different
combination of solar modules. The proposed PSO scheme is Fig. 1 Equivalent circuit model of a SPVA with bypass diode
also used to select the optimum configuration under partial
shaded conditions. Both functions that are reconfigurations as After improving the accuracy of equivalent circuit of SPV
well GP tracking can be possible using PSO by considering its module model for all operating conditions, the performance of
capability to handle parallel processing. different interconnected SPV arrays are investigated by
including bypass diode under different mismatch conditions.
II. CHARACTERISTICS OF SPVA UNDER PARTIAL SHADED The model is developed using MATLAB M-file. The detailed
CONDITIONS explanation of the effect of bypass diodes in the characteristics
The commonly used electrical equivalent circuit for SPV is given in [14], [42]. Fig. 2 shows the electrical characteristics
module is shown in Fig. 1 [17], [31]. The relationship between of SPVA with bypass diodes under different partial shaded
solar cell’s current and voltage has both the implicit and conditions [43].Number of peaks in the V-P characteristics is
nonlinear mathematical equations. Therefore, determination of less than or equal to the number of zones receiving different
the equivalent circuit parameters requires more computational insolation.
effort for each operating condition when electrical
performance is analyzed [32]-[35]. In most studies, only the
photo-current and the diode saturation current are changed

International Scholarly and Scientific Research & Innovation 8(1) 2014 90 scholar.waset.org/1307-6892/9997238
 World Academy of Science, Engineering and Technology
International Journal of Electrical and Computer Engineering
Vol:8, No:1, 2014

behavior when they are set equal; rand1 and rand2 are two
randomly generated numbers with a range of [0, 1] added in
the model to introduce stochastic nature in particle’s
movement; and w is the inertia weight and it keeps a balance
between exploration and exploitation. In our case, it is a
linearly decreasing function of the iteration index:

w − w min 
w (k ) = w max −  max  × iter
 (3)
 itermax 

where itermax is the maximum number of iteration, iter is the

current iteration number, wmax is the initial weight and wmin is
Fig. 2 Characteristics of four series connected SPV modules under the final weight. In conclusion, an initial value of w around 1,
partial shaded conditions
with a gradual decline toward 0 is considered as a proper
choice. The most important factor that governs the PSO
From Fig. 2, it is observed that the V-P characteristics have
International Science Index, Electrical and Computer Engineering Vol:8, No:1, 2014 waset.org/Publication/9997238

performance in its search for optimal solution is to maintain a

multiple peaks due to partial shading. Among the multiple
peaks one is global peak (GP) and others are local peak power balance between exploration and exploitation. Recently, PSO
developments and applications have been widely explored in
points. In this situation the conventional MPPT algorithm
engineering and science mainly due to its distinct favorable
could fail to determine the actual GP or even traps into one of
the local peaks. Therefore, considerable amount of possible characteristics [30].Just like in the case of other evolutionary
algorithms, PSO has many key features that attracted many
SPV power is not utilized.
researchers to employ it in different applications in which
conventional optimization algorithms might fail [43].
III. OVERVIEW OF PARTICLE SWARM OPTIMIZATION
Recently, a number of heuristic optimization techniques IV. APPLICATION OF PSO TO GP TRACKING OF SPVA
such as genetic algorithms (GA), ant colony algorithm (ACO),
Nonlinear optimization problem can be stated in
PSO and recently biogeography-based optimization (BBO),
are developed to solve a variety of complex engineering mathematical terms as follows:
Find X = (x1 ,x2,...,xn ) such that F(X) is minimum or
problems that are difficult to be solved using conventional
optimization methods. PSO is developed by Kennedy and maximum
Eberhart [27]. It was found to be reliable in solving non-linear Subject to g j (X) ≥ 0, j = 1, 2… m and xli ≤ xi ≤ xui, i = 1,
2... n,
problems with multiple optima. In PSO, a number of particles
form a ‘‘swarm” that evolve or fly throughout the feasible where F is the objective function to be minimized or
hyperspace to search for fruitful regions in which optimal maximized, xi’s are variables, gj is constraint function, xli and
xui are the lower and upper bounds on the variables.
solution may exist. Each particle has two vectors associated
with it, the position (Xi) and velocity (Vi) vectors. In N- In this work the objective function considered is
F(X) =Maximization of SPVA power, PPV
dimensional search space, Xi = [xi1, xi2, . . .,xiN] and Vi = [vi1,
The variable x1= SPVA current, IPV
vi2, . . ., viN] are the two vectors associated with each particle i.
During their search, members of the swarm interact with each The constraint is IPVmax≥ IPV≥IPVmin.
Here, IPVmax=Isc, short circuit current of SPVA and
others in a certain way to optimize their search experience.
There are different variants of particle swarm paradigms but IPVmin=0.
In this work, the PSO is used to find SPVA current which
the most commonly used one is the gbest model where the
ensure that the function F(X) has a maximum value. The
whole population is considered as a single neighborhood
throughout the flying experience [27], [28]. In each iteration, procedure to get the optimum power of SPVA under partial
shaded condition is given below:
particle with the best solution shares its position coordinates
1. Read number of modules connected, insolation pattern
(gbest) information with the rest of the swarm. Each particle
updates its coordinates based on its own best search and temperature for each module
2. Initialize PSO parameters such as wmax, wmin, c1, c2 and
experience (pbest) and gbest according to the following
equations: Itermax
3. Generate initial population of N particles (design
( ) ( )
v ik +1 = wvik + c1rand1 pbest ik − x ik + c 2 rand 2 gbest ik − x ik (1)
variables) with random positions and velocities;
4. Compute objective value, current and power
5. Measure the fitness of each particle
x ik +1 = x ik + v ik +1 (2) 6. Update personal best: Compare the fitness value of each
particle with its pbests. If the current value is better than
where c1 and c2 are two positive acceleration constants, they pbest, then set pbest value to the current value;
keep balance between the particle’s individual and social

International Scholarly and Scientific Research & Innovation 8(1) 2014 91 scholar.waset.org/1307-6892/9997238
 World Academy of Science, Engineering and Technology
International Journal of Electrical and Computer Engineering
Vol:8, No:1, 2014

7. Update global best:: Compare the fitness value of each

particle with gbest. If the current value is better than
gbest, set gbest to the current particle’s value;
8. Update velocities:: Calculate velocities Vk+1 using (1)
9. Update positions:: Calculate positions Xk+1 using (2)
10. Return to step 4 till the current iteration reaches the
maximum iteration number;
11. Output the optimal value of SPVA current and
corresponding SPVA power in thee last iteration.
The code for PSO has been written in MATLAB and the
simulation results are presented for five ive set of different
shading patterns for SPVA (Fig. 3). The performance of the
PSO is validated graphically by comparing its output (marked
in green color) with that of the binary search method (marked
in red color). In all the cases, PSO gives the optimum power
(global peak) which is matched with the result of binary
International Science Index, Electrical and Computer Engineering Vol:8, No:1, 2014 waset.org/Publication/9997238

search.

Fig. 4 Schematic diagram of SPVA configurations

From the above discussion, it may be concluded that, while

Fig. 3 Validation of PSO for GP tracking of partial shaded SPVA
it is very important to model, study, and understand the effects
of shading on SPV
PV arrays, a simple tool is not available for the
V. SIMULATION OF DIFFERENT SPVA CONFIGURATIONS AND purpose. Therefore,, it is felt that there is a need for a flexible,
THEIR PERFORMANCE COMPARISON interactive, and comprehensive simulation model capable to
predict the PV characteristics (including multiple peaks) and
Practically
ractically partial shade has great impact on larger arrays.
output power under partially shaded conditions. The improved
For understanding the practical cases, it is required to go for
model [17] developed by authors have been taken for the
larger arrays with different configurations. Several S SPVA
analysis. Modeling of a large array with shading patterns is
configurations have been proposed in the literature as shown
very complex. In this work, a generalized code has been
in Fig. 4. [1], [8], [15]-[17].. They are series, parallel, series
series-
developed for all the configurations with and without
parallel (SP) total cross tied (TCT) and bridge linked (BL)
considering the effect of changing Rsh. This program gives the
configurations. Series and parallel configurations are the basic
output power, voltage and current values for any irradiance
configurations and the performance of these configurations
and temperature patterns. The architecture of the developed
has been discussed in detail by [14]. The major drawbacks of
software [17] is shown in Fig. 5.
using the series or parallel configuration are that the current
and voltage is less respectively. The other derived
configurations are shown from Figs.. 4 c to f [17]. The
importance of selecting the proper size of the PV array and
batteries in partially shade SPV systems has been discussed by
[44]. It is required for the stable operation of SPV system with
a sudden and large change in SPV power because of insolation
variation, caused by shading etc. Shading caused due to
passing clouds also has a financial claim on the utility. Jewell
and Unruh [45] have carried out an economic analysis to
estimate the cost of the fluctuations in power gene
generation from
a PV source. Based on the literature it is understood that not
only the size of the SPVA but also its configuration that
significantly affects its power output, and therefore, the Fig. 5 Architecture of the developed software
performance of the system under partially shaded conditions.

International Scholarly and Scientific Research & Innovation 8(1) 2014 92 scholar.waset.org/1307-6892/9997238
 World Academy of Science, Engineering and Technology
International Journal of Electrical and Computer Engineering
Vol:8, No:1, 2014

VI. RMSDOF DIFFERENT SPVA CONFIGURATIONS UNDER TABLE I

COMPARISON OF RMSD AND MEAN VALUE FOR DIFFERENT
DIFFERENT SHADING SCENARIOS
CONFIGURATIONS WITH DIFFERENT SIZES
In this section a comparison is made amongst the SPVA Maximum Power (W) Voltage at MPP (V)
Array
configurations of different array sizes in terms of Root mean Configuration Mean RMSD Mean RMSD
Size
square deviation (RMSD) and mean value of both voltage and Value Value Value Value
power. The RMSD has been calculated using (1): 2X4 SP 92.76 172.40 20.29 10.75
2X4 TCT 94.67 170.96 22.92 9.01
n 2 2X4 BL 105.58 161.79 24.23 7.94
∑ (Pmax( unshaded ) − Pmax( shaded ) ) 2X4 MB 121.98 146.44 27.28 4.80
i =1 4X2 SP 103.96 162.45 44.35 17.79
RMSD = (4)
n 4X2 TCT 117.7 149.56 41.59 20.96
4X2 BL 104.00 162.41 44.33 19.40
Here the case where one bypass diode across group of 36 4X2 MB 114.02 153.46 42.44 19.39
cells (one module) has been considered. The array sizes are: 2X6 SP 145.51 252.85 20.38 10.69
2x6, 6x2, 2x4, 4x2, 3x4, 4x3, 4X6, 6x4, 3x3and 4x4. Fifteen 2X6 TCT 149.78 249 22.25 9.57
different random shading profiles are generated for each of the 2X6 BL 175.42 223.74 27.39 4.89
International Science Index, Electrical and Computer Engineering Vol:8, No:1, 2014 waset.org/Publication/9997238

ten different array sizes. The power, voltage and current value 2X6 MB 187.23 213.27 27.39 4.87
of each configuration is obtained for each of these shading 6X2 SP 143.13 254.32 56.59 31.86
patterns. The mean value of power and voltage is calculated 6X2 TCT 160.09 237.43 57.01 32.5
for each configuration for each size. The mean value denotes 6X2 BL 143.11 254.35 56.58 31.86
the mean power and voltage which can be obtained from any 6X2 MB 165.34 233.52 65.97 25.05
configuration. Similarly RMSD is calculated for each 3X4 SP 128.14 269.75 27.69 30.11
configuration by primarily subtracting the uniform irradiance 3X4 TCT 142.58 255.72 29.88 28.47
value from the partial shaded value. These values are squared 3X4 BL 146.70 251.35 33.43 24.78
3X4 MB 144.93 253.70 35.57 22.87
and then the square root of its average is obtained. This gives
4X3 SP 125.79 272.99 34.25 11.51
the RMSD value for a configuration of a particular size. Table
4X3 TCT 132.37 263.53 34.11 23.36
I gives the mean value and RMSD value for all the
4X3 BL 132.89 263.01 36.41 21.06
configurations with different sizes. From Table I it can be
4X3 MB 137.95 257.95 40.40 17.07
inferred that depending on the size of array and type of
3X3 SP 85.72 212.85 27.89 16.05
shading scenario different configurations are dominant. But in
3X3 TCT 92.79 206.06 26.70 17.76
most of the cases TCT seems to be dominant closely followed 3X3 BL 84.5 214.41 29.21 15.33
by MB. It is also note that wherever the modules with same 3X3 MB 83.22 215.21 28.83 15.46
shade are grouped in a string, MB is dominant in which ties 4X4 SP 145.31 384.64 33.30 26.61
are less compared to TCT. 4X4 TCT 164.62 366.41 34.05 25.71
25 different random shading profiles are generated for each 4X4 BL 145.03 384.56 36.51 24.46
array size. The power, voltage and current value of each 4X4 MB 144.14 385.53 40.93 19.65
configuration is obtained for each of these shading patterns. 4X6 SP 186.66 606.83 22.25 35.43
The mean value of power and voltage is calculated for each 4X6 TCT 211.54 583.56 26.51 30.46
configuration for each size. The mean value denotes the mean 4X6 BL 186.06 606.71 28.69 29.52
power and voltage which can be obtained from any 4X6 MB 227.06 572.72 39.42 23.4
configuration. Similarly RMSD is calculated for each 6X4 SP 197.86 596.88 46.31 42.47
configuration by primarily subtracting the uniform insolation 6X4 TCT 234.57 561.86 45.18 42.41
value from the partial shaded value. These values are squared 6X4 BL 184.48 607.33 48.79 40.98
and then the square root of its average is obtained. This gives 6X4 MB 219.1 579.74 54.58 37.99
the RMSD value for a configuration of a particular size. Table
I gives the mean value and RMSD value for all the VII. PRACTICAL VERIFICATION
configurations with different sizes. For different array sizes To verify the outputs obtained from software simulation,
the comparison of mean and RMSD values of all the practical verification are performed. 3X3 SPVA (SOLKAR
configurations are given from Table I. solar module) setup is used. The specifications are given in
From Table I it can be inferred that depending on the size of appendix. The terminals from this setup are drawn to the panel
array different configurations are dominant. But in most of the board. This is shown in Fig. 6 along with the electronic load to
cases TCT seems to be dominant closely followed by MB. trace the characteristics. The electronic load [14] is used to
verify the characteristics. A sample snap shot of the CRO
screen shows V-I characteristics of the different SPVA
configurations for a particular shading pattern (Fig. 10).
GWINSTEK GDS-1022 Digital Storage Oscilloscope (DSO)
is used to trace the practical characteristics. It is calibrated

International Scholarly and Scientific Research & Innovation 8(1) 2014 93 scholar.waset.org/1307-6892/9997238
 World Academy of Science, Engineering and Technology
International Journal of Electrical and Computer Engineering
Vol:8, No:1, 2014

using Fluke5500A Multi-Product Calibrator. For different

insolation and temperature the practical characteristics are
easily traced out using electronic load method and the relevant
data traced by DSO are stored in Excel spreadsheet to
calculate V-P characteristics and for comparison of model
parameters.

Fig. 6 Practical setup of a 3X3 array with Panel board and Electronic
load
International Science Index, Electrical and Computer Engineering Vol:8, No:1, 2014 waset.org/Publication/9997238

Outputs were verified for uniform as well as partial shaded

conditions. Artificial shadings were created. Multiple peaks
were observed on introduction of bypass diodes. Also value of
power obtained increased considerably by inclusion of bypass
diodes. The practical verification was done for several input
shading patterns. The outputs obtained were closer to the
outputs obtained from simulation which took into Fig. 7 Schematic of Electrical Reconfiguration of SPVA
consideration the effect of varying Rsh.
Fig. 8 plots the convergence of total power computed by
VIII. OPTIMUM SELECTION OF SPVA UNDER PARTIAL PSO over the number of iterations for different shading
SHADED CONDITIONS USING PSO patterns [48]. Initially, the particles are randomly initialized.
This section deals with a dynamical electrical array Therefore, the initial power is always high. This initial power
reconfiguration of SPVA under partial shaded conditions to corresponds to the 0th iteration. As the algorithm progresses,
improve its energy production. The reconfiguration strategy is the convergence is drastic and it finds a global maxima very
carried out by inserting a controllable switch between the SPV quickly. The number of iterations needed for the convergence
modules in an array which allows electrical reconnection of is seen to be 5-10, for this application environment.
SPV modules [46] depending on the input insolation and
temperature conditions. As a result, the SPV system exhibits a
self-capacity for real time adaptation to the SPV generator
external operating conditions in order to improve the energy
extraction of the system. SP configuration is the most
commonly existing/used SPVA configuration due to its
simplicity and low cost per kWpeak. As discussed in the
previous section, series connection of SPV modules in SP
configuration can adversely affect the maximum available
power when they are operating at different conditions/partially
shaded. Fig. 7 shows a set-up for improving the extraction of
energy from SPVA under partial shaded conditions using PSO
based array reconfigurable system. Fig. 8 The trend of convergence of PSO with the number of iterations
The control signals for the switches to make proper array for different shading patterns
reconfiguration is based on the output from PSO algorithm.
The insolation and temperature data has been fed to the PC After finding the global peak produced by optimum
using corresponding sensors and data acquisition system. configuration, corresponding pulses will be given to the
(Dynalab weather tech data logger with sensors) PSO analog switch. Fig. 9 shows the pulse pattern produced for
algorithm is used to find out the maximum power in each different configuration based on the result computed by PSO.
configuration as well as global best from the output of
different array configurations. All the computations have been
done using MATLAB software. The corresponding signals
have been given to the switch to reconfigure the array. For
practical case, LTC201A (quad switch) switcher IC [47] can
be used.

International Scholarly and Scientific Research & Innovation 8(1) 2014 94 scholar.waset.org/1307-6892/9997238
 World Academy of Science, Engineering and Technology
International Journal of Electrical and Computer Engineering
Vol:8, No:1, 2014

developing a more realistic model using MATLAB M-file has

been presented. From the comparison of different
configurations under different random shading patterns, it is
found that different configurations are dominant under
different shading patterns. To get more energy output from
SPVA reconfiguration of SPVA has been implemented using
PSO. This reconfiguration method offers a greater output
power than the PV systems with static configuration.

APPENDIX
The Parameters of the SOLKAR solar Module used for
practical verification at STC (1000 W/m2 and T=250C) is
given below:

Rated Power( P ) : 37.08W

International Science Index, Electrical and Computer Engineering Vol:8, No:1, 2014 waset.org/Publication/9997238

Voltage at Maximum power( Vm ) : 16.56 V

Current at Maximum power ( Im ) : 2.25 A
Open circuit voltage ( Voc) : 21.24 V
Short circuit current ( Isc ) : 2.55A

Fig. 9 Pulse pattern given to switches for reconfiguration No. of Series Cells (Ns) : 36

After reconfiguring the PSO give the gbest value of the ACKNOWLEDGMENT
selected configuration which will be the reference GP power
The author wishes to thank the management of SSN college
value. This reference value is compared with the actual power
of Engineering for providing all the experimental and
of SPVA and error signal is processed with PI controller. The
computational facilities to carry out this work.
output of the controller is then compared with high frequency
triangular carrier to produce PWM pulses to chopper to
REFERENCES
transfer maximum power to the load. The block diagram of the
[1] V. Quaschningand R. Hanitsch, “Numerical simulation of current–
proposed system is shown in Fig. 10 voltage characteristics of photovoltaic systems with shaded solar cells”,
Solar Energy, Vol. 56, No. 6, pp. 513–520, 1996.
[2] M. Wolf and H. Rauschenbach, “Advanced Energy Conversation”,
No.3, p.455, 1963.
[3] J. A. Gow and C. D. Manning, “Development of a photovoltaic array
model for use in power-electronics simulation studies”, IEEE Proc.
Electr. Power Appl., Vol. 146, No.2, pp.193-200, 1999.
[4] H. Kawamura, K. Naka, N. Yonekura, S. Yamanaka, H. Kawamura, H.
Ohna, and K. Naito, “Simulation of I–V characteristics of a PV module
with shaded PV cells”, Solar Energy Materials and Solar Cells 75, 613–
621, 2003.
[5] E. E. V. Dyk, E. L. Meyer, F. J. Vorster and A. W. R. Leitch, “Long-
term monitoring of photovoltaic devices”, Renewable Energy 25, 183–
197, 2002.
[6] D. Weinstock and J. Appelbaum, “Shadow Variation on Photovoltaic
Collectors in a Solar Field”, in the proc. of 23rd IEEE Convention of
Electrical and Electronics Engineers, pp. 354–357, 2004.
[7] A. Woyte, J. Nijsand R. Belmans, “Partial shadowing of photo voltaic
arrays with different system configurations: literature review and field
test results”, Solar Energy74, pp 217-233, 2003.
Fig. 10 Block diagram of proposed PSO based system [8] N. D. Kaushikaand N. K. Gautam, “Energy yield simulations of
interconnected solar PV arrays”, IEEE Transactions on Energy
Conversion, Vol. 18, No.1, pp 127-134, 2003.
IX. CONCLUSION [9] A. W. Y. Ho and S. R. Wenham, “Intelligent strategies for minimizing
In this paper, PSO technique is used to track the global mismatch losses in photovoltaic modules and systems”, in the proc. of
17th European Photovoltaic Solar Energy Conference and Exhibition,
maximum power point of SPVA that are exposed to unequal Munich-Germany, 2001.
solar irradiation which occurs in the real solar power systems. [10] Hans. S. Rauschenbach, “Electrical output of Shadowed Solar Arrays”,
The algorithm is tested for different set of input insolation and IEEE Transactions on Electron Devices, Vol. 18, No. 8, pp 483-490,
1971.
temperature patterns. The obtained results prove that the
[11] M. C. Alonso-García, J. M. Ruiz and W. Herrmann, “Computer
proposed system is able to track efficiently the GP with high simulation of shading effects in photovoltaic arrays”, Renewable
accuracy and reliability. Moreover, analysis of various SPVA Energy, Vol. 31, No. 12,pp 1986-1993, 2006.
configurations with respect to environmental parameters by

International Scholarly and Scientific Research & Innovation 8(1) 2014 95 scholar.waset.org/1307-6892/9997238
 World Academy of Science, Engineering and Technology
International Journal of Electrical and Computer Engineering
Vol:8, No:1, 2014

[12] M. C. Alonso-García, J. M. Ruiz and F. Chenlo, “Experimental study of [34] T. Ikegami, T.Maezono, F. Nakanishi, Y. Yamagata K. Ebihara,
mismatch and shading effects in the I-V characteristic of a photovoltaic “Estimation of equivalent circuit parameters of PV module and its
module”, Solar Energy Materials & Solar Cells 90,pp 329– 340, 2006. application to optimal operation of PV system”, Solar Energy Materials
[13] Engin Karatepe, Mutlu Boztepe and Metin Colak., “Development of and Solar Cells 67, pp 389–395, 2001.
suitable model for characterizing photovoltaic arrays with shaded solar [35] J. Merten, J. M. Asensi, C. Voz, A. V. Shah, R. Platz and J.Andreu, “
cells”, Solar Energy, pp 329-340, 2007. Improved equivalent circuit and analytical model for amorphous silicon
[14] R. Ramaprabha and B. L. Mathur, “Effect of Shading on Series and solar cells and modules”, IEEE Transactions on Electron Devices,
Parallel Connected Solar PV Modules”, Journal of Modern Applied Vol.45, No. 2, pp 423–429, 1998.
Science, pp 32-41, 2009. [36] M. A. Blas, J. L. Torres, E. Prieto and A. Garcia, “Selecting a suitable
[15] C. Gonzalez and R. Weaver, “Circuit design considerations for model for characterizing photovoltaic devices”, Renewable Energy 25,
photovoltaic modules and systems,” in Proc. 14thIEEE Photovolt. Spec. pp 371–380, 2002.
conf., pp. 528–535, 1980. [37] L. Lu and H.X.Yang, “A study on simulations of the power output and
[16] N. F. Shephard and R. S. Sugimura, “The integration of bypass diode practical models for building systems”, Journal of Solar Energy
with terrestrial photovoltaic modules and arrays,” in Proc. Engineering, 126, pp 929–935, 2004.
17thIEEEPhotovolt. Spec. Conf., pp. 676, 1984. [38] M. K. El-Adawi and I. A. Al-Nuaim, “A method to determine the solar
[17] R. Ramaprabhaand B. L. Mathur, “A Comprehensive Review and cells series resistance from a single I–V. Characteristic curve
Analysis of Solar Photovoltaic Array Configurations under Partial considering its shunt resistance – new approach”, Vacuum 64, pp 33–36,
Shaded Conditions”, International Journal of Photo energy, Special 2002.
Issue on Recent Developments in Solar Energy Harvesting and Photo [39] A. K. Sharma, R. Dwivedi and S. K. Srivastava, “Performance analysis
catalysis, Vol. 2012, pp 1-16, Feb 2012. of a solar array under shadow condition”, IEE Proceedings-G 138 (3), pp
[18] N. Kasa, T. Lida and L. Chen, “Flyback inverter controlled by sensorless 301–306, 1991.
International Science Index, Electrical and Computer Engineering Vol:8, No:1, 2014 waset.org/Publication/9997238

current MPPT for photovoltaic power system” , IEEE Transactions on [40] Engin Karatepe, Mutlu Boztepe and Metin Colak, “Development of
Industrial Electronics, Vol. 52, No.4 , pp 1145–1152, 2005. suitable model for characterizing photovoltaic arrays with shaded solar
[19] N. Femia, G. Petrone, G. Spagnuolo, and M.Vitelli, “Optimization of cells”, Solar Energy, pp 329-340, 2007.
perturb and observe maximum power point tracking method” ,IEEE [41] Marcelo Gradella Villalva, Jonas Rafael Gazoli, and Ernesto Ruppert
Transaction on Power Electronics, Vol. 20, No. 4, pp 963–973, 2005. Filho, “Comprehensive Approach to Modeling and Simulation of
[20] M. Veerachary, “Power tracking for non-linear PV sources with coupled Photovoltaic Arrays”, IEEE Transactions on Power Electronics, vol. 24,
inductor SEPIC converter”, IEEE Transactions on Aerospace and no. 5, pp 1198-1208, 2009.
Electronic Systems 41 (3), pp 1019–1029, 2005. [42] S. Silvestre, A. Boronat and A. Chouder, “Study of bypass diodes
[21] Joung Hu Park, Jun-YounAhn, Bo-HyungCho and Gwon-Jong Yu, configuration on PV modules”, Applied Energy 86, pp 1632–1640, 2009.
“Dual module based maximum power point tracking control of [43] R.Ramaprabha, N. Balamurugan and B.L.Mathur, “ Implementation of
photovoltaic systems”, IEEE Transactions on Industrial Electronics, Vol. Particle Swarm Optimization based Maximum Power Point Tracking of
53, No. 4, pp 1036–1047, 2006 Solar Photovoltaic Array under Non uniform Insolation Conditions”,
[22] E. V. Solodovnik, S. Liu and R. A. Dougal, “Power controller design for International Review of Modelling and Simulation, Vol.6, No.3, pp
maximum power point tracking in solar insolations”, IEEE Transactions 1503-1510, 2011.
on Power Electronics, Vol. 19, No. 5, pp 1295–1304, 2004. [44] Hiren Patel and VivekAgarwal, “Maximum Power Point Tracking
[23] K. Kobayashi, I. Takano, and Y. Sawada, “A study on a two stage Scheme for PV Systems operating under Partially Shaded Conditions,”
maximum power point tracking control of a photovoltaic system under IEEE Transactions on Industrial Electronics, Vol. 55, No. 4,pp 1689-
partially shaded insolation conditions”, Proceedings of the IEEE Power 1698, 2008.
Engineering Society General Meeting, vol. 4,pp. 2612–2617, 2003. [45] W. T. Jewell and T.D. Unruh, “Limits on cloud-induced fluctuation in
[24] A. Kajihara and T. Harakawa, “On considerations of equivalent model photovoltaic generation,” IEEE Transactions on Energy Conversion,
about PV cell under partial shading”, Proceedings of Japan Industry vol. 5, no. 1, pp. 8–14, 1990.
Applications Society Conference IEE of Japan, Vol. 1, No. 71, pp 289– [46] G. Velasco, J. J. Negroni, F. Guinjoan and R. Pique, “Irradiance
292, 2005. Equalization Method for Output Power Optimization in Plant Oriented
[25] Hiren Patel and VivekAgarwal, “MATLAB-Based Modeling to Study Grid-Connected PV Generators”, 11th European Conference on Power
the Effects of Partial Shading on PV Array Characteristics”, IEEE Electronics and Applications, (EPE 05). Dresden, September 2005.
Transactions on Energy Conversion, Vol. 23, No. 1, pp 302-310, 2008. [47] Linear Technology LTC201A Datasheet. Linear Technology
[26] R. Ramaprabha, M. Balaji and B. L. Mathur, “Maximum Power Point Corporation, 1991.
Tracking Of Partially Shaded Solar PV System Using Modified [48] www.mathworks.com.
Fibonacci Search Method With Fuzzy Controller”, International Journal
of Electrical Power and Energy Systems, Elsevier publications, Vol. 43,
pp. 754-765, 2012.
[27] J. Kennedy and R. Eberhart, “Particle swarm optimization”, Proceedings
of IEEE International Conference on Neural Networks (ICNN’95), Vol.
4, pp.1942–1948, 1995.
[28] Y. B. Wang, XinPeng and Ben-Zheng Wei, “A new particle swarm
optimization based auto tuning of PID controller”, Proceedings of IEEE
International Conference on Machine Learning and Cybernetics, July,
pp.1818–1823, 2008.
[29] R. C. Eberhart and Y. Shi, “Comparing Inertia Weights and Constriction
Factors in Particle Swarm Optimization”, Proceedings of the IEEE
International Congress Evolutionary Computation, San Diego, Vol. 1,
pp. 84-88,2000.
[30] K. T. Chaturvedi, M. Pandit and L.Srivastava, “Particle Swarm
Optimization with Time Varying Acceleration Coefficients for Non-
Convex Economic Power Dispatch”, Electrical Power and Energy
Systems, Vol. 31, No. 6, pp. 249-257, 2009.
[31] J. A. Duffie and W. A. Beckman, “Solar Engineering and Thermal
Processes”, John Wiley & Sons Inc., New York, 1991.
[32] K. F. Teng and P. Wu, “PV module characterization using Q-R
decomposition based on the least square method”, IEEE Transactions on
Industrial Electronics, Vol. 36, No. 1, pp 71–75, 1989.
[33] K. Araki and M. Yamaguchi, “Novel equivalent circuit model and
statistical analysis in parameters identification”, Solar Energy Materials
and Solar Cells 75, pp 457–466, 2003.

International Scholarly and Scientific Research & Innovation 8(1) 2014 96 scholar.waset.org/1307-6892/9997238