An Assessment Framework for PV Parallel MPPT Configuration with a
New Utilization for UIPS Loads
MOSTAFA EL-SAYED, AHMED HUZAYYIN, ABDELMOMEN MAHGOUB,
ESSAM ABULZAHAB
Electrical Power Engineering,
Faculty of Engineering,
Cairo University,
Cairo University Road - Giza,
EGYPT
Abstract: - The prevalence rate of photovoltaics (PV)-based generation systems has increased by more than 15
folds in the last decade, putting it on the top compared to any other power generation system from the
expandability point of view. A portion of this huge expansion serves to energize standalone remote areas.
Seeking improvements from different aspects of PV systems has been the focus of many studies. In the track of
these improvements, parallel MPPT configuration for PV standalone systems have been introduced in the
literature as an alternative to a series configuration to improve the overall efficiency of standalone PV systems.
However, this efficiency improvement of the parallel MPPT configuration over the series one is not valid for
any standalone application, therefore an assessment procedure is required to determine the most efficient MPPT
configuration for different standalone applications. Therefore, in this study, an assessment procedure of parallel
MPPT is conducted to demonstrate the suitability of utilizing such a configuration compared to series one,
based on load daytime energy contributions. This assessment will help PV system designers to determine which
MPPT configuration should be selected for applications under study. Furthermore, a new utilization of parallel
MPPT configuration is introduced for operating universal input power supply (UIPS) loads to eliminate the
inverter stage, thereby increasing the overall system efficiency and reliability. Finally, a systematic procedure
to size the complete system is introduced and reinforced by a sizing example.
Key-Words: - Batteries, Off-Grid, Parallel MPPT, PV, Standalone and Universal Input Power Supplies.
Received: March 10, 2021. Revised: November 20, 2021. Accepted: December 23, 2021. Published: January 5, 2022.
1 Introduction
Off-grid standalone applications depending on
photovoltaic (PV) modules as the main and sole
power sources are increasing worldwide. This
increase is due to many factors with the most
important ones: the continuous reduction in PV
system component prices, especially PV module
prices, and continuous developments in PV power
converter performance [1].In the literature,
improvements in standalone PV systems have been
addressed from many aspects in order to improve
the overall system performance and open frontier for
new applications [2-6]. Developments on the scale
of optimizing MPPT algorithms has been addressed
as shown in [7-9]. One of the addressed
developments in the literature is the use of a parallel
MPPT configuration instead of a series
configuration [5]. In this configuration, the load is
directly connected to the PV source, and a bi-
directional DC-DC MPPT converter is connected in
parallel with the main PV bus instead of the
typically used series connection. The traditional
series MPPT-based off-grid PV system is shown in
Fig.1, and the parallel one is shown in Fig.2. The
parallel configuration allows for direct flow of
energy from the PV array directly to the load
without passing through the MPPT converter during
the daytime period. While, in series configuration
the converter process all of the energy generated by
the PV array. Therefore, losses associated within the
series configuration during the daytime mode of
operation are eliminated in the parallel converter,
which in turn improves the overall system
efficiency. However, this efficiency improvement in
parallel configuration is only applicable for energy
portion processed to the load during daytime. Since
the amount of load daytime consumed energy varies
from one standalone PV application to another,
therefore, an assessment procedure is required in
order to select the suitable MPPT configuration for
each application.
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2022.17.2
Mostafa El-Sayed, Ahmed Huzayyin,
Abdelmomen Mahgoub, Essam Abulzahab
E-ISSN: 2224-350X
9
Volume 17, 2022
Fig. 1: Typical Off-grid system with Series MPPT
Configuration
This assessment is important as it determines
whether efficiency improvements of the parallel
MPPT configuration compared to the series
configuration could be achieved for a certain load
profile or not.
PV standalone systems were used to feed different
types of loads. These loads were categorized as AC
and DC loads. One of the main types of AC loads
currently used is loads that adopt universal input
power supplies (UIPS) at their input power stage;
these loads will be termed as UIPS loads. UIPS
loads are continuously increasing, ranging from
mobile phones, LED lamps, laptop chargers, battery
chargers, LCD TVs, printers, scanners, network
switches, routers, and desktop computers to big data
servers. Currently, many residential and commercial
non-motor loads are UIPS loads. The most
important unique feature of these power supplies is
their ability to tolerate large input supply voltage
variations that could vary in the range of 90Vac up
to 260Vac. Knowing that UIPS loads are equipped
with an AC/DC rectifier at their input power stage,
the UIPS loads can also be operated from the DC
supply. These features make them good candidates
to be fed by a parallel MPPT configuration because
in this configuration, the DC link voltage could be
designed to match the UIPS operating voltage
window. Feeding UIPS loads directly from
standalone PV systems based on parallel MPPT
configuration facilitates the elimination of front-end
DC/AC inverters typically used with UIPS loads in
typical standalone PV installations. Hence, the
overall system efficiency and reliability can be
improved.
Fig. 2: Typical Off-grid system With Parallel
MPPT. Configuration
In this study, an assessment framework is developed
to assess the feasibility of using a parallel MPPT
configuration over a series configuration based on
the load daytime energy contribution. This
assessment is important to provide a guide for off-
grid PV system designers for selection between
series and parallel MPPT configurations. In
addition, a new utilisation of a parallel MPPT
configuration to feed UIPS loads while eliminating
the DC/AC inverter is proposed. Finally, a
systematic design procedure is proposed for sizing
off-grid PV systems based on a parallel
configuration supplying UIPS loads.
2 Parallel versus Series Standalone
Configuration
Fig.1 shows the traditional off-grid PV
configuration for standalone off-grid PV powered
AC loads with a series MPPT configuration. The
main electrical components of this system consist of
PV modules, series-connected MPPT DC-DC
converter, battery bank, DC/AC inverter, and
interconnecting cables. In this configuration, the
total PV power is always processed through the
series MPPT DC-DC converter. The output of the
MPPT converter is connected to the batteries and
the input side of the DC-AC inverter. Finally, the
output of the inverter is connected to the load. This
configuration is widely used in the implementation
of standalone PV systems [10], [11].
The series MPPT DC-DC converter is responsible
for achieving multiple functions. The first is to
operate the PV string at maximum power. The
second is to charge the battery bank during sun
hours with excess PV power greater than the load
consumed power. The third function is to safely
operate the battery by limiting the battery charging
when a full battery state of charge (SOC) is
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2022.17.2
Mostafa El-Sayed, Ahmed Huzayyin,
Abdelmomen Mahgoub, Essam Abulzahab
E-ISSN: 2224-350X
Volume 17, 2022
achieved and disconnects the loads when the
minimum state of charge is achieved.
Typically, the rated DC voltage of the battery banks
used in this configuration is 12, 24, and 48 V, with a
higher battery voltage selected for large loads to
reduce conduction cable losses [12]. The DC/AC
inverter is used to provide two functions: voltage
stepping up in addition to DC to AC inversion.
Fig.2 shows the parallel MPPT configuration for
standalone off-grid PV-powered loads. In the
parallel configuration, the system components are
quite similar to those used in the series
configuration. However, a bidirectional DC-DC
converter is used in a parallel configuration instead
of a unidirectional one used in a series
configuration. Because the PV array maximum
power point voltage varies over the day, the main
DC bus and load voltage vary accordingly. Hence, it
is important to use a load front-end converter to
adapt the input voltage to feed the load in a parallel
MPPT configuration. The main advantage of
parallel configuration over series MPPT is that it
reduces the amount of losses incorporated inside the
converter by allowing the direct flow of load energy
from the PV array to the load during the daytime
period.
3 Parallel MPPT Modes of Operation
The modes of operation of the parallel MPPT
configuration used for supplying UIPS loads are
summarized in Table1.
During mode 1, the available maximum PV power,
denoted as PPVmax, is greater than the load power,
denoted as PLoad, and the battery state of charge,
denoted as SOC, is less than 100%, the total load
power will be supplied directly from the PV string
power, and the remaining PV power will be handled
by the bi-directional DC–DC converter and stored in
the batteries.
Table 1. Modes of Operation of Parallel MPPT
Configuration
Mode
no.
Modes of operation
1
PPVmax > PLoad, Battery SOC less than 100%.
2
PPVmax > PLoad, Battery SOC equals 100%.
3
PPVmax = PLoad
4
PPVmax < PLoad, Battery SOC greater than
DODmax.
5
PPVmax < PLoad, Battery SOC less than
DODmax.
During mode 2, the available maximum PV power is
greater than the load power, and the battery state of
charge SOC is equal to 100%, the total load power
will be supplied directly from the PV string, and the
bi-directional DC–DC converter will be in idle
mode as the battery cannot be overcharged. In this
mode, the MPPT function is deactivated.
During mode no.3, the available maximum PV
power is equal to the load power; independent of the
battery SOC, the total load power will be supplied
directly from the PV string power and the bi-
directional DC–DC converter will be in idle mode,
but this mode is rare for a long period during the
day.
During mode no.4, the available maximum PV
power is less than the load power, and the battery
SOC is higher than the maximum allowable depth of
discharge, denoted as DODmax. In this mode, a
portion of the load power equal to the available
maximum PV power will be supplied directly from
the PV string power, and the remaining required
load power is supplied by the bidirectional DC–DC
converter and absorbed from the stored energy of
the batteries.
During mode no.5, the available maximum PV
power is less than the load power, and the battery
SOC is lower than DODmax; thus, forcing the
battery to discharge more power beyond its
maximum DOD limit will reduce battery lifetime
and system reliability, and hence either switching
off some loads or the total load will be
disconnected.
4 Assessment of MPPT Configurations
Power profiles of loads supplied by standalone PV
systems can be categorized into four main types:
constant, night operating loads, day time operating
loads, and mixed profile loads, as indicated in Fig.3.
Examples for these load profiles could be PV
powered; street lighting for night operated loads as
in [13], remote medical clinic for day time operated
loads as in [6, 14].
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2022.17.2
Mostafa El-Sayed, Ahmed Huzayyin,
Abdelmomen Mahgoub, Essam Abulzahab
E-ISSN: 2224-350X
Volume 17, 2022
Fig. 3: Different load power profiles categories, (a)
constant load, (b) night load, (c) daytime, (d) mixed
load.
From an operational point of view, in the parallel
MPPT configuration, the efficiency improvement
over series one depends mainly on the amount of
daytime energy consumed by the load directly from
the PV array. This means that variation in the
amount of daytime energy from one application to
another would alter the optimum selection between
parallel and series MPPT configurations. Therefore,
to determine which MPPT configuration is more
effective from the energy efficiency perspective, an
assessment procedure for the efficiency benefits of
the MPPT configuration used should be performed
based on the load power profile. In this section, an
assessment procedure is introduced to evaluate the
parallel MPPT configuration efficiency
improvements in comparison to the series one based
on the load day time energy consumption.
In Fig.4, an arbitrary load and solar power profiles
are shown for an arbitrary day. The load profile is
divided into three portions: the first energy portion
(E1) and third energy portion (E3) (pink shaded
area) lay at night, while the second energy portion
(E2) (yellow shaded area) lay in the daytime.
An energy flow diagram for each energy portion
presented in the arbitrary load profile is shown in
Fig.5 and Fig.6 for parallel and series MPPT
configurations, respectively, where the energy
distribution through each system component is
indicated.
Fig. 4: Arbitrary daily load power profile (curve
with pink and yellow shaded area, daily sun power
profile (black curve).
From Fig.5 and Fig.6, the overall efficiency of
series, denoted as (ηser), and parallel, denoted as
(ηpar), can be calculated using (1) and (2),
respectively.
Fig. 5: Energy flow diagram of parallel MPPT
configuration.
The component efficiencies used in the calculation
of both efficiencies are the MPPT converter
efficiency, denoted as ηDC, DC/AC inverter
efficiency, denoted as (ηinv), and battery efficiency
(ηb).
PV
ARRAY Inverter
BI-Directional
DC-DC
converter
Battery
{ E2/ η(inv) }
{ E1+E3/ [ η(dc) x η(inv)] }
{ E1+E3/[η(dc)^2 x η(Bat) x η(inv) ] }
AC
Loads
{ E2}
{ E1 + E3}
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2022.17.2
Mostafa El-Sayed, Ahmed Huzayyin,
Abdelmomen Mahgoub, Essam Abulzahab
E-ISSN: 2224-350X
Volume 17, 2022
(1)
(2)
It can be seen from (1) and (2) that there are two
main differences between the calculated series and
parallel efficiencies. The first one is that in
calculating ηpar, the daytime energy E2 is divided
only by the inverter efficiency, while the DC-DC
converter efficiency is eliminated as this energy
portion E2 does not pass by the DC-DC converter
while transferring from the PV array to the load.
The second difference, also in ηpar, is that the
nighttime energy, E1 and E3, are divided by the
square of ηDC. Qualitatively speaking, the effect of
both differences is that as the daytime energy
portion E2 increases, ηpar increases compared to ηser.
The graphical interpretation curves of (1) and (2) are
shown in Fig.7, where reasonable converters and
battery efficiencies are assumed. In this curve, the
efficiencies of both parallel and series
configurations are plotted versus the daytime energy
portion E2, where E2 is normalized with respect to
the total daily energy. The Fig. shows that
increasing E2 results in an increase in both the
parallel and series configuration efficiencies;
however, the rate of efficiency enhancement in the
parallel configuration is higher than that in the series
configuration.
The curves also show that beyond a certain
percentage of daytime energy contribution, the
parallel configuration is more efficient than the
series configuration. Hence, using a parallel MPPT
configuration is more efficient than using series one,
provided that the daily load energy exceeded a
certain percentage of the total load daily energy
consumption. This percentage, as seen from the
curve shown in Fig.7, lies between 55% and 60%.
One implication of this result is that the series MPPT
configuration would be the optimum selection from
an energy efficiency perspective for night-only
operated loads.
PV
ARRAY
Uni-Directional
DC-DC
converter
{ E1+E3/ η(dc) x η(Bat) x η(inv)} +
{ E2/ η(dc) x η(inv)}
{ E1+E2+E3/ η(inv)}
{ E1+E2+E3}
AC
Loads Inverter
Battery
Fig. 6: Energy flow diagram of series MPPT
configuration.
Simultaneously, the parallel MPPT configuration
would be the optimum choice for daytime only
operated loads. For loads with constant or mixed
power profiles, the efficiency assessment using (1)
and (2) should be performed using the actual
efficiencies of each converter used to decide which
MPPT configuration is more efficient. The
presumed efficiencies for plotting these curves were
ηDC= 0.95, ηinv = 0.95, ηb = 0.85).
Fig. 7: Parallel and series configuration efficiencies
versus daily loading portion.
Although the DC –DC converter efficiencies vary
from one converter topology to another, in this
analysis , the efficiencies of both unidirectional
converters , used in series, and bidirectional
converters used in parallel, were assumed to be the
same. This was done to neutralize the comparison
between series and parallel configurations from
efficiency differences due to converter topology and
concentrate on the load curve effects on the overall
system efficiency. However, the actual efficiencies
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2022.17.2
Mostafa El-Sayed, Ahmed Huzayyin,
Abdelmomen Mahgoub, Essam Abulzahab
E-ISSN: 2224-350X
Volume 17, 2022
of the selected converter can be considered in the
configuration efficiency assessment framework
presented by applying the actual efficiency value for
each configuration.
5 New Utilization of Parallel MPPT
Configuration
In this section, a new utilization of the parallel
MPPT configuration for supplying UIPS loads is
introduced, as shown in Fig.8. This proposal is
driven by four main factors, the first factor is the
wide spreading of UIPS loads with their superior
feature of accepting wide AC as well as DC supply
voltage variations. The second factor is the
continuous increase in the market share of PV
powered off-grid applications, which is a direct
result of the continuous price reduction in PV
system costs. The third factor is the capability of the
parallel MPPT configuration to enhance the overall
system efficiency compared to the series
configuration for certain applications with high day
time energy contributions. The fourth factor is the
capability to eliminate one of the main system
components encountered in the design of off-grid
PV systems feeding AC loads, which is the AC/DC
inverter, which in turn has a positive impact on the
overall system cost, system efficiency, and
reliability.
Fig. 8: Proposed System Configuration
The advantages of reduced overall system costs and
enhanced system efficiency and reliability would
pave the way for the development of MPPT
bidirectional DC/DC converters dedicated to
standalone PV systems on a commercial scale.
This new utilization imposes a new design
constraint while sizing the PV array. This constraint
is that the variation of the PV string voltage window
during day operating hours should fall within the
UIPS load operating voltage window. This design
constraint ensures the proper operation of the load.
In the next section, this constraint is considered
while sizing a full standalone system feeding UIPS
loads.
6 System Sizing
In this section, the sizing of standalone PV systems
for the proposed parallel MPPT configuration
feeding UIPS loads is introduced. Similar to any
standalone PV system, in the proposed
configuration, the system components to be sized
are as follows:
1) The PV module data includes the module
maximum power, Pmax, number of modules per
string, number of parallel strings, and module
voltage ratings (the maximum power point
voltage, Vmpp , and open-circuit voltage Voc).
2) Battery bank ratings include battery
chemistry type, battery nominal voltage, and
battery Ampere-Hour Capacity (AH).
3) The ratings of the DC/DC MPPT charge
controller.
4) Cable sizing per each stage including cable
cross-sectional area and lengths.
The required design input data to determine and size
the above-mentioned system components are as
follows.
1) Estimated load power profile versus time for
the worst design month. This month has the
lowest daily sun-peak hours for constant load
power profiles. However, with varying load
profiles, this month should be the month with the
highest ratio of daily energy consumption
divided by its associated sun peak hours.
2) Required days of autonomy to account for
cloudy and sunless days, Nauto.
In the proposed configuration, a modified design
methodology is used to size and select the complete
system components listed above. This modified
methodology guarantees matching between the sized
PV string Vmpp window and the allowed UIPS
voltage window. This matching between both
voltage windows is mandatory to allow for proper
load operation.
6.1 PV Array Sizing
In this section, the PV array sizing parameters are
determined, which are the total PV array power,
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2022.17.2
Mostafa El-Sayed, Ahmed Huzayyin,
Abdelmomen Mahgoub, Essam Abulzahab
E-ISSN: 2224-350X
Volume 17, 2022
denoted as Pt, number of series modules per string,
denoted as Ns , and number of parallel strings,
denoted as Np.
The first sizing parameter which is the total PV
array power, Pt, is determined based on three
factors, as shown in (3). The first factor is the
reflected load power, denoted as . where is
calculated by dividing the daily load energy
consumption ( ) by the corresponding daily
sun peak hours (SPH), as shown in (4). The second
factor is the overall efficiency factor , which
includes module temperature losses, soiling losses,
cabling losses, DC-DC converter losses, and battery
cyclic losses. The last factor is the autonomy
oversizing factor ( ), which is used to account for
the extra charging energy required to charge
batteries after battery deep discharging during
cloudy days. In calculating Pt, sizing based on the
maximum value of (i.e. ) ensures
the capability of the PV array to supply load energy
requirements during the worst day.
(3)
(4)
The next step in sizing the PV array is to
determine the number of series modules per string,
denoted as Ns .During the day, the PV string output
voltage, denoted as Vstring, should be within the
allowed voltage window dictated by the UIPS load.
Because Vstring equals the module voltage, denoted
as Vmod, multiplied by Ns, as shown in (5), selecting
Ns will depend on the UIPS voltage window limits
and the output voltage per module.
(5)
The module output voltage Vmod varies depending
on many factors. The first factor is the module
technology used, whether it is crystalline silicon, C-
Si, or thin-film, and the second factor is the number
of cells per module. The third factor is the
environmental operating conditions and, more
specifically, the module temperature. The last factor
affecting Vmod was the electrical loading condition.
The first three factors determine the I/V curve of the
PV module, while the latter determines the point at
which the PV module is operating. It is obvious that
the module output voltage that is considered in the
calculations, while selecting Ns, is the maximum
power point voltage, denoted as , as operating
at this voltage ensures maximising the output power
of the PV array. It should be noted that module
will also vary during the daytime because of
module temperature and irradiance variations. The
benefit of utilizing UIPS loads which have a unique
feature of accepting a wide range of input supply
voltages that are capable of covering voltage
variations during daytime. To fit the voltage
variation into the UIPS voltage window, Ns should
be appropriately selected. The optimum Ns is
bounded by Nsmax and Nsmin, as shown in (6).
Where Nsmax and Nsmin are the maximum and the
minimum allowed number of series-connected
modules per string that ensure fitting of the in
to the UIPS voltage window.
(6)
Formulas for determining Nsmin and Nsmax are
shown in (7) and (8), respectively. A new DC bus
voltage window is defined based on the UIPS input
voltage window to account for Vmpp variation during
daytime and to allow for some voltage safety
margin. The two limits defining this new voltage
window are denoted as VSmin and VSmax, representing
the lower and upper voltage window limits,
respectively.
(7)
(8)
This new window is related to the UIPS input
voltage window by a voltage variation factor,
denoted as Kvf, which accounts for the deviation of
the operating Vmpp from its nominal STC value ,
denoted as Vmpp_STC, as shown in (9) and (10).
Moreover, a 10% margin is considered to leave a
suitable safety margin over the UIPS window limits.
The voltage variation factor Kvf depends on the
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2022.17.2
Mostafa El-Sayed, Ahmed Huzayyin,
Abdelmomen Mahgoub, Essam Abulzahab
E-ISSN: 2224-350X
Volume 17, 2022
range of variation of the module temperature and
voltage temperature coefficient of the module used.
(9)
(10)
From (7) and (8), the allowed range for is
determined.
The next step in sizing the PV array is to
determine the number of parallel PV strings in the
array, denoted as Np. For applications with
relatively high-power requirements, paralleling of
PV strings is needed. To check whether parallel PV
strings are needed, the minimum required power per
PV module, denoted as PMod_min, should be
calculated. An iterative procedure was used to
determine PMod_min using (11).
(11)
For the first iteration, the number of parallel
strings, Np, is set equal to one, and then PMod_min is
calculated. If PV modules with this power do not
exist in the market, paralleling of PV strings will be
required. In this case, Np is incremented by one and
PMod_min is recalculated until PMod_min is achieved,
which is less than or equal to the market available
PV module power. Then, the number of parallel
strings Np is set equal to the last increment iteration.
After setting the value of Np, Pmod will be selected
from market-available modules such that its value is
greater than or equal to PMod_min. Finally, Ns was
calculated using (12).
(12)
6.2 Battery Sizing
The first design parameter to consider in battery
bank sizing is deciding the battery chemistry to be
used. The battery chemistry selected in any
application is based on the battery performance
parameters of interest required by the application.
Generally speaking, the battery’s main performance
parameters are battery specific energy and power
densities, internal discharge rate, memory effect
behavior, number of charge/discharge cycles of the
battery, and maximum allowed depth of discharge,
denoted as . For standalone PV
applications, the last two performance parameters
are the most important parameters to be considered
while selecting the battery chemistry used. Lead
acid batteries are the most adopted battery chemistry
in many off-grid PV installations [15], as they are
considered to be the cheapest rechargeable battery
technology [16]. However, it is expected that
lithium-ion batteries will replace lead-acid batteries
in the near future because of their continuous price
reduction [17] and, of course, their superior
performance over lead acid.
After selecting the battery chemistry, the second
parameter to consider is the battery energy capacity
which, which is determined based on the
maximum daily load energy, denoted as .
Because, in the proposed parallel configuration, the
battery is positioned after the DC to DC converter,
the converter efficiency along with the ,
and days of autonomy, denoted as , should be
considered while determining the battery capacity,
as shown in (13).
(13)
Next to the calculation of the battery bank
capacity, the battery bank voltage is selected. In the
proposed configuration, the converter used should
be bidirectional, in contrast to the unidirectional
converter found in the traditional configuration. The
battery bank voltage should be selected to be
suitable for all operation modes of the selected
bidirectional DC–DC converter. In a series MPPT
configuration, it is common to select a battery bank
voltage of 12, 24 or 48V. The same voltage levels
can be selected in the parallel MPPT configuration
or even higher, depending on the bidirectional DC-
DC converter voltage ratings.
6.3 Converter Sizing
The proposed configuration utilizes a parallel
bidirectional DC-DC converter instead of a series
unidirectional converter used in traditional stand-
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2022.17.2
Mostafa El-Sayed, Ahmed Huzayyin,
Abdelmomen Mahgoub, Essam Abulzahab
E-ISSN: 2224-350X
Volume 17, 2022
alone PV system configurations. Although the
commercial availability of the bidirectional DC–DC
converter is relatively lower than that of the
unidirectional converter, its market share is
estimated to increase in the near future, mainly
because of continuous development in the e-
mobility and micro-grid sectors. Some of the
commercially available bidirectional DC-DC
converters can be found in the referenced datasheets
[18], [19].
The main sizing specifications of the
bidirectional DC-DC converter are the power rating,
low voltage, and high-voltage side voltage window
limits. The converter power rating is determined
based on the maximum of the charging, denoted as
PCh_max, and the discharging power of the converter,
denoted as PLoad_max. The maximum charging and
discharging powers depend on the load power
profile versus time and PV array power. It is
recommended to account for a 10% oversizing
safety factor in calculating the converter maximum
power, as shown in (14).
(14)
The converter low-voltage side was selected as
the battery side, and its voltage window limits were
Vbmax and Vbmin. Where Vbmax is the maximum
attainable battery voltage under fully charged
conditions, and Vbmin is the minimum operating
battery voltage at the minimum allowed state of
charge. These battery voltage limits are determined
from the battery datasheet; however, it is common to
be 15%–20% away from the nominal battery voltage
values. The high-voltage side of the converter is
connected to the main DC bus; hence, its voltage
window should be equal to or wider than that of the
UIPS voltage window.
6.4 Cable Sizing
Cables represent a main component in the off-grid
PV system and share a considerable percentage of
the total system losses. Designers typically consider
that their losses share between 2% and 5% relative
to the system peak power [20], [21]. Transferring
the same amount of power over a higher voltage
level decreases the operating current, thereby
reducing losses and allowing for lower cable cross-
sectional area, csa, usage. Typical off-grid PV
systems operate with standardized DC bus voltages
of 12, 24, or 48 V. The selection of a higher
operating voltage is usually recommended for larger
load powers to reduce operating current which
reduces cable losses, assuming constant current
density for cable sizing. In the proposed
configuration, power is transferred to the load
through a voltage level that lies in the range of 130–
320V. This range is greater than the typical DC bus
voltage used for traditional off-grid PV systems;
hence, the cable transmission efficiency in the
proposed configuration is enhanced. In addition, this
feature can be used to use cables with a lower cross-
sectional area which decreases the overall system
cost.
The proposed configuration comprises three
main cables for which the csa and length should be
determined. The first cable is the main DC bus cable
connecting the PV array to the bidirectional
converter, the second one is the load connection
cable to the main DC bus, and the third one is the
battery connection cable to the bidirectional
converter. The length of the cable is determined
based on the relative physical allocation of the
system components. Component allocation that
allows shorter cable lengths is preferable for
reducing the voltage drop and power losses.
From the cable manufacturer catalogues, the cable
csa is selected based on the calculated maximum
cable current and maximum allowed resistance per
unit length that satisfies the maximum allowed
voltage drop. The calculation of the maximum
current and maximum allowed resistance per unit
length (km) is shown in (15) and (16), respectively.
(15)
(16)
6.5 Case Study
In this section, the sizing of the complete parallel
MPPT off-grid PV system feeding the UIPS load is
determined following the steps illustrated in the
sizing section. The sizing of the PV array, batteries,
DC-DC converter, and cables will be determined.
For a given off-grid area utilizing UIPS loads,
the daily energy consumption for a standalone
application on the worst design day was 12 kWh,
and the corresponding SPH was 4. The days of the
autonomy factor are assumed to be 1.25.
From this value of the batteries will take 12
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2022.17.2
Mostafa El-Sayed, Ahmed Huzayyin,
Abdelmomen Mahgoub, Essam Abulzahab
E-ISSN: 2224-350X
Volume 17, 2022
days to recharge the autonomy battery capacity after
a full 3 days autonomy period. The overall
efficiency was assumed to be 0.75. Therefore,
substituting in (3), the required PV power at STC
equals the 5 kW peak.
The UIPS allowed the DC supply window to
typically range from 130 to 320 VDC. The voltage
variation factor, Kvf, was assumed to be 1.2. By
substituting in (7) and (8), Vsmin and Vsmax are
172 V and 242 V, respectively. The selected number
of PV modules per string, Ns, should produce a
voltage that lies within this range (from 172 to 242
V) under different operating conditions. From the
PV module type selected, Vmpp_STC is determined
and substituted in (11) to obtain Pmod_min while
assuming NP = 1. Owing to its high commercial
availability and hence relatively lower price per
watt, the 60 cell C-Si module will be selected. Sixty
cell PV module has a Vmpp_STC that varies from
30 to 35 V depending on the module design; in
calculations, 33 V will be used. From the
substitution in (11), Pmin equals 682 W. Because Pmin
is higher than the maximum of any available
commercial 60 cell module, parallel strings are
required. By incrementing Np to two, Pmin equals
341 W. This value of Pmin is commercially available
in 60 cell modules range. From commercially
available modules, JAM60s20-365/MR [22] was
selected. This module is rated at 365 W and
Vmpp_STC equals 33.96V. The number of modules
per string, Ns, is then calculated using (12) which
equals seven modules per string. Therefore, the PV
array consists of two parallel strings with seven
series modules per string with a total power of 5110
W under STC conditions.
To size the battery, first, the battery chemistry
will be selected. Owing to its wide market
availability and high number of charge/discharge
cycles, deep-cycle lead acid chemistry will be
selected. Nauto is assumed to be 3 days, is
assumed to be 0.95. DODmax was chosen to be 80%.
Using (13), the required battery capacity was 48
kWh. The operating battery voltage was selected to
be 48 V. Hence, the AH capacity of the battery was
1000 AH.
The next step is to size the bi-directional DC–
DC converter. From the load power profile and size
of the PV array, the maximum power of both is
equal to 6.3 Kw. Therefore, from (3), the power
rating of the converter is 7 kW. The voltage ratings
of the converter battery side are rated the same as
the battery nominal voltage at 48 V. However, the
converter should be capable of operating in a
window of ± 20% around this voltage to allow for
battery voltage variations during charging and
discharging. As stated earlier in the design section,
the voltage ratings of the converter DC side
connected to the main DC bus are rated based on
UIPS voltage window limits which are 130 and
320V DC. Therefore, the Bidirectional DC-DC
converter ratings are as follows: power rating of 7
kW, battery side voltage of 48 V (± 20%), 240V (±
50%).
After sizing the converter, the cable sizes were
determined. Typically, the cable lengths are dictated
by the physical orientation of the system
components relative to each other. This, in turn, is
dependent on the locations of off-grid sites ready for
system installation. In the case under study, only the
main DC bus cable is sized, and the other two
system cables can be sized using the same
procedure. To determine the csa of this cable, first,
the maximum current flowing through the circuit
should be determined as shown in (15).Based on this
calculation, the maximum current of the main DC
bus is approximately 39 A. This current is calculated
based on the UIPS minimum voltage window limit
which is equal to 130V DC at the maximum PV
power. From one of the PV cable manufacturer’s
catalogue [23], 2.5 mm2 or 4 mm2 csa would be
sufficient to handle this amount of current.
However, assuming that the cable length is to 40m
(two ways), this csa would produce a high voltage
drop and hence high power losses which is equal to
10% approximately at the lowest allowed supply
voltage condition. By setting the allowed voltage
drop to 3%, using (16), the maximum acceptable
resistance per km is equals to 2.5Ω /km. This
resistance could be achieved using a 10 mm2 cable.
In this step, the sizing of the off-grid PV system
supplying the UIPS based on parallel MPPT is
performed.
7 Discussion
The system efficiency analysis presented in section
4 showed that the efficiency improvements in
parallel MPPT configuration over series one is not
justified for all applications. It was shown that the
efficiency improvements depends on the amount of
load daytime energy contribution with respect to the
total daily energy consumption. So, for standalone
applications with high daytime energy contributions
it is advised to use the parallel MPPT configuration,
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2022.17.2
Mostafa El-Sayed, Ahmed Huzayyin,
Abdelmomen Mahgoub, Essam Abulzahab
E-ISSN: 2224-350X
Volume 17, 2022
this amount is roughly around 50%. The accurate
percentage could be calculated as shown in section
4.
The nature of the input interfacing power
converter of the UIPS loads , such as LED [24],
allows for two operating features which was well
utilized in this paper to match the operating
behavior of parallel MPPT configuration and
directly connect UIPS load to the main DC bus.
These two features are the wide operating input
voltage window and the capability to be operated
from DC source. These two features allows for
directly suppling the UIPS loads from the DC bus of
the parallel MPPT configuration, provided that well
sizing of PV string is done, as shown in section 6.1.
Suppling UIPS load from the main DC bus directly,
allows the elimination of the DC/AC inverter stage
which is typically used in standalone PV systems
feeding the UIPS loads, which enhances overall
system efficiency and reliability.
8 Conclusion
In this paper, an assessment procedure to assess
efficiency improvements of parallel MPPT
configuration over series one based on the load
power profile is introduced. The assessment
procedure showed that if the daytime load energy
consumption is lower than 50% of the total load
daily energy, the series MPPT configuration will be
preferred over parallel MPPT from an energy
efficiency perspective. Moreover, a new utilization
of the parallel MPPT configuration is introduced to
supply the UIPS load. This utilization eliminates the
DC/AC inverter stage which increases the overall
system efficiency and reliability, however this
elimination is valid only for UIPS loads. Finally, a
detailed design procedure for a parallel MPPT
feeding UIPS load is presented, along with a
detailed case study.
The future work for this research is to implement
hardware setup for a small prototype with UIPS
loads.
References:
[1] G. Kavlak, J. McNerney, and J. E. Trancik,
“Evaluating the causes of cost reduction in
photovoltaic modules,” Energy Policy, vol. 123,
pp. 700–710, 2018, doi:
10.1016/j.enpol.2018.08.015.
[2] L. A. Lamont and L. El Chaar, “Enhancement of a
stand-alone photovoltaic system’s performance:
Reduction of soft and hard shading,” Renew.
Energy, vol. 36, no. 4, pp. 1306–1310, 2011,
doi: 10.1016/j.renene.2010.09.018.
[3] M. Das and V. Agarwal, “Novel High-
Performance Stand-Alone Solar PV System
With High-Gain High-Efficiency DC-DC
Converter Power Stages,” IEEE Trans. Ind.
Appl., vol. 51, no. 6, pp. 4718–4728, 2015, doi:
10.1109/TIA.2015.2454488.
[4] L. A. Lamont and L. El Chaar, “Enhancement of a
stand-alone photovoltaic system’s performance:
Reduction of soft and hard shading,” Renew.
Energy, vol. 36, no. 4, pp. 1306–1310, 2011,
doi: 10.1016/j.renene.2010.09.018.
[5] R. Gules, J. D. P. Pacheco, H. L. Hey, and J.
Imhoff, “A maximum power point tracking
system with parallel connection for PV stand-
alone applications,” IEEE Trans. Ind. Electron.,
vol. 55, no. 7, pp. 2674–2683, 2008, doi:
10.1109/TIE.2008.924033.
[6] Hanaa M. Farghally, Ninet M. Ahmed, Abd El-
Shafy A. Nafeh , Faten H. Fahmy, Amal A.
Hassan, Emaad A. Sweelem, Water from Air
Production System based on Solar Energy in
Egypt, International Journal of Mechanics,
Volume 14, 2020, pp. 215-220.
[7] M. M. Eshak, M. A. Khafagy, P. Makeen and S.
O. Abdellatif, "Optimizing the performance of a
stand-alone PV system under non-uniform
irradiance using Gray-Wolf and hybrid neural
network AI-MPPT algorithms," 2020 2nd Novel
Intelligent and Leading Emerging Sciences
Conference (NILES), 2020, pp. 600-605, doi:
10.1109/NILES50944.2020.9257965.
[8] Windarko, N.A.; Nizar Habibi, M.; Sumantri, B.;
Prasetyono, E.; Efendi, M.Z.; Taufik. A New
MPPT Algorithm for Photovoltaic Power
Generation under Uniform and Partial Shading
Conditions. Energies 2021, 14, 483.
https://doi.org/10.3390/en14020483.
[9] Vipin Kumar, Sandip Ghosh, N.K. Swami Naidu,
Shyam Kamal, R.K. Saket, S.K. Nagar,Load
voltage-based MPPT technique for standalone
PV systems using adaptive step,
International Journal of Electrical Power & Energy
Systems,
Volume 128,2021,106732,ISSN 0142-0615,
[10] H. El-Houari et al., “Design, simulation, and
economic optimization of an off-grid
photovoltaic system for rural electrification,”
Energies, vol. 12, no. 24, 2019, doi:
10.3390/en12244735.
[11] J. Song and Y. Choi, “Design of photovoltaic
systems to power aerators for natural
purification of acid mine drainage,” Renew.
Energy, vol. 83, pp. 759–766, 2015, doi:
10.1016/j.renene.2015.05.014.
[12] N. D. Nordin and H. Abdul Rahman, “A novel
optimization method for designing standalone
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2022.17.2
Mostafa El-Sayed, Ahmed Huzayyin,
Abdelmomen Mahgoub, Essam Abulzahab
E-ISSN: 2224-350X
Volume 17, 2022
photovoltaic system,” Renew. Energy, vol. 89,
pp. 706–715, 2016, doi:
10.1016/j.renene.2015.12.001.
[13] E. I. Archibong, S. Ozuomba and E. Ekott,
"Internet of Things (IoT)-based, Solar Powered
Street Light System with Anti-vandalisation
Mechanism," 2020 International Conference in
Mathematics, Computer Engineering and
Computer Science (ICMCECS), 2020, pp. 1-6,
doi: 10.1109/ICMCECS47690.2020.240867.
[14] Hanaa M. Farghally, Ninet M. Ahmed, Abd El-
Shafy A. Nafeh , Faten H. Fahmy, Amal A.
Hassan, Emaad A. Sweelem, Water from Air
Production System based on Solar Energy in
Egypt, International Journal of Mechanics,
Volume 14, 2020, pp. 215-220.
[15] S. Paul Ayeng’o, T. Schirmer, K. P. Kairies, H.
Axelsen, and D. Uwe Sauer, “Comparison of
off-grid power supply systems using lead-acid
and lithium-ion batteries,” Sol. Energy, vol. 162,
no. December 2017, pp. 140–152, 2018, doi:
10.1016/j.solener.2017.12.049.
[16] Pavlov, D. (2011). In Lead-acid batteries:
science and
technology: a handbook of lead-acid battery
technology and its influence on the product.
Essay, Elsevier Science Ltd.
[17] Y. Zhao et al., “A Review on Battery Market
Trends, Second-Life Reuse, and Recycling,”
Sustain. Chem., vol. 2, no. 1, pp. 167–205,
2021, doi: 10.3390/suschem2010011.
[18] Epic Power bidirectional DC/DC converters
brochure, https://epicpower.es/wp-
content/uploads/2019/05/EPC_ENG_2019_broc
hure_.pdf
[19] TDK-Lambda Corporation, product information,
https://eu.mouser.com/pdfdocs/TDKLambdaAp
plicationNote.PDf.
[20] S. Jogunuri, R. Kumar, and D. Kumar, “Sizing
an off-grid photovoltaic system (A case study),”
2017 Int. Conf. Energy, Commun. Data Anal.
Soft Comput. ICECDS 2017, pp. 2618–2622,
2018, doi: 10.1109/ICECDS.2017.8389927.
[21] W. Ali, H. Farooq, A. U. Rehman, Q. Awais, M.
Jamil, and A. Noman, “Design considerations of
stand-alone solar photovoltaic systems,” 2018
Int. Conf. Comput. Electron. Electr. Eng. ICE
Cube 2018, no. November, pp. 1–6, 2019, doi:
10.1109/ICECUBE.2018.8610970.
[22]https://www.jasolar.com.cn/uploadfile/2020/0619
/20200619041705631.pdf
[23]https://www.smsystems.co.in/POLYCAB-
SOLAR-Cable.pdf
[24] Deepak Agrawal, Rajneesh Kumar Karn,
Deepak Verma, Rakeshwri Agrawal, DC-DC
Converter Topologies for LED Driver Circuit: A
Review, International Journal of Circuits,
Systems and Signal Processing, Volume 14,
2020, pp. 542-547.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.en
_US
WSEAS TRANSACTIONS on POWER SYSTEMS
DOI: 10.37394/232016.2022.17.2
Mostafa El-Sayed, Ahmed Huzayyin,
Abdelmomen Mahgoub, Essam Abulzahab
E-ISSN: 2224-350X
Volume 17, 2022