Two sites were selected for the collection of fine particles intended for testing as part of this research work. These are two sites located in the northern zone of Cameroon which is an area south of the Sahara and where the dry season lasts around seven months [19] - [22] . A period long enough for a large quantity of dust to accumulate on the surface of photovoltaic panels. It is a sufficiently sunny area with an average irradiance estimated at 5.8 kW/m²/day [23] [24] . The first site is located in the Adamaoua region with geographic coordinates: 7˚20' north, 13˚30' east and whose dust is characteristic of the nature of the soil which is lateritic. The second site is located in the Northern region with coordinates: 8˚30' North, 14˚00' East and whose dust results from silty soil. The average annual temperature in the Adamaoua region is 22˚C with a peak reaching 33˚C during the month of March which is the hottest month. The average annual temperature in the Northern region is 29˚C with a peak reaching 40˚C during the month of March which also represents the hottest month [25] . Figure 1 shows the dust samples taken from the two study sites. The dust collected in Ngaoundere is of a lateritic type and that of Garoua is of a silty type. These dust characteristics are specific to the type of soil found in the two study areas respectively.

The characteristics of the photovoltaic panels used for the experiments are recorded in Table 1 below. The electrical characteristics of the two panels are almost identical in order to facilitate comparison studies. The two types of panels are those which mainly equip photovoltaic systems in the study regions, monocrystalline and polycrystalline.

STC: 1000 W/m²; 25˚C; AM 1.5.

Figure 2 presents the experimental setup used for tests. The experiment was conducted using a 100 W monocrystalline panel and a 100 W polycrystalline panel, the characteristics of which are presented in Table 1 . The measuring devices used during the tests can be identified on the experiment set-up as follows:

(3): Spectroradiometer

(4): Thermocouple

(5): Solar Power Meter

(6): Multimeters

(7): Solar Panel multimeter

For measurements of electrical parameters, multimeters are used. Temperature probes connected to thermocouple are also used to measure the surface temperatures of photovoltaic panels. But the ambient temperature is measured by the spectroradiometer. The solar Power meter is used to measure the solar radiation. The Solar Panel Multimeter allowed to check the operation of our panels and measure the parameters specific to electrical panels. It can test the maximum power point and open-circuit voltage. It provides automatic and manual MPPT detection, over-voltage, over-temperature, over-current protection. Table 2 summarizes the specifications of measuring equipment that is used in this study.

The main objective of this work is to assess the impact of fine particles deposition on the surface of monocrystalline and polycrystalline panels. Two dust samples are collected from two different sites. These are lateritic dust and silty dust. Results from experimental work aimed at assessing the density of dust deposited on the surface of panels in areas located in the south of the Sahara show that this density can reach 30 mg/cm² [17] [26] . Based on the results of this experimental work, and following preliminary tests, the dust densities selected for testing in this work cover the range from 0 mg/cm² to 25 mg/cm² in increments of 5 mg/cm². The samples are prepared and packaged in boxes, each containing a quantity corresponding to 5 mg/cm² for both samples: i.e. 24.2 g per boxe. A test day begins by cleaning the panels with clean water and a soft sponge. The panels are then exposed to sunlight and oriented due south at a 22˚ inclination. The measuring devices and loads are then connected to read the following parameters: current, voltage, ambient temperature, panel surface temperature, and irradiance. The dust particles are uniformly spread on the flour strainer whose sieve size is less than 160 micron meter. A test day starts at 9 a.m. and ends at 4 p.m. The chosen time range corresponds to the period of a day of sunshine in the study areas (approximately seven hours of sunshine) [27] [28] . Each test day corresponds to a dust density applied to the panel surfaces. One test series covers exactly six days, which corresponds to densities 0, 5, 10, 15, 20 and 25 mg/cm². A total of five test series were conducted:

Equation (1) allows to determine the power loss factor due to dust deposition as follow:

$P_{\left(\%\right)reduction}=100\times \frac{P_{clean}-P_{dust}}{P_{clean}}$(1)

where $P_{pp}$ represents the power generated by the clean photovoltaic panel and $P_{dp}$ is the power generated by the panel covered with fine particles. The actual conversion efficiency of the panel is given by Equation (2) below:

${\eta }_{module}=\frac{P_{\max }}{\left(G\times A\right)}$(2)

where $P_{\max }$: is the maximum power produced by the panel in Watt; G: irradiance in W/m²; A: the surface area of the panel in m².

The percentage reduction in yield is given by the following Equation (3) [29] :

${\eta }_{\left(\%\right)reduction}=\frac{{\eta }_{clean}-{\eta }_{dust}}{{\eta }_{clean}}\times 100$(3)

where ${\eta }_{clean}$ is the efficiency of the pure panel then ${\eta }_{dust}$ is the efficiency of the panel covered by fine particles.

We know that temperature and irradiance are two essential weather factors that have a direct influence on the performance of solar panels. However, as much as the increase in irradiance increases the short-circuit current generated by the photovoltaic cell, the temperature on the surface of the cell leads to a relatively significant drop in the open circuit voltage, and consequently a drop in the power generated. In the area south of the Sahara where we are carrying out this study, the increase in solar irradiance is accompanied by an increase in ambient temperature, unlike in humid tropical areas where the two phenomena do not necessarily have a linked behavior. Analysis of the curves in Figure 3 shows that the ambient temperature during the test period is an increasing function from 9 a.m., the start time of the readings, until 4 p.m. The irradiance is an increasing function from 9 a.m. until 12 p.m. where it reaches a value of 800 W/m² and it decreases to 320 W/cm² around 4 p.m. We note that from 1:15 p.m., while the irradiance decreases more quickly, the ambient temperature varies around an average of 35˚C, higher than 25˚C, the standard temperature of the test conditions, which contributes to the continual heating of the photovoltaic panel. We can conclude from the analysis of these curves and in the test conditions that beyond 1:15 p.m., the weather conditions are not favorable for optimal power production because the combined effects of temperature and irradiance are not favorable.

Figure 4 shows the ambient temperature and surface temperature profiles of the monocrystalline and polycrystalline panels over time in the study area. We notice that the cell temperatures increase from the start of the tests to reach an optimum around 12 h for the monocrystalline and at 12:15 pm for the polycrystalline. Afterwards, these temperatures gradually drop until the end of the tests at 4 p.m. It can be seen that throughout the testing period, the surface temperature of the monocrystalline panel is higher than that of the polycrystalline panel. The peak temperature on the surface of the monocrystalline panel is 58.1˚C for an ambient temperature of 34.1˚C while the peak temperature of the polycrystalline panel is 53.6˚C. Over the entire test day, the largest temperature difference between the monocrystalline and the polycrystalline is 6.8˚C and recorded at 11 a.m. Considering these results, we note that when the two types of panels are exposed to the same ambient temperature, the rise in temperature on the surface of the monocrystalline panel is greater and faster than that of the polycrystalline panel. This shows that in the study area, the power delivered by the monocrystalline panel will be more negatively affected compared to that delivered by the polycrystalline panel.

Figure 5 , consisting of two rows of 6 graphs, represents the power profiles generated for different densities of lateritic and silty type dust applied to the surface of the two types of monocrystalline and polycrystalline panels. The curve in blue is that of monocrystalline and the curve in red is that of polycrystalline. The curves in the left row are those obtained with lateritic dust while those in the right row are those obtained with silty dust. We observe that for all the graphs, the power generated by the polycrystalline panel is greater than that generated by the monocrystalline panel. This observation is also valid for empty tests, i.e. with clean panels. These results tend to call into question the literature data which stipulates that the power generated by the monocrystalline is greater than that generated by the polycrystalline panel with the same electrical characteristics and under the same test conditions. These results can be explained by the influence of the higher temperature gradient on the surface of the monocrystalline panel which would lead to the consequent drop in the power generated compared to that of the polycrystalline panel. We clearly observe the impact of the density of dust deposited on the surface of the panels through the power levels generated.

Figure 6 and Figure 7 present the profiles of the conversion efficiencies of the monocrystalline and polycrystalline panels covered by lateritic and silty dust respectively. For both types of dust, we note that the conversion efficiency of polycrystalline is higher than that of monocrystalline. We can say in this case that regardless of the type of dust, the monocrystalline panel is more impacted by the deposition of dust than the polycrystalline panel. This result confirms those of the work in [30] , but contradicts the results obtained in [18] . Some particularities emerge from the analysis of the profiles of the curves depending on whether it is lateritic or silty dust. Indeed, in the presence of lateritic dust, the gap between the conversion yields between mono and poly gradually narrows and the two

yields become almost equal and very low for densities greater than 15 mg/cm². As for silty dust, we also notice that the gap between the conversion yields is narrowing but they become almost equal and very low at 25 mg/cm². Considering these results, we can see that lateritic dust has a greater impact on the conversion efficiency than silty dust.