Comparative Performance Analysis of PV Module Positions in a Solar PV Array Under Partial Shading Conditions

The irradiation intensity is the primary catalyst for energy production in photovoltaic (PV) systems. Nevertheless, it alleviates partial shading (PS) by inhibiting power production, primarily in habitable metropolitan areas, owing to surrounding objects or high soiling rates, particularly in dry and semi-arid climates. Hence, the PV module’s location, orientation, and tilt angle significantly affect its performance and shorten its durability. Thus, this study undertakes a benchmark analysis of the two distinct panel positions’ performance using PSIM and MATLAB software simulations and an actual practical test. A visual inspection of a solar power plant in Morocco’s Green Energy Park is the origin of the survey concept; it observes that the soil density consistently covers certain PV panel spots. In this context, the solar energy data of an intelligent home deployed in the earlier-mentioned research platform has been deployed to emulate the performance of future buildings challenged with various obstacles. Therefore, the roof-mounted PV system satisfied the home’s energy demands. In the first scenario, the PV module is in portrait mode. In the second case, they are in landscape orientation. The findings of this investigation demonstrate that standard PV panels produce more power when arranged in a landscape configuration than in a portrait configuration, exhibiting a discrepancy of up to $1010 Wh$ for a modest PV system. Finally, this study suggests appropriate orientation and design recommendations for standard and advanced solar modules, mainly those deployed in arid and heavily populated urban regions featuring severe shading constraints.


I. INTRODUCTION
The world's skyrocketing energy needs, concern over pollution levels, and depletion of fossil fuels are hallmarks of the contemporary era, promoting the integration of renewable energies to tackle the issues mentioned earlier. However, their contribution to meeting the world's energy demands remains uncertain [1], [2]. Consequently, universities, governments, The associate editor coordinating the review of this manuscript and approving it for publication was Tarek H. M. EL-Fouly . and industries have discovered several sustainable energy resources that can steer away from traditional sources. Moreover, accelerate the energy transition to clean resources that satisfy the energy requirements of multiple sectors, including logistics, supply chain operations, and power grid integration [3]. According to the International Energy Agency Renewables 2020 prediction study, renewable energy will account for 95 % of the newly installed power production by 2025. At the same time, solar energy will account for 60 % of all newly established renewable energy sources.
To date, global solar PV capacity addition is expected to reach approximately 160 × 10 9 W [4], [5]. Morocco is among the Middle East and North African (MENA) countries that receive tremendous solar irradiation, with around 3000 hours of sunlight per year and approximately 5000 Wh/m 2 /day of irradiation, especially during summer [6]. Moreover, the geographical location of Morocco makes it particularly appealing for solar, wind, and hydroelectric power production. Furthermore, it is among the African nations, involving a power transmission link to Europe. Therefore, it has attracted significant interest as a promising energy source by investing in its efforts and providing financial incentives to establish a gigantic renewable energy plant to fulfill its energy needs and international ''Net Zero'' targets by 2050 [7]. Unfortunately, most Moroccan territories experience harsh outdoor conditions, notably humidity, temperature, and high airborne dust rates [8]. Consequently, it suffers from dust deposition on the PV module surface. Hence, a photovoltaic system would deliver a relatively moderate power output because the soiling issue not only obstructs or scatters incoming irradiation but also absorbs heat, thus increasing the cell temperature. Indeed, the cells immediately positioned at the rear of the dust collection are severely damaged [9]. Sulaiman et al. conducted experiments demonstrating that dust deposited on a glass plate tilted at 45 • inhibited transmissions by 18% and 30% on average after ten and 30 days, respectively [10]. Furthermore, Kichou et al.'s research indicated that the impact of pollution on the optical and electrical degradation of PV modules strongly depends on the technology utilized [11], [12], [13]. A recent study highlighted that the average yield degradation rates are 1.19 %, 1.17 %, and 1.67 % per year for polycrystalline, monocrystalline, and thin-film cadmium chloride, respectively [14]. Numerous experts have cited dust accumulation on PV panel glass as a primary concern [15]. Depending on the PV module orientation and tilt angle, they may spread evenly or irregularly across the solar module surface, collecting at the edges and corners [16], [17], [18]. Specifically, humidity, wind patterns, and the nature of dust vary according to the region's morphology, chemistry, and composition [19], [20], [21]. For instance, solar power plants in the MENA region differ from Nordic ones because deserts are the world's most outstanding mineral dust reserves. In 2014, Ghazi investigated global dust accumulation patterns and proved that the MENA region had substantial dust accumulation zones [22]. Chanchangi et al. investigated the correlation between climatic conditions, dust deposition, rebound, resuspension, and adverse effects on PV module performance [23]. As a consequence of aerosols absorbing moisture and gravitationally dropping onto surfaces, the above-described conditions promote dust settling on the PV modules. Furthermore, once the air is dry, the sand droplets transform into dust particles, generating a thin layer that covers the PV surface. Therefore, dust accumulation acts as a direct barrier to partial shading (PS) issues, representing a severe barrier to solar energy production. PS is a significant mismatch loss that may create hotspots owing to the uneven radiation exposure throughout the PV module [24]. As a result, shaded cells provide insufficient power and are constrained to flow the same current as units capable of producing high currents. Thus, the shaded units act as loads rather than generators, increasing the temperature and harming the PV array [25]. Moreover, statistical analysis demonstrated that the annual energy production reduction associated with shadows in residential applications ranged from 10% to 20% [26]. Besides, the shading effects on the edges and pillars waste approximately 18% to 35% of energy [27]. Hence, a regular and widespread practice to mitigate hotspot failure is to insert a bypass diode across a single cell or a set of cells to provide alternative flow paths for the current stemming from unshaded subgroup units. However, the reliability of bypass diodes is still being determined owing to scarce evidence, apart from infrequent experimental research on electrical and thermal characterization [28], [29], together assessing overlapping and nonoverlapping bypass diode topologies [30]. In addition, the multiple peaks in the power-voltage (P-V) curve make maximum power point tracking (MPPT) challenging. Therefore, although a strategy based on I-V characteristics may adequately explore PV array performance, it must typically explain the MPPT process [31], [32], [33]. Thus, the PS restriction distorts the related maximum power point (MPP) voltage using global maximum power point tracking methods [34], [35].
Similarly, cleaning is an effective strategy appended to the improved methodology list that removes impurities and extends the PV module's lifetime. However, the expense of such operations and proper planning are significant constraints [36], [37]. To obtain the best payback, PV system size, dusty climate, and seasonal sequence are critical considerations for establishing an optimal cleaning program. In this context. Extensive research has also addressed cleaning issues, including lists of methods and particle resuspension theories for assessing dust-cleaning procedures. The wind direction and velocity determine the dust distribution and removal from the PV module surfaces [38], [39]. Conventional use prefers interconnecting PV panels in a series-parallel (SP) topology to provide the required electrical parameters. Nevertheless, a solar PV plant may experience partial shading during operation, drastically lowering its power production. For this purpose, literature reviews have revealed several strategies for altering solar panel configurations, enhancing their efficiency, and mitigating shadowing effects [40]. The least obtrusive shading scenarios may be shown through experimentation with mathematical puzzles, unlike those inspired by classic approaches, such as Su-Do-Ku, Latin Square, and Magic Square [41], [42]. Considering the aforementioned studies in the literature. Figure 1 provides an overview of the techniques for enhancing the energy output of solar panels and their direct and indirect impacts on partial shadowing and dust buildup on the PV panel surface [43]. Ultimately, the framework aims to achieve an optimum solar panel arrangement for increased energy production under partial shade or in a dusty environment. The main contributions of this study are as follows: First, it sheds new light on PV systems in urban environments, including an unexpected increase in the number of buildings likely to cast shadows. Second, it urges researchers to redesign PV modules by rearranging the PV cells inside the PV modules in such a way as to remove the partial shading effect. Indeed, to reconsider the half-cut module and bifacial PV system. Third, the solar panels are rehabilitated in confined areas and rooftop spaces. Finally, we prioritize using machine-learning techniques to estimate and develop a PV reference model without requiring significant equipment.

A. PAPER ORGANIZATION
The rest of the paper is briefly organized as follows: • The second section covers current solar cell modeling in the literature and demonstrates various PV panel interconnections.
• The third section is devoted to the simulation process to investigate the PV module behavior under uniform and non-uniform radiation distributions.
• The fourth section delves into the experimental methods for orientation situations using a 3D model and a PV array reference model description using an artificial neural network tool.
• Finally, a complete and concise finding, a discussion section, and a clearly stated conclusion outlining the merits and drawbacks of such PV orientations highlight prospective avenues for further research (see Figure 2).

II. EQUIVALENT CIRCUIT OF PHOTOVOLTAIC CELL
A photovoltaic (PV) cell, also called a solar cell, is a renewable energy technology that directly converts solar radiation into electricity by adopting the physical mechanism of the PV effect. The PV cell consists of two independent thin layers, each differently doped with electrons, and holes overloaded with semiconductor materials. Once the P-N junction is exposed to sunlight, it absorbs photon energy exceeding its bandgap energy to form barriers, namely electron-hole pairs [44]. In contrast, excess photon energy is converted to heat. Modeling PV cells is, thus, an indispensable step and a deciding factor in PV module simulation, granting an in-depth analysis of system management and the assessment of the temperature and irradiance impact on the panel's behavior. Various PV models, including implicit and explicit models, have been published and applied in the scientific literature to determine the current-voltage (I-V) and power-voltage (P-V) curves. However, explicit options require more computational work than implicit options, which require little computational effort [45]. Furthermore, they involve basic analytical formulas that provide researchers with a means to approximate the critical solar cell characteristics (Eq 1, 2, 3, and 4). As shown in Figure 3, the photovoltaic cell equivalent circuits most commonly cited in the literature are as follows (1D1R, 1D2R, 2D4R, 2D2R, 3D2R, 3D5R, and xD2R). Nevertheless, 1D2R (designated as a single-diode model) is the recommended option among the previously stated PV diode variants [46], [47]. Consequently, the current study selected a single-diode model because of its simplicity, conceptual ease, and a few vital components (IP V, Id, Rsh, Rs, and a2 ). The equivalent circuit comprises a parallel-mounted diode with a shunt resistor to mimic a PN junction. It was then coupled with a series resistor to reflect the internal resistance of the PV cell (see Figure 3 (b)). (2) where I PV : Represents the photocurrent., The PV module is made up of PV cell packages that are connected serially to increase the output voltage. The cells are grouped into subparts that are connected in parallel with the bypass diode. Therefore, panels in typical large-scale PV arrays may be configured in series, parallel, or a combination of both topologies. Several interconnection arrangements exist in the literature, including (S, P, SP, HC, BL, TCT, and forced TCT, as well as hybrid topologies such as (SP-TCT, BL-TCT, and BL-H). Figure 4 illustrates basic topologies [48], [49].

III. SOLAR RADIATION DISTRIBUTION EFFECT A. PV CHARACTERISTICS UNDER UNIFORM IRRADIANCE
The commercial PV module considered in this subsection is a standard Almaden PV panel comprising 72 cells connected in series and arranged on three substrates, each of which has 24 units. The module's cell connectivity layouts are shown in Figure 5. The Almaden SEA P72T polycrystalline panel used in this research is mounted on the test bench. Figure 6 depicts the P-V and I-V curves of the PV module under different levels of homogeneous irradiation, ranging from 200 W /m 2 to 1000 W /m 2 . Table 1 describes the characteristics of the module under standard test conditions (STC).

B. PV CHARACTERISTICS UNDER PS CONDITIONS AND THE SIMULATION PROCEDURE
Photovoltaic panels behave as loads rather than energy sources when temporarily or permanently darkened by dust or partial shading. Consequently, long-term shading could lead to destruction. As a remedy, the current of the unshaded cells is routed via an antiparallel mounted bypass diode in each cell subarray, preventing the overheating of the uncovered cells [50]. In addition, the P-V curve exhibits multiple peaks during the bypass diode activation, where the global peak corresponds to the highest power produced. As part of the simulation process, the electrical characteristics are measured while evaluating the recommended landscape and portrait modes of PV panels under extreme soiling and shading conditions using a single diode model designed in the PSIM software. In the first shaded case, the PV module is mounted in portrait orientation and receives 1000 W /m 2 of irradiation. In contrast, the shaded cells absorbed 200 W /m 2 , as shown in Figure 7. Similarly, the second shading scenario necessitates mounting the solar panel in landscape orientation. Furthermore, the shaded cells underneath received 200 W /m 2 irradiation. In contrast, unshaded cells are subjected to 1000 W /m 2 of irradiation. Figures 8 and 9 show the simulation of cell shading on the underside of the PV modules mounted in the landscape orientation. In addition, the Figure 7 depicts the cell-shading view at the bottom of the PV modules arranged in portrait orientation.
Consequently, these actions occasionally occur at the Green Power Park research platform. For instance, Figure 10 highlights the dust deposition on the PV module surface, including wet and dry accumulations. Furthermore, dust settles on the module's bottom surface owing to gravitational force and the panel frame capturing impurities. Likewise, Figure 11 shows the afternoon shadow cast from the first PV string on the remaining strings of an intelligent home built on the Green Energy Park research platform. The P-V curve showed a single peak under normal operating conditions, depending on the acquired results presented in Figure 6. Nevertheless, it produces several peaks owing to mismatched circumstances caused mainly by non-uniform irradiation. A limited current is produced by the darkened  cells. Therefore, the current can freely circulate through illuminated cells without difficulty. However, the current produced by the well-lit cells is sufficient to harm the poorly lit cells, which mandates the alternative channel. The operating principle of the PV module and bypass diode behavior will drive our investigation regarding the scenarios mentioned above for the following explanation: • First Scenario: This study adopts solar panels consisting of three cell clusters, represented by units called subparts cells, connected serially. The shading pattern applied at this instant covered only a single subpart, resulting in two peaks in the P-V curve, as shown in Figure (8 (c)). Each height corresponds to the power output level. The darkened sub-part corresponds to the local peak power. Therefore, it exhibited a mild current that could flow freely through the unshaded cells without enabling the bypass diode. Similarly, the voltage correlating to this current equals the sum of the subpart's voltages. Figure (8 (a)) emphasizes (in green) the paths of the current produced by the failing subpart cell. The Figure (8(b)) Figure (8 (a)) emphasizes (in green) the paths of the current produced by the failing subpart cell. Figure (8 (b)) depicts the P-V curve of the shaded subparts, where the well-illuminated subfragment components generate a significant current that may induce irreparable damage to the shaded cells. Therefore, the bypass diode acts as an alternating current canal. Furthermore, the voltage proportional to the current generated by a well-lit subpart is the sum of the voltages omitted by the darker subpart. A red line in Figure (8 (a)) marks the current path. Besides, Figure (8 (b)) depicts the P-V curve of the lit subparts. Figure 9 illustrates the second pattern, wherein the partial shading mode is identical to the preceding stage. However, the global power peak is significantly lower than in the prior scenario, owing to the influence of the single irradiated subpart on the remaining two highly illuminated subparts.
• Second Scenario: In the second shading scenario, displayed in Figure 7, the shading is spread over the entire cell row underneath the PV module, reflecting the actual dust deposition and surrounding objects that undoubtedly provoke partial shading. Nevertheless, the power achieved is modest, and the resulting P-V curve incorporates a single peak, denoted as the panel that undergoes full shading. In addition, the mismatch among the subparts is non-existent, allowing the current to flow through the cells without enabling a bypass diode owing to the insufficient current. The green line indicates the current flow route. At the same time, Figure (7 (b)) represents the P-V curve of the PV module under partial shading conditions.
According to the simulation results of the two shading scenarios applied to a typical PV module, shading a single subpart in a landscape-mounting orientation led to almost one-third of the module's total power. Similarly, twothirds of the module's overall capacity is lost when the two cell subparts are shaded. Thus, integrating practical MPPT metaheuristic algorithms is required to track the global maximum power point (GMPP) in such situations. In contrast, the second solar panel is mounted in portrait orientation. As a result, the shadow surrounding the module's footer cells causes the P-V curve to possess a single peak, delivering a maximum output power proportional to the full shading of the PV module. This results in a drastic decrease in power generation compared to the previous state. Typically, the MPPT strategy recommended for such circumstances is conventional MPPT.

IV. EXPERIMENTAL PROCEDURE
The experimental procedure of this investigation was to establish a photovoltaic array consisting of nine PV modules connected serially to meet the inverter's current and voltage requirements to support the simulation. Furthermore, it fulfills the daily energy requirements of an intelligent house; the experiment proceeds in two stages. The first step involves mounting the PV modules in portrait orientation for two  months, starting on November 7, 2021, and ending on January 2, 2022. After that, the modules will operate in landscape orientation starting January 3, 2022. The overall capacity of the considered PV system is 3000 W . Table 2 lists their technical and mechanical characteristics. Figure 12 presents an overview of the test bed architecture. Eventually, the PV modules are installed in both orientations to ensure proper alignment. Furthermore, as previously mentioned, the test coincides with the solstice period, when the sun's path is at its lowest position. Consequently, the shade afforded by the strings is crucial, particularly in tiny metropolitan areas. The soiling issue is also included indirectly in the following section, as the shading covers premises identical to the soiling observed in solar units at the Green Power Park facility. Likewise, shading is transient, unlike soiling, which is permanent as long as specific prerequisites such as precipitation are missing. Similarly, solar system architecture and sizing are extensively evaluated to determine the most delicate PV module placements. However, this geographical area occasionally experiences unusual wind speeds, which may destroy the modules on the rooftops of the buildings. Furthermore, authorized interspacing between strings is     unrespected due to space challenges. The interspacing chosen for this case is appropriate based on preliminary 3D tests using sketch software, as shown in Figure 13.

A. EXPERIMENTAL SETUP FOR PORTRAIT ORIENTATION
In this scenario, the modules are arranged in portrait orientation, as shown in Figure 11. Four days are scheduled for display during data assignment. Namely, most days during this period are cloudy, and the designation of these days is neither random nor planned. Therefore, choosing such a date is unavoidable because of the relative insolation. Furthermore, the panels are adequately cleaned during testing to prevent further environmental effects and to focus exclusively on a single delimitation, bearing in mind that the evaluation depends on any fate that directly or indirectly induces shadowing on the photovoltaic cells within the same model. Figure 14 shows the power curves generated during the experimental period.

B. EXPERIMENTAL SETUP FOR LANDSCAPE ORIENTATION
The PV units that experienced shade exposure during the previous test are rearranged in landscape orientation, as depicted in Figure 15. Consequently, four days are allotted for examination, as previously performed for the portrait orientation trial. Hence, selecting these days has proven highly challenging. Moreover, the experiment is conducted at a specified time because the string's shadow appears most prominently when the sun's path is at its lowest point. Therefore, the changeover phase is identified as the solstice period. Figure 16 depicts the power output profiles of the PV plant.

C. REFERENCE MODEL DEVELOPMENT
To achieve a baseline model of the photovoltaic power plant considered in this framework, the performance of each scenario has been evaluated by comparing the power output for the instances mentioned earlier. Hence, the ANN model is applied with a certain degree of confidence to quantify the power profile generated by the photovoltaic power plant without experiencing mismatch failure. Also, the ANN model used several key performance metrics to figure out the recommended PV module position.

1) DATA ACQUISITION
The dataset utilized in this investigation spans the winter months of November 2021 to February 2022, during which significant temperature and solar irradiance fluctuations are anticipated. Figure 17 illustrates the test bench's electrical and meteorological data acquisition structures. At the same time, Figure 18 provides an aerial view of the electrical and meteorological station locations, which are close to 350 m. Thus, it may enhance the efficiency and accuracy of the ANN model.

2) DATA PRE-PROCESSING
Noise, missing values, and even inappropriate formats are standard features of real-world data that may or may not be directly relevant to an ANN model [51]. Therefore, VOLUME 11, 2023   data preparation involves cleaning and restoring the data to acceptable requirements to enhance the accuracy and efficacy of the machine learning process. Although data preparation methodologies are readily available for the PV domain, they typically provide relatively generic procedures and concepts that require no in-depth understanding [52], [53]. Consequently, the first stage included resampling and synchronizing the data sources. The raw data is then filtered to identify and eliminate outliers before being replaced with nan values. Figure 19 shows the data preparation steps involved.

3) ARTIFICIAL NEURAL NETWORK MODEL
An artificial neural network (ANN) is a distributed parallel process acting as a single unit capable of sorting and processing information from experience. A neural network is analogous to the human brain in two aspects. First, the network learns from inputs during the process [54], [55]. Second, the information is recorded via synaptic weights, indicating the connection strength between the neurons. The neural network model used in this study included multilayered perceptrons. The input layer is the starting layer and includes various input variables. The final layer is the  output layer that generates the expected results of the model. Multiple hidden layers between the levels interconnect the input and output layers. Once the training and test samples are established, a neural network design can be developed. The number of hidden layers and neurons in each input and hidden layer are the key attributes of ANN configurations. The selected ANN paradigm consisted of one input layer with ten neurons and two hidden layers with ten and five neurons, respectively. Thus, the root mean square error (RMSE), mean square error (MSE), and mean absolute error (MAE) are assessed to ascertain the optimal neural network model. The values are 0.096, 0.098, and 0.08, respectively. The figure illustrates the architecture of the ANN model. In addition, this project integrated a backpropagation (BP) algorithm to enhance learning. The BP has been identified as a robust supervised learning algorithm. Therefore, with an accuracy of 98.8%, the proposed ANN model is close to the reference model for both PV panel instances. Figures 21 and 22 show the results obtained when the ANN reference model is applied to test bed operation under trouble-free circumstances. VOLUME 11, 2023

D. RESULTS AND DISCUSSION
The analysis of the energy curves, modeling, and experimental results for the mounting orientation scenarios are presented in the following section. Besides applying key performance indicators, including energy production and loss, the results acquired for each orientation mode are compared to the ANN model's results. Thus, the current subsection summarizes the effective area influenced by the four module designs under the foot row shadow in both portrait and landscape modes. Finally, the ''Appendix'' section briefly summarizes the simulation results for all the PV orientations.

1) PORTRAIT ORIENTATION
The daily energy curves and bar graphs produced under shadowing demonstrate the energy profiles, as illustrated in Figure 23. This divulges a considerable fall in energy during the afternoon after the shade obscures the row footer of the PV panel. Under such conditions, the regular energy generated between 14 : 50 and 16 : 00 (hour) is significantly mitigated. For instance, the power generated on December 11, October 12, January 17, and January 18 are (733 Wh and     respectively, as illustrated in Figure 25. Meanwhile, the following maximum power point voltages (216 V , 218 V , 213 V , and 217 V ) were derived via a computation procedure utilizing the easy-to-implement GMPP model theory published by chalh at al. [56], once the system was undergoing a significant drop. To determine the voltage values associated with maximum power under the partial-shading effect (see Equations 5, 6, and 7) [56], [57]. In counterpoint, the voltages associated with the extracted power over a predisposing period are (266 V , 268 V , 281 V , and 259 V ), respectively. Therefore, PV systems undoubtedly track the local maximum power point.
2) LANDSCAPE ORIENTATION The resorting daily energy curves and bar graph show the energy profile, shown in Figures 22 and 24, accompanied by significant growth in energy compared to the previous   scenario. Indeed, the energy produced between 14 : 50 and 16 : 00 (hour) is sufficient. For instance, the power produced on January 3, January 7, January 12, and January 14, respectively (2020Wh and 1140Wh), (1841Wh and 1044Wh), (1983 Wh and 1089 Wh), (2184 Wh and 1304 Wh) ). Bayside, the reported energy loss on the aforesaid dates is 120 Wh, 250 Wh, 293 Wh, as seen in Figure 25. There is a limpid improvement in the amount of power lost in the landscape model case. During the period under consideration, the maximum power point voltage extracted is consistent with that acquired by performing a computation utilizing the easyto-implement GMPP model theory, with some exceptions.

3) ANALYSIS
The extensive deliberation afforded by the power curve and bar chart has led the authors to highlight their impact on the performance of mounted panels in landscape orientation. The latter instance runs perfectly, addressing shadows, including those thrown on the PV panel's underside, dust deposition VOLUME 11, 2023 along the edge, and any other factors that may involve the formation of an anti-radiation layer on the PV panels. Furthermore, the power loss closely corresponds to the shaded area once the PV module foot is shaded. Unlike in portrait mode, when the PV panel beneath is obscured, the panel operates as fully shaded. Furthermore, the photovoltaic unit design successfully mitigated the shading consequences on power production, independent of the shadow source. This is shown by a series of simulations on different PV modules, including traditional and newer ones, as shown in Tables 3 and 4 (see appendix section).

V. CONCLUSION
In this contribution, the author introduces a theoretical, simulated, and concrete investigation of the different orientations of solar modules under partial shading and dust constraints. Furthermore, a simulation assignment is performed for conventional and advanced models to simulate the resulting energy required to ensure the injection of photovoltaic energy into new-generation power grids. This research focuses on photovoltaic installations located in areas with mandatory panel inclination and shading issues. This survey highlights certain main conclusions that can be summarized as follows: • Identifying the orientation of the cell subparts is recommended as a preliminary step to properly mount the PV modules under partial shading and dust accumulation conditions.
• Improper orientation can lead to substantial energy losses exceeding 1000 Wh for a PV plant up to 3000 W when the PV panels are mounted in portrait mode.
• Advanced MPPT inverter technology may provide different energy loss rates than those indicated in this study; however, the landscape is significantly superior to the portrait mode. Ultimately, this study introduces a perspective on the optimal PV panel mounting orientation, whereby the authors recommend using a landscape orientation based on the acquired results. The future of renewable energy will undoubtedly become everywhere to fulfill energy demands; hence, this research will serve as a powerful reference for scholars planning to conduct research in the MENA region or metropolitan areas. Furthermore, such research remains valid when the PV panel is tilted.

APPENDIX
See Tables 3 and 4.