SEARCH

SEARCH BY CITATION

Keywords:

  • dust;
  • spectral response;
  • amorphous silicon;
  • crystalline silicon;
  • transmittance;
  • soiling ratio

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODOLOGY
  5. RESULTS AND DISCUSSION
  6. CONCLUSION
  7. REFERENCES

The effect of dust on photovoltaic modules is investigated with respect to concentration and spectral transmittance. Samples were collected in the form of raw dust as well as accumulated dust on exposed sheets of glass at different tilt angles. Spectral transmittance of the samples was determined. Transmittance variation between top, middle and bottom was identified for samples collected at different inclinations, where the worst case was seen at a tilt angle of 30o with a non-uniformity of 4.4% in comparison with 0.2% for the 90° tilt. The measured data showed a decrease in transmittance at wavelengths <570 nm. Integrating this with measured spectral responses of different technologies demonstrates that wide band-gap thin-film technologies are affected more than, for example crystalline silicon technologies. The worst case is amorphous silicon, where a 33% reduction in photocurrent is predicted for a dust concentration of 4.25 mg/cm2. Similarly, crystalline silicon and CIGS technologies are predicted to be less affected, with 28.6% and 28.5% reductions in photocurrent, respectively. The same procedure was repeated with varying Air Mass (AM), tilt angle and dust concentration values to produce a soiling ratio table for different technologies under different AM, tilt angle and dust concentration values. Copyright © 2012 John Wiley & Sons, Ltd.

INTRODUCTION

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODOLOGY
  5. RESULTS AND DISCUSSION
  6. CONCLUSION
  7. REFERENCES

Sand dust is expected to be a detrimental agent in most arid zones of the world, at least as far as solar energy applications are concerned. When particles are deposited on photovoltaic (PV) modules, they interfere with illumination quality by both attenuating and scattering light. The degree to which the particles interfere depends on their constitution, density and size distribution [1]. Particles impinge onto a surface due to gravity, electrostatic charge or mechanical effects (wind or water droplets). After deposition, they are held by a charge double layer, surface energy effects and capillary effects, in addition to gravity and electrostatic forces [2, 3].

Dust is one of the natural elements present in most environments. The particle size and composition depend on the location [4]. In some regions, dusty weather conditions tend to be more severe than in others. For example in Kuwait, dust is present in 27% of daylight hours throughout May to August (Figure 1). It causes deterioration in visibility during dusty days [2, 3]. Dust eventually settles on exposed surfaces, creating a fine layer of accumulated dust. Different parameters are reported to influence dust accumulation such as gravitational forces, wind speed, wind direction, electrostatic charges and the wetness of the surface [5]. Of those parameters, the most dominating ones are the gravitational effect, particle size and wind speed [6-8]. Slow wind speeds increase the deposition of dust, whereas high wind speeds help to remove dust if the wind is incident in an appropriate direction [5, 9].

image

Figure 1. Falling dust in Kuwait International Airport. Note here is that visibility does not always relate to falling dust. It can be a factor of humidity, rising dust and suspended dust.

Download figure to PowerPoint

The random accumulation of dust on the PV module surface area can produce spots with varying concentrations of dust particles, as illustrated in Figure 2. These spots vary in shape, location and dust density. The variation in dust accumulation can lead to different transmittance of light into the module, thus leading to small random areas on the PV module with partial shading from incident solar radiation [10].

image

Figure 2. Accumulated dust on different PV modules installed in Kuwait.

Download figure to PowerPoint

It appears obvious that dust has a direct effect of reducing the performance of solar PV modules. A progressive effect has been reported for certain weather parameters such as relative humidity, rain and ambient temperature [11]. Both settled as well as airborne dust reduces the amount of solar radiation incident on the surface of a PV module [12]. Goosens and Van Kerschaever provide a relation between wind speed, airborne dust, settled dust and the reduction in PV cell short-circuit current [9] using a wind tunnel experiment where they exposed a PV module to different dust concentration under variable wind speed. They showed that there is an aerodynamic relation between airborne dust, accumulated dust and the reduction in PV power output. Others have reported a relation between dust particle size, particle distribution, tilt angle and the reduction in transmittance of solar radiation [1, 9, 10, 13, 14]. Recently, Garcia et al. reported increase in the angle of incidence losses due to dust accumulation where a method is proposed to calculate the losses based on the difference between measured irradiance on reference cell and calculated from two pyranometer measurements [15].

Most of the work reported in relation to dust has considered the modules output performance. Some papers suggested that effective remedies could be casual cleaning, static devices and optimization for tilt angle according to the dust information in the region [5, 16-18]. Others report a special glass coating that promotes self-cleaning [19, 20]. The costs and effectiveness of these methods are not yet known and will require further investigations. The variation of dust density on the tilted PV modules and the effects of dust on different PV technologies have yet to be investigated. In this paper, a relation between spectral transmittance and dust density is established. This will be used to identify the variation of dust density in dependence of module inclination. The effect of dust on different PV technologies is demonstrated by calculating the short-circuit current using the effective spectral response.

METHODOLOGY

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODOLOGY
  5. RESULTS AND DISCUSSION
  6. CONCLUSION
  7. REFERENCES

Different approaches have been reported where correlations between dust and PV system output are shown [5]. In most cases, the measurements taken were the output of PV modules in relation to the percentage of accumulated dust, or it was related to a dust density. For our work, more quantitative measurements of dust were needed to be able to correlate it to module performance, which is why a new approach was used and is explained in the following sections.

Dust sample was collected in Kuwait during two time intervals, May 2010 and May 2011 for a period of 30 days each for sediment characterisation. The sand's material composition was analysed using a microscope and image processing software to determine the grain size distribution. This was parameterised using Φ as the sorting criteria.

  • display math(1)

where D is the particle diameter in millimetre (mm) [21]. The next step was to determine the material composition of the sample using an x-ray diffraction (XRD) analysis method. The collected dust was then deposited in the lab onto a 2.0 × 2.0 cm area using a 5.0 × 5.0 cm soda lime glass with a thickness of 1.0 mm. The deposition of dust was performed by free fall from 1.0-m height using a cylindrical tube to minimize the effect of wind currents in the lab. The weight of the samples was measured using a balance with sensitivity of 0.1 mg. Samples were then encapsulated by adding another sheet of glass and sealing the edges with adhesive epoxy. Finally the transmittance of each sample was measured using a spectrophotometer with an integrating sphere and spot size of 4.0 cm2. The dust sample transmittances were measured over a wavelength range of 300–1200 nm because any values under 300 nm or higher than 1200 nm are outside the spectral response of most PV technologies. Fifty per cent of the total encapsulated sample spectral transmittance were randomly selected and retaken to ensure repeatability and to measure possible deviation in the measurements due to electrostatic charges collected on the glass during deposition and measurement process. It was noted that when the sample edges are electrostatically shielded and grounded during the measurements, the average percentage difference between the shielded and non-shielded samples was 1.9%, and the average percentage deviation between the repeated samples measurement improved from 7.5% to 1.8%. The transmittance of a clean (non-dusty) encapsulated glass sample was measured and used to correct the sample measurements (avoiding wavelengths below 300 nm due to the filtering property of soda lime glass).

To investigate the effect of tilt angle, a number of 4.0 × 4.0 cm tempered glass samples were exposed outdoors in Kuwait for a period of 1 month (September–October 2009). The samples were placed facing south, with tilt angles of 0°, 15°, 30°, 45°, 60° and 90°. The samples were then encapsulated by another sheet of glass and the edges were sealed with adhesive epoxy and shipped to the UK for further analysis. They were divided conceptually for analysis into three sections: top, middle and bottom. The transmittance of each section of each sample was measured using a spectrophotometer. The non-uniformity of the samples was calculated by means of calculating the area under the spectral transmittance curve (At) measured from the samples at the top, bottom and middle. Then the non-uniformity was calculated from the following equation:

  • display math(2)

The spectral transmittance data obtained from the dust samples were then used to investigate spectral effects of the dust accumulation. The effect on solar devices was investigated by modifying measured spectral response data from European Solar Test Installation for nine crystalline silicon (c-Si), three amorphous silicon (a-Si), two copper indium gallium diselenide (CIGS) and one cadmium telluride (CdTe) modules and calculating the short-circuit current that would have been generated by the devices under AM1.5. The effect of different incident spectra on device performance is then investigated by use of artificial spectra. These were generated using a ‘Simple Model of the Atmospheric Radiative Transfer of Sunshine (SMARTS)’ code to account for the tilt angle of the surface plane and combining it with a matlab code (MathWorks, Natick, MA, USA) to generate soiling ratio, which is the ratio of dusty to clean photocurrent [15] for different technologies under variable tilt angle, air mass (AM) and dust concentration (Figure 3).

image

Figure 3. Detailed analysis procedure used in this work.

Download figure to PowerPoint

RESULTS AND DISCUSSION

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODOLOGY
  5. RESULTS AND DISCUSSION
  6. CONCLUSION
  7. REFERENCES

Sediment characterization

The collected dust sample size distributions were analysed and are shown in Figure 4. The two samples' dominating sediments type was of clay and the rest of different kinds of silt. The used sample in our experiments was of sample two (collected in 2011), with major grain size distribution was found to be of clay and fine silt. The other sediments types were distributed between coarse, medium and very fine silt (Table 1). An XRD elemental analysis was used to determine the dust samples material components. The dominating component of the two samples was of quartz followed by calcite and albite, and the results are shown in Table 2.

image

Figure 4. Probability distribution of the counted samples.

Download figure to PowerPoint

Table 1. Dust grain distribution and sedmint types.
Φ*D(µm)% of the total sampleGrain type
  • *

    Φ = −Log2 × (diameter of the particle in mm).

0.0–1.01000–5000.00Coarse grained
1.0–2.0500–2500.00Medium grained
2.0–3.0250–1250.82Fine grained
3.0–4.0125–634.78Very fine grained
4.0–5.063–318.16Coarse silt
5.0–6.031–1616.47Medium silt
6.0–7.016–823.82Fine silt
7.0–8.08–420.19Very fine silt
<8.0<425.75Clay
Table 2. Dust sample material composition.
Compound nameChemical formula
QuartzSiO2
CalciteCaCO3
Albite, calcian, ordered(Na, Ca) Al(Si, Al)3 O8
DolomiteCaMg (C O3)2
MuscoviteKAl3Si3 O10 (OH)2
PalygorskiteMg5 (Si, Al)8 O20 (OH)2 !8 H2O
Lizardite-1 TMg3Si2 O5 (OH)4
Kaolinite 1MdAl2Si2O5 (OH)4

Uncertainty of the measurements

Because of the unstable nature of dust particle placement when deposited on a glass, it becomes necessary to use a method of encapsulation. Another sheet of glass to encapsulate the dust preserves it during the transportation and movement in the measurements stage. This may affect the magnitude of transmittance and thus values relative to a control are measured.

There is a small gap between the glass sheets that may change the spectral transmission due to multiple reflections and scattering within the optical system. This effect was quantified by comparing measurements of dust samples deposited on one sheet and then encapsulated with the second sheet, and the results are shown in Figure 5. The features of the transmittance curve were conserved when adding the second sheet of glass for dust concentration of ≤7.5 mg/cm2. Dust densities of 9.7 mg/cm2 and higher exhibit a slight change. The average variation between the one and two sheets measured sample transmittance was found to be 1.9% with maximum deviation of 3.4% and minimum of 0.15%, which was found in the lower density measurements. This is marginal in comparison with the magnitude of effects being reported in this paper, thus confirming the validity of the chosen approach of encapsulating the samples with a second sheet of glass.

image

Figure 5. Dust transmittance measurements with one and two sheets of glass.

Download figure to PowerPoint

Dust concentration

The transmission of the deposited dust on the glass reduces more for lower wavelengths (300–570 nm regions) than for higher wavelengths, as shown in Figure 6. The small discontinuity in the transmittance curve at 350 and 800 nm happens during the detector change in the spectrophotometer and is a normal measurement uncertainty for the device used.

image

Figure 6. Spectral transmittance curves for different dust density samples.

Download figure to PowerPoint

It is shown in Figure 7 that for dust densities above 19 mg/cm2, the wavelength-dependence of the transmission reduction is significantly reduced. For wavelengths >570 nm, the variation between density–transmittance curves with respect to average is 2.5% whereas for wavelengths <570 nm, it is 11%. Thus, dust affects lower wavelengths more severely. This is shown in Figure 7 by means of distance variation between the curves at the specified wavelength.

image

Figure 7. Variation of dust density with transmittance at different wavelengths.

Download figure to PowerPoint

In order to verify if this behaviour is reasonable and the spectral effect is according to expectations, this has been modelled as Mie scatter of light on dust particles. The simulation was carried out with matlab. The base material used in the simulation was of quartz with the real refractive index of 1.458 and a range of complex parts of the refractive indices identified from literature [22]. Particle size was varied from 1 up to 100 µm diameter in size. The attenuation effect with respect to the wavelength was clearly shown in the lower particle size (1–3 µm; Figure 8). The transmittance curve was obtained from the simulated absorption data. The simulation demonstrates that Mie scatter happens largely for smaller particle size (<6 µm) where the attenuation of the wavelength becomes more apparent especially at lower wavelength. Although not much could be performed to confirm the internal scatter between the particles layering on top of each other due to the limitation of the simulation tools, this approximation seems to confirm the measurements. Therefore, it can be confirmed that the spectral-dependent effect of the transmittance curve obtained from the dust samples is due to the lesser particle size and material refractive index mixture of the sample.

image

Figure 8. Transmittance of dust particles with different sizes simulated with Mie scatter.

Download figure to PowerPoint

Tilt angle

The variation of dust accumulation on tilted surfaces was shown in the spectral transmittance data obtained from measuring the dust samples collected in Kuwait in the period from September to October 2009. The dust samples were encapsulated and spectral transmittance was measured at three different areas, top, middle and bottom (Figure 9).

image

Figure 9. Spectral transmittance curves for different tilted samples at 0°, 15°, 30°, 45°, 60° and 90°. The locations of the measured transmittance data are marked by T = Top, B = Bottom and M = Middle.

Download figure to PowerPoint

Two general trends were observed in the spectral transmittance data shown in Figure 9. The first trend shows that with increased tilt angle, dust accumulation decreases; this can be explained as gravity affects dust samples more at higher tilt angles. The second trend is shown more clearly in Table 3, where transmittance through the tilted samples decreases towards the bottom. In Figure 10, the 90° showed a non-uniformity of only 0.21%, in comparison with the 30° that showed 4.39% non-uniformity between the top, middle and bottom sections. The 0° showed a higher variation than that at 15°. This can be attributed to the higher dust density of the 0° sample that made it more sensitive to environmental effects such as wind direction and rain water accumulation in comparison with the tilted samples. In addition, the 30° allowed water to slide down, allowing more dust to settle at the bottom than at the top area, thus creating a very clean area at the top of the sample and a very dusty area at the bottom.

Table 3. Fitted dust density (mg/cm2) values for the tilted dust samples.
 0° (mg/cm2)15° (mg/cm2)30° (mg/cm2)
 BMTBMTBMT
AVG3.63.03.52.42.12.42.71.81.3
MAX4.33.83.93.12.83.13.32.41.9
MIN3.22.63.22.01.82.22.41.51.0
STD0.30.30.20.30.30.20.20.20.2
 45o (mg/cm2)60o (mg/cm2)90o (mg/cm2)
 BMTBMTBMT
AVG1.11.51.31.31.21.00.30.20.2
MAX1.72.01.91.71.71.50.50.40.4
MIN0.91.21.11.11.00.80.030.040.05
STD0.20.20.20.20.20.20.10.10.1
image

Figure 10. Non-uniformity of transmittance at different tilt angles [%].

Download figure to PowerPoint

The variation of the transmittance through the tilted samples is affected by the variation of dust density at different positions on the sample. The spectral transmittance data for the tilted samples were fitted to the data, measured in Section 3.3, of the dust density range of 1–4.5 mg/cm2 to identify the dust density of the tilted samples at the different measured locations using linear regression between the selected dust density ranges. The calculated density transmittance curves for the tilted samples are shown in Figure 11, and the calculated values for the different locations for tilted samples are shown in Table 3. The results in Table 3 agree with the results obtained in Figure 10 where the sample tilted at 30° showed the worst case variation of 1.4 mg/cm2 in dust density, in comparison with the 90° with only 0.1 mg/cm2 of variation between top, middle and bottom sections of the sample.

image

Figure 11. Transmittance density curves obtained by fitting the spectral transmittance curves for the tilted samples into measured dust density curves obtained in Section 3.3.

Download figure to PowerPoint

Spectral effect

The first step to quantify the effect that dust has on PV cells is to examine the effect of the transmittance curves on the spectrum. The AM 1.5 G standard spectrum [23] was multiplied with the different spectral transmittance curves at 4.25, 14, 19 and 30 mg/cm2 dust densities.

Spectral response data for different PV technologies supplied by ESTI were corrected for the different spectral transmittance curves obtained in Sections 3.3 and Sections 3.4 in order to identify if different technologies are affected differently. The corrected curves show a variation between different technologies with regard to the same spectral transmittance dust curves as shown in Figures 12 and 13. In Figure 12, samples of c-Si module spectral response data were corrected with four dust-spectral transmittance curves, whereas Figure 13 shows the same transmittance curves applied to a-Si, CIGS and CdTe PV modules. The spectral photocurrents shown in Table 4 were obtained by integrating the area under the product curve of AM 1.5 and the modified spectral responses in Figures 12 and 13.

image

Figure 12. Spectral response of c-Si modules corrected for four different spectral transmittance dust curves, D1 = 4.25, D2 = 14, D3 = 19 and D4 = 30 mg/cm2.

Download figure to PowerPoint

image

Figure 13. Spectral response of thin-film modules corrected for four different spectral transmittance dust curves, D1 = 4.25, D2 = 14, D3 = 19 and D4 = 30 mg/cm2.

Download figure to PowerPoint

Table 4. Percentage difference variation between clean module data and those corrected for the spectral photocurrent.
Density (mg/cm2)a-Si (%)CIGS (%)CdTe (%)c-Si (%)
1.2−10.8−9.1−9.7−9.1
4.25−33.0−28.5−30.1−28.6
14−66.0−59.6−61.9−59.6
19−77.4−70.6−73.1−70.6
30−98.4−97.8−98.1−97.8

From Table 4, it can be seen that wide band-gap materials (a-Si and CdTe) are affected more than the c-Si and CIGS modules when they are covered with dust. This can be correlated to the high band-gap of the affected modules that have effective spectral response ranges between 300 and 800 nm where the spectral transmittance through dust decreases more strongly than at longer wavelengths.

Soiling ratio

Because dust has spectrally selective attenuation, the short-circuit current is a good choice to represent the effect of dust on the PV module. Short-circuit current of a PV module is affected largely by irradiance and less by temperature. To further quantify the effects of dust, the effects of different spectra were simulated. Using SMARTS [24] to account for the effect of tilt angle, different spectrum data were generated under tilt angle from 0° to 90° and AM values from 1 to 10. matlab code was generated to integrate all the data together and generate a different photocurrent value under a range of dust density, AM and tilt angle values. Finally, the soiling ratio was obtained by dividing the dusty photocurrent over the clean value. The full processing procedure is shown in the flow chart in Figure 3. This procedure was repeated for tilt angles 0°–90°, AM 1–10 and different PV technologies spectral response data.

The results for AM 1 and AM 6 at with 0° tilt and 90° tilt are shown in Figures 14 and 15, respectively. The values obtained from Figures 14 and 15 can be used to further verify outdoor dust effect on outdoor modules by comparing the short-circuit current of a clean module with the short-circuit current of a module affected by dust. Narrow band-gap technologies such as monocrystalline or polycrystalline silicon have a better soiling ratio when compared with other technologies such as CdTe and a-Si. This is clearly seen when looking at the variation between the curves at Figures 14 and 15, where the lowest soiling ratios obtained at a fixed dust concentration of 2.3 mg/cm2 and AM 1.0 for c-Si, a-Si and CdTe are 0.908, 0.894 and 0.904, respectively. On the other hand at 2.3 mg/cm2 and AM 6.0, the soiling ratio for c-Si is 0.914, a-Si is 0.9 and CdTe is 0.908. Those values in addition to the clear variation between the soiling ratios in Figure 15 confirm the spectral dependency of the dust effect and its different effect on different PV modules, where the losses are at highest for wider band-gap technologies under lower air mass values. The effect of tilt angle is also more apparent at lower air mass values and is even worse for wide band-gap technologies such as amorphous silicon as shown in Figure 15, where lower values at AM 1 than that of other technologies are shown. Thus the soiling ratio data can be used to determine the absolute orientation to be used for different technologies under different air mass values.

image

Figure 14. Soiling ratio for different technologies at AM 1 and AM 6 at 0° tilt.

Download figure to PowerPoint

image

Figure 15. Soiling ratio for different technology at AM 1 and AM 6 at 90° tilt.

Download figure to PowerPoint

CONCLUSION

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODOLOGY
  5. RESULTS AND DISCUSSION
  6. CONCLUSION
  7. REFERENCES

It is shown that there is a relation between dust density and spectral transmittance, especially for lower amounts of dust deposited on photovoltaic panels. At higher dust concentrations (>19 mg/cm2), the effect of wavelength becomes marginal. It is also shown that the effect of dust predominantly affects the lower wavelength regions and, as such, affects wide band-gap materials disproportionally. Transmittance is less spectrally selective at higher wavelengths (>600 nm). This effect is explained by Mie scatter, where smaller dust particle size <3 µm plays a major role in attenuating the transmittance.

The effect of tilt angle is directly related to the amount of dust density variation on the surface. The 90° tilt angle showed the least variation of dust accumulation of 0.1 mg/cm2 due to the fact that gravity affects it the most and supports the process of dust removal over time. The sample tilted at 30°, which is the optimal tilt angle for PV modules in Kuwait regarding sun position, showed the highest variation of dust accumulation of 1.4 mg/cm2 with most of dust settling towards the bottom of the sample.

Dust contributes to the reduction of PV output by attenuating irradiance in a spectrally dependent manner. This can be seen in the effect on the spectral response data. The effect is not the same magnitude for all types of PV technology because the spectral transmittance affects various spectral response shapes differently. The effect is worse for the PV modules with wider band-gap such as a-Si and CdTe technologies, which showed 33% reduction in photocurrent when a concentration of 4.25 mg/cm2 of dust was applied. In comparison, c-Si and CIGS technologies showed 28.6% and 28.5% reductions at the same dust density.

The effect of dust on PV modules can be monitored in a manner of the ratio between short-circuit current of the module affected by dust over the short-circuit current of the clean module. This value can be reproduced in the lab with the combination of dust samples obtained from the specific region that the measurement of dust effect is desired, spectral response data of the technology in question and with the combination of simulated data from SMARTS. This can be used as an identifier to show the absolute effect of dust density and to clearly differentiate this effect for various PV technologies.

REFERENCES

  1. Top of page
  2. ABSTRACT
  3. INTRODUCTION
  4. METHODOLOGY
  5. RESULTS AND DISCUSSION
  6. CONCLUSION
  7. REFERENCES
  • 1
    Al-Hasan AY. A new correlation for direct beam solar radiation received by photovoltaic panel with sand dust accumulated on its surface. Solar Energy 11, 1998; 63: 323333.
  • 2
    M. I. Safar and Kuwait Meteorological Dept. Dust and Dust Storms in Kuwait. State of Kuwait, Directorate General of Civil Aviation, Meteorological Dept.: Kuwait, 1980.
  • 3
    M. I. Safar and Kuwait Meteorological Dept. Climatological Section. Frequency of Dust in Day-Time Summer in Kuwait. Directorate General of Civil Aviation, Meteorological Department, Climatological Section: Kuwait, 1980.
  • 4
    Biryukov S. An experimental study of the dry deposition mechanism for airborne dust. Journal of Aerosol Science 2, 1998; 29: 129–139.
  • 5
    Mani M, Pillai R. Impact of dust on solar photovoltaic (PV) performance: research status, challenges and recommendations. Renewable and Sustainable Energy Reviews 12, 2010; 14: 31243131.
  • 6
    Said SAM. Effects of dust accumulation on performances of thermal and photovoltaic flat-plate collectors. Applied Energy 1990; 37: 7384.
  • 7
    El-Shobokshy MS, Hussein FM. Degradation of photovoltaic cell performance due to dust deposition on to its surface. Renewable Energy 10, 1993; 3: 585590.
  • 8
    Elminir HK, Ghitas AE, Hamid RH, El-Hussainy F, Beheary MM, Abdel-Moneim KM. Effect of dust on the transparent cover of solar collectors. Energy Conversion and Management 11, 2006; 47: 31923203.
  • 9
    Goossens D, Van Kerschaever E. Aeolian dust deposition on photovoltaic solar cells: the effects of wind velocity and airborne dust concentration on cell performance. Solar Energy 7, 1999; 66: 277289.
  • 10
    Qasem H, Betts TR, Gottschalg R. Effect of shading caused by dust on cadmium telluride photovoltaic modules. In 36th IEEE Photovoltaic Specialists Conference, Seattle, WS, 2011.
  • 11
    AlBusairi1 HA, Möller HJ. Performance evaluation of CdTe PV modules under natural outdoor conditions in Kuwait. In 25th European Photovoltaic Solar Energy Conference and Exhibition, Valencia, Spain, 2010; 34683470.
  • 12
    Goossens D, Offer ZY, Zangvil A. Wind tunnel experiments and field investigations of eolian dust deposition on photovoltaic solar collectors. Solar Energy 1, 1993; 50: 7584.
  • 13
    El-Shobokshy MS, Hussein FM. Effect of dust with different physical properties on the performance of photovoltaic cells. Solar Energy 12, 1993; 51: 505511.
  • 14
    Pang H, Close J, Lam K. Study on effect of urban pollution to performance of commercial copper indium diselenide module. In IEEE 4th World Conf. on Photovoltaic Energy Conversion, Waikoloa, HI, 2007; 21952198.
  • 15
    Garcia M, Marroyo L, Lorenzo E, Perez M. Soiling and other optical losses in solar-tracking PV plants in Navarra. Progress in Photovoltaics: Research and Applications 2011; 19: 211217.
  • 16
    Hammond R, Srinivasan D, Harris A, Whitfield K, Wohlgemuth J. Effects of soiling on PV module and radiometer performance In 26th IEEE photovoltaic Specialists Conference, 1997; 11211124.
  • 17
    Rolland JM, Astier S, Protin L. Static device for improving a high voltage photovoltaic generator working under dusty conditions. Solar Cells 5, 1990; 28: 277286.
  • 18
    Rabii AB, Jraidi M, Bouazzi AS. Investigation of the Degradation in Field-Aged Photovoltaic Modules. WCPEC-3 Organizing Committee: Tokyo Univ Agriculture and Technology; Tokyo, Koganei 184–8588, Japan, 2003.
  • 19
    Bruynooghe S, Spinzig S, Fliedner M, hsu GJ. Realization of high performance AR-coatings with dust- and water-repellent properties. Vakuum in Forschung Und Praxis 2008; 20: 2529.
  • 20
    Nishioka K, Nabemoto K, Sakurada Y, Ota Y, Takami K, Nagai H, Araki K. Anti-soiling layer coated on PMMA Fresnel lens for concentrator photovoltaic modules. In 36th IEEE Photovoltaic Specialists Conference, Seattle, WS, 2011.
  • 21
    Boggs S. Siliciclastic sedimentary rocks. In Petrology of Sedimentary of rocks, 2nd edn. Cambridge University Press: Cambridge, UK., 2009; 19310 sec. 2.
  • 22
    Kitamura R, Pilon L, Jonasz M. Optical constants of fused quartz from extreme ultraviolet to far infrared at near room temperatures. Applied Optics 2007; 46(33): 81188133.
  • 23
    Principles for terrestrial photovoltaic (PV) solar devices with reference spectral irradiance data, IEC Standard, 60904–3, 2008.
  • 24
    Gueymard C. SMARTS, A simple model of the atmospheric radiative transfer of sunshine: algorithms and performance assessment. Professional Paper FSEC-PF-270-95. Florida Solar Energy Center, 1679 Clearlake Road, Cocoa, FL 32922, 1995.