IET Renewable Power Generation

Special Issue: Selected Papers from the 5th IET Renewable Power
Generation Conference 2016

String level optimisation on grid-tied solar PV ISSN 1752-1416

Received on 3rd April 2017
Revised 19th October 2017
systems to reduce partial shading loss Accepted on 13th November 2017
E-First on 30th November 2017
doi: 10.1049/iet-rpg.2017.0229
www.ietdl.org

Ahsan Sarwar Rana1,2, Mashood Nasir1, Hassan Abbas Khan1

1Department of Electrical Engineering, Lahore University of Management Sciences, Lahore, Pakistan
2Department of Electrical Engineering, Information Technology University of the Punjab, Lahore, Pakistan
E-mail: hassan.khan@.lums.edu.pk

Abstract: Partial shading, commonly observed in domestic rooftop solar photovoltaic (PV) deployments, can be highly
detrimental to the performance ratio (PR) of a PV system. Typically, for domestic installations, string-inverter or module micro-
inverter configurations are deployed. While module level micro-inverters generally present a better response to non-uniform
distributions of sunlight, they are still less common and therefore, costly in many emerging markets. String-level
implementations, on the other hand, are widely deployed as they are less complex and cost efficient. In this work, the authors
present an analytical and simulation framework for improving PR under partial shading conditions through alteration of string
connections in a string-level inverter system. Results show up to 4.6% higher PR in winter months for a 42.24 kWp system
installed at Lahore University of Management Sciences, Lahore, Pakistan.

1 Introduction PV string. According to Zaihidee et al. [20], dust accumulation of

20 g/m2 on a PV panel reduces short circuit current, open circuit
There is a growing shift from fossil fuels to renewable resources
voltage and efficiency by 15–21, 2–6 and 15–35%, respectively.
for electricity generation worldwide. Renewable resources,
Typically, in the case of soft shade, the maximum power point
particularly solar energy has a huge potential in many countries and
tracking (MPPT) algorithm of the inverter may reduce the current
can contribute significantly to the overall electricity mix [1, 2].
in the entire string to take advantage of overall voltage contribution
Solar energy can be produced through (a) solar thermal energy
of the string to maximise the power output. In hard shades, where a
extraction or (b) photovoltaic (PV) extraction using solar PV
panel is hard shaded (bird dropping or other reasons where the
modules/panels. The solar thermal process harnesses the solar
input irradiance is fully blocked), the bypass diode becomes active,
energy by extracting heat from sunlight which can then be used to
completely bypassing the panel/module. This lowers the overall
make steam to drive a turbine to produce electricity. On the other
operating voltage of the string. In this work, we focus only on ‘soft
hand, PV technology extracts the energy of photons in sunlight
shading losses’ (also referred as ‘partial shading losses’) due to the
through solar cells to generate electron/hole pairs which flow in the
structures in the vicinity of the PV deployment.
outer circuit to generate electricity. Solar PV technologies have
Most residential sites, where PV panels are installed, are usually
seen a much higher growth in the last decade due to decreasing
surrounded by other structures or buildings and have a variable
costs of solar panels and balance-of-system components [3–5].
pattern of shade with respect to the position of the sun. When
Most commonly found urban domestic PV systems include
shade(s) reach a PV installation, it decreases the output of the
grid-tied topologies where many of the panels are connected to a
panels by lowering the current generation of its shaded cell as cells
central inverter feeding directly to the grid. Many factors affect the
are connected in series. As a remedy, solar panels are equipped
output of PV system; these mainly include temperature [6, 7], low
with bypass diodes which (a) lower the losses by providing a
irradiance [8, 9], pitch and orientation of PV panels [10], efficiency
parallel path for the shaded cells/panels and (b) prevent against
of inverters and batteries (in the case of systems with backups [11],
hot-spots which could permanently damage a panel. However,
generally installed in areas with intermittent grids), wiring losses
depending on the type of shade, the shading losses could still be
and shading [9, 12, 13]. The shadowing or shading loss can be very
very significant.
significant for urban settings affecting the performance ratio (PR)
The PR can be improved by minimising the effect of shading
for central inverter orientation [14–16]. For instance, Deline et al.
through various schemes such as modifying interconnections of
[16] showed that the PR of a c-Si panel based PV system could
modules and strings, reconfigurable arrays and string level
range from as little as 20% to 80% for a 30% shading.
optimisation [21–30]. The possible modules and string
There are various classifications of shading of which
interconnection schemes include total cross tied (TCT) and branch
dichotomist classification, i.e. ‘objective’ and ‘subjective’ shading,
linked (BL) discussed in [24–26]. These schemes generally
is more prevalent. Objective shade is due to cloudy weather or it
distribute the effect of partial shading evenly and minimise the
simply can be a time of the day when there is sparse irradiance
power degradation due to the shadows. BL and TCT are less
available. Objective shading cannot be avoided as the sun gets
susceptible to partial shading problems; however, large
blocked in it and whole PV installation is likely to get evenly
interconnection redundancy requires extra conductors, resulting in
affected. The subjective shading can be classified into ‘static’ and
expensive cabling and a reduced return-on-investments index.
‘dynamic’ shading [17]. Static shading occurs due to an anomaly in
Several other techniques using dynamically reconfigurable PV
the vicinity of a PV system (such as dirt, bird droppings, etc.) and
arrays to mitigate the effects of partial shading have been presented
is also referred to as hard shading [9]. Dynamic (soft) shading can
in the literature [27–32]. These schemes utilise complex control
be in the shape of distant buildings, structures or trees causing a
algorithms to control the switches responsible for reconfiguration
shade on the PV installation. Hard shading can be improved by
of the array. Computational complexity along with real-time
cleaning panels [18, 19], whereas multiple techniques are
sensing requirements along with decreasing solar module prices
employed to reduce soft shading loss. Hard shading affects both
makes these schemes costly and largely unviable for small-scale
open operating open circuit voltage and short circuit currents of a
implementations. In addition, a reliability issue of switches is often

IET Renew. Power Gener., 2018, Vol. 12 Iss. 2, pp. 143-148 143
© The Institution of Engineering and Technology 2017
 to stay shade free for larger intervals. The only cost of this
alteration is the extra conductor requirement which is significantly
less than TCT and BL modified reconnection schemes used in the
literature. Moreover, retrofitting of the existing systems to TCT or
BL orientation requires complex interconnections (from
implementation point-of-view) may be highly challenging.
Therefore, the presented framework is suitable for planning new
installations as well as retrofitting of the existing installations with
minor modifications in the string structure.

2 Methodology
Typically, in rooftop implementations, string level inverters are
commonly implemented. A simple arrangement of this scheme
Fig. 1  Three strings of 22 panels each connected to a central string
with three parallel strings of 22 panels each connected to a central
inverter
inverter is shown in Fig. 1. Each panel generally contains a number
of bypass diodes which play a central role in minimising shading
losses. Typically, three bypass diodes are used in a panel of 60
cells, which distributes one diode per block of 20 cells as shown in
Fig. 2. If one cell is shaded in a block (e.g. cell 1–20), an alternate
path for current is provided by the bypass diode (BD1). While,
under partial shading condition, the current may remain the same in
a panel, the power output of the system is affected due to the
exclusion of the ‘bypassed’ block. Further shading of cells within
the same block will not affect the power output as the block is
already being bypassed. However, if one cell from another block
(e.g. cell 21–40) also gets shaded then two blocks from the panel
are (typically) bypassed resulting in one-third of the production.
This is particularly critical in the performance of these systems and
various efficient MPPT algorithms tackle this by appropriately
Fig. 2  Basic schematic of a PV panel with 60 cells and three bypass decreasing the current levels to maximise the power output [33–
diodes making three blocks of 20 cells 35].
In this work, the key task is to analyse the system's shading
an important concern for these systems. Therefore, for small- and pattern and evaluate possible gains through the possible
medium-scale installations, a simplified, computationally less restructuring of the strings. The resulting gains can be analysed
extensive and cost effective strategy with minimum hardware through a software (such as PSIM) or analytically. In this work, we
(switches, cables, and conductors) requirements is highly desirable have used both approaches to ascertain the efficiency
to mitigate the power degradation effects of shading. improvement. For analytical evaluation of the partially shaded
An interesting technique based on the SU DO KU configuration system, it is important to summarise some basic PV cell parameters
of modules to enhance the power output of the PV array is [36]
discussed in [32], However, in such a scheme, physical locations of
the modules are changed, while the electrical interconnection
q V + IRs V + IRs
I = Isc − I01e nkT − , (1)
remains unaltered. Such a scheme based upon module relocation is Rsh
sub-optimal due to (a) relocation of modules requires labour and
physical resources to realise the physical relocation and (b) since kT Isc
V oc = ln +1 , (2)
electrical interconnections are unaltered, while the position of the q Io
module has been changed, therefore, an extra conductor may be
required for the module to be located at another position. This extra where I is the  output current at the terminal, Isc is the short-circuit
conductor will not only increase the cost of the system but also cell current, I01 is the reverse saturation current, q is the charge of
enhance the associated distribution and wiring losses. Moreover, the electron, V is the voltage at the terminal, Rs is the series
the SU DO KU based method does not take the site-specific
resistance of a cell, n is the ideality factor, k is the Boltzmann
shading patterns and incident irradiance into consideration for
constant, T is  the temperature under standard test conditions
maximising the PV array output power. In contrast, genetic
(STC), Rsh is the shunt resistance of a cell and Voc is the open-
algorithm (GA)-based electrical interconnection optimisation of
various modules in PV arrays is utilised such that their physical circuit cell voltage.
location remains unaltered as discussed in [31]. Although the Equations (1) and (2) evaluate the current and voltage under
labour requirements associated with the relocation of panels and STC for a solar cell. However, in order to incorporate the effect of
complexity of interconnections resulted from physical relocation changing irradiance and changing temperature, further translation
may be reduced by using GA-based optimisation of equations are established. For a typical Si-based solar panel,
interconnections. However, such GA-based schemes have the equations are summarised [37]
tendency to converge to local maximum rather than global
maximum, which may result in reduced output power. Moreover, E
Isc = Isc, o (1 + α T − T o , (3)
the convergence of the GA algorithm is highly dependent on Eo
parameter selection which limits its widespread use.
In this work, we devise a method for enhancing PR in a string- E
V oc = V oc, o 1 + aln + β T − To , (4)
level implementation through shading analysis at the time of Eo
installation or one-time rearrangement of string structures in the
existing PV systems to achieve a higher PR. It should be noted that Isc
Imp = Imp, o , (5)
the modification does not include physical relocation of the panels, Isc, o
but only involves re-stringing with minor alterations whereby
several shaded panels in neighbouring strings are swapped with un- V mp = V mp, o + V oc − V oc, o + Rs Imp, o − Imp , (6)
shaded panels to increase the combined output of two strings. In
essence, the efficiency gains are achieved through allowing strings

144 IET Renew. Power Gener., 2018, Vol. 12 Iss. 2, pp. 143-148
© The Institution of Engineering and Technology 2017
 where S is the total number of shaded blocks in a string, RBD is the
forward resistance of a bypass diode, VT is the diode threshold
voltage drop and M is the total number of modules/panels in a
string.
If the condition presented in (8) is false, then the MPPT
algorithm will not bypass the blocks, therefore, to attain the
maximum power from the string, each block will contribute
towards the net power from the string. Such a string is classified as
an inactive bypass string and its important parameters including
short circuit current Ia°sc, instantaneous current at maximum power
point Ia°mp, open circuit voltage Va°oc, and the instantaneous
voltage at maximum power point Va°mp, must be modified and are
given by (11)–(14). In this case, short circuit current will vary in
Fig. 3  Typical grid-tied solar PV system having R strings with M modules
direct proportion with the shade irradiance Es normalised over
in each string
standard irradiance Eo, while open circuit voltage will show a
Pcal = V mpImp, (7) logarithmic dependence with shade irradiance Es normalised over
standard irradiance Eo as shown by (11) and (13), where shade
where Isc,o is the short-circuit current under STC, α is the short- irradiance assumption has already been explained above
circuit current temperature coefficient, T is the operating
temperature, To is the temperature under STC, E is the a∘ Es
Isc = Isc, o (1 + α T − T o , (11)
instantaneous irradiance, Eo is the standard Irradiance (1000 W/ Eo
m2), Voc,o is the open-circuit voltage under STC, a is the irradiance ∘
a
correction factor of Voc, β is the open-circuit voltage temperature a∘ Isc
Imp = Imp, o , (12)
coefficient, Imp is the instantaneous current at maximum power, Isc, o
Imp,o is the current at maximum power under STC, Vmp is the
instantaneous voltage at maximum power, Vmp,o is the voltage at ∘
a Es
V oc = V oc, o 1 + aln + β T − To , (13)
maximum power under STC, Rs is the series resistance and Pcal is Eo
the calculated maximum power. ∘ ∘
a a a0
Equations (3)–(7) quantify the response of a solar panel to V mp = V mp, o + V oc − V oc, o + Rs Imp, o − Imp M. (14)
changing parameters. As every PV installation is different due to
its location, design, number of panels installed, and manufacture of These equations have been used in conjunction with software
the panels, the aforementioned equations cannot be linearised for simulation to evaluate system performance for the observed pattern
MPPT operation under shaded conditions. What is needed is a of shading.
generic set of equations which could quantify the response of the
PV system even in shaded conditions under normal MPPT
operation. For instance, if some of the blocks/panels are being 3 Optimisation frame-work for string level
bypassed due to non-uniform shading then the MPPT algorithm optimisation
must be able to account for that in power estimation. We, therefore, For a generalised solar PV system having R strings with M
deduce the following condition as a reference for ascertaining modules in each string as shown in Fig. 3, the total number of
maximum attainable power from a string blocks Nt can be calculated depending upon the number of bypass
diodes D per module
U
E ≥ Es, (8)
N Nt = R × M × D . (15)
where U is the total number of un-shaded blocks in a string, N is Based upon the incident irradiance on each block, these blocks
the total number of blocks in a string and Es is the irradiance in can further be classified as shaded blocks Si and un-shaded blocks
shade. Ui in each string i. Therefore, for each string i the total number of
‘Blocks’ basically represent the number of bypass diodes in a
string. If a bypass is active, it will be counted as a shaded block blocks per string N, the total number of un-shaded blocks N1 and
and if not it will be counted as an un-shaded block. Irradiance is the total number of shaded blocks N2 can be represented by (16),
measured through an irradiance sensor (SMA Sunny Sensor Box) (17) and (18), respectively
with data-logging and provides the values of irradiance on a 15 min
interval. The irradiance of shaded panels is also measured through N = Ui + Si ∀i ∈ [1, R], (16)
a reference irradiance sensor (KEWTECHPV1). For any string, if
R
(8) is true, then the instantaneous current at maximum power Iampis
the same as (5) while instantaneous voltage at maximum power
N1 = ∑ U i, (17)
i=1
Vamp is given by (10) [38]. In this case, short circuit current will
vary in direct proportion with the incident irradiance E normalised R
over standard irradiance Eo, while open circuit voltage will show a N2 = ∑ Si . (18)
logarithmic dependence with incident irradiance E normalised over i=1

standard irradiance Eo as shown by (3) and (4)

While the total number of blocks N given by (15) can be
represented in terms of (17) and (18) by (19)
a Isc
Imp = Imp, o , (9)
Isc, o R

a
Nt = ∑ Ui + Si . (19)
V mp = V mp, o + V oc − V oc, o + Rs Imp, o − Imp i=1

U a
(10)
× M − Imp S RBD − S V T,
N

IET Renew. Power Gener., 2018, Vol. 12 Iss. 2, pp. 143-148 145
© The Institution of Engineering and Technology 2017
 Subject to the constraints given by (15)–(19).
This optimisation problem is solved using standard linear
optimisation technique in MATLAB to find the values of Ui and Si
for each string i. Based upon the found values, connections of
modules in the strings are modified to obtain the optimised output
from the installed system capacity. The schematic diagram of the
system and one optimised reconnection, after optimisation has been
shown in Figs. 3 and 4, respectively.

4 Conductor requirements for reconfiguration of

connections
Fig. 5 shows a typical case of PV installation with various lengths
(xa–xd) shown. The extra conductor required Condx for optimised
Fig. 4  Schematic diagram for the optimized system with modified structure interconnection of panels in terms of xa, xb, xc and xd can be
in neighboring strings to maximize PR expressed as a function of numbers of re-connections λ calculated
through optimisation framework discussed in Section 3. Therefore,
an extra conductor is required to obtain enhanced PR through inter
string reconnections of PV modules is given as

M
Condx = λ 2xd + 4xa λ ∈ 0, . (23)
2

The total conductor Condt required to ensure optimised

operation is given as

Condt = M − 1 xb + λ 2xd + 4xa , (24)

Fig. 5  Typical series connected panels with various lengths nominated for
overall conductor requirement calculations where xa is the length of ground to the top pane of junction box, xb 
is the length between two junction boxes of two connecting panels
Table 1 Rating of installed panel BYD240P6–30 in series, xc is the length of cable from a particular panel to the
Rated maximum power (Pmax) 240 Wp sheath provided for cable integrity and xd is the length from one
tolerance 0–5 W row of panels to next row of panels in an installation.
voltage at Pmax (Vmp) 29.64 V
vurrent at Pmax (Imp) 8.10 A 5 System implementation (case study)
open-circuit voltage (Voc) 37.3 V The proposed methodology is tested through a 42.24 kWp system
short-circuit current (Isc) 8.57 A installed at Lahore University of Management Sciences (LUMS),
Lahore, Pakistan. In this system, three central inverters are
nominal operating cell temp. 45°C ± 2°C
connected to eight strings (5.28 kWp each) with 22 modules (panel)
per string. Out of three inverters, the strings connected to the
second inverter remain completely shade-free, which acts as a
For a given shade pattern, Si and Ui may vary in each string i, reference for other strings (shaded) due to a neighbouring building.
therefore, the output power PSi(t) of each string at any time t will Specifications for installed panels are given in Table 1 and detailed
vary accordingly and is given as system description is presented in our earlier work [14].
The top view and the building level installation design of the
a a
V mp , i t Imp, i t → Ei t Ui t ≥ NEs, i t system are depicted in Fig. 6 with string connection design shown
in Fig. 7. Different colours in Fig. 7 represent separate strings of 22
PSi t = a∘
V mp a∘
, i t Imp, i t → Ei t Ui t < NEs, i t
, (20)
panels, for instance, IJK (orange) and FHI (dark green) are two out
; ∀t ∈ [1, T], ∀i ∈ [1, R] of eight strings of 22 panels in series. DEG (grey) strings are the
ones which stay shadow free at all times and serve as a reference
Based upon the information of Ui and Si, in each string, for loss characterisation. The string structure is fixed and generally
connections of blocks and associated modules can be modified optimised for best performance in summer months when the
such that most of the un-shaded blocks are in the same string for sunlight is at its maximum. However, shading pattern differs in
longer intervals such that the overall output power is maximised. winter and the performance of the system decline as a result.
Theoretically, the maximum attainable power Pmax(t) at any Therefore, an optimised solution with optimum string connections
time instant t for the installed system at a given shading profile is is key to maximise PR throughout the year.
given by the summation of individual maximum power point
operation of all the modules and is given as 6 Results and discussion
R M
To quantify the overall gains, it is important to analyse the PR of
Pmax t = ∑ ∑ Pcal i j t , . (21) the system which is defined as
i=1 j=1
Measured AC output KWh
PR = , (25)
The optimisation function is developed to minimise the cumulative Theoretical DC production without losses
sum of the difference between the maximum attainable power and
the possible attainable power through re-connections of the blocks where theoretical direct current (DC) production is calculated by
in the strings over a defined time period Ts is finding equivalent peak sunlight hours of the day through local
measurement or through the National Renewable Energy
Ts Ts R Laboratory (NREL) data [39] which when multiplied by panel
min ∑P max t − ∑ ∑ PSi t . (22) nameplate capacity and a number of panels gives the theoretical
Ui, Si t = 1 t=1i=1 DC string production without losses. This with reference to actual

146 IET Renew. Power Gener., 2018, Vol. 12 Iss. 2, pp. 143-148
© The Institution of Engineering and Technology 2017
 Fig. 9  PV inverter power output (simulated) for the baseline compared
with modified string structure for a typical winter day
Fig. 6  Building level view (left) and the top view (right) of the installation
at LUMS library building with four obstructing structures causing soft
shading at various times of the day

Fig. 10  PV inverter power output (analytically calculated) for the baseline
compared with modified string structure for a typical winter day

Fig. 7  Sting level installation design for the rooftop system

Fig. 11  PR improvement after the proposed restringing for 12-month

period

discussed in Section 3 for the two shaded strings, the

rearrangement gives higher power output for a typical day as
shown by Figs. 9 and 10.
Fig. 8  Typical winter day measured data along with simulated and To quantify the annual gains, we evaluate the PR (on monthly
calculated results basis) which is shown in Fig. 11. This is done through modelling
the building structure along with obstructions in PVSOL premium
accounted energy units added to the grid (alternating current) gives and irradiance data are taken from the NREL [39]. Once the
the PR. For the current implementation, for a typical winter day, shading patterns are known, the processing could be done
the power produced by an inverter (combination of two entire accordingly. Unlike active schemes (such as reconfigurable arrays
strings) is shown in Fig. 8 along with the simulated (PSIM) and which require real-time information for processing), our work is
calculated (analytical model discussed in Section 2) data. Measured based on offline processing (with standard computing resources) of
output and irradiance data corresponding to the observations are the information for one-time alteration of strings. Results show a
taken at a 15 min interval through the data logging system. This higher gain in the PR (up to 4.6%) in winter months as the system
averaging, along with variations in the shade irradiance values due encounters large shade for these months. However, the re-stringing
to reflections from neighbouring mumty accounts for the slight does not have a negative impact on the summer months largely due
discrepancy in measured and simulated/calculated results in Fig. 8. to the fact that the shades are minimal in these months.
The average factor of shade on various panels may vary due to It is important to note that the extra conductor required to
ambient reflections and some variation is therefore observed. achieve this enhanced PR is calculated through (23). In general, the
Simulations were performed using PSIM software to evaluate viability of the proposed optimisation can be assessed through the
the system performance for the observed pattern of shading. comparison of the cost associated with the extra conductor and the
Variable shading was added to the simulation through C-Block savings associated with the enhanced utilisation of the grid-tied
generating varying irradiance to the solar physical module in the system after optimisation. The savings can be calculated by
software. Data for temperature were also added in the C-block. multiplying per unit cost ($/kWh) of electricity to the difference of
Analytical results have been achieved using the model elaborated a number of generated units after and before optimisation. Thereby,
in Section 2. After performing optimisation using the framework payback time for the cost of the conductor can also be calculated.

IET Renew. Power Gener., 2018, Vol. 12 Iss. 2, pp. 143-148 147
© The Institution of Engineering and Technology 2017
For the current installation, a 63.4 m extra conductor is required to [14] Ahmed, J., Salam, Z.: ‘A critical evaluation on maximum power point
tracking methods for partial shading in PV systems’, Renew. Sustain. Energy
make optimised interconnections, while approx. 114 extra units Rev., 2015, 47, pp. 933–953
(kWh/year) would be generated after the optimisation. Thereby, [15] Batzelis, E.I., Georgilakis, P.S., Papathanassiou, S.A.: ‘Energy models for
taking into account the basket electricity price ($0.15/kWh), the photovoltaic systems under partial shading conditions: a comprehensive
payback time for the extra cost of the conductor (approx. 1$/m) review’, IET Renew. Power Gener., 2015, 9, pp. 340–349
[16] Deline, C.: ‘Partially shaded operation of a grid-tied PV system’. 2009 34th
comes out to be <4 years for this installation. Similarly, for any IEEE Photovoltaic Specialists Conf. (PVSC), 2009, pp. 001268–001273
other installation, this analysis must be done for any possible [17] Bulanyi, P., Zhang, R.: ‘Shading analysis & improvement for distributed
restringing based upon the outcomes of optimisation discussed in residential grid-connected photovoltaics systems’. Solar 2014, 2014
Section 3 to get the maximum efficiency from the system. [18] Lamont, L.A., El Chaar, L.: ‘Enhancement of a stand-alone photovoltaic
system's performance: reduction of soft and hard shading’, Renew. Energy,
2011, 36, pp. 1306–1310
7 Conclusion [19] Ramli, M.A., Prasetyono, E., Wicaksana, R.W., et al.: ‘On the investigation of
photovoltaic output power reduction due to dust accumulation and weather
The losses due to partial shading are not proportional to the shaded conditions’, Renew. Energy, 2016, 99, pp. 836–844
area but depend on the shading pattern, array configuration and [20] Zaihidee, F.M., Mekhilef, S., Seyedmahmoudian, M., et al.: ‘Dust as an
unalterable deteriorative factor affecting PV panel's efficiency: why and how’,
physical location of shaded modules in the array. As shading Renew. Sustain. Energy Rev., 2016, 65, pp. 1267–1278
patterns vary throughout the year, due to relative sun position, the [21] Chaaban, M.A., El Chaar, L., Alahmad, M.: ‘An adaptive photovoltaic
shading on the panels varies affecting the system's PR. We analyse topology to overcome shading effect in PV systems’, Int. J. Photoenergy,
shading losses for a central inverter PV system and evaluate gains 2015, 2015, pp. 1–9
[22] Hasan, M., Parida, S.: ‘An overview of solar photovoltaic panel modeling
in PR due to minor re-stringing of neighbouring panels. In our based on analytical and experimental viewpoint’, Renew. Sustain. Energy
installation, this scheme produces a higher PR of up to 4.6% in Rev., 2016, 60, pp. 75–83
winter whereas a minute increase in PR is seen in summer months. [23] La Manna, D., Vigni, V.L., Sanseverino, E.R., et al.: ‘Reconfigurable
This work is particularly relevant for domestic rooftop electrical interconnection strategies for photovoltaic arrays: a review’, Renew.
Sustain. Energy Rev., 2014, 33, pp. 412–426
deployments where PR reduction is commonly observed due to [24] Kaushika, N.D., Gautam, N.K.: ‘Energy yield simulations of interconnected
partial shading losses. solar PV arrays’, IEEE Trans. Energy Convers., 2003, 18, pp. 127–134
[25] El-Dein, M.S., Kazerani, M., Salama, M.: ‘An optimal total cross tied
interconnection for reducing mismatch losses in photovoltaic arrays’, IEEE
8 Acknowledgments Trans. Sustain. Energy, 2013, 4, pp. 99–107
[26] Wang, Y.-J., Lin, S.-S.: ‘Analysis of a partially shaded PV array considering
Hassan khan would like to acknowledge the support of (a) LUMS different module connection schemes and effects of bypass diodes’. 2011 Int.
FIF grant (216) for RA funding and (b) HEC, Pakistan for initial Conf. and Utility Exhibition on Power and Energy Systems: Issues &
conference funding (252.102). Prospects for Asia (ICUE), 2011, pp. 1–7
[27] El-Dein, M. S., Kazerani, M., Salama, M.: ‘Optimal photovoltaic array
reconfiguration to reduce partial shading losses’, IEEE Trans. Sustain.
9 References Energy, 2013, 4, pp. 145–153
[28] Nguyen, D., Lehman, B.: ‘An adaptive solar photovoltaic array using model-
[1] Harijan, K., Uqaili, M.A., Mirza, U.K.: ‘Assessment of solar PV power based reconfiguration algorithm’, IEEE Trans. Ind. Electron., 2008, 55, pp.
generation potential in Pakistan’, J. Clean Energy Technol., 2015, 3, pp. 54– 2644–2654
56 [29] Spagnuolo, G., Petrone, G., Lehman, B., et al.: ‘Control of photovoltaic
[2] Li, D., Liu, G., Liao, S.: ‘Solar potential in urban residential buildings’, Sol. arrays: dynamical reconfiguration for fighting mismatched conditions and
Energy, 2015, 111, pp. 225–235 meeting load requests’, IEEE Ind. Electron. Mag., 2015, 9, pp. 62–76
[3] Sahu, B.K.: ‘A study on global solar PV energy developments and policies [30] Storey, J.P., Wilson, P.R., Bagnall, D.: ‘Improved optimization strategy for
with special focus on the top ten solar PV power producing countries’, Renew. irradiance equalization in dynamic photovoltaic arrays’, IEEE Trans. Power
Sustain. Energy Rev., 2015, 43, pp. 621–634 Electron., 2013, 28, pp. 2946–2956
[4] Khan, J., Arsalan, M.H.: ‘Solar power technologies for sustainable electricity [31] Niranjan, D.S., Dhale, S., Mukherjee, J., et al.: ‘Solar PV array
generation – a review’, Renew. Sustain. Energy Rev., 2016, 55, pp. 414–425 reconfiguration under partial shading conditions for maximum power
[5] Hirth, L.: ‘Market value of solar power: is photovoltaics cost-competitive?’, extraction using genetic algorithm’, Renew. Sustain. Energy Rev., 2015, 43,
IET Renew. Power Gener., 2015, 9, pp. 37–45 pp. 102–110
[6] Nordmann, T., Clavadetscher, L.: ‘Understanding temperature effects on PV [32] Rani, B.I., Ilango, G.S., Nagamani, C.: ‘Enhanced power generation from PV
system performance’. Proc. 3rd World Conf. on Photovoltaic Energy array under partial shading conditions by shade dispersion using Su Do Ku
Conversion, 2003, 2003, pp. 2243–2246 configuration’, IEEE Trans. Sustain. Energy, 2013, 4, pp. 594–601
[7] Khan, H.A., Pervaiz, S.: ‘Technological review on solar PV in Pakistan: [33] Chin, C.S., Neelakantan, P., Yang, S.S., et al.: ‘Effect of partially shaded
scope, practices and recommendations for optimized system design’, Renew. conditions on photovoltaic array's maximum power point tracking’, IEEE
Sustain. Energy Rev., 2013, 23, pp. 147–154 Transactions on Industrial Electronics, 2011, 12, pp. 52–59
[8] Pervaiz, S., Khan, H.A.: ‘Low irradiance loss quantification in c-Si panels for [34] Wang, Y., Li, Y., Ruan, X.: ‘High-Accuracy and fast-Speed MPPT methods
photovoltaic systems’, J. Renew. Sustain. Energy, 2015, 7, p. 013129 for PV string under partially shaded conditions’, IEEE Trans. Ind. Electron.,
[9] Thevenard, D., Pelland, S.: ‘Estimating the uncertainty in long-term 2016, 63, pp. 235–245
photovoltaic yield predictions’, Sol. Energy, 2013, 91, pp. 432–445 [35] Nasir, M., Zia, M.F.: ‘Global maximum power point tracking algorithm for
[10] de Wild-Scholten, M., Alsema, E., Ter Horst, E., et al.: ‘A cost and photovoltaic systems under partial shading conditions’. 2014 16th Int. Power
environmental impact comparison of grid-connected rooftop and ground- Electronics and Motion Control Conf. and Exposition (PEMC), 2014, pp.
based PV systems’. 21st European Photovoltaic Solar Energy Conf., 2006, pp. 667–672
3167–3173 [36] Solanki, C.S.: ‘Solar photovoltaics: fundamentals, technologies and
[11] Mundada, A.S., Shah, K.K., Pearce, J.: ‘Levelized cost of electricity for solar applications’ (PHI Learning Pvt. Ltd, India, 2015)
photovoltaic, battery and cogen hybrid systems’, Renew. Sustain. Energy Rev., [37] Ding, K., Zhang, J., Bian, X., et al.: ‘A simplified model for photovoltaic
2016, 57, pp. 692–703 modules based on improved translation equations’, Sol. Energy, 2014, 101,
[12] Pervaiz, S., Khan, H.A., Qureshi, M.A.: ‘A practical perspective on grid-tied pp. 40–52
PV systems in low reliability grids’. 2014 IEEE 40th Photovoltaic Specialist [38] Rana, A.S., Khan, H.A.: ‘String level optimization on grid-tied solar PV
Conf. (PVSC), 2014, pp. 1969–1972 systems for minimizing soft shading power loss’. 5th IET Int. Conf. on
[13] Wang, Y.J., Hsu, P.C.: ‘Analytical modelling of partial shading and different Renewable Power Generation, 2016
orientation of photovoltaic modules’, IET Renew. Power Gener., 2010, 4, pp. [39] National Solar Radiation Database (NSRDB). https://nsrdb.nrel.gov/
272–282

148 IET Renew. Power Gener., 2018, Vol. 12 Iss. 2, pp. 143-148
© The Institution of Engineering and Technology 2017