See discussions, stats, and author profiles for this publication at: [Link]

net/publication/371779600

Improvement of PV System Performances using Novel PV Array

Reconfiguration Based on Balancing Voltage Algorithm under Voltage
Mismatch Conditions

Conference Paper · October 2022

CITATIONS READS

0 45

5 authors, including:

Abdelouadoud Loukriz Djamel Saigaa

University of Sciences and Technology Houari Boumediene
12 PUBLICATIONS 21 CITATIONS
33 PUBLICATIONS 44 CITATIONS
SEE PROFILE
SEE PROFILE

Ahmed Bendib Drif Mahmoud

University of Blida Université de M'Sila
38 PUBLICATIONS 165 CITATIONS 22 PUBLICATIONS 57 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Abdelouadoud Loukriz on 22 June 2023.

The user has requested enhancement of the downloaded file.

 1

Research Paper

Improvement of PV System Performances using Novel PV Array

Reconfiguration Based on Balancing Voltage Algorithm under Voltage
Mismatch Conditions
[Link]*a ,[Link], [Link] , Ahmed Bendibb , and [Link] b
a
Laboratoire de Modélisation des Systèmes Energétiques (LMSE), Mohamed KHIDER University, Biskra, Algeria.
b
Department of Electrical Engineering, Mohamed BOUDIAF University, M’sila, Algeria

ARTICLE INFO ABST RACT

Article history : In this paper, a new PV system reconfiguration based on balancing voltage mismatch is
proposed and investigated. It is well-known that the mismatch issue reflects a reduction
Received 00 January 20
of the voltage or/and current of the concerned PV modules. Where most techniques in
Received in revised form 00 February 20 the literature have treated these phenomena only for PV arrays interconnected in TCT.
This paper proposes a new advanced technique based on voltage balancing to improve
Accepted 00 February 20
the PV array connected in Serie-Parallel (SP) under voltage mismatch effects. The
design process of the proposed algorithm is investigated and verified under different
voltage mismatch scenarios. A comparative study with SP and SuDoKu methods is
Keywords: carried out where a hugely important impact of PV array output is demonstrated using
the proposed method. Further, a comprehensivee analysis is performed where many
Dynamic reconfiguration; Switching
indexes are used such as the mismatch power loss, fill factor and percentage of power
matrix; Novel voltage balancing method;
loss.
PV module; Comparative study (SP,
SuDoKu, and VB method)

1. Introduction
The generation of electricity by photovoltaic (PV) systems has gained popularity among the general population due to
awareness and technological advances that have significantly reduced the cost per watt [1]. Despite technological progress,
there are few bottlenecks [2]: low conversion efficiency, high dependence on irradiation, temperature variation, cell ageing,
and other environmental conditions that degrade the performance of the PV system. The basic configuration scheme
includes series and parallel interconnection schemes. The series configuration provides a higher voltage at a relatively
lower current, while the parallel configuration provides a higher current at a relatively lower voltage level. The mismatch
phenomenon and associated losses can be significantly reduced by changing the interconnection scheme or the location of
the PV modules in order to suit the adverse conditions encountered [3]. When connecting PV modules in parallel under
voltage mismatch, the affected PV modules force the healthy ones to operate in the negative voltage region, which results
in a clear voltage loss in the system and causes hot spots [4]. Much work has been done on optimizing the energy output of
PV systems, including electrical reconfiguration methods and physical reallocation of PV modules. The dynamic methods
require a larger number of switches, sensors, and sophisticated control algorithms [5][6]. In addition, the reconfiguration
scheme can be made adaptive by including an adaptive panel bank. The appropriate number of modules from the adaptive
bank is added to the fixed panel bank depending on the prevailing shading conditions. However, this strategy requires a
larger adaptive bank to accommodate all possible shading types, which increases considerably the cost of the system. In
addition, physical configuration methods have two limitations, namely poor shadow dispersion and wiring complexity
[7][8]. In 2021, Loukriz and al[9] have introduced a new dynamic method to reconfigure the PV array using a switching

* Corresponding author. Tel.: +213670453778.

E-mail address: [Link]@[Link]

Recueil De Mécanique © 2020

2

matrix. However, the developed method involves the simulation of a PV array consists of of 16 modules interconnected
initially in TCT. Indeed, the author did not test the effectiveness against the physical SuDoKu method interconnected in SP.
Here, a new strategy to maximize the power output of PV arrays is discussed based on a dynamic reconfiguration of PV
modules interconnected in SP to combat voltage mismatch problems using the voltage balancing approach. For this
purpose, a new algorithm is utilized to reconfigure the PV array under voltage mismatch conditions. The proposed
approach involves the simulation of a PV array of 16 PV modules interconnected in SP (serie-parallel), as well as an
algorithm that controls the overall performance improvement under different voltage imbalance conditions. The authors
performed a comparative study with the most well-known PV reconfiguration method called SuDoKu in order to validate
the suggested approach. The obtained results by using MATLAB/Simulink simulation demonstrate a huge rate of
improvement varying between 26.71%, and 195.92 % in all reported voltage mismatch cases, which is enormous relative
to static configuration (SP) and physical reconfiguration (SuDoKu) methods operating below the total available power.
The main contributions of this paper are:
- The voltage balancing method is investigated under voltage mismatch for PV array interconnected in SP.
- The proposed approach can offer a better output power under voltage mismatch conditions.
- Comparative study with SP and SuDoKu methods is established.
- A comprehensive statically analysis is performed using the mismatched power loss, the fill factor, and the
percentage of power loss.

2. Modeling of solar PV panel

In this study, the well-known used single-diode model based on the Shockley diode equation is considered. The
photovoltaic cell is represented by the electrical circuit which is shown in figure 1. Given that the current source Iph
represents the cell photo-current, Rs and Rsh are the series and intrinsic shunt resistances of the cell, respectively [10].

Fig.1- Equivalent single-diode model of solar cell

Based on the single-diode model given in figure 1, the output current of a solar cell can be defined as follows [10],[11]:
𝑞(𝑉𝑝𝑣 + 𝐼𝑝𝑣 . 𝑅𝑠 ) 𝑞(𝑉𝑝𝑣 + 𝐼𝑝𝑣 . 𝑅𝑠 )
𝐼𝑝𝑣 = 𝑁𝑃 𝐼𝑝ℎ − 𝑁𝑃 𝐼𝑠 (exp ( ) − 1) − 𝑁𝑃 ( ) (1)
𝐴 .𝐾 .𝑇 .𝑛 .𝑁 𝑆 𝑁 . 𝑅
𝑆 𝑠ℎ
being Is the cell reverse saturation current; q the electron charge (1.60217646 9 10-19 C); K the Boltzmann’s constant
(1.3806503 9 10-23 J/K); A the diode ideality constant; V the output voltage of the cell; T the cell temperature; Np the
number of PV cells connected in parallel; and Ns the number of PV cells connected in series.
Besides, the expression relating the generated photocurrent Iph to the solar irradiation can be written as follows [10]:
𝑆 (2)
𝐼𝑝ℎ = [𝐼𝑠𝑐 + 𝑘𝑖 (𝑇 − 𝑇𝑟 )]
1000
where Isc is the short-circuit current of the cell at reference temperature; S is the solar irradiation in W/m², ki is the
coefficient of the short-circuit current temperature; Tr is the reference temperature of the cell.

3. Proposed voltage-balancing approach for SP interconnection

To avoid the limitations of the SP configuration, a new approach is proposed by using a new strategy to reduce the
mismatching effects, in particular the voltage mismatch. It is worth mentioning that under voltage mismatch conditions, a
PV array provides a better output power when the affected modules are grouped in the same row or a limited number of
rows. Figure 2(a) depicts the operation principle of the proposed balancing algorithm based on the output voltage
measurements of each PV module. In the first step, the voltage values of each PV module of a PV array are extracted and
collected into a matrix (n×m). Then, this voltages’ matrix is converted into a state vector. Next, the voltages of the state
vector are arranged based on their values in increasing order. The produced vector that contains the numerical values of
voltages is then truncated on the number of columns n of the data matrix to obtain m lines of values of nearby parallel
voltages. The resulted combination is involved in a truth table that defines the state of switches to be piloted.
 3

(a) (b)
Fig.2- Proposed approach,(a): main steps of the voltage balancing algorithm and (b): proposed switching matrix
The switching matrix for this particular application is carried out to provide a circuit switch so that with a correct
operation of opening and closing. The position of the solar modules in the SP connection circuit can be changed with less
number of switches. The used switching matrix is designed in figure 2(b) so that each PV module needs one double-pole
switch on each row. The number of switches installed is NPV× NROW, where NPV and NROW denote the number of PV
modules, and the number of rows, respectively.

4. Results and discussion

Simulations in MATLAB/Simulink software are provided to validate our proposed approach and analyze the results
using PV modules interconnected in SP (4×4), this system is investigated and studied under various voltage mismatch
scenarios such as temperature and partial ageing variations. In addition, the PV module type is (1 Soltech 1 STH-215-P),
with MPP 215[W] at STC being used. To study the efficiency of the proposed method, two scenarios are considered and
studied, the first is caused by the very high temperature and the second is the Potential Induced Degradation PID failure
(Low Rsh value). Firstly, the authors begin with the voltage mismatch that occurs with varying the temperature, potential
induced degradation (PID) loss of solar cells, and the other failures like diode failure, cells interconnect failure breakage
..etc. Those causes affect the voltage output of the PV module. The results of the proposed algorithm (VB) are tested with
a 4 × 4 PV array with different voltage mismatches. The investigated cases are considered as follows:
- PV array with 43.75 % degradation rate.
- PV array with 56.25 % degradation rate.
- PV array with 75 % degradation rate.
The obtained results have been compared with some interconnected schemes such as SP, and SuDoKu methods. Besides,
an intensive analysis considering mismatch power loss, fill factor, and percentage of power loss and power improvement is
included.
Case 1 : 2 PV modules M(1,1) and M(2,1) are aged (low Rsh ), with high temperature in M(1,4) and M(4,3)
In figure 3(a), the first test is conducted by applying the ageing scenario and the temperature increasing (60°C) on
solar PV system initially interconnected in SP, where the ageing rate represents 25 % (4 PV modules / 16 PV modules), and
the temperature increasing rate also represents 25 %. The obtained arrangement with the developed VB method is
illustrated in figure 3(c), as well as the scheme by SuDoKu has been inserted in figure 3(b) for comparison. It is shown that
the proposed algorithm rearranges the modules realizing a PV module voltage balancing. The main objective is to achieve,
if possible, PV modules with the same or close voltage values on the same row.
4

(a) (b) (c) (d)

Fig.3- PV array under 43.75% of infected modules, (a) SP configuration, (b) SuDoKu arrangement (c) VB algorithm
arrangement, and (d) I-V and P-V characteristic curve for case 1
From figure 3(d), it is very clear that the output power obtained by the proposed method is greater than both other
methods; SP, and SuDoKu. The gained power using the developed algorithm in the case of PV array under 43.75% of
infected modules is 510.8 W representing an improvement of 26.71 %, and 516.1 W, representing an improvement of
27.06 % compared to SP, and SuDoKu, respectively. With a significant improvement in terms of the output voltage. The
finding mentioned above is summarized in table 1.

Table 1 -Simulation results for case 1 Table 2- Simulation results for case 2

SP SuDoKu Proposed SP SuDoKu Proposed VB

VB

Output 1912 1906.7 2422.8 Output Power 1433.1 1432.7 2357.5

Power PPM (W)
PPM(W)

Output 76.72 76.59 82.34 Output Voltage 76.45 75.35 80.49

Voltage Vm Vm (V)
(V)

Gained 26.71 27.06 Gained power 64.5 64.54

power using using VB
VB method method (%)
(%)

Case 2 : 4 pv modules m(1,1) , m(1,2),m(2.1) and m(2.2) are aged (low rsh ), with high temperature in m(3,1),
m(3,4), m(4,2) and m(4,3)
In this case, as shown in figure 4(a) represents a solar PV system initially interconnected in SP under voltage
mismatch conditions, by increasing the temperature (60°C) of 5 PV modules, where the rate represents 31.25%, and also
insertion of 4 PV modules (25%) with a low parallel resistance value (ageing issue). Figure 4(c) shows the arrangement
achieved using the developed VB method, and also the scheme by SuDoKu has been included in figure 4(B) for
comparison. One can be noticed that the developed method rearranges the modules realizing a PV module voltage
balancing. The aim is to have, if possible, PV modules with the close voltage value on the same row.
As observed in figure 4(d), it is very clear that the output power obtained by the proposed method is greater than both
reported methods; SP, and SuDoKu. The gained power using the developed algorithm in the case of PV array under 56.26
% of infected PV modules is 924.4 W representing an improvement of 64.5 %, and 924.8 W, representing an improvement
of 64.54 % compared to SP, and SuDoKu, respectively. Where a significant improvement in terms of the output voltage.
The reported finding above is summarized in table 2.
 5

(a) (b) (c) (d)

Fig.4- PV array under 56.25% of infected modules, (a) SP configuration, (b) SuDoKu arrangement, (c) VB algorithm
arrangement, and (d) I-V and P-V characteristic curve of case 2.

Case 3 : 2 PV modules M(1,1) , M(1,2),M(2.1) M(2,2), M(3,3), M(3,4), M(4,3), and M(4.4) are aged (low Rsh ),
with a high temperature in M(1,4), M(2,3), M(4,1) and M(4,4).

For further validation of the proposed algorithm, and as shown in figure 5 (a), a totally degraded PV array initially
interconnected in SP has been chosen, with 50% of the PV modules having the lowest parallel resistance (PID dilemma ),
and an increase of the ambient temperature inside the PV modules. The obtained arrangement by the proposed VB method
is depicted in figure 5(c), as well as the scheme by SuDoKu has been inserted in figure 5(b) for comparison. It is clearly
shown that the proposed algorithm rearranges the modules realizing a PV module voltage balancing. The aim is to have, if
possible, PV modules with the same or close voltage value on the same row.

(a) (b) (c) (d)

Fig.5- PV array under more than 75% of infected modules,(a) SP configuration,(b) SuDoKu arrangement, (c) VB algorithm
arrangement, and (d) I-Vand P-V characteristic curve for case 3.
As shown in Figure 5(d), a rate of more than 200 % of power is gained after applying the proposed VB algorithm
compared to both methods; SP and SuDoKu. The gained power using the proposed algorithm in the case of PV array under
75% of infected PV modules is 1011.14W representing an improvement of 195.93%, and 1018.39 W, representing an
improvement of 203.11% compared to SP, and SuDoKu, respectively. Whereas a huge increase in the output voltage. The
mentioned finding above is summarized in table 3.
6

Table 3- Simulation results for case 3

SP SuDoKu Proposed VB

Output Power PPM 516.06 508.81 1527.2

(W)

Output Voltage Vm 38.13 33.30 52.32

(V)

Gained power using 195.93 200.15

the VB method (%)

As a conclusion of this simulation study, and after discussing the results obtained, it can be concluded that the reliability of
the proposed VB algorithm increases when we treat a PV array with a large number of PV modules infected by the
phenomenon of voltage mismatch like the aging or/and variation in the ambient temperature.

5. Comparative study between SP, SuDoKu, and proposed VB methods

A comprehensive analysis is performed based on the mismatch power loss, fill factor, and percentage of power loss
and power improvement, in order to evaluate the overall performance improvement obtained with the application of the VB
method compared to the interconnected SP and SuDoKu schemes. The analysis is performed by using 4×4 interconnected
PV modules described for three cases of different voltage mismatch conditions considered in the previous subsections. The
details of the mathematical formulation for the measurements used are as follows [12]:

Mismatch power loss (PMismatchloss) = PMax(HC)-PGMPP_(PC) (3)

( 𝑉𝑚 ×𝐼𝑚 )
Fill Factor (FF) =
𝑉𝑜𝑐 ×𝐼𝑠𝑐
(4)
𝐺𝑀𝑃𝑃𝐻𝐶 −𝐺𝑀𝑃𝑃𝑃𝐶
Percentage Power loss (Ploss(%)) = (5)
𝐺𝑀𝑃𝑃𝐻𝐶
𝐺𝑀𝑃𝑃𝑝𝑒𝑒𝑟−𝐺𝑀𝑃𝑃𝑉𝐵
Percentage of power enhancement (Penhancement%) = 𝐺𝑀𝑃𝑃𝑝𝑒𝑒𝑟
(6)

whereaHCais healthy conditions, GMPP_HC and GMPP_PC are the global maximum power point at healthy conditions
(HC) and poorly conditions (PC), respectively. Vm and Im are the maximum value of voltage and current in poor conditions.
Isc and Voc are the short circuit current and open-circuit voltage, respectively. GMPPpeer denotes the global maximum power
point obtained by SP and SuDoKu interconnected systems, while GMPPVBis is the global maximum power point given by
the proposed technique VB .
Figure 6 shows the global maximum power generated by the PV array under fully healthy and poor conditions, as well as
the mismatch power loss among them for each arrangement in the studied three cases. Looking at the graphs, it can be
observed that the VB arrangement achieved the lowest values of unbalanced power loss in the bar graphs over the three
voltage mismatch cases, which enhances the reliability of the proposed VB-based scheme.
The remaining metrics (FF, %Ploss, and %Penhancement) of the introduced and competitive techniques based on
interconnected systems under the three voltage mismatch cases are presented in Figures 7 and 8. The figures of the fill
factor (FF) show that the VB-based PV reconfiguration has the highest FF compared to the other approaches, which means
that the proposed VB scheme produces the highest maximum power for any degradation case. According to the percentage
power loss given in Figures 8(a), 8(b), and 8(c), the previous observation is improved as the VB has the minimum % power
losses of 29.56, 31.46, and 55.60 % over the voltage mismatch conditions, respectively, whereas SP takes the second place
regarding the maximum power generation in all the studied cases. While the SuDoKu method presents a high power loss
value in all cases. Based on the percentage of power enhancement bars in Figures 8(d), the VB arrangement shows a
significant enhancement in power generation compared to SP, and SuDoKu schemes, and noteworthy behavior in
comparison to the SuDoKu interconnected system in cases 2 and 3. For the power improvement in the case of SP, the VB
approach improves the produced power by 26.71, 64.49, and 195.92 % over the three voltage mismatch cases, respectively.
The above analysis highlights the superiority of the proposed VB arrangement under different degradation cases in
generating the highest power with the greatest regularity in PV characteristics with the shortest runtime and fastest
convergence.
 7

(a) (b)

(c)

Fig-6- Mismatcha power loss for the studied voltage mismatch cases, (a) case 1, (b) case 2, and (c) case 3.

(a) (b)

(c)

Fig.7- Filla factor for the studied voltage mismatch cases, (a) case 1, (b) case 2, and (c) case 3.
8

(a) (b)

(c) (d)

Fig.8- Percentage values, (a) power losses for case 1, (b) power losses for case 2, (c) power losses for case 3, and (d)
power enhancement between VB arrangement and the other approaches.

6. Conclusion
In this work, a new reconfiguration algorithm for PV systems was adopted to attain the maximum power
production under voltage mismatch effects. For this, a VB-based reconfiguration method was developed for automatic
reconfiguration of the SP structure by considering the continuous rearrangement of the solar modules. The main idea of the
adopted VB method is based on the relocation of the installed PV modules in order to balance their voltage values. In brief,
The main contribution of the present work was the improvement of PV system performance using Novel PV system
reconfiguration based on a voltage balancing algorithm under voltage mismatch conditions. Software validation has been
conducted using MATLAB/Simulink. The proposed approach has provided good results. In addition, a comparative study
has been established with the SP and SuDoKu methods. In which the simulations demonstrate that the VB approach
improves the generated power by 26.71%, 64.49%, and 195.92 % for the three voltage mismatch cases, respectively, with
minimum power loss and higher fill factor. The obtained results have shown the superiority of the proposed method in the
output PV power under voltage mismatch. Moreover, the proposed approach is dynamic and does not need extensive
wiring compared to SuDoKu method. As perspectives, the extension of this research work can be envisaged; further
research is required to assure the practicality to realize the proposed technique.

REFERENCES

[1] T. Hong, C. Koo, J. Park, and H. S. Park, “A GIS (geographic information system)-based optimization model for
estimating the electricity generation of the rooftop PV (photovoltaic) system,” Energy, vol. 65, pp. 190–199, Feb.
2014, doi: 10.1016/[Link].2013.11.082.
[2] T. Hong, C. Koo, J. Park, and H. S. Park, “A GIS (geographic information system)-based optimization model for
estimating the electricity generation of the rooftop PV (photovoltaic) system,” Energy, vol. 65, pp. 190–199, Nov.
2014, doi: 10.1016/[Link].2013.11.082.
[3] R. Candela, V. di Dio, E. R. Sanseverino, and P. Romano, “Reconfiguration Techniques of Partial Shaded PV
 9

Systems for the Maximization of Electrical Energy Production,” in 2007 International Conference on Clean
Electrical Power, May 2007, pp. 716–719, doi: 10.1109/ICCEP.2007.384290.
[4] C. M. Amaral Da Luz, P. Dos Santos Vicente, F. L. Tofoli, and E. M. Vicente, “Influence of series and shunt
resistances in the ideality factor of photovoltaic modules,” in 2016 12th IEEE International Conference on Industry
Applications, INDUSCON 2016, Nov. 2017, pp. 1–6, doi: 10.1109/INDUSCON.2016.7874588.
[5] A. Harrag and S. Messalti, “Adaptive GA-based reconfiguration of photovoltaic array combating partial shading
conditions,” Neural Comput. Appl., vol. 30, no. 4, pp. 1145–1170, Aug. 2018, doi: 10.1007/s00521-016-2757-y.
[6] S. T. BUDDHA, “Topology reconfiguration to improve the photovoltaic (PV) array performance,” Arizona State
Univ., vol. 11, no. 1, pp. 147–173, 2011.
[7] B. I. Rani, G. S. Ilango, and C. Nagamani, “Enhanced Power Generation From PV Array Under Partial Shading
Conditions by Shade Dispersion Using Su Do Ku Configuration,” IEEE Trans. Sustain. Energy, vol. 4, no. 3, pp.
594–601, Jul. 2013, doi: 10.1109/TSTE.2012.2230033.
[8] B. Dhanalakshmi and N. Rajasekar, “A novel Competence Square based PV array reconfiguration technique for
solar PV maximum power extraction,” Energy Convers. Manag., vol. 174, pp. 897–912, Oct. 2018, doi:
10.1016/[Link].2018.08.077.
[9] A. Loukriz et al., “A New Simplified Algorithm for Real-Time Power Optimization of TCT Interconnected PV
Array under Any Mismatch Conditions,” J. Eur. des Systèmes Autom., vol. 54, no. 6, pp. 805–817, Dec. 2021, doi:
10.18280/jesa.540602.
[10] S. A. Rahman, R. K. Varma, and T. Vanderheide, “Generalised model of a photovoltaic panel,” IET Renew. Power
Gener., vol. 8, no. 3, pp. 217–229, Apr. 2014, doi: 10.1049/iet-rpg.2013.0094.
[11] A. S. H. van der Heide, A. Schönecker, J. H. Bultman, and W. C. Sinke, “Explanation of high solar cell diode
factors by nonuniform contact resistance,” Prog. Photovoltaics Res. Appl., vol. 13, no. 1, pp. 3–16, Jan. 2005, doi:
10.1002/pip.556.
[12] D. Yousri, D. Allam, and M. B. Eteiba, “Optimal photovoltaic array reconfiguration for alleviating the partial
shading influence based on a modified harris hawks optimizer,” Energy Convers. Manag., vol. 206, p. 112470, Feb.
2020, doi: 10.1016/[Link].2020.112470.

View publication stats