Renewable Energy 216 (2023) 119117

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Experimental investigation of soiling impact on PV module performance in

Yanbu Al Sinaiyah, Saudi Arabia
Khaled S. AlZahrani *
Department of Mechanical Engineering, Yanbu Industrial College, 41912, Yanbu Alsinaiyah, Saudi Arabia

A R T I C L E I N F O A B S T R A C T

Keywords: For many years, photovoltaic (PV) modules have seen widespread use as one of the most effective solutions to
Photovoltaic meet the growing demand for clean energy. However, compared with other power generation technologies, PV
Soiling effect modules still face operational challenges, especially in industrial cities and the harsh environmental conditions of
Solar panel performance
arid coastal regions. This study examines the effects of dust, sand grains and soiling characterisation on the
Tilt angle
Dust characteristics
efficiency of PV modules under different environmental conditions. The researcher used PV modules of two
different technologies (Mono-Si and Poly-Si), two different tilt angles (10◦ and 24◦ ), and two different surface
conditions (clean and soiled) for their experiments. The study was conducted over a period of two years in
Yanbu, an industrial city on the western coast of the Kingdom of Saudi Arabia. It was the first study of its kind in
this city. The main findings indicate that the more the PV module was able to capture solar radiation, the less it
resisted soiling effect. The mono-Si module outperformed the poly-Si module by 13.7% on average for both clean
and soiled modules over the period of the study. However, the losses in power output due to the impact of soiling
on the poly-Si modules were less on average by 11% when compared to the mono-Si modules. The overall
average power losses for both mono-Si and poly-Si modules were lower by around 40% when installed at the 24◦
tilt angle. The study of dust particles’ morphology and chemical compositions, along with the collected envi­
ronmental data, confirmed that the dust-induced shade effect could contribute to the soiling impact on PV
modules.

1. Introduction 32⁰20′N and longitudes 34⁰-50′E and 54⁰20′E. Thus, it is an arid envi­
ronment known for high temperatures, low precipitation, drought and
The Kingdom of Saudi Arabia has abundant environmentally friendly winds that raise dust and sandstorms [3].
energy resources, including solar, wind and geothermal energy, that are Solar photovoltaic (PV) panels are designed to collect this abundant
feasible to integrate into the power grid to use as stand-alone or hybrid solar radiation. The amount of insolation hitting the panels largely de­
systems. The optimum use of these renewable energy resources is pro­ termines their energy performance and output [4]. However, many
moted by Saudi Vision 2030 to minimise environmental impacts, pre­ factors affect the performance of solar PV panels and prevent them from
serve natural resources, minimise waste and improve future economic capturing maximum solar radiation or operating at their nominal effi­
and social needs. Saudi Arabia began using solar energy as a renewable ciencies to achieve their rated capacities. PV module designs, array
energy technology to increase power generation, reduce oil consump­ configurations, installation methods and environmental conditions are
tion and achieve a more environmentally friendly future energy mix. the major factors that affect system performance, longevity and reli­
This trend has been motivated by the remarkable cutting-edge tech­ ability [5]. For example, elevating the modules and facing the optimum
nology, high performance and decline in the price of solar technology. tilt angles and orientations are configuration factors that play significant
Saudi’s climate is well suited for solar energy projects and creates an roles in improving solar PV performance [6]. For optimum performance,
excellent opportunity to utilise such technology in power generation. the solar industry experts recommend mounting modules at least six
The country’s insolation is among the highest in the world, with an inches from the roof or as high as possible from the ground [7]. Raising
average of 6.474 kWh/m2 per day [1,2]. Additionally, Saudi Arabia is the arrays allows air to circulate along the back surface of the panels to
one of the Sunbelt countries located between latitudes 16⁰-40′N and cool them, thus increasing efficiency and electrical output.

* Corresponding authir.
E-mail address: [email protected].

https://doi.org/10.1016/j.renene.2023.119117
Received 26 April 2023; Received in revised form 12 July 2023; Accepted 3 August 2023
Available online 3 August 2023
0960-1481/© 2023 Elsevier Ltd. All rights reserved.
K.S. AlZahrani Renewable Energy 216 (2023) 119117

The tilt angle of the solar panels dramatically influences how much Alnasser et al. [26] conducted an outdoor experiment to examine the
solar radiation strikes the panels. The optimum tilt angle is dependent composition of accumulated dust on solar PV. They reported that more
on the season and the latitude where the panel is located [8]. Solar PV than 50% of the soiling was silicon oxide, followed by cement and in­
panels receive the maximum irradiance and thus produce the most en­ dustrial gypsum. The result also revealed that industrial gypsum, natural
ergy when they are perpendicular to the sunlight incident angle [9]. and white cement exhibited the greatest impact on the loss of PV energy
Typically, fixed-tilt solar panels are tilted at an angle equal to the lati­ production, while sand had the least effect. Moreover, they recorded
tude location of the panels to maximise the annually received irradiance that industrial gypsum had a high degree of adhesion on PV cells, while
[10]. Likewise, the orientation – the azimuth angle – contributes to how natural cement was the most difficult to clean when using a dry-cleaning
much solar radiation can be captured by the PV panels and significantly method due to the miniature sizes of the particles. Moderate rainfall
affects solar PV performance. As a rule of thumb, the ideal orientation of cannot effectively clean soiled panels when the particles are in the range
fixed solar PV panels to provide the best energy production is typically of 2–10 μm.
towards the South in the northern hemisphere and towards the North in Hachicha et al. [27] performed outdoor and indoor experiments in
the southern hemisphere [11]. Sharjah (UAE) to examine dust characteristics and their impacts on solar
Environmental conditions such as soiling and dust [12], incident PV performance. The morphology of the dust samples they collected
irradiance [13], ambient temperature [14], wind speed [15], conden­ over a period of five months indicated that the average particle size of
sation and humidity [16,17] are other critical parameters that signifi­ deposited dust was between 1.61 and 38.40 μm. They confirmed that
cantly affect solar PV performance. The PV modules give the nominal dust particle collection was a function of the PV slope, and higher PV
performance when they are at their rated conditions of 1000 W/m2 of module inclinations resulted in lower dust accumulations. The results of
solar intensity, a temperature of 25 ◦ C, and an air mass of 1.5, according three different tilt angles (45◦ , 25◦ , and 0◦ ) were tested on PV models,
to the standard test condition of PV rating established by the Interna­ and they found that the dust deposition increased by 10.95%, 14.11%
tional Electrotechnical Commission (IEC 61853–1) [18]. This means and 37.63%, respectively, as the slope decreased. In addition, their
that the module performance and efficiency will increase with the in­ experiment result showed that there was a direct connection between
crease of solar irradiance and decrease proportionally as the PV module the PV power production and the density of the dust, which was 5.44
surface temperature increases. Soiling is the focus of this study, as it is g/m2, accumulated over the five months; the PV performance loss
considered the foremost challenge facing solar PV projects worldwide, increased by 12.7%. Other studies addressed the impact of soiling on the
particularly in arid climate regions such as Saudi Arabia. The arid performance of solar PV in different geographical locations and atmo­
climate countries have been experiencing heightened sand and dust spheric conditions. Adinoyi and Said [28] studied the effect of dust
storms in the last few decades, with major impacts on both solar projects accumulation on the power outputs of solar PV modules in the Eastern
and human health, and today they stand as the worst dust accumulation province (Dhahran) of Saudi Arabia. Their investigation demonstrated
zones in the world [19,20]. that when soiling accumulated on a PV array surface and was not
Soiling is the specks of dirt and debris that build up on the surfaces of cleaned for six months or more, the PV power output decreased by 50%.
PV panels. It negatively affects the performance of PV modules by The results also indicated that the PV performance losses due to soiling
blocking the irradiance incident onto the PV modules’ surfaces; conse­ did not depend on the duration of module exposure to outdoor condi­
quently, the power output of the PV array is reduced. The soiling for­ tions, but depended on the type and intensity of the dust. They stated
mation typically results from dust storms, dirt, bird droppings, bacteria that one single dust storm could reduce the array’s electrical power
biofilms, algae, lichen, mosses, fungi, plant debris, pollen, engine output by 20% or more.
exhaust fumes, nearby busy roadways, airports, construction sites, Benghanem et al. [29] analysed the effect of dust accumulation on
agricultural zones, industrial emissions and other atmospheric particles. PV array surfaces in the non-coastal area of western Saudi Arabia. They
The different densities, sizes and compositions of this dirt and dust build observed a 28% power output decrease due to the soiling in two months.
up on the PV panels’ surfaces [21,22]; when they accumulate, they The study used very low-quality solar modules that had approximately a
obstruct the illumination of the PV panels by attenuating, reflecting, 60% fill factor. It is believed that the typical solar PV cell technologies
scattering and absorbing the incoming solar radiation. The degree to available for commercial use that have a 70% fill factor would have
which the soil particles block the sunlight depends upon soil size, type, different outcomes. Additionally, the duration of the study was rela­
density and the timespan of the soiling settlement [23]. tively short and failed to determine the worst possible dust conglom­
Many field and laboratory studies have focused on investigating the erate effect. Baras et al. [30] evaluated soiling deprivation and its
characteristics of deposited dust on PV panels and its effect on PV per­ economic impact on PV power plants in central Saudi Arabia. They re­
formance. A study carried out by W. Javed et al. in Doha, the capital of ported seasonal power losses due to the soiling accumulation, with the
Qatar [24], investigated the characterisation of dust accumulated on PV highest loss rates identified in the spring months, which is also when the
panels installed in arid climates and its impact on the PV systems’ per­ highest daily average radiation occurs in this region. Karmouch and El
formance. X-ray diffraction (XRD), X-ray fluorescence (XRF) and scan­ Hor [31] conducted an outdoor investigation of PV performance due to
ning electron microscopy (SEM) particle characterisation were used. the effect of dirt and dust using two 30W solar panels mounted on a
The study noted that calcium, silicon, iron, magnesium and aluminium rooftop in the southern region of Saudi Arabia (Jazan). It was found that
were the most dominant elements in the accumulated dust, with calcium dust agglomeration reduced the performance of the solar cells by about
constituting the highest percentage of particles. They claimed that the 10% for an exposure time of four months.
dust accumulation rate decreased with increasing time of exposure; they One of the limitations of previous studies is their short durations. As
also argued that soiled panel surfaces retained less dust than recently far as the author is aware, most of the studies that focused on measuring
cleaned panels without providing any further detail. Kazem and Chai­ the impact of soiling on solar PV modules in the Gulf region did not
chan [25] experimentally studied the impact of weather agents such as exceed six months. Such a relatively short experiment duration doesn’t
relative humidity and wind speed on the dust accumulation on PV sur­ account for the variations of seasonal weather patterns throughout the
faces in six cities in the north of Oman, as well as the influence of the year. Moreover, no studies investigated the performance of solar PV
cleaning techniques. The authors reported that the industrial cities have modules in the industrial city of Yanbu, the coastal city that hosts many
higher percentages of dust, while the cities far from the industrial zones petroleum and petrochemical industries and other energy-intensive in­
have lower percentages of soiling on PV panels. The result of the study dustries that may affect the dust and soil composition. To overcome
showed that the most conspicuous pollutants were particulate matter these limitations in the current study, experimental work has been
(PM) that originated from industrial facilities that burned fossil fuels carried out to examine the impact of pollution on the performance of PV
such as coal, oil and natural gas. solar modules under climatic conditions in the industrial city of Yanbu

2
K.S. AlZahrani Renewable Energy 216 (2023) 119117

for two years’ duration. Two different PV module technologies and two
tilt angles have been investigated indoors and outdoors. In addition,
samples of the accumulated dust on the surfaces of PV modules have
been analysed to discover the influence of nearby industries on the
composition and characterisation of dust and its impact on the PV
module performance.

2. Methodology

This study was conducted at the Research Centre of the Royal

Commission Colleges and Institutes in Yanbu. Yanbu Industrial City is
located on the western coast of the Kingdom of Saudi Arabia at the co­
ordinates 24◦ N and 38◦ E. Two different technologies of PV modules
were investigated in this study, namely: Mono-Si, and Poly-Si. These
first-generation c-Si PV technologies were selected because they are
highly efficient and more affordable compared to other PV modules;
they are also the most utilised modules for residential solar rooftops and
large-scale PV projects in arid regions. The electrical specifications for
those PV module technologies are listed in Table 1.
Since the quantities and types of soil deposits, air pollutants, hu­
midity variations, wind speeds and dusting on the solar modules vary
dramatically with each season, the study lasted for two years [32]. The
experimental work was initiated with an indoor test at the standard test Fig. 1. The indoor experimental setup.
conditions (STC) to determine and compare the performance of the
employed modules in the experiment. The PV module was cleaned,
measure actual power losses due to uniform or non-uniform soiling. The
installed horizontally under an artificial light source consisting of eight
PV modules were exposed to the Yanbu climatic conditions, which are
500W halogen bulbs and mounted over a custom-built structure (Fig. 1).
hot, with high humidity and frequent sandstorms. For each tilted angle
The indoor investigations were performed to compare the modules
set, one of the modules was cleaned once every two days in the early
and evaluate their mismatch while avoiding the outdoor environment’s
morning, the time at which there was little or no wind or dust and power
influence and inconsistency in solar radiation. After the indoor test, the
production was low. The other module was never cleaned. Regular
system was placed in an outdoor area that was open and shade free.
municipal water was used for the module cleaning process, followed by
Modules were mounted at their optimum orientations where they
wiping manually with non-abrasive cloths to ensure the removal of
received the maximum sunlight for outdoor module performance as­
water residue, dust and bird droppings.
sessments. Both the indoor and the outdoor tests were conducted for 24
The weather data and the PV modules’ Voc and Isc readings were
h, and the readings were collected at 10 min intervals. Based on those
recorded every 10 min using eight module backside temperature sensors
results, the modules were sorted according to performance to ensure
for each module. A CR1000 data logger that was connected to a desktop
that similar panels were tested together. Two sets of PV modules of each
computer through a USB communication port was used to record the
technology were mounted on an open rack, exposed to the natural at­
performance of the modules. This collected data was exported auto­
mosphere and placed facing south. The first set was adjusted at a tilt
matically to an online tool and an excel spreadsheet for processing. The
angle of 10◦ , and the other set at 24◦ . The PV panels were elevated 1 m
module specifications that were used are shown in Table 2.
from the ground, which is close to the standards being followed in Saudi
Furthermore, at the end of the experiment period, three dust samples
Arabia for ground mounting of PV panels. The 10◦ angle is the standard
were collected from the modules’ surfaces using a soft brush to deter­
rooftop tilt angle used in Saudi Arabia to allow dust, debris and water to
mine the dust characteristics, i.e. morphologies, particle sizes, compo­
run off properly; it also allows sufficient ventilation space behind
sitions and weight percentages. The dust samples were prepared by
modules in order to reduce array temperature via air circulation, since
depositing sand grains directly onto carbon tape over a standard
almost all of the building roofs are flat in Saudi Arabia. The other chosen
aluminium stub. Morphological analysis and elemental mapping tech­
tilt-angle in this study was 24◦ , corresponding with the latitude of the
niques were then performed. The SEM imaging was performed with a
project’s geographical location at Yanbu, Saudi Arabia. This angle is the
Thermo-Fisher Teneo microscope equipped with a field emission elec­
typical tilt angle for fixed arrays, especially ground mount systems, to
tron gun, and an energy-dispersive X-ray spectroscopy (EDX) detector
receive the maximum solar radiation and produce the most energy
was used for the chemical analysis. The SEM images were acquired at 3
throughout the year. Fig. 2 summarises the experimental procedure and
kV with a beam current of 25 pA by using a standard Everhart-Thornley
PV modules’ configurations in a flowchart diagram. A total of eight
detector (ETD) for the secondary electrons acquisition and to show the
modules were used in the study; first, all modules were tested indoors,
morphological features. The EDX analysis was carried out with an EDX
then outdoors. The results of these two studies were used to sort modules
SDD detector and treated with EDX TEAM™ software v. 4.5 [33].
into two pairs for each technology. The first pair of each was installed at
The weather data for the study’s location was obtained from air
a tilt angle of 10◦ , while the second was installed at 24◦ . One module of
quality and meteorological stations of the Royal Commission in Yanbu
each pair was kept clean, while the other was left soiled.
and solar energy monitoring and measurement stations of King Abdullah
As Fig. 3 shows, all modules were installed in portrait mode to
City for Atomic and Renewable Energy. The collected data were solar

Table 1
The electrical specifications for PV modules technologies used in the study.
Module Technology Power (Pmax) Rated Voltage (Vmp) Rated Current (Imp) Open-Circuit Voltage (Voc) Short-Circuit Current (Isc)

Mono-Si 100 Watts 18.9 V 5.29 A 22.5 V 5.57 A

Poly-Si 100 Watts 18.0 V 5.56 A 22.0 V 6.06 A

3
K.S. AlZahrani Renewable Energy 216 (2023) 119117

Fig. 2. Experimental procedure and PV modules configuration flowchart.

Table 2
PV modules specification (cell-data).
Parameters (Unit) Module 1 Module 2

Module Technology Mono-Si Poly-Si

Tilt Angle 10◦ & 24◦ 10◦ & 24◦
Total area (m2) 0.73 0.67
Aperture area (m2) 0.68 0.65
Cell area (m2) 0.63 0.61
Number of cells in series 36 36
FF (%) 72.8% 75.1%

3. Results and discussion

Solar radiation and ambient temperature are the most significant and
directly related parameters to the output power of the PV solar cells
[34]. Over the period of the study, the ambient temperature varied be­
tween 12 ◦ C and 45.3 ◦ C. Solar radiation varied between 970.4 W/m2 in
July 2020 and 423.00 W/m2 in December 2020.

Fig. 3. Outdoor experimental setup. 3.1. Indoor and outdoor study

radiation, ambient temperature, particulate matter data (PM2.5) and Two sets of experiments were conducted to verify the performance
(PM10), relative humidity, wind speed and direction and barometric conformity of the employed modules: indoors under controlled condi­
pressure. tions, and outdoors under environmental conditions. The indoor and
outdoor investigations demonstrated relatively identical performances
with a minimal natural variation of 5% or less in average power output
for all modules of the same technology. The results neutralise any

4
K.S. AlZahrani Renewable Energy 216 (2023) 119117

Fig. 4. Indoor and outdoor study graphs.

possibility of errors due to the manufacturing differences between the Based on the recorded data over the period ranging from June 2020 to
cells used in both types. The average power generation for two of the July 2022, the overall monthly average losses in the performance of the
used modules in the indoor study is depicted in Fig. 4a, and Fig. 4b soiled Mono-Si PV module were 24.2%, compared with the clean
depicts the average power generation in the outdoor study. All eight module.
modules performed similarly with insignificant variation. Except for the effect of dust accumulation on one of the two modules,
The results verified the rated performance of all modules according the responses of both modules to climatic conditions were identical to all
to the manufacturers’ specifications. Based on the results of those tests, other climate conditions. The sharp dip in power output during the
the modules were sorted according to performance to ensure that month of September 2020 can be seen in both modules; they both
modules of similar performance were tested together. recorded the lowest readings throughout the study period. This response
can be traced back to the observed elevation in the particulate matter
3.2. Effect of soiling data for both PM2.5 and PM10. The maximum reading was recorded
during the month of September 2020; the monthly average reading was
As previously stated, two sets of PV modules were used in this study. 14.19 μg/m3 for PM 2.5 and 106.97 μg/m3 for PM10. The ambient
Each set was comprised of two Mono-Si PV modules and two Poly-Si PV temperature was also recorded at its highest reading throughout the
modules. The first was installed at a tilt angle of 10◦ , while the second study period, reaching 34 ◦ C on average, and the highest humidity value
was installed at 24◦ . One module of each type was cleaned regularly was an average of 66% during the same month. A similar response to the
following the cleaning protocol previously noted. The acquired data climate conditions was observed during the month of Mar 2021, where
allowed comparison and evaluation of the efficiency and durability of both modules achieved their best performance throughout the study
both technologies and the best installation tilt angle was able to be period, driven by the low ambient parameter readings. The monthly
found. The results indicated identical performance of the modules for average reading for the solar radiation during March 2021 was 461.73
both tested tilt angles. Therefore, for brevity, the data acquired from the W/m2, the ambient temperature was 24 ◦ C, the humidity was 53% and
set installed at a 10◦ tilt angle will be presented to clarify the effect of the particulate matter was 8.66 μg/m3 and 76.91 μg/m3 for PM2.5 and
soiling on the performance of both employed technologies. PM10, respectively.
Therefore, the power generated by the Mono-Si PV module was
3.2.1. Mono-Si PV modules diminished by the particulate of soil due to hindrance of the radiation
The first set of analyses examined the effect of soiling on Mono-Si PV received. Cleaning dust, debris, and contamination from the modules
modules at the 10◦ tilt angle. Fig. 5 shows the pattern of the average according to the suggested protocol is recommended to ensure stable
monthly output power in watts over the two-year period of the experi­ performance and to prevent losses in energy production that may reach
ment. Initial visual observation of the graph clearly shows that the up to 24%, on monthly average.
average output power of the soiled module was lower than that of the
clean module. In the early months of the study (June 2020 to May 2021), 3.2.2. Poly-Si PV modules
the monthly average power output losses of the soiled PV module The results of the Poly-Si PV module that was installed at a tilt angle
increased from 12.7% to 29.6%. During the latter months of the study of 10◦ were analysed and compared to the Mono-Si PV module at the
(June 2021 to July 2022), these losses increased to 39.04% in July 2022. same tilt angle. Fig. 6 shows the pattern for average monthly power

Fig. 5. Mono-Si PV modules (clean vs soiled) graph. Fig. 6. Poly-Si PV modules (clean vs soiled) graph.

5
K.S. AlZahrani Renewable Energy 216 (2023) 119117

output in watts against date. Similar performance is observed for the air circulation, since almost all the building roofs are flat in Saudi
Poly-Si PV module, but with a slight difference in the performance under Arabia.
the impact of soiling in the short and long terms. However, the difference in the monthly average output power was
The output power of the soiled module was distinctly lower than that not distinct and clear, as indicated by the comparison of the pattern of
of the clean module throughout this period. At the start of the study the soiled module at both tilt angles. The trend seen in the output power
(June 2020 to December 2020), the average monthly power output of for soiled PV modules at both tilt angles suggests that the performance of
the soiled PV module declined by 15%. During the last six months of the soiled PV modules at a 24◦ tilt angle was slightly better than the module
study, losses almost doubled to 29%. The response to the different at the 10◦ tilt angle, with a value of 4.6% in the latter months of the
environmental conditions for the Poly-Si modules was the same as the study (September 2021 to July 2022). A similar pattern was seen in the
Mono-Si modules, but the losses in power output for the Poly-Si modules months between October 2020 to March 2021; the highest difference in
were found to be less than those of the Mono-Si modules under the performance was recorded for the module installed at 24◦ in October
impact of soiling. On a monthly average over the period of the study, the 2020, by 28.16%. During those periods, the solar radiation was also
soiled Poly-Si modules underperformed the cleaned modules by 13.4% recorded at its highest readings, reaching up to 602 W/m2. The incli­
at the tilt angle of 10◦ . In comparison with the performance of the Mono- nation angle of 24◦ is the standard inclination angle used in the Kingdom
Si soiled modules, the Mono-Si module outperformed the Poly-Si module of Saudi Arabia for the ground installation of photovoltaic panels. This is
by 13.7%, on average, for both cleaned and soiled modules over the also the latitude for the geographical location of Yanbu. Hence, this tilt
period of the study. However, the losses in power output of the Poly-Si angle will allow the maximum capture of solar rays [43].
modules due to the impact of soiling were less than those of the Mono-Si By looking at the entire study period and using the average output
modules by 11%, on average. The higher manufacturing efficiency of the power normalised values, it was found that the performance of both
Mono-Si modules is the main reason for their better performance, while units at the two tested tilt angles was similar at 0.6. This might be
the material selection and protection layers for the Poly-Si modules may explained by the fact that each inclination angle has a specific feature
be the reason behind their better handling of the impact of soiling [35]. that enhances its performance during certain weather conditions. The
These findings are in line with other studies from the MENA region ability to maximise the utilisation of solar rays with the 24◦ angle during
[36–42]. high solar radiation days was counterbalanced by the ability of the 10◦
tilt angle to cool the cells on hot summer days by promoting air circu­
3.3. Effect of tilt angle lation during low solar radiation and high-speed wind. Even though the
overall average power losses of the Mono-Si modules installed at a tilt
The tilt angle of the modules was a crucial factor for defining the angle of 24◦ were lower by 41% than the power losses for the modules
efficiency of the PV modules. The first set of analyses was performed on installed at a tilt angle of 10◦ , the module at a tilt angle of 10◦ produced
the Mono-Si PV modules to examine the effect of both predefined tilt more power than the modules at 24◦ by 5.1% for the soiled modules and
angles (10◦ and 24◦ ) on power generation under different operating by 16% for the cleaned modules, on average, as shown in Fig. 8.
conditions and to locate the best tilt angle that would withstand the
impact of soiling in the study location.

3.3.1. Mono-Si PV modules

Fig. 7 shows the line graph of clean and soiled modules for Mono-Si
PV at 10◦ and 24◦ tilt angles. The patterns of the cleaned modules at both
tilt angles reveal that the average monthly power outputs of the modules
at 10◦ are significantly higher than the modules at 24◦ by 16%, on
average. The modules at the 24◦ angle were outperformed by the
modules at the 10◦ tilt angle throughout the study. The highest recorded
difference was in May 2022, when the power output of the module at the
10◦ tilt angle was 20% higher than the module at the 24◦ angle. During
that month, the solar radiation was at a maximum monthly average
value of 579.58 w/m2. From March 2021 to September 2021, this dif­
ference was continually above 10%. A similar performance was evident
from April 2022 to July 2022. Even though the 10◦ tilt angle allowed for
dust to accumulate over the surface of the PV module, it permitted the
dust, debris and water to run off properly and created sufficient venti­
Fig. 8. Mono-Si PV modules (10◦ vs 24◦ tilt) graph.
lation space behind the modules to reduce array temperatures through

Fig. 7. Mono-Si PV modules (10◦ vs 24◦ tilt) graph.

6
K.S. AlZahrani Renewable Energy 216 (2023) 119117

Though the recommended tilt angle is 24◦ at the location where the
study was conducted, this study reveals that the weather conditions and
environmental characteristics make this tilt angle less efficient when
compared with higher tilt angles for Mono-Si PV modules.

3.3.2. Poly-Si PV modules

The effect of tilt angle on the performance of a Poly-Si PV module
under similar configurations and conditions was analysed using the
same procedure as for the Mono-Si PV module. The analysed data from
the comparison of average monthly output power generation by the
Poly-Si PV modules at the two predefined tilt angles is represented in
Fig. 9. A comparison of the clean PV modules results at both tilt angles
reveals that the impact of the change in tilt angle was not very distinc­
tive in the Poly-Si PV modules. The trend of the monthly average output
power recorded for both tilt angles was irregular throughout the study
period and revealed a mixed pattern in the performance.
The results showed that each module was outperformed by the other Fig. 10. Poly-Si PV modules (10◦ & 24◦ tilt).
module by 10%, on average, for 12 non-consecutive months. For
example, in September 2020, the clean module installed at 10◦ was a modules at 10◦ and 24◦ , but at the expense of the generated output
better performer than the module installed at 24◦ , by 43%. However, in power.
April 2022, the cleaned module at 24◦ was producing a 27% higher
power output than the module at 10◦ . Such fluctuating performances
3.4. Dust characterisation
cannot be explained by the collected environmental data or air quality
data, as both modules were tested at the same times and locations,
The dust characterisation investigation’s results framed the compo­
exposed to the same conditions and cleaned regularly following the
sition and characterisation of the dust grains in the three dust samples
same protocol. This may signify that the poly-Si PV module technology
collected from the PV modules’ surfaces. Evidence of the morphology of
was able to produce almost similar performance for both tested tilt an­
the dust particles was provided by the SEM micrograph images for each
gles, unlike the Mono-Si PV module technology, which showed a greater
dust sample and included texture, shape and size. The dust particles
response to the change in the tilt angle. However, the overall average
were characterised by irregular, rough, and non-spherical morphologies
percentage difference for the clean poly-Si PV module at a 24◦ tilt angle
and sizes ranging from 2 μm to 65 μm. The majority of the particle sizes
was almost 1% higher than the one at a 10◦ inclination. This indistinct
in the population were smaller than 10 μm. The small particle size forms
result was confirmed by comparing the performances of the soiled Poly-
adhesion forces with the PV module surface that hinders the natural
Si modules installed at 10◦ and 24◦ tilt angles. The module installed at
cleaning process by strong wind or rain [45]. Therefore, it is more
24◦ was superior throughout the study period, except for three non-
difficult for the smaller dust particles to come off or roll down the tilted
consecutive months. On average, the performance of the module
module’s surface, and they require labour-intensive cleaning. Moreover,
installed at a tilt angle of 24◦ was 10% higher than the module installed
humidity plays a significant role in the adhesion force between indi­
at 10◦ .
vidual dust particles and between the dust particles and the surfaces of
Overall, the Poly-Si module installed at a tilt angle of 24◦ maintained
the PV modules [46,47]. In the study site, the monthly average of the
slightly higher performance when kept cleaned. It also outperformed the
humidity was between 40% and 60% throughout the study period.
module installed at the 10◦ tilt angle by 10% when exposed to the
Fig. 11 shows one of the SEM micrograph images. Fig. 11a shows that
climate conditions without cleaning and reduced the power losses by
the dust particles were irregular in shapes and sizes. Fig. 11b shows the
40%, as shown in Fig. 10.
adhesion of small dust particles to large dust particles due to the elec­
Overall, it can be concluded that the more the PV module was able to
trostatic charges of the dust particles [48]. Such formations enhance the
utilise solar radiation, the less it resisted climatic conditions. The
reflection and scattering of incoming solar radiation. Fig. 11c and
different protective films and other coating materials that were used to
d shows the non-spherical and sharp edges of the dust particles.
improve the modules’ thermal and dust resistance reduced the number
The EDX chemical analysis of the dust samples generated a detailed
of photons that reached the photovoltaic cells [44]. For example,
report for each dust sample. The analysis revealed the presence of
Mono-Si modules outperformed all other modules when they were kept
identical elements and compounds with varying percentage composi­
clean at a 10◦ tilt angle but suffered the most losses when they were left
tions that were not highly deviating from each other. The main elements
soiled. Fewer power losses were manifested for both of the Poly-Si
identified from the accumulated dust were Oxygen, Sodium,

Fig. 9. Poly-Si PV modules (10◦ & 24◦ tilt).

7
K.S. AlZahrani Renewable Energy 216 (2023) 119117

Fig. 11. SEM images of dust particles.

Magnesium, Aluminium, Silicon, Sulphur, Titanium, Calcium, Iron and The Energy dispersive spectra of the dust particles spectra are pre­
Carbon. Oxygen at 46.5% and Silicon at 14.8% were the main constit­ sented in Fig. 12. Carbon was observed in two of the three samples,
uents of the samples with the highest weight percentage of particles. possibly due to the presence of a petroleum industry nearby. It is
Traces of other elements such as Chlorine 2.8%, Potassium 1.6%, were possible that the carbon seen in the spectra is due to the use of carbon
also present. This was confirmed by the eZAF smart quant results ob­ tape. Therefore, there is a certain amount of error associated with the
tained from the EDX chemical analysis. measurement of carbon [49].

Fig. 12. Weight % of organic components of the polluting material.

8
K.S. AlZahrani Renewable Energy 216 (2023) 119117

These readings were confirmed by air quality data, which indicated 4. Nomenclature
the existence of some other fine particles (PM2.5 and PM10) enriched in
the oxidation of volatile organic compounds (VOCs, Sulphur, and Ni­ PV Photovoltaic
trogen Oxides) originating from fossil fuel combustion. The desert C–Si Crystalline silicon
landscape of the study site explains the presence of silicate, as the Mono-Si Monocrystalline silicon
windblown dust from the desert adds a large quantity of silicate com­ Poly-Si Polycrystalline silicon
pounds. Potassium, sodium and chloride can be associated with sea salt XRD X-ray diffraction
since the study site is located on the coast of the Red Sea [50]. The dust EDX energy-dispersive X-ray spectroscopy
particles’ morphology and chemical compositions, with the help of other SEM Scanning electron microscopy
environmental conditions such as humidity, defined the dust particles’ ETD Everhart-Thornley detector
adhesion force to the PV module surface. Dust particles then caused SDD Silicon Drift Detector
dust-induced shading that scattered sunlight by reflecting it or forcing it PM Particulate matter
to deviate from the straight light path to the solar modules. Thus, the STC Standard test conditions
power output is negatively affected [51,52]. MENA Middle East and North Africa
The results obtained from both studies emphasize the need for an Pmax Maximum Power (W)
auxiliary cleaning system to mitigate the impact of soiling on the per­ Isc Short-Circuit Current (A)
formance of the PV panels. These systems vary from conventional to Voc Open-Circuit Voltage (V)
advanced cleaning systems. Conventional systems include manual water Imp Current at Maximum Power (A)
cleaning, sprinkle water cleaning, brush cleaning, and wiper cleaning, Vmp Voltage at Maximum Power (V)
while advanced cleaning systems include robotic cleaning, anti-soiling FF Failure factor (%)
coatings (self-cleaning surfaces), electrostatic cleaning, and EDS (Elec­
trodynamic System for Cleaning), etc. Due to labor issues and the in­ References
efficiency of conventional cleaning methods, advanced cleaning
methods are being employed on a large scale. [1] E. Zell, S. Gasim, S. Wilcox, S. Katamoura, T. Stoffel, H. Shibli, et al., Assessment of
Solar Radiation Resources in Saudi Arabia, Solar Energy, 2015, pp. 422–438,
https://doi.org/10.1016/j.solener.2015.06.031.
4. Conclusion [2] A. Hepbasli, Z. Alsuhaibani, A key review on present status and future directions of
solar energy studies and applications in Saudi Arabia, Renew. Sustain. Energy Rev.
(2011) 5021–5050, https://doi.org/10.1016/j.rser.2011.07.052.
The impact of soiling on the performance of two different technol­ [3] M. Almazroui, Rainfall trends and extremes in Saudi Arabia in recent decades,
ogies of PV modules installed at two different tilt angles was studied for Atmosphere 11 (9) (2020) 964.
[4] M.K. Al-Ghezi, R.T. Ahmed, M.T. Chaichan, The influence of temperature and
a period of two years. The data acquired were processed and analysed. It
irradiance on performance of the photovoltaic panel in the middle of Iraq, Int. J.
can be concluded that the more the PV module was able to utilise solar Renew. Energy Dev. 11 (2) (2022) 501.
radiation, the less it resisted climatic conditions. The different com­ [5] K.V. Vidyanandan, An overview of factors affecting the performance of solar PV
parison studies performed indicated that the Mono-Si module out­ systems, Energy Scan 27 (2017) 2–8. A house journal of Corporate Planning, NTPC
Ltd.
performed the Poly-Si module by 13.7%, on average, for both the [6] K.E. N’Tsoukpoe, Effect of orientation and tilt angles of solar collectors on their
cleaned and the soiled modules over the period of the study. However, performance: analysis of the relevance of general recommendations in the West
the losses in power output due to the impact of soiling on the Poly-Si and Central African context, Scientific African 15 (2022), e01069.
[7] S.P. Aly, S. Ahzi, N. Barth, Effect of physical and environmental factors on the
modules were less than those for the Mono-Si modules by 11%, on performance of a photovoltaic panel, Sol. Energy Mater. Sol. Cell. (2019), 109948,
average. And the overall average power losses of the Mono-Si modules https://doi.org/10.1016/j.solmat.2019.109948.
installed at a tilt angle of 24◦ were lower by 41% than those installed at a [8] T.O. Kaddoura, M.A. Ramli, Y.A. Al-Turki, On the estimation of the optimum tilt
angle of PV panel in Saudi Arabia, Renew. Sustain. Energy Rev. 65 (2016)
10◦ tilt angle. However, the modules at a tilt angle of 10◦ produced more 626–634.
power than the modules at 24◦ by an average of 5.1% for the soiled [9] K. Sukarno, A.S.A. Hamid, C.H. Jackson, C.F. Pien, J. Dayou, Comparison of power
modules and by 16% for the cleaned modules. Furthermore, the Poly-Si output between fixed and perpendicular solar photovoltaic PV panel in tropical
climate region, Adv. Sci. Lett. 23 (2) (2017) 1259–1263.
module installed at a tilt angle of 24◦ maintained slightly higher per­
[10] G. Stapleton, S. Neill, Grid-connected Solar Electric Systems: the Earthscan Expert
formance when kept cleaned, and it outperformed the module installed Handbook for Planning, Design and Installation, Routledge, 2012.
at the 10◦ tilt angle by 10% when exposed to climate conditions without [11] F. Jafarkazemi, S. Ali Saadabadi, Optimum tilt angle and orientation of solar
surfaces in Abu Dhabi, UAE. Renewable Energy. (2013) 44–49, https://doi.org/
cleaning and reduced the power losses by 40%. While the study of the
10.1016/j.renene.2012.10.036.
dust particles’ morphology and chemical compositions revealed typical [12] H.K. Elminir, A.E. Ghitas, R.H. Hamid, F. El-Hussainy, M.M. Beheary, K.M. Abdel-
characteristics aligned with the region. The results, along with the Moneim, Effect of Dust on the Transparent Cover of Solar Collectors, Energy
collected environmental data, confirmed the dust-induced shade effect Conversion and Management, 2006, pp. 3192–3203, https://doi.org/10.1016/j.
enconman.2006.02.014.
as a mechanism for the soiling impact on PV modules. Further research [13] M.A. Bashir, H.M. Ali, S. Khalil, M. Ali, A.M. Siddiqui, Comparison of performance
is essential for extensive results to acquire a broad understanding of the measurements of photovoltaic modules during winter months in taxila, Pakistan,
adhesion properties, which could improve the PV modules’ surface Int. J. Photoenergy (2014) 1–8, https://doi.org/10.1155/2014/898414.
[14] V.J. Fesharaki, I. Arbaban, J.J. Fesharaki, M. Jomardi, The effect of environment
treatments for use in industrial, dusty, and coastal regions. on photovoltaic module efficiency, Adv. Mater. Res. (2012) 1283–1286. https://do
i.org/10.4028/www.scientific.net/amr.535-537.1283.
[15] R. Siddiqui, U. Bajpai, Correlation between thicknesses of dust collected on
CRediT authorship contribution statement
photovoltaic module and difference in efficiencies in composite climate,
International Journal of Energy and Environmental Engineering (2012) 26,
Khaled S. AlZahrani: The author confirms sole responsibility for the https://doi.org/10.1186/2251-6832-3-26.
following: study conception and design, data collection, analysis and [16] S. Mekhilef, R. Saidur, M. Kamalisarvestani, Effect of dust, humidity and air
velocity on efficiency of photovoltaic cells, Renew. Sustain. Energy Rev. (2012)
interpretation of results, and manuscript preparation. 2920–2925, https://doi.org/10.1016/j.rser.2012.02.012.
[17] M. Koehl, M. Heck, S. Wiesmeier, Modelling of conditions for accelerated lifetime
testing of Humidity impact on PV-modules based on monitoring of climatic data,
Declaration of competing interest Sol. Energy Mater. Sol. Cell. (2012) 282–291, https://doi.org/10.1016/j.
solmat.2011.12.011.
[18] P.J. Verlinden, J.B. Lasich, Energy rating of Concentrator PV systems using multi-
The authors declare that they have no known competing financial junction III–V solar cells, IEEE Photovolatic Specialists Conference (2008), https://
interests or personal relationships that could have appeared to influence doi.org/10.1109/pvsc.2008.4922912, 2008 33rd.
the work reported in this paper.

9
K.S. AlZahrani Renewable Energy 216 (2023) 119117

[19] S.C.S. Costa, A. Antonia Sonia, L.L. Kazmerski, Solar Energy Dust and Soiling R&D [35] V. Komoni, A. Gebremedhin, N. Ibrahimi, A comparison of the performance of
Progress: Literature Review Update for, Renewable and Sustainable Energy mono-Si and poly-Si photovoltaic modules operating under Kosovo climate
Reviews, 2016, pp. 2504–2536, https://doi.org/10.1016/j.rser.2017.09.015, condition, J. Renew. Sustain. Energy 10 (1) (2018), 013504.
2018. [36] Rima Isaifan, Daniel Johnson, Luis Ackermann, Ben Figgis, Mohammed Ayoub,
[20] J.K.W. Chesnutt, H. Ashkanani, B. Guo, C.-Y. Wu, Simulation of microscale particle Evaluation of the adhesion forces between dust particles and photovoltaic module
interactions for optimization of an electrodynamic dust shield to clean desert dust surfaces, Sol. Energy Mater. Sol. Cell. 191 (2019) 413–421, https://doi.org/
from solar panels, Sol. Energy (2017) 1197–1207, https://doi.org/10.1016/j. 10.1016/j.solmat.2018.11.031.
solener.2017.07.064. [37] C. Schill, S. Brachmann, M. Koehl, Impact of soiling on IV-curves and efficiency of
[21] M.A. Shirakawa, R. Zilles, A. Mocelin, C.C. Gaylarde, A. Gorbushina, G. Heidrich, PV-modules, Sol. Energy 112 (2015) 259–262.
et al., Microbial colonization affects the efficiency of photovoltaic panels in a [38] G.P. Smestad, T.A. Germer, H. Alrashidi, E.F. Fernández, S. Dey, H. Brahma,
tropical environment, J. Environ. Manag. (2015) 160–167, https://doi.org/ L. Micheli, Modelling photovoltaic soiling losses through optical characterization,
10.1016/j.jenvman.2015.03.050. Sci. Rep. 10 (1) (2020) 1–13.
[22] J. Tanesab, D. Parlevliet, J. Whale, T. Urmee, T. Pryor, The contribution of dust to [39] H.Z. Al Garni, The impact of soiling on PV module performance in Saudi Arabia,
performance degradation of PV modules in a temperate climate zone, Sol. Energy Energies 15 (21) (2022) 8033.
(2015) 147–157, https://doi.org/10.1016/j.solener.2015.06.052. [40] S.A. Sadat, J. Faraji, M. Nazififard, A. Ketabi, The experimental analysis of dust
[23] K. Ilse, L. Micheli, B.W. Figgis, K. Lange, D. Daßler, H. Hanifi, et al., Techno- deposition effect on solar photovoltaic panels in Iran’s desert environment,
economic assessment of soiling losses and mitigation strategies for solar power Sustain. Energy Technol. Assessments 47 (2021), 101542.
generation, Joule (2019) 2303–2321, https://doi.org/10.1016/j. [41] N. Nader, W. Al-Kouz, S. Al-Dahidi, Assessment of existing photovoltaic system
joule.2019.08.019. with cooling and cleaning system: case study at Al-Khobar city, SAVE Proc. (2019).
[24] W. Javed, Y. Wubulikasimu, B. Figgis, B. Guo, Characterization of dust [42] Mohamed Benghanem, Abdullah Almohammedi, Mohd Khan, Al-Mashraqi Ahmed,
accumulated on photovoltaic panels in Doha, Qatar. Solar Energy. (2017) Effect of dust accumulation on the performance of photovoltaic panels in desert
123–135, https://doi.org/10.1016/j.solener.2016.11.053. countries: a case study for Madinah, Saudi Arabia, Int. J. Power Electron. Drive
[25] H.A. Kazem, M.T. Chaichan, The effect of dust accumulation and cleaning methods Syst. 9 (2018) 1356–1366, https://doi.org/10.11591/ijpeds.v9n3.pp1356-1366.
on PV panels’ outcomes based on an experimental study of six locations in [43] K. Ramaneti, P. Kakani, S. Prakash, June). Improving solar panel efficiency by solar
Northern Oman, Sol. Energy (2019) 30–38, https://doi.org/10.1016/j. tracking and tilt angle optimization with deep learning, in: 2021 5th International
solener.2019.05.036. Conference on Smart Grid and Smart Cities (ICSGSC), IEEE, 2021, pp. 102–106.
[26] T.M.A. Alnasser, A.M.J. Mahdy, K.I. Abass, M.T. Chaichan, H.A. Kazem, Impact of [44] Abdelnabi Ali, Samah Ahmed, Effect of glass and plastic covers on the performance
Dust Ingredient on Photovoltaic Performance: an Experimental Study, Solar of solar photovoltaic cells in the presence of mud, Omdurman Islamic University
Energy, 2020, pp. 651–659, https://doi.org/10.1016/j.solener.2019.12.008. Journal 18 (2022) 30–40, https://doi.org/10.52981/oiuj.v18i1.1884.
[27] A.A. Hachicha, I. Al-Sawafta, Z. Said, Impact of dust on the performance of solar [45] K.L. Johnson, Kevin Kendall, A.D. Roberts, Surface energy and contact of elastic
photovoltaic (PV) systems under United Arab Emirates weather conditions, Renew. solids, Proc. Math. Phys. Eng. Sci. 324 (1971) 301–313, https://doi.org/10.1098/
Energy (2019) 287–297, https://doi.org/10.1016/j.renene.2019.04.004. rspa.1971.0141.
[28] M.J. Adinoyi, S.A.M. Said, Effect of dust accumulation on the power outputs of [46] L.L. Kazmerski, Fundamental studies of adhesion of dust to PV module surfaces:
solar photovoltaic modules, Renew. Energy (2013) 633–636, https://doi.org/ chemical and physical relationships at the microscale, IEEE J. Photovoltaics 6 (3)
10.1016/j.renene.2013.06.014. (May 2016) 719–729.
[29] M. Benghanem, A. Almohammedi, M.T. Khan, A. Al-Masraqi, Effect of dust [47] G. Hassan, B.S. Yilbas, S.A.M. Said, N. Al-Aqeeli, A. Matin, Chemo-mechanical
accumulation on the performance of photovoltaic panels in desert countries: a case characteristics of mud formed from environmental dust particles in humid ambient
study for Madinah, Saudi Arabia, Int. J. Power Electron. Drive Syst. (2018) 1356, air, Sci. Rep. 6 (1) (Sep. 2016), 30253.
https://doi.org/10.11591/ijpeds.v9.i3.pp1356-1366. [48] G. Hassan, B.S. Yilbas, S.A.M. Said, A. Matin, Chemo-mechanical characteristics of
[30] Abdulaziz Baras, Russell Jones, Ayman Alqahtani, Maher Alodan, King Abdullah, mud formed from environmental dust particles in humid ambient air, Nat. Publ. Gr.
Measured Soiling Loss and its Economic Impact for PV Plants in Central Saudi (2016) 1–14. May.
Arabia, 2016, pp. 1–7, https://doi.org/10.1109/SASG.2016.7849657. [49] Alexander Laskin, James Cowin, Automated single-particle SEM/EDX analysis of
[31] R. Karmouch, H. El Hor, Solar cells performance reduction under the effect of dust submicrometer particles down to 0.1 μm, Anal. Chem. 73 (2001) 1023–1029,
in jazan region, J. Fund. Renew. Energy Appl. (2017), https://doi.org/10.4172/ https://doi.org/10.1021/ac0009604.
2090-4541.1000228. [50] E.A. Reid, et al., Characterization of African dust transported to Puerto Rico by
[32] Wasim Javed, Bing Guo, Ben Figgis, Brahim Aïssa, Dust potency in the context of individual particle and size segregated bulk analysis, J. Geophys. Res. 108 (2003)
solar photovoltaic (PV) soiling loss, Sol. Energy (2021) 1040–1052, https://doi. 8591–8598.
org/10.1016/j.solener.2021.04.015, 2020. [51] Mohamed Benghanem, Abdullah Almohammedi, Mohammed Khan, Al-Masraqi,
[33] EDAX Inc, Texture and Elemental Analytical Microscopy, 2017, Version 4.5. Ahmad, Effect of dust accumulation on the performance of photovoltaic panels in
https://www.edax.com. desert countries: a case study for Madinah, Saudi Arabia, Int. J. Power Electron.
[34] Al-Ghezi, Moafaq Kaseim, Ahmed, Roshen Tariq, Chaichan, Miqdam Tariq, The Drive Syst. 9 (2018) 1356, https://doi.org/10.11591/ijpeds.v9.i3.pp1356-1366.
influence of temperature and irradiance on performance of the photovoltaic panel [52] S.A.M. Said, N. Al-Aqeeli, H.M. Walwil, The potential of using textured and anti-
in the middle of Iraq, Int. J. Renew. Energy Dev. 11 (2) (2022) 501. reflective coated glasses in minimizing dust fouling, Sol. Energy 113 (2015)
295–302.

10