Solar Energy 183 (2019) 120–131

Contents lists available at ScienceDirect

Solar Energy
journal homepage: www.elsevier.com/locate/solener

A model for direct-coupled PV systems with batteries depending on solar T

radiation, temperature and number of serial connected PV cells
Sarah Paul Ayeng'oa,b, Hendrik Axelsena,b, David Haberschusza,b, Dirk Uwe Sauera,b
a
Electrochemical Energy Conversion and Storage Systems Group, Institute for Power Electronics and Electrical Drives (ISEA), RWTH Aachen University, Jaegerstrasse 17-
19, 52066 Aachen, Germany
b
Institute for Power Generation and Storage Systems (PGS), E.ON ERC, RWTH Aachen University, Germany

ARTICLE INFO ABSTRACT

Keywords: In the designing of solar photovoltaic (PV) systems, which are directly coupled to batteries, it is of great im-
PV cells portance to consider voltage-operating points of PV modules and batteries. Usually, batteries operating points
Solar home system are supposed to be near to the PV maximum power points to maximize the PV output.
Battery PV output voltage is affected by incident radiation, air temperature, shading, module orientation, PV tech-
Optimization
nology; tilt angle and number of serial connected PV cells. This paper presents a PV model, which takes into
account the number of PV cells connected in series, incident solar radiation and temperature as main factors
affecting PV systems that are directly coupled to the batteries. The developed model was then validated by
conducting some laboratory tests.
In this study, an amorphous silicon thin film PV module was used. Electrical load data collected in Dodoma,
Tanzania and lead acid battery with 24 V were also used in the designing and simulation of PV systems.
Optimization and the laboratory test results have revealed how good the optimization tool is. The tool can
select the optimal number of PV cells required for the given system. The selected optimal number of PV cells in
series always matches with the battery operating voltage, which helps in the reduction of losses that usually
occur in direct-coupled systems, hence maximizing PV output.
The results presented in this paper are based on solar home system simulated in Dodoma, Tanzania; however,
the model can be adapted to any other region and any type of PV module by changing input data like solar
radiation, air temperature, longitude, latitude, load profile and STC parameters of the PV module respectively.

1. Introduction yield of the PV system (Elmetennani et al., 2016; Ezinwanne et al.,

2017; Paul Ayeng'o et al., 2018). On the other hand, the use of an MPP-
Solar PV systems have become one of the solutions for electricity Tracker increases the system investment cost making it unaffordable to
supply in urban and off-grid areas. This is due to the availability of solar many low income populations (Petreuş et al., 2013; Ezinwanne et al.,
resource in most of the regions (Petit et al., 2016). Not only that, but it 2017). MPP-Tracker increases PV system complexity and losses.
is also a clean resource which attracts people to use it (Fara and Moreover, during its operation lifetime, it can fail and increases
Craciunescu, 2017). The decrease of the costs of PV panels and batteries maintenance cost.
have made these systems more affordable even for people with low A direct-coupled PV system with a battery can be a cost effective
income especially in off-grid areas (Baurzhan and Jenkins, 2016; Okoye alternative where the MPP-Tracker is not used. In this topology, PV
and Solyalı, 2017; Sani Hassan et al., 2017). Recently, conventional modules are connected in parallel with batteries through a charge
fuels prices have increased very rapidly hence motivating people to use controller. This topology is simple and cheap to install (Qazi, 2016).
other forms of energy resources like solar (Al-Addous et al., 2017). Especially for system, which has to be installed in off grid areas, a less
PV battery systems can be installed in two main topologies, which complex, reliable and high efficient topology is required.
are: a PV system directly coupled to a battery through a charge con- The major drawback of this topology is that, when battery voltage is
troller and a PV system connected to a battery through an MPP-Tracker. operated far from the maximum power point (UMPP ) of the PV module,
A major advantage of a system with MPP-Tracker is that it can deliver the output power of the PV module is lowered. Different loads con-
maximum available PV power at any time, which can lead to higher nected in the system can cause a mismatch between battery and PV

E-mail address: [email protected] (S. Paul Ayeng'o).

https://doi.org/10.1016/j.solener.2019.03.010
Received 22 May 2018; Received in revised form 26 February 2019; Accepted 4 March 2019
0038-092X/ © 2019 International Solar Energy Society. Published by Elsevier Ltd. All rights reserved.
S. Paul Ayeng'o, et al. Solar Energy 183 (2019) 120–131

Nomenclature P-PVdir.coup output PV power in a direct-coupled topology

P-PVMPP PV output power at maximum power point
DC direct current P-PVpeak peak PV power of the module
GA genetic algorithm PV photovoltaic
Isc short circuit current SHS solar home systems
LCOE levelized cost of electricity SOC state of charge
Max maximum STC standard test condition at 25 °C and 1000 W/m2
Min minimum Ubatt battery voltage
MPP maximum power point Ubatt_nom nominal battery voltage
NPV net present value UMPP_STC maximum power point voltage at STC
Pb lead acid UMPP voltage at maximum power point
Pmax maximum power UOC open circuit voltage

module. Not only that but also there will always be a great mismatch of The optimized number of PV cells in series has reduced PV modules
the operating voltages if the difference between UMPP and battery op- cost, increased system efficiency and reliability.
erating voltage is high. This leads to wastage of PV output which could From the proposed module, it can be seen at first that its im-
be utilized (Pearsall, 2017). For these systems, appropriate matching of plementation might be expensive especially at PV cell manufacturer
battery and PV voltages is very important for high efficiencies. level. On the other hand, there is huge market especially in the devel-
Normally, a PV module consists of cells that are connected in series oping countries. The results presented here can be considered in the
to increase output voltage. In order to increase the amount of output future by PV manufacturers, especially when the cost per watt peak of
current, the cells are connected in parallel. The number of PV cells in producing solar modules with different numbers of cells for different
series determines also theUMPP . Since the PV modules are connected in region will be similar.
parallel with batteries, their operating voltages are determined by the In order to have confidence in the optimization results, validation of
batteries. The battery operating voltages and the number of PV cells in the model was done through laboratory tests. A PV simulator, a load
series affect the PV module’s output. simulator and lead acid batteries were used.
There are a number of factors, which also affect PV output when
operated at both topologies. These include intensity of solar radiation,
2. Methodology
air temperature, shading, wind speed and type of the material used to
manufacture PV panels (Green, 1998; Schwingshackl et al., 2013; Bellia
2.1. Modelling of the PV system
et al., 2014; Fialho et al., 2014; Rao et al., 2014; Eke and Demircan,
2015; Berthod et al., 2016; Menoufi et al., 2017; Vasel and Iakovidis,
The model of the direct-coupled PV system consists of PV module,
2017).
battery, charge controller and load. It operates in a way that the output
Large numbers of studies were done in the development of PV
from the PV covers first the load and the excess power is used to charge
models that consider different factors affecting the PV output. For ex-
the battery (Merei et al., 2013). The flow of energy to and from the
ample; Pendem and Mikkili (2018) developed PV model by taking into
battery is controlled by a charge controller (Singh, 2013; Müller et al.,
consideration the effect of shading on the electrical characteristics of
2014).
the PV arrays. The developed model was run in Matlab/Simulink en-
vironment. Different shading conditions like short and narrow, short
and wide, long and narrow and diagonal shading conditions were 2.2. PV model
considered to compare the performance of PV module output.
Qi and Ming (2012) also presented a stand-alone PV model in which The PV model uses solar radiation data, which were measured at
PV output was affected by intensity of irradiation and air temperature. horizontal plane (beam, reflected and diffuse) as input. These data were
Increase of temperature affected negatively the output voltage whereas taken from the SoDa website for the region of Dodoma, Tanzania (SoDa,
increase of irradiation increases the PV output. Schwingshackl et al. 2004). The data were at an interval of 10 min that was required by the
(2013), Vasel and Iakovidis (2017) and Kaldellis et al. (2014) presented developed model.
PV models and impact of wind on the performance of the PV modules. The model calculates global solar radiation at an inclined plane so
In this study, the number of PV cells connected in series is con- as to maximize the output from the PV module as presented by
sidered as the major factor affecting PV output for the direct- coupled Ritzenhoff (1992), Shen et al. (2008) Merei (2016) and Ibrahim et al.
systems. A model for direct-coupled PV system with batteries, which (2013) in the equations below.
takes different numbers of PV cells in series into consideration, was
Ht = Hb + Hd + Hr (1)
developed. Other factors like temperature and solar radiation are also
taken into account. An optimization tool for such a system was also where Hb , Hd and Hr are beam, diffuse and reflected radiation at in-
developed. The developed model was run using Matlab/Simulink. clined plane (W/m2) and Ht is the total solar radiation at inclined plane
The developed tool has a capability of selecting the optimal number (W/m2).
of PV cells in series and parallel. It considers changes of solar radiation, In order to get the values of solar radiation at inclined plane, the
temperature and load profiles in selecting the optimal results. The op- following Eqs. (2)–(4) were used. For the case of beam (direct) solar
timal number of PV cells in series always makes the PV operating vol- radiation at inclined plane, consider the following:
tage to be near to the maximum power points. By doing so, the direct-
coupled PV systems can have less losses during their operation. Hb = Hb,0. cos ( )/cos( z ) (2)
Furthermore, the selected optimal results have low levelized cost of
Hb,0 is the beam (direct) solar radiation measured at horizontal
electricity (LCOE).
plane (W/m2), and ( z ) are the incidence angle of direct radiation on
A range of optimal number of PV cells in series and ratio of nominal
tilted surface and horizontal surface, respectively (°).
battery voltage to maximum power point voltage are proposed. Within
Diffuse solar radiation at an inclined plane was calculated using the
this range, a direct-coupled PV system is operated at high efficiencies.
following equation:

121
S. Paul Ayeng'o, et al. Solar Energy 183 (2019) 120–131

Hd = Hd,0
1
(1 + cos ( )) maximum power point (MPP) results to a drastic decrease of PV per-
2 (3) formance (PV power) and thus causing a reduction in efficiency. In
is the diffuse solar radiation measured at horizontal plane (W/ systems with large coupling ratios, the operating points range mostly in
m2) and is the inclination (tilt) angle (°). area 1 and for small coupling ratios the operating points range mostly in
For the case of reflected solar radiation at an inclined plane, the area 2. Coupling ration is defined as
following equation was used. UMPPSTC
cr =
1 Ubat (9)
Hr = Ht,0 (1 cos ( ))
s
2 (4)
where UMPPSTC is the maximum power point PV module voltage (V) at
is the ground reflectivity (albedo) and Ht,0 is the global solar
s STC, Ubat is the battery operating voltage (V).
radiation at horizontal plane (W/m2). PV output for the direct-coupled PV model calculated in the current
study is a function of solar radiationHt , temperature T , battery voltage
2.3. PV output at maximum power point Ubatt and number of serial connected PV cells NPVserial
(P (Ht , T , Ubatt , NPVserial )) .
After the calculation of the global solar radiation at an inclined The direct-coupled PV model is based on the following approach:
plane from the data measured at horizontal plane, this value was then
used to calculate the PV current and voltage at maximum power point 1. For standard test conditions, a characteristic of the solar cell over
(IMPP and UMPP ). The equations presented by Merei (2016), Sauer the entire voltage range is known.
(1994), Hansen et al. (2001) were used and are presented as follows. 2. Maximum power point current, voltage and power at each operating
For the current at maximum power point: point are available.
3. By suitable mapping functions, any operating conditions
( 1 Ht ) 2
IMPP = (Ht , T , Ubatt , NPVserial) of the PV module can be mapped to a char-
( + ( 1 Ht ) (5)
acteristic under standard test conditions.
whereby 1 and are PV module constants. Details of the calculations of
these constants are attached in Appendix A. In this study, three mapping functions were used. Consider the I-V
The voltage at maximum power point depends strongly on the PV curves presented in Figs. 1 and 2. For the drawn graphs, there is a
module temperature and the solar radiation. The PV module tempera- difference between the graph at STC and other operating conditions in
ture was calculated using the following equation. voltage and current. In order to estimate voltages at any operating
point, a mapping factor corresponding to a ratio of maximum power
TPV = Tamb + (Ht TPVcoeff ) (6)
point voltage under standard test condition and maximum power point
U
whereby TPV is the PV module temperature (°C), Tamb is the ambient voltages at any operating points ( MPPSTC ) was used. Consider Eq. (11).
UMPP
temperature (°C), Ht is the global solar radiation at inclined plane (W/ In order to estimate power at any operating point, a mapping factor
m2) and TPVcoeff is the coefficient temperature of the PV module set corresponding to power at maximum power point at any operating
between 0.02 °C/W/m2 for completely free standing panels with good condition and power at maximum power point at standard test condi-
heat dissipation and 0.04 °C/W/m2 for poor cooling, directly mounted tion ( was used (refer Eq. (20)).
to the roof. In the direct-coupled PV system with a battery, usually the PV
Now, modules are connected in parallel with a battery. That means PV op-
a2ln (Ht + 1) + a3 erating voltages depends strongly on battery operating voltages. In
UMPP = a1ln2 (Ht + 1) + + Ct (TPV Tref ) order to map battery operating voltages to a new I-V curve of the new
ln (Ht + 1)+a4 (7)
module so as its relative position on the new I-V curve to be the same as
where UMPP is the PV voltage at maximum power point (V), Tref is the at the original curve, a mapping factor corresponding to number of
reference temperature (25 °C), Ct is the temperature coefficient de- serial connected PV cells of the original module and number of the
scribing the voltage decrease with the increase in temperature, a1, a2 , a3 serial connected PV cells of the new module ( N
NPVorign
) was implemented
and a 4 are constants and the details of which are in Appendix A. For PVserial
(consider Eq. (12)).
(Ht + 1) , 1 is in W/m2.
Consider the following I-V-P curve (Fig. 2) of the original modelled
The theoretical power of a single PV module operated at maximum
PV module and new optimal module with 72 and 45 cells connected in
power point is then calculated as
series respectively. It can be shown that, battery-operating window for
PMPP = UMPP IMPP (8) the new module (45 cells in series) and original module is similar.
The only difference, which can be seen from these graphs, is that,
for the new module, battery-operating window is near to the maximum
2.4. PV output at direct-coupled system with batteries
power point whereas with the original module, battery-operating
window is far from the maximum power point. Making the new module
In order to understand the methodology used in calculating the
(with 45 cells in series) to operate in a more efficient way compared to
output of PV system in a direct-coupled topology, consider Fig. 1 as a
the original one.
sample of an I-V-P curve.
Usually, air temperatures and solar radiation affect PV systems. Low
temperatures and high solar radiations favour high PV output while
high temperatures and low solar radiations favour low PV output.
Therefore, in the development of direct-coupled PV model, the changes
in temperature and solar radiation at each time-step were taken into
consideration.
Two operating regions are presented in Fig. 1: linear (area 1) and
polynomial (area 2). Area 1 shows a nearly constant current, a voltage
increase within this range directly leads to an increase in power (in-
crease in efficiency also).
For the case of area 2, an increase in the voltage exceeding the Fig. 1. Sample I-V-P curve.

122
S. Paul Ayeng'o, et al. Solar Energy 183 (2019) 120–131

When the PV module is operated in this area, operating voltages are

always less or equal to UAN (U UAN ) . This region occurs especially
when PV temperature is low and solar radiation is high. UAN is the
voltage at which the transition from linear to polynomial takes place.
Sauer (1994) presented that, UAN precisely occurs when
UAN = 0.75UMPP . In this region, the current is almost constant and its
values correspond to the short circuit current. From Eq. (11),
P (Ubatt , Nserial ) = ISC Ucorr forU UAN (13)
Scenario 2: Polynomial area (area 2)
When the PV module is operated at area 2, operating voltages are
always high or equal to UAN (U UAN ) . During this time, PV temperature
is high and solar radiation is low. For this case, fourth degree poly-
nomial calculations were selected for the calculation of the PV output
power (Sauer, 1994). For the parameterization of the polynomial,
Fig. 2. I-V-P curve for original PV module (72 cells) and optimized PV module
UOCV ,UMPP ,UAN , ISC and PMPP must be available. The following conditions
(45 cells) with battery operating window.
are paramount.

Since PV operating points at standard test condition were known

(always taken from the manufacturer’s data sheet), any other PV op-
• At the point ofU AN , the function should be continuous.

erating points were estimated by mapping the PV operating points to

P (UAN ) = ISC UAN (14)
the standard test condition points using the mapping functions which
are stated above. The new operating points follow the shape of the • In addition to above point, the function at the point U AN should be
operating points at standard test condition (see Fig. 2). continuously differentiable
In order to calculate PV output power for the direct-coupled PV dP
system with the battery, it is important to consider two operation areas |UAN = ISC
dU (15)
as presented in Fig. 1. The output power can be either in area 1 or in
area 2.
The model presented by Sauer (1994) was adopted. However, in this
• The polynomial should pass through point (U MPP , PMPP )

research there is a provision of changing the PV characteristic; by P (UMPP ) = PMPP (16)

changing the number of serial connected PV cells.
The following equations were used in modelling the direct-coupled • The maximum power point should also be a local maximum of the
PV system with the battery. Consider the condition when PV operating polynomial
points were mapped to the standard test condition. By using voltage- dP
UMPP |UMPP = 0
mapping ratio ( U STC ) , any PV operating voltages were mapped to the dU (17)
MPP
standard test condition voltage. Eqs. (10) and (11) were used.
UMPPSTC • The polynomial should satisfy the idle condition at the point (U OCV ,
Ucorr = Ubatt as presented also by Sauer(Sauer, 1994) 0).
UMPP
(10) P (UOCV ) = 0 (18)
where Ucorr is the corrected or mapped PV voltage (V), UMPPSTC is the Therefore in scenario 2, the output power was calculated as follows:
maximum power point voltage (V) at standard test condition (25 °C
with 1000 W/m2; this value is always taken from the manufacturer’s P (Ubatt , Nserial ) = b1 + b2 Ucorr + b3 Ucorr 2 + b4 Ucorr 3
data sheet), UMPP is the maximum power point voltages (V) at any + b5 Ucorr 4 forU UAN (19)
U
operating points, Ubatt is the battery voltage (V) and MPPSTC is the
UMPP whereby P(Ubatt , Nserial) is the PV output power mapped at standard test
mapping factor. condition depending on battery voltage and number of PV cells con-
The developed optimisation tool can select the optimal number of nected in series (W) and b1, b2 , b3 , b4 and b5 are constants obtained by
PV cells in series different from the original parametrized PV module. solving Eq. (19) using conditions presented in Eqs. (14)–(18). The de-
The optimal number of serial connected PV module always makes the tails of b1, b2 , b3 , b4 and b5 are as shown in Appendix A.
PV system to operate near to the maximum power point. After mapping different operating conditions to the standard test
In order to map battery-operating voltages to any new PV module so conditions by considering voltage and number of PV cells connected in
as its relative position remain the same as that from original modelled series, mapping any PV output power to the standard test condition
PV module, Eq. (12) was used. power, the power factor ( was used.
UMPPSTC PMPP
Ucorr = U P = P (Ubatt , Nserial )
UMPP (11) PMPPSTC (20)
NPVorign I = P /U (21)
U = (Ubatt )
NPVserial (12)
where P is the direct-coupled PV output power at any operating point
whereby NPVorign and NPVserial are the numbers of serial connected PV cells (W), PMPP is the power at maximum power point at each operation
of the original PV module and of the new PV module respectively. condition (W), PMPPSTC is the maximum power at standard test condition
Output power depending on the mapped voltage and number of (W) and I is the output PV current at each operation condition mapped
serial connected PV cells was calculated by considering the following at standard test conditions.
two scenarios. By referring to Fig. 1, output power can either be cal- Therefore, from the above equations, the PV output power and
culated from scenario 1 or 2. current for direct-coupled PV module depending on temperature, solar
Scenario 1: Linear area (area 1) radiation, numbers of serial connected PV cells were calculated. In

123
S. Paul Ayeng'o, et al. Solar Energy 183 (2019) 120–131

order to increase the output power, several of these modules should be Table 1
connected in parallel (NPVparallel ). The following Eq. (22) was used. Electrical data sheet and theoretical values for amorphous silicon module.

Ptotal = P NPVparallel Electrical data (at STC) Value Theoretical (at STC) Value
(22)
Max. power [Pmax ] 235.00 W Max. power [Pmax ] 23.50 W
In brief, the structure of the PV model is described in Fig. 3. The Max. power voltage [UMPP ] 43.50 V Max. power voltage [UMPP ] 43.50 V
output voltage of the PV panel, which depends on the battery operating Max. power current [IMPP ] 5.43 A Max. power current [IMPP ] 0.54 A
voltages for a direct-coupled system and the current, were calculated by Open circuit voltage [Uoc ] 52.40 V Open circuit voltage [Uoc ] 52.40 V
using the Eqs. (1)–(22) presented above. Short circuit current [Isc ] 5.78 A Short circuit current [Isc ] 0.58 A

2.5. Used PV module data sheet t

SOC (t ) = SOC0 + (Ieff (t )/ Cact (t ) dt 23
0
An amorphous silicon PV module with 72 cells in series was ori- SOC (t ) is the actual state of charge in percentage, SOC0 is the initial
ginally parametrized. The parameters taken from the data sheet state of charge in percentage, Ieff (t ) is the effective battery current that
(Panasonic, 2012) are shown in Table 1. According to the presented is used to charge or discharge the battery and Cact (t ) is the actual bat-
data sheet, a single PV cell has a voltage of about 0.6 V. tery capacity in Ah.
For increasing accuracy in the optimization process, a smaller the- The adopted model included different chemical processes, which are
oretical PV cell was used by reducing the PV cell’s size to one tenth of taking place within the battery such as corrosion, sulfation and acid
the original. Changing the number of PV cells of original modelled PV stratification for the estimation of battery operation and lifetime.
module during optimization process caused large incremental steps in
PV output power. 2.7. Load profiles
For a good optimization process, smaller incremental steps were
applied when either increasing or decreasing the number of cells. In For the case of load profile data, load profile measured in one of the
order to achieve this, original modelled PV module was reduced to one households in Tanzania (rafikipower, 2017) was used. The load for the
tenth as it can be seen in Table 1. The reduction in size only affects the SHS amounts to 119 kWh per year, corresponding to an average of daily
output current and hence the output PV power, but not the cell voltage. consumption of just 0.3 kWh. The most commonly used electrical
However from the final optimal results, the values ofUoc , Isc , UMPP and equipment are lights, radios and phone chargers.
IMPP of a new optimal PV module at standard test condition can be es- The profile had a very similar characteristic for most days, with no
timated. major differences between weekdays and weekends. An example of a
one week load profile for SHS is shown in Fig. 4.
2.6. Battery model
2.8. Solar radiation data and temperature data
The battery is a key component in a direct-coupled PV system.
During periods when there is no electricity generated by the PV panels The daily global solar radiation in Dodoma for the year 2005 is
such as during the nights or cloudy days, the batteries can be used to shown in Fig. 5. The graph shows that there is a high amount of solar
cover the load (Yahiaoui et al., 2016; Makibar et al., 2017; Sani Hassan radiation throughout the year with positive effects on photovoltaic
et al., 2017). For the case of lead acid batteries, the model presented by electricity generation.
Schiffer et al. (2007), Merei et al. (2013) was adopted respectively. On average, Dodoma has a mean air temperature of 22.8 °C (see
The battery’s state of charge was calculated by using Coulomb- Fig. 6). Sometimes the temperature rises above 30 °C, negatively af-
counting equation as presented in Eq. (23) below. fecting the yield of solar panels due to decreased efficiency.

Horizontal Longitude,
Ambient
measured solar latitude,
temperature
radiation data inclination angle

Calculate temperature Calculate solar radiation

of the PV cell at inclined plane

Battery voltage,
Calculate voltage, current and power at MPP number of serial
connected PV cells

Calculate voltage, current and power for the direct coupled PV system

Fig. 3. Structure of the PV model for direct-coupled PV systems.

124
S. Paul Ayeng'o, et al. Solar Energy 183 (2019) 120–131

Table 2
Used economic parameters.
Component Price

PV generator 1250 €/kWp

Charge controller 80 €
Lead-acid battery 140 €/kWh
Interest rate (with inflation) 6.67%

Fig. 4. Load profile for SHS. equipment costs and the real energy supplied to cover the load.

2.10. Economic and optimization parameters

For optimization, several parameters within a certain range were

defined and optimized by using GA. In the Table 2 below, important
cost assumptions are presented. For the cost of equipment, the prices of
the year 2017 were used although the costs of equipment may change at
any time depending on different manufacturers. The interest rate for
Tanzania including inflation rate was used.
Levelized cost of electricity was calculated as a fitness function in
each of the iterations. To have an assurance of 100% load coverage for
selected parameters, a penalty value was added to the fitness function
Fig. 5. Daily global solar radiation in Dodoma in 2005 (Paul Ayeng'o et al.,
(LCOE) if the load coverage was below 100%, thus preventing the tool
2018).
from selecting higher fitness values. Consider Eq. (28).

Load covered [%]

Penalty = 1000 / kWh. + 1000 / kWh
100 (28)

Table 3 presents different optimization parameters. The upper and

lower boundaries were set prior to the optimization. Since Dodoma is
located in the southern hemisphere, an Azimuth angle of 180° means
that a panel faces exactly north. At an angle other than 180°, it faces
either eastwards or westwards.

2.11. Laboratory test

The simulated model was validated by performing laboratory test.

Fig. 6. Daily air temperature data for Dodoma for the year 2005 (Paul Ayeng'o The following equipment was used: PV simulator, load simulator, lead
et al., 2018).
acid battery and charge controller.
In the laboratory test, Elgar PV simulator was used. TerraSAS soft-
2.9. Levelized cost of electricity ware was installed in the desktop to control the simulator. By using this
software, different PV panels were modelled based on the optimal re-
For the calculation of the levelized cost of electricity (LCOE), the sults. The parameters which were changed included Uoc , Isc , UMPP , IMPP ,
following Eqs. (24)–(27) were used. β (voltage temperature coefficient) and γ (power temperature coeffi-
n cient) according to different types of PV modules used. The used data
NPVsum = NPVtotal + NPVOC (24)
0 sheet can be found in the Table 6. The software furthermore uses global
solar radiation and PV temperature profiles as input to simulate real-
where NPVtotal is the net present value of the total investment, NPVOC is
time profiles.
the net present value of the total operation costs and n is the number of
The load simulator from Höcherl and Hackl (2018) was used. It
the years.
enables the electronic loads to simulate load profiles saved in CSV files.
The annuity factor was calculated as
The validation took one day for each test. Each test was performed by
i (1 + 1)tref taking the following steps.
Afactor =
((1 + i)tref ) 1 (25)
Table 3
The annuity was then calculated by multiplying the annuity factor Optimization parameters.
to the NPVsum as follows
Parameter Unit Increment Lower Upper
A = Afactor NPVsum (26) boundary boundary

In order to get the LCOE, the annuity A was divided by the real load Inclination angle ° 1 0 90
that was covered by the PV system Ereal_supplied . This means, the excess Azimuth angle ° 1 90 (East) 270 (West)
SOC max % 1 0 100
generated energy from PV was not considered. SOC min % 1 0 100
A Battery capacity Wh 1 5 5000
LCOE = Number of PV cells in – 1 1 200
Ereal _supplied (27) serial
Number of PV modules in – 1 1 500
The LCOE was considered as a fitness function during optimization parallel
process. Low LCOE was preferred but this depended strongly on the

125
S. Paul Ayeng'o, et al. Solar Energy 183 (2019) 120–131

Fig. 7. I-V-P Curve at constant solar radiation and different air temperatures.

Fig. 9. I-V-P Curve at constant solar radiation and different air temperatures
1. Prepare input data: global solar radiation, PV temperature and load
with battery operating window.
that are saved as CSV data and PV data sheets.
2. Prepare and connect the required battery and charge controller with
the PV and load simulators. In this study, lead acid battery with and low solar radiation decreases the PV output current and voltage.
70 Ah was used due to their availability in the laboratory. However, the output voltage can change if the air temperature is not
3. Change the PV characteristic according to the given PV data sheet. maintained constant. High solar radiations have high temperatures,
4. Load the global solar radiation, PV temperature, load data and save which affects negatively the output voltage.
them in the required folders.
5. Switch on the load simulator. 3.2. Solar home system with lead acid battery
6. Set the execution time. Make sure the set execution time and in-
tervals of the measured data are the same. For the simulated system, the number of battery cells in serial was
7. Start the test by allowing the PV simulator to generate power that is kept constant to maintain the 24 V DC system. For the case of lead acid
used first by the load and the excess is saved in the battery. batteries, 12 cells in series were kept constant. The nominal battery cell
voltage used for simulation was 2 V.
I-V curve with the battery-operating window is described in Fig. 9.
3. Results and discussion By using the original PV module with 72 cells and a 24 V battery, it can
be shown that the battery-operating window is in the low power region
3.1. PV module characteristics leading to unnecessary losses.
This PV module was used in order to test how good the optimization
In this part, the PV characteristics are described after simulating the tool was. Any type of PV module can be parameterized and the tool
PV model at different conditions. The PV model was simulated at dif- selects the optimal number of serial connected PV cells. For the purpose
ferent air temperatures (0 °C, 25 °C and 60 °C) maintaining a constant of this study, a PV module with a higher number of PV cells connected
solar radiation of 1000 W/m2 and different solar radiations at constant in series was selected and the optimization was done.
temperature. If this PV module was used in a 24 V battery system, many losses
Consider a PV module simulated at constant solar radiation of would occur in the system. Normally, the battery operating voltages
1000 W/m2 and different air temperatures (0 °C, 25 °C and 60 °C). The I- affect PV operating voltages hence making the PV modules to be op-
V-P curve below (Fig. 7) presents original solar module. erated at low power region. A correct match of the number of PV cells
PV output was higher at lower temperatures than at higher tem- in serial and the battery operating voltages should be found to
peratures (refer Fig. 7). Higher temperatures negatively affect the
output of the PV panel (Ye et al., 2013; Spertino et al., 2016; Peng et al., Table 4
2017). With increasing solar radiation, air temperature also increases. Optimization results.
This leads to increasing temperatures of PV modules, which affect PV
Parameter Unit Value
operations.
Now consider a PV module operated at different solar radiation and Inclination angle ° 15
constant air temperature of 25 °C (refer Fig. 8). The simulation results Azimuth angle ° 187
SOC max % 100
have shown that high solar radiation increases the current and voltage
SOC min % 45
Battery capacity Wh 744
Number of PV cells in serial – 45
Number of PV modules in parallel – 7
Peak Power for theoretical PV module Wp 14.68
Peak Power for a PV system Wp 102.76

Table 5
Electrical data sheet of the optimized PV module.
Electrical data (at STC) Value

Max. power [Pmax ] 14.68 W

Max. power voltage [UMPP ] 27.03 V
Max. power current [IMPP ] 0.54 A
Open circuit voltage [Uoc ] 32.75 V
Short circuit current [Isc ] 0.58 A
Fig. 8. I-V-P curve at constant air temperature and different solar radiation.

126
S. Paul Ayeng'o, et al. Solar Energy 183 (2019) 120–131

(a) (b)
2
Fig. 10. I-V-P curve at constant 1000 W/m solar radiation (a) and at constant air temperature 25 °C (b) for theoretical PV module.

Fig. 11. Comparison between PV modules operated at MPP and at a direct-

coupled topology. Fig. 14. Losses corresponding to the number of serial connected PV cells.

Fig. 12. Comparison of the PV power output with battery output and load.
Fig. 15. Cost comparison of a system with and without MPP-Tracker.

Table 6
New data sheet used for optimal amorphous silicon PV modules.
No. No. cells in serial UMPP IMPP Uoc Isc

1 72 (original) 43.40 1.62 52.80 1.73

2 45 27.13 1.62 32.75 1.73
3 47 28.33 1.62 34.21 1.73

45 cells in series and a peak power of 14.68 Wp (see Tables 4 and 5). In
order to cover 100% of the load and to enable fast charging of the
battery, 7 of these theoretical modules were connected in parallel.
Fig. 13. Output voltage comparison between direct-coupled system and a
Characteristics of the theoretical PV modules are shown in
system operated at MPP using Pb acid battery.
Fig. 10(a) and (b). The graphs show how different temperatures and
solar radiation affect the PV module output.
maximize the PV output. A comparison was done between a PV system with lead acid battery
The optimization results are presented in Table 4. Table 5 presents operated at its maximum power points and the other one operated at a
the STC of the new optimized PV module. The system was optimized for direct-coupled topology. Consider Fig. 11, which shows, that there are
an operational life of 20 years. The optimal theoretical PV module has some operating points, which are close to the diagonal line meaning

127
S. Paul Ayeng'o, et al. Solar Energy 183 (2019) 120–131

Fig. 16. Power output at directly coupled system and at MPP with 72 cells.

that at these moments the system is operated close to the maximum See also Eq. (27).
power point. Optimization results have shown that a SHS with a lead acid battery
However on the other hand there are also some points which are far has LCOE of 0.37 €/kWh. Low lifetime of lead acid battery (4 years)
from the diagonal line meaning that at those moments, the system was increased investment cost by increasing battery replacement costs.
operated at low power region (far from the maximum power points).
The decrease of the PV output power occurs especially in the 3.3. Comparison with a system using MPP-Tracker
afternoons (see Figs. 12 and 13) when the battery is fully charged. At
these times, PV operating voltages (battery voltages) were higher than Economic comparison has been done between a SHS, which uses an
theUMPP , hence making a system to operate at a lower power level. MPP-Tracker and the other one that is directly coupled to a battery. For
The UMPP of the PV module was effectively lowered during the such a system, the cost of MPP-Tracker should be below 12.5% of the
afternoon due to higher ambient temperatures and thus the mis- total investment cost for it to compete with the direct-coupled system
matching between PV operating voltage and UMPP of the module was (see Fig. 15). This means, the cost of MPP-Tracker should be very cheap
increased. (below 80€) for it to be cost effective. Since the losses in this system can
In Fig. 13, the battery voltage against the theoretical maximum be minimized, the use of MPP-Tracker is not relevant.
power point voltages is presented. It can be seen from the graph that At the moment the cost of MPP-Tracker for small PV systems (up to
when the solar radiation is zero (UMPP=0) , the battery was at a certain 120 Wp) is above 100 Euro depending on different manufactures.
state of charge which was enough to cover the load at night. The op- Direct-coupled PV systems with a battery can be used, however a good
erating points below the diagonal line present that UMPP is greater than combination of battery operating voltage and number of PV cells in
the battery voltages. This mainly occurs in the morning. At this point, serial (UMPP) should be chosen in order to reduce the losses in these
the battery charging only if when there is excess from the PV module. systems.
The points above the diagonal line occur when the battery voltages
are higher than theUMPP . This condition occurs mainly in the afternoon.
4. Validation of the PV model
In order to avoid such a mismatch, another possible alternative
would be avoiding charging the battery up to 100% SOC so as to keep
The validation was done for the optimal number of serial connected
the battery operating points below theUMPP . However, since the con-
PV cells. For the tests, which were done, 72, 45 and 47 cells in serial
dition is set for the load to be covered 100%, a higher battery capacity
and 3 modules in parallel were used, with 72 cells being the original
would be leading to higher investment costs for the whole system.
parametrized modules and 45 and 47 being the optimal results. A 70 Ah
Additionally, a low average SOC results in fast calendrical aging for
lead acid battery was connected in parallel to a PV simulator. This
lead acid batteries (Lujano-Rojas et al., 2016; Kujundžić et al., 2017).
battery was used due to its availability in the laboratory.
The observed decrease of PV generation due to fully charged bat-
The used solar radiation data were from Dodoma. For the first tests,
teries is not a permanent situation. It can change if different factors like
a day with high solar radiation was used. Later, solar radiation and
number of PV cells or load profile are altered. In this study, losses have
temperature were changed. Table 6 shows the STC parameters which
been defined as in Eq. (29).
were used as input to the PV simulator. (Table 6 is supposed to be after
Empp Edircoup this paragraph.)
100% Starting with the PV module with 72 cells in serial, Figs. 16 and 17
Empp (29)
show the measurement results. This is actually the original amorphous
where Edircoup is the yearly energy generated from the direct-coupled silicon PV module, which can be found on the market. Fig. 16 shows
system and Empp is the yearly energy generated when the PV system is that most of the points are far from the diagonal line, meaning that the
operated at its maximum power points. PV operating points are far from the maximum power points. This is
From Fig. 14, it can be shown clearly that, when a direct-coupled PV because the used PV module had more PV cells in serial and hence high
system is operated at a range of 45–55 number of serial connected PV values of UMPP points compared to the real PV operating voltages.
cells, fewer losses can be encountered. In general, for selection of high It can also be seen that there are some points that are below the
efficient direct-coupled system, a range of a ratio of nominal battery diagonal line. This was due to an equipment error: Between 0 A and
voltage to PV maximum power point voltage at STC is considered. From 0.5 A, the PV simulator was not operating well, only either at 0 A or
0.67 to 0.9, a direct-coupled PV system can operate at high efficiencies above 0.5 A. In calculation of the losses incorporated in a system, the
greater or equal to 90%. error was not considered.
The levelized cost of electricity was calculated by using the eco- A comparison between the battery operating voltages (PV output
nomic parameters listed in Table 2. The model was simulated at a voltage at direct-coupled topology) with the theoreticalUMPP is pre-
period of 20 years. The fitness function (LCOE) was calculated by di- sented in Fig. 17. It is worth mentioning that the tests were started
viding the annuity with the covered load (Magnor and Sauer, 2016). when the solar radiation was zero and the battery was fully charged.

128
S. Paul Ayeng'o, et al. Solar Energy 183 (2019) 120–131

Fig. 19. Battery voltage and UMPP with 45 PV cells.

Fig. 17. Battery voltage and UMPP with 72 PV cells.

Table 7
Efficiency with different serial connected PV cells for amorphous silicon
module.
Item Number of cells Efficiency (%)

Sunny day 45 97.00

47 95.38
72 64.50
Cloudy 45 95.35

Fig. 18. Power output at directly coupled system and at MPP with 45 cells. amorphous silicon PV module have shown that the optimal number of
PV cells in series is between 45 and 55.
In Fig. 17, different scenarios can be shown. When UMPP was zero, An efficiency of 98% (loss 2%) can be reached if there is a good
the battery was at a certain state of charge. This was at night when match between load, battery and PV maximum voltage. The results
there was no solar radiation, making UMPP to be zero. During this time, presented here are specific for the simulated PV system operated in
the battery was used to cover the load making its voltage to decrease regions with high solar radiation like Dodoma, Tanzania. However, the
from full charge voltage of 28 V to around 24.8 V. model can be easily adapted and used to other regions and types of PV
These battery voltages also correspond to the PV operating voltages modules by changing the input data including solar radiation, air
at direct-coupled topology. As soon as there was solar radiation, UMPP temperature, latitude, longitude, load profile and STC parameters of the
increased and the PV operating voltage (battery voltage) also increased used PV module.
thus charging the battery by solar radiation. In general, for maximum output of the direct-coupled PV system
With the non-optimized system, the operating points were far from with battery, a range of a ratio of nominal battery voltage to PV max-
the ideal line. This module was not suitable for a 24 V system; it was imum power point voltage at STC between 0.67 and 0.9 is proposed.
therefore only used to prove the optimization tool’s quality. Any type of Within this range, a direct-coupled PV system can be operated at higher
PV module can be used and the tool can select the best optimal serial efficiency greater than 90%.
number of PV cells for the used system. An optimized PV module is also proposed. The implementation of
For the selected optimal results of the system with 45 PV cells in the proposed module might be expensive at the beginning from man-
series, the results are presented in the following graphs. In Fig. 18, it ufacturer side by changing the PV specification in the production line,
can be clearly seen that many points are very near to the diagonal line. however there is still a promising and huge market for the PV modules
This shows that the PV operating voltage of a direct-coupled system is especially in the developing countries.
very near to the theoretical maximum power point voltages (UMPP ) (see For the presented optimal system, the LCOE was 0.37 €/kWh when
also Fig. 19). The points below the diagonal line in Fig. 18 are due to the PV modules were operated at an inclination angle of 15° and azi-
the PV simulator error. Apart from this error, the system operated at a muth angle of 187° west. Lower LCOE could be achieved if the battery
high efficiency. lifetime would be long, interest rate and cost of PV system (battery, PV
Table 7 below shows the different levels of efficiency reached in the modules and charge controller) would be low. Since a high efficiency
tests that were conducted when the solar radiation was high and low. can be achieved by the direct-coupled system, the use of MPP-Trackers
High efficiencies were reached when the correct number of serially in SHS operated in Tanzania is not relevant. The cost of an MPP-Tracker
connected PV cells was used. would have to be very low for it to compete with a direct-coupled PV
system.
5. Conclusion and recommendation The results from laboratory tests have shown that the losses in the
direct-coupled system can be minimized if an appropriate number of
In this study, a direct-coupled PV model that takes into considera- serial connected PV cells is chosen, hence making it to operate at high
tion solar radiation, temperature and number of serial connected PV efficiency. A high efficiency of 97% was reached for the PV system with
cells is presented. The presented results for a solar home system with an amorphous PV modules coupled to a lead acid battery.

Appendix A. Parameters used in the PV model

The methodology presented here is as presented by Merei (2016), Sauer (1994). In the developed PV model, four sampling data points are
required for solar radiation (H1, H2 , H3 and H 4 ), maximum current (I1, I2 , I3 and I 4 ) and maximum voltage (U1, U2, U3 and U4 ) as input to the model.
These sample data points can be extracted from the data sheet. These can be the values measured for example at 25 °C for 0 W/m2, 200 W/m2,
400 W/m2 and 1000 W/m2 respectively.
Now consider Eq. (5), where 1 and were used. These constants are calculated using the following equations.

129
S. Paul Ayeng'o, et al. Solar Energy 183 (2019) 120–131

I1 I2 (H1 H2)
1 =
(H21I2 H22 I1) (A.1)

H1H2 I1 I2 (H2 I1 H1I2)(H1 H2)

=
(H21I2 H22 I1)2 (A.2)
For the case of Eq. (7), the parameters a1, a2 , a3 and a 4 were used. These were calculated as follows:
((a2 U1)ln(H1 + 1) + (U2 a2)ln(H2 + 1) + a 4 (U2 U1))
a1 =
(ln2 (H2 + 1) ln2 (H1 + 1)) (A.3)

h 4 k3 h3 k 4
a2 =
i3k 4 i 4 k3 (A.4)

a3 = (U1 a2)ln(H1 + 1) a1(ln2 (H1 + 1)) + a 4U1 (A.5)

i3 h 4 h3 i 4
a4 =
i3 k 4 i 4 k3 (A.6)

k3 = (U2 U3)ln2 (H1+ 1)+(U3 U1)ln2 (H2 + 1) + (U1 U2)ln2 (H3 + 1) (A.7)

k 4 = (U2 U4)ln2 (H1 + 1) + (U4 U1)ln2 (H2 + 1) + (U1 U2)ln2 (H 4 + 1) (A.8)

i3 = ln(H1+ 1)(ln2 (H3 + 1) ln2 (H2 + 1)) + ln(H2 + 1)(ln2 (H1 + 1) ln2 (H3 + 1)) + ln(H3 + 1)(ln2 (H2 + 1) + ln2 (H1 + 1)) (A.9)

i 4 = ln(H1 + 1)(ln2 (H 4 + 1) ln2 (H2 + 1))ln(H2 + 1) + (ln2 (H1 + 1) ln2 (H 4 + 1)) + ln(H 4 + 1)(ln2 (H2 + 1)ln2 (H1 + 1)) (A.10)

h3 = ln2 (H1 + 1)(U3ln(H3 + 1) U2ln(H2 + 1)) + ln2 (H2 + 1)(U1ln(H1 + 1) U3ln(H3 + 1)) + (U2ln(H2 + 1) U1ln(H1 + 1) (A.11)

h4 = ln2 (H1 + 1)(U4 ln(H 4 + 1) U2 ln(H2 + 1)) + ln2 (H2 + 1)(U1ln(H1 + 1) U4 ln(H 4 + 1)) + ln2 (H 4 + 1)(U2ln(H2 + 1) U1ln(H1 + 1) (A.12)
For the case of Eq. (19), the used parametersb1, b2 , b3 , b4 and b5 were calculated as follows:
((U2AN . (2ISC U2MPP 3PMPP UMPP + PMPP UAN)) (ISC U3AN U2MPP ) (PMPP U2AN U2MPP )
b1 = ( ) ( ) ( )
(UAN UMPP )3 (UAN UMPP ) 2 (UAN UOC)2 (UAN UMPP ) 2 (UMPP UOC )2
(2PMPP U2AN U2MPP ) (ISC U2AN U2MPP (UAN + UMPP )
+( ) ( )
(UAN UMPP )3 (UAN UOC) (UAN UMPP )3 (UAN UOC) (A.13)

2ISC UAN UMPP (UAN + UMPP ) 2 ((UMPP (4ISC U2AN + ISC UAN UMPP 6PMPP UAN + ISC U2MPP 2ISC U2AN UMPP (UAN + UMPP )
b2 = +
(UAN UMPP )3 (UAN UOC) (UAN UMPP )3 (UAN UMPP ) 2 (UAN UOC)2
2PMPP UAN UMPP (UAN + UMPP ) 4PMPP UAN UMPP (UAN + UMPP )
+
(UAN UMPP ) 2 (UAN UOC )2 (UAN UMPP )3 (UAN UOC ) (A.14)

(2ISC U2AN + 2ISC UAN UMPP 3PMPP UAN + 2ISC U2MPP 3PMPP UMPP) ((PMPP (U2AN + 4UAN UMPP + U2MPP 2PMPP (U2AN + 4UAN UMPP + U2MPP )
b3 = +
(UAN UMPP )3 (UAN UMPP )2 (UMPP UOC)2 (UAN UMPP )3 (UMPP UOC)
ISC (U3AN + 5U2AN UMPP + 5UAN U2MPP ISC (UAN (U2AN + 4UAN UMPP + U2MPP
(UAN UMPP )3 (UAN UOC) (UAN UMPP ) 2 (UAN UOC)2 (A.15)

(3ISC U2MPP 2
6ISC UMPP UOC ISC UOC + 4P_MPPUOC) 2(PMPP ISc UOC) 2(3UMPP + UOC)(PMPP ISC UMPP )
b4 = 2
+
(UAN UMPP ) 2 (UMPP UOC ) (UAN UMPP )(UMPP UOC) 2 (UAN UMPP )3(UMPP UOC)
2ISC UOC 2ISC UOC (UMPP + UOC)
+ +
(UAN UOC)(UMPP UOC) 2 (UAN UOC )2 (UMPP UOC)2 (A.16)

b5 = (U3OC (ISC UAN 2PMPP + ISC UMPP) PMPP U3AN + UOC (ISC U3MPP UAN ( ISC U2MPP + 6PMPP UMPP) + 4ISC U2AN UMPP) + U2AN ( 2ISC U2MPP + 3PMPP UMPP)
2
+ UOC (3PMPP UMPP 2ISC U2AN 2ISC U2MPP + UAN (3PMPP 2ISC UMPP)))
/((U2AN 2UAN UOC + 2
UOC)(U ANUMPP UAN UOC + UMPP UOC U2MPP)( U2AN UMPP + UOCU2AN + 2UAN U2MPP 2UOCUAN UMPP U3MPP
+ UOCU2MPP)) (A.17)

References Procedia 92, 2–9.

Eke, R., Demircan, C., 2015. Shading effect on the energy rating of two identical Pv
systems on a building façade. Sol. Energy 122, 48–57.
Al-Addous, M., Dalala, Z., Class, C.B., Alawneh, F., Al-Taani, H., 2017. Performance Elmetennani, S., Laleg-Kirati, T.M., Djemai, M., Tadjine, M., 2016. New Mppt algorithm
analysis of off-grid pv systems in the Jordan valley. Renew. Energy 113, 930–941. for pv applications based on hybrid dynamical approach. J. Process Control 48,
Baurzhan, S., Jenkins, G.P., 2016. Off-grid solar Pv. Is it an affordable or appropriate 14–24.
solution for rural electrification in sub-Saharan African Countries? Renew. Sustain. Ezinwanne, O., Zhongwen, F., Zhijun, L., 2017. Energy performance and cost comparison
Energy Rev. 60, 1405–1418. of Mppt techniques for photovoltaics and other applications. Energy Procedia 107,
Bellia, L., Marino, C., Minichiello, F., Pedace, A., 2014. An overview on solar shading 297–303.
systems for buildings. Energy Procedia 62, 309–317. Fara, L., Craciunescu, D., 2017. Output analysis of stand-alone Pv systems. Modeling,
Berthod, C., Strandberg, R., Yordanov, G.H., Beyer, H.G., Odden, J.O., 2016. On the simulation and control. Energy Procedia 112, 595–605.
variability of the temperature coefficients of Mc-Si solar cells with irradiance. Energy Fialho, L., Melicio, R., Mendes, V.M.F., Figueiredo, J., Collares-Pereira, M., 2014. Effect of

130
S. Paul Ayeng'o, et al. Solar Energy 183 (2019) 120–131

shading on series solar modules. Simulation and experimental results. Procedia Petreuş, D., Daraban, S., Ciocan, I., Patarau, T., Morel, C., Machmoum, M., 2013. Low cost
Technol. 17, 295–302. single stage micro-inverter with Mppt for grid connected applications. Sol. Energy 92,
Green, M.A., 1998. Solar CellsOperating Principles, Technology And System Applications. 241–255.
University Of New South Wales, Kensington, N.S.W. Xiv, 274. Qazi, S., 2016. Fixed standalone Pv systems for disaster relief and remote areas. In: Qazi,
Hansen, A.D., Sørensen, P.E., Hansen, L.H., Bindner, H.W., 2001. Models For A Stand- S. (Ed.), Photovoltaic (Pv) Systems For Disaster Relief And Remote Areas, 1st Ed.
Alone Pv System. < http://Orbit.Dtu.Dk/Files/7727175/Ris_R_1219.Pdf > . Elsevier, Amsterdam, pp. 139–175.
(Accessed 31 October 2018). Qi, C., Ming, Z., 2012. Photovoltaic module simulink model for a stand-alone Pv system.
Höcherl And Hackl Load Simulator – Google-Suche. < https://Www.Google.De/Search? Physics Procedia 24, 94–100.
Rlz=1c1gcea_Ende751de751&Ei=Fb1vwqq5figykwxytox4aw&Q=H%C3%B6cherl Rafikipower, 2017. < http://Www.Rafikipower.Com/ - Google-Suche > . < https://
+And+Hackl+Load+Simulator&Oq=H%C3%B6cherl+And+Hackl+Load Www.Google.De/Search?Rlz=1c1gcea_Ende751de751&Ei=
+Simulator&Gs_L=Psy-Ab.3...8129.19889.0.22857.17.17.0.0.0.0.82.1041.17.17. A5nlwr6adc6vsaek8b2icq&Q=Http%3a%2f%2fwww.Rafikipower.Com%2f&Oq=
0...0...1c.1.64.Psy-Ab.0.1.81...0i7i30k1j0i7i5i30k1.0.Czme8ozr5ri > . (Accessed 29 Http%3a%2f%2fwww.Rafikipower.Com%2f&Gs_L=Psy-Ab.3...4571.4571.0.4896.1.
January 2018). 1.0.0.0.0.60.60.1.1.0...0...1c.1.64.Psy-Ab.0.0.0...0.39idh2_K4oy > . (Accessed 2
Ibrahim, A., El-Sebaii, A.A., Ramadan, M.R.I., El-Broullesy, S.M., 2013. Estimation of January 2018).
solar irradiance on tilted surfaces facing south for tanta, Egypt. Int. J. Sustain. Energy Rao, A., Pillai, R., Mani, M., Ramamurthy, P., 2014. Influence of dust deposition on
32 (2), 111–120. photovoltaic panel performance. Energy Procedia 54, 690–700.
Kaldellis, J.K., Kapsali, M., Kavadias, K.A., 2014. Temperature and wind speed impact on Ritzenhoff, P., 1992. Erstellung Eines Modells Zur Simulation Der Solarstrahlung Auf
the efficiency of Pv installations. experience obtained from outdoor measurements in Beliebig Orientierte Flächen Und Deren Trennung In Diffusund - Google-Suche.
Greece. Renew. Energy 66, 612–624. < https://Www.Google.De/Search?Q=Erstellung+Eines+Modells+Zur
Kujundžić, G., Ileš, Š., Matuško, J., Vašak, M., 2017. Optimal charging of valve-regulated +Simulation+Der+Solarstrahlung+Auf+Beliebig+Orientierte+Fl%C3%A4chen
lead-acid batteries based on model predictive control. Appl. Energy 187, 189–202. +Und+Deren+Trennung+In+Diffusund&Rlz=1c1gcea_Ende751de751&Oq=
Lujano-Rojas, J.M., Dufo-López, R., Atencio-Guerra, J.L., Rodrigues, E.M.G., Bernal- Erstellung+Eines+Modells+Zur+Simulation+Der+Solarstrahlung+Auf
Agustín, J.L., Catalão, J.P.S., 2016. Operating conditions of lead-acid batteries in the +Beliebig+Orientierte+Fl%C3%A4chen+Und+Deren+Trennung+In+Diffusund
optimization of hybrid energy systems and microgrids. Appl. Energy 179, 590–600. &Aqs=Chrome.69i57.1868j0j8&Sourceid=Chrome&Ie=Utf-8 > . (Accessed 2
Magnor, D., Sauer, D.U., 2016. Optimization of Pv battery systems using genetic algo- January 2018).
rithms. Energy Procedia 99, 332–340. Sani Hassan, A., Cipcigan, L., Jenkins, N., 2017. Optimal battery storage operation for Pv
Makibar, A., Narvarte, L., Lorenzo, E., 2017. On the relation between battery size and Pv systems with tariff incentives. Appl. Energy 203, 422–441.
power ramp rate limitation. Sol. Energy 142, 182–193. Sauer, 1994. Untersuchungen Zum Einsatz Undentwicklung Von Simulationsmodellen
Menoufi, K., Farghal, H.F.M., Farghali, A.A., Khedr, M.H., 2017. Dust accumulation on Fürdie Auslegung Von Photovoltaik-Systemen. < https://Www.Researchgate.Net/
photovoltaic panels. A case study at the east bank of the Nile (Beni-Suef, Egypt). Profile/Dirk_Sauer/Publication/287332644_Untersuchungen_Zum_Einsatz_Und_
Energy Procedia 128, 24–31. Entwicklung_Von_Simulationsmodellen_Fur_Die_Auslegung_Von_Photovoltaik-
Merei, G., 2016. Optimization of Off-Grid Hybrid Pv-Wind-Diesel Power Supplies With Systemen/Links/5675c0c608ae502c99ce09f1/Untersuchungen-Zum-Einsatz-Und-
Multi-Technology Battery Systems Taking Into Account Battery Aging. Shaker Verlag, Entwicklung-Von-Simulationsmodellen-Fuer-Die-Auslegung-Von-Photovoltaik-
Aachen, Iii 184 Seiten. Systemen.Pdf > . (Accessed 7 December 2017).
Merei, G., Berger, C., Sauer, D.U., 2013. Optimization of an off-grid hybrid Schiffer, J., Sauer, D.U., Bindner, H., Cronin, T., Lundsager, P., Kaiser, R., 2007. Model
Pv–wind–diesel system with different battery technologies using genetic algorithm. prediction for ranking lead-acid batteries according to expected lifetime in renewable
Sol. Energy 97, 460–473. energy systems and autonomous power-supply systems. J. Power Sources 168 (1),
Müller, M., Bründlinger, R., Arz, O., Miller, W., Schulz, J., Lauss, G., 2014. Pv-Off-grid 66–78.
hybrid systems and mppt charge controllers, a state of the art analyses. Energy Schwingshackl, C., Petitta, M., Wagner, J.E., Belluardo, G., Moser, D., Castelli, M.,
Procedia 57, 1421–1430. Zebisch, M., Tetzlaff, A., 2013. Wind effect on Pv module temperature. Analysis of
Okoye, C.O., Solyalı, O., 2017. Optimal sizing of stand-alone photovoltaic systems in different techniques for an accurate estimation. Energy Procedia 40, 77–86.
residential buildings. Energy 126, 573–584. Shen, C., He, Y.-L., Liu, Y.-W., Tao, W.-Q., 2008. Modelling and simulation of solar ra-
Panasonic, 2012. Electrical and mechanical characteristics N235b N230b. < http://Pdf. diation data processing with simulink. Simulation Modell. Practice Theory 16 (7),
Archiexpo.Com/Pdf/Panasonic-Solar/Hit-Photovoltaic-Module-Black-Back-Sheet- 721–735.
N235b-N230b/83914-246153.Html > . (Accessed 31 October 2018). Singh, G.K., 2013. Solar power generation by Pv (Photovoltaic) technology. A review.
Paul Ayeng'o, S., Schirmer, T., Kairies, K.-P., Axelsen, H., Uwe Sauer, D., 2018. Energy 53, 1–13.
Comparison of off-grid power supply systems using lead-acid and lithium-ion bat- Soda, 2004. Free Time-Series Of Solar Radiation Data. < http://Www.Soda-Is.Com/Eng/
teries. Sol. Energy 162, 140–152. Services/Services_Radiation_Free_Eng.Php > . (Accessed 2 January 2018).
Pearsall, N., 2017. The Performance Of Photovoltaic (Pv) Systems. Modelling, Spertino, F., D’angola, A., Enescu, D., Di Leo, P., Fracastoro, G.V., Zaffina, R., 2016.
Measurement and Assessment. Woodhead Publishing, Cambridge 1 Online Resource. Thermal-electrical model for energy estimation of a water cooled photovoltaic
Pendem, S.R., Mikkili, S., 2018. Modeling, simulation and performance analysis of solar module. Sol. Energy 133, 119–140.
Pv array configurations (series, series-parallel and honey-comb) to extract maximum Vasel, A., Iakovidis, F., 2017. The effect of wind direction on the performance of solar Pv
power under partial shading conditions. Energy Rep. 4, 274–287. plants. Energy Convers. Manage. 153, 455–461.
Peng, Z., Herfatmanesh, M.R., Liu, Y., 2017. Cooled solar Pv panels for output energy Yahiaoui, A., Benmansour, K., Tadjine, M., 2016. Control, analysis and optimization of
efficiency optimisation. Energy Convers. Manage. 150, 949–955. hybrid Pv-diesel-battery systems for isolated rural city in Algeria. Sol. Energy 137,
Petit, M., Prada, E., Sauvant-Moynot, V., 2016. Development of an empirical aging model 1–10.
for Li-ion batteries and application to assess the impact of vehicle-to-grid strategies Ye, Z., Nobre, A., Reindl, T., Luther, J., Reise, C., 2013. On Pv module temperatures in
on battery lifetime. Appl. Energy 172, 398–407. tropical regions. Sol. Energy 88, 80–87.

131