Siddiqui and Bajpai International Journal of Energy and Environmental Engineering 2012, 3:26

http://www.journal-ijeee.com/content/3/1/26

ORIGINAL RESEARCH Open Access

Correlation between thicknesses of dust collected

on photovoltaic module and difference in
efficiencies in composite climate
Rahnuma Siddiqui* and Usha Bajpai

Abstract
An effort has been made to develop an equation with the given data for all seasons for a location (Lucknow, India)
consisting of composite climate, which is further helpful in developing a relation between difference in efficiencies
of module with respect to thicknesses of dust collected on the module. This equation that is developed
mathematically is in good correlation with the measured data. Here, data are shown for a whole year (from 2010 to
2011); the study done gives us a broad view of finding out the difference in efficiencies of module when dust
collects on it.
Keywords: (Atmospheric) dust, regression analysis, efficiency of solar photovoltaic module

Background atmosphere affect the amount and properties of the radi-

In solar cells, lots of power get lost due to various rea- ation finally reaching the collectors. The accumulation
sons such as reflection losses at the top surface, incom- of dust particles on the surface of PV module greatly
plete absorption of photon energy due to limited cell affects its performance especially in desert areas. Some
thickness, series and shunt resistance loss, curve or fill correlation studies have done direct beam solar radiation
factor loss, etc. Results obtained show that there is a dir- received by photovoltaic panel with sand dust accumu-
ect proportionality between solar flux, output current, lated on its surface as given by Al-Hasan [6].
and efficiency of the photovoltaic module as given by Desert countries are of course best suited to photo-
Omubo-Pepple et al. [1]. voltaic power generation due to the abundant avail-
Dust contributes to as much as 40% degradation in ability of sunlight throughout the year. Nowadays, the
peak power of photovoltaics; there is surprisingly little ideas of setting up vast solar arrays in desert countries
scientific work done on the subject. Since no information and exporting the power to other countries are being dis-
about the type of dust, density, and rate of accumulation cussed. In a bigger PV solar plant, more work force and
of dust was noted, no general understanding of the machines will be needed to help make the rounds and
underlying physical principles could be deduced. Dust clean the panels, especially after a stand storm.
accumulation on the photovoltaic (PV) panel surface Many research results discuss about performance of
depends on different parameters such as PV panel inclin- panel with dust concentration on the surface, but for a
ation and kind of installation as given by Del Cueto [2]. common PV user, it is important to know how fre-
The research done previously include degradation ana- quently the panel has been cleaned. In case if frequent
lysis of silicon photovoltaic modules and effective effi- cleaning is not possible, it is important to know the per-
ciency of PV modules under field conditions as given formance loss due to dust for additional estimation to
[3,4]. Some studies have also been done on corrosion compensate the loss.
effects in thin film photovoltaic modules as given by It has been concluded in many studies that dust accu-
Carlson et al. [5]. The airborne particles in the mulation considerably deteriorates the performance of
photovoltaic cells. However, in carrying out the investi-
* Correspondence: rmsphy@gmail.com gation on the effect of dust and particulate pollution, the
Renewable Energy Research Laboratory, Department of Physics, University of
Lucknow, Lucknow 226007, India

© 2012 Siddiqui and Bajpai; licensee Springer. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.
Siddiqui and Bajpai International Journal of Energy and Environmental Engineering 2012, 3:26 Page 2 of 7
http://www.journal-ijeee.com/content/3/1/26

physical characteristics of dust must be determined and to 50,000 lux in three ranges of 0 to 2,000, 0 to 20,000,
correlated to the observed effects [7]. and 0 to 50,000 lux; accuracy ±5%, ±2 least count) to
Research done also include characteristic distribution measure solar flux. Data are taken for one whole year
of total, diffuse and direct solar radiation at given loca- from 2010 to 2011.
tions, qualification testing of modules, etc. [8-11]. Differ- The commercial photovoltaic module used has the
ent analysis, performance, and reliability testing of following components:
photovoltaic modules and arrays have been done as given
[12-14].
Solar type is polycrystalline silicon.
As we know that module performance is greatly
Cell area = 0.216 m2.
affected by the amount of dust collected on it [15], some-
Electrical efficiency STP = 12% with a power peak of
times when we want to get the amount of difference 36 W.
value in efficiency of modules, the instruments for meas-
Electrical specifications is at standard test conditions
uring it at that instant are not found. The main aim of of 100 MW/cm2, AM 1.5, and at 25°C cell
this paper is to solve these kinds of problems. Here, an temperature within normal production of tolerance
attempt has been made to obtain a relation between ±3%.
‘thicknesses of dust collected and difference in efficien-
Area of the panel = 0.340 m2
cies’ for all seasons so that difference in efficiencies of
modules can be easily estimated by measuring the thick- Development of correlation between thicknesses of dust
nesses of dust collected using simple devices. collected on photovoltaic module and the difference
Here, a statistical analysis has been done in finding out in efficiencies
the correlation between thicknesses of dust collected on As stated earlier, we have tried to verify the data taken
photovoltaic module and the difference in efficiencies of for the whole year while considering different values,
the module for a whole year considering all seasons of a depending on the variation of climate. This fluctuation
composite climate. resulted from different factors affecting the perform-
ance of solar photovoltaic module. For example, these
Methods factors may be temperature, wind velocity, humidity,
Instrumentation cloud cover, and dust.
The SPV module (number 2007.20.685) which is manu- The aim of this study is to find out the relation of the
factured in the year 2007 by M/s Rajasthan Electronics difference in efficiencies with respect to thicknesses of
and Instruments Ltd, Jaipur, India is studied. The mod- dust collected. Here, reading is taken three times for
ule has been connected to a rheostat ammeter digital each month after every 10 days, and finally, the average
multimeter (DT9205A, Agronic Ltd.). Other instruments of the three readings is taken as measured value of the
used are the digital thermometer (SE-221-P-K, Agronic data for both the variables (difference in efficiencies and
Ltd.) to measure cell temperature, the Thermo-Hygro thicknesses of dust collected on the module) for that
Clock (288-CTH, Agronic Ltd.) to measure ambient particular month. In this way, each month has its indi-
temperature and humidity and digital luxmeter (range 0 vidual reading taken throughout the year. In Table 1,

Table 1 Difference in efficiencies of solar photovoltaic module and thicknesses of dust collected
Months Monthly average of difference in efficiencies Monthly average of thicknesses of dust
(before and after collection of dust)(Δη) (%) collected on the module (t) (mm)
January 1.5666 0.0002400800
February 1.9574 0.0010534300
March 1.3707 0.0007545513
April 1.9563 0.0007006547
May 1.6332 0.0014258080
June 1.8172 0.0005536642
July 0.9202 0.0003919747
August Data unavailable(rainy month) Data unavailable
September Data unavailable(rainy month) Data unavailable
October 1.8790 0.0009505386
November 1.5544 0.0006369589
December 2.0198 0.0006516579
Siddiqui and Bajpai International Journal of Energy and Environmental Engineering 2012, 3:26 Page 3 of 7
http://www.journal-ijeee.com/content/3/1/26

both variables are taken as monthly average value of dif- Calculation

ference in efficiencies and monthly average value of For calculations of the relation, we have used the process
thicknesses of dust collected on the module. of linear regression as it is given by the method of least
Thus, the efficiencies of the module and their corre- squares solved for two variables, as shown in Table 2.
sponding thicknesses of dust particles are taken into ac- Problems involving more than two variables can be trea-
count. Efficiency of module has been found with the ted in a manner analogous to that for two variables.
help of the maximum power, solar flux, and area of solar
panel. On the other hand, thickness of dust particles col- Now; Δη ¼ ao þ a1 t; ð1Þ
lected on the module is calculated using the following X X
Δη ¼ n ao þ a1 t; ð2Þ
the basic formula:
X X X
Thickness of dust collected Δη:t ¼ ao t þ a1 t2: ð3Þ
¼ Volume of dust collected=Area of solar panel; P P P P
Substituting the values of t, Δη, t2, and Δη.t
where in Equations 2 and 3 when n =10, we get the following:

Volume of dust collected 16:6748 ¼ 10 ao þ a1 73:593183  104 ; ð4Þ

¼ Mass of dust collected=Density of dust
126:30879920  104 ¼ 73:593183  104 ao
collected on module:
þ a1 645:4632795  108 : ð5Þ
The mass of dust collected is measured for every
Now, solving Equations 4 and 5 to get the values of ao
month using a chemical balance. The area of solar
and a1,
panel = 0.340 m2, and the density of dust collected =
3g/cm3. 1:66748 ¼ ao þ a1 7:3593183  104 ; ð6Þ
As we know that density of dust is governed by the
diameter of dust particles, dust consists of multiple com- 1:71631 ¼ ao þ a1 8:770693877  104 ; ð7Þ
posite particles ranging from 0.1 to 80 μm in diameter;
sand consists of loose siliceous particles whose size range a1 ¼ 345:9471:
is approximately 0.08 to 1 mm in diameter (full text is Keeping the value of a1 to find out ao in Equation 6,
found in the handbook [16]); thus, knowing the average
characteristic of dust, density is known to be the above a1 ¼ 345:9471:
given value).
General variation in the difference in efficiencies and Therefore, the equation comes out as follows:
thicknesses of dust has been recorded which has been Δη ¼ 1:4128 þ 345:9471t: ð8Þ
observed to be different for different months. Here,
readings have been taken consecutively for 12 months, The corresponding relation between the ‘monthly
as given in columns 2 and 3 in Table 1. average of thicknesses of dust collected and monthly

Table 2 Method of least squares solved for two variables

Monthly average of Monthly average of difference t2 Δη.t
thicknesses of dust in efficiencies (Δη) (%)
collected (t) (mm)
0.0002400800 1.5666 5:7638406  108 3:76109328  104
8
0.0010534300 1.9574 110:9714765  10 20:61983882  104
0.0007545513 1.3707 56:9347664  108 10:34263460  104
8
0.0007006547 1.9563 49:0917008  10 13:70690790  104
0.0014258080 1.6332 203:2928453  108 23:28629600  104
8
0.0005536642 1.8172 30:6544046  10 10:06118580  104
0.0003919747 0.9202 15:3644165  108 3:60714717  104
8
0.0009505386 1.8790 90:3523630  10 17:86062029  104
0.0006369589 1.5544 40:5716640  108 9:90088914  104
8
0.0006516579 2.0198 42:4658018  10 13:16218620  104
P P P2 P
t ¼ 0:0073593183 Δη ¼ 16:6748 t ¼ 645:4632795  108 Δη:t ¼ 126:30879920  104
Siddiqui and Bajpai International Journal of Energy and Environmental Engineering 2012, 3:26 Page 4 of 7
http://www.journal-ijeee.com/content/3/1/26

Figure 1 Correlation between thicknesses of dust and difference in efficiencies. Correlation between the monthly average of thicknesses of
dust collected on photovoltaic module and the monthly average of difference in efficiencies for a whole year (2010 to 2011).

P
average of difference in efficiencies’ graph for the data t ¼ estimated value þ u=n ¼ 73:593183  104 =10;
are given in Figure 1. ¼ 7:3593183  104 :
Here, t and Δη are two variables; t is considered as an ð11Þ
independent variable, and Δη is considered as an
dependent variable (t = monthly average of thicknesses of P
Δη ¼ estimated value þ v=n ¼ 16:6748=10;
dust collected) and (Δη = monthly average of difference ð12Þ
in efficiencies). In statistics, linear regression is an ap- ¼ 1:66748:
proach to model the relationship between the two vari- Substituting the above values in the regression equa-
ables. Problems involving more than two or three tion of t on Δη, we get the following:
variables can be treated in a manner analogous to that
for two variables. t  t ¼ btΔη ðΔη  ΔηÞ:
To further reconfirm the relation between thicknesses
Similarly, with the equation of the line of regression of
of dust collected and the difference in efficiencies, equa-
Δη on t, we get the following:
tions to the lines of regression and regression coefficients
have been found with the help of the deviation method Δη  Δη ¼ bΔη:t ðt  t Þ:
using Pearson's formula as given in Table 3. We have
taken the estimated value of thicknesses of dust collected Therefore, the final two equations come out to be as
as 0 mm and therefore accordingly estimated the value of follows:
difference in efficiencies of solar photovoltaic module
t ¼ 3:4916  104 Δη þ 1:5371  104 ; ð13Þ
used as 0%, respectively.
Substituting the above values of Table 3 in the for- Δη ¼ 1:4128 þ 346 t: ð14Þ
mula, we get the following:
Using Pearson's formula, coefficient of correlation has
P P P been found out as follows:
uv  u v=n
Regression coefficient ðΔη on t Þ ¼ P P ; P P P
u2  ð uÞ2 =n uv  u v=n
r ¼ P P  P P
√ u2  ð uÞ2 =n √ v2  ð vÞ2 =n;

bΔη:t ¼ 3:5936  104 =103:8674 ¼ 3:5936  104 =√103:8674  108 √1:0292

ð9Þ ¼ 0:34756839:
¼ 0:0346  104 :
Note. We have seen that there are two lines of regres-
P PP sion (Equations 13 and 14): one of Δη on t and the other
uv  u v=n of t on Δη. In the first one, Δη is the dependent variable
Regression coefficient ðt on ΔηÞ ¼ P P ;
v2  ð vÞ2 =n and t is the independent variable, while in the other one,
Δη is the independent variable and t is the dependent
variable. Since the assumptions for deriving the equations
btΔη ¼ 3:5936  104 =1:0292; of two lines of regression are different, the two equations
ð10Þ
¼ 3:4916  104 : are not reversible. In other words, one equation gives the
 http://www.journal-ijeee.com/content/3/1/26
Siddiqui and Bajpai International Journal of Energy and Environmental Engineering 2012, 3:26
Table 3 Equations to the lines of regression and regression coefficients
Monthly average of thicknesses of dust collected on the module Monthly average of difference in efficiencies Product of deviation
Monthly average Deviation from Square of deviation For a whole year Deviation from Square of deviation
of thicknesses 2of estimated value estimated value
dust for a year
t u u2 Δη v v2 uv
8
0.0002400800 0.0002400800 5:76384064  10 1.5666 1.5666 2.4542 3:76109328  104
8
0.0010534300 0.0010534300 110:9714765  10 1.9574 1.9574 3.8314 20:61983882  104
0.0007545513 0.0007545513 56:93476643  108 1.3707 1.3707 1.8788 10:34263460  104
8
0.0007006547 0.0007006547 49:09170081  10 1.9563 1.9563 3.8271 13:70690790  104
0.0014258080 0.0014258080 203:2928453  108 1.6332 1.6332 2.6673 23:28629600  104
8
0.0005536642 0.0005536642 30:6544046  10 1.8172 1.8172 3.3022 10:06118580  104
0.0003919747 0.0003919747 15:3644165  108 0.9202 0.9202 0.8468 3:60714717  104
8
0.0009505386 0.0009505386 90:3523630  10 1.8790 1.8790 3.5306 17:86062029  104
0.0006369589 0.0006369589 40:5716640  108 1.5544 1.5544 2.4161 9:90088914  104
8
0.0006516579 0.0006516579 42:4658018  10 2.0198 2.0198 4.0796 13:16218620  104
P P P 2 P P P 2 P
t ¼ 73:593183  104 u ¼ 73:593183  104 u ¼ 645:46327  108 Δη ¼ 16:6748 v ¼ 16:6748 v ¼ 28:8341 uv ¼ 126:30879920  104

Page 5 of 7
Siddiqui and Bajpai International Journal of Energy and Environmental Engineering 2012, 3:26 Page 6 of 7
http://www.journal-ijeee.com/content/3/1/26

Table 4 Values of Δη measured and calculated

Δη measured Δη calculated Difference of the two Δη values Difference2
(from measured values) (calculated from equation)
1.5666 1.4959 0.0707 0.0049
1.9574 1.7772 0.1802 0.0324
1.3707 1.6738 −0.3031 0.0918
1.9563 1.6552 0.3011 0.0906
1.6332 1.9060 −0.2728 0.0744
1.8172 1.6043 0.2129 0.0453
0.9202 1.5480 −0.6277 0.3940
Data unavailable Data unavailable - -
Data unavailable Data unavailable - -
1.8790 1.7416 0.1374 0.0188
1.5544 1.6331 −0.0787 0.0062
2.0198 1.6382 0.3816 0.1456
P P P P
Δηmeasured ¼ 16:6748 Δηcalculated ¼ 16:6733 difference ¼ 0:0016 difference2 ¼ 0:9040

estimate of Δη, while the other gives the estimate of t. Conclusions

However, if r is equal to (±) 1, the two lines of regression Silicon solar cell operating efficiency is at optimum level
can be reduced to the same form, but from the above in the absence of dust which is an ideal condition. Prac-
value of r, it is clear that r = 0, i.e., the estimated value of tically, it is seen that there is in average a reduction in
Δη (or t) is the same for all values of t (or Δη). Thus, the output, near 10% to 20%, when heavy layers of dust are
degree of correlation between the two variables is much accumulated on terrestrial modules [17].
greater but not perfect; there are other climatic variables Dust consists of multiple composite particles ranging
influencing the correlation between the two. approximately from 0.1 to 80 μm in diameter. Dust par-
ticles may be electrically conducting and are usually sol-
Results and discussion uble in water. Sand dust is most severe in low humidity
We have made an effort to test the validity of the regions. Dust becomes airborne with slight winds and
equation deduced. Table 4 gives the value of Δη mea- may remain suspended for hours as dust clouds [16].
sured and calculated from Equation 8 and also from During wind storms, dust particles penetrate almost
Equation 14. any enclosure which is not hermetically sealed, and it
While keeping manual error in mind, the values may be hygroscopic. Its presence on metallic surfaces
obtained (Δη measured and Δη calculated) are quite in may aggravate corrosion.
agreement with each other as seen from the table and It is reported that a small amount of dust on silicon
from calculating the error of prediction given as follows: solar photovoltaic module covers has a negligible effect
on the sunlight transmission to the silicon photovoltaic
module. An attempt has been made to develop an equa-
Estimate of error
tion to calculate difference in efficiencies from thicknesses
¼ ðåΔη measured  åΔη calculatedÞ1=2 =n of dust particles collected for a whole year. The standard
equation obtained is very good to measure the correlation
¼ ð0:9040Þ1=2 =10 ¼ 0:0950789;
with obtained results for almost every month. This is clear
from the calculated percentage error from Table 4. There-
where the value 0.0950789 signifies the amount of devi- fore, the standard equation helps to evaluate the difference
ation of the predicted value from the observed value, in efficiencies of module while knowing thicknesses of
which is considerably very small. dust collected on it for any climatic conditions of the area.
Thus, the equation obtained (Equation 8) may be taken This result has been obtained with a particular module
as a valid equation, and using it, Δη (difference in effi- studied the whole year; a combined study can be done
ciencies of module) can be calculated using the known by taking different modules with their respective climatic
values of t (thicknesses of dust collected on module) for conditions, and difference in output of solar photovoltaic
a location. modules can be easily evaluated.
Siddiqui and Bajpai International Journal of Energy and Environmental Engineering 2012, 3:26 Page 7 of 7
http://www.journal-ijeee.com/content/3/1/26

Abbreviations 15. Shaharin, A., Sulaiman, S.A., Hussain, H.H.: Nik Leh, NSH, Mohd SI, R: Effects
Δη: monthly average of difference in efficiencies of solar photovoltaic of dust on the performance of PV panels. World Academy of Science,
module (before and after collection of dust) (percentage); ao: first constant; Engineering and Technol 58, 588–592 (2011)
t: monthly average of thicknesses of dust collected (thickness in millimeter); 16. TRW Systems Group, Power Sources Engineering Department,
u: deviation from the estimated value of thicknesses of dust collected; Rauschenbach HS, Jet Propulsion Laboratory, United States National
v: deviation from the estimated value of difference in efficiencies; Aeronautics and Space Administration: Solar Cell Array Design Handbook,
Δη: arithmetic average of the thicknesses of dust collected on module; 1st edn. NASA, Washington, DC (1976)
Δηmeasured: measured value of difference in efficiencies of module; 17. Monto, M., Rohit, P.: Impact of dust on solar photovoltaic (PV) performance:
Δηcalculated: calculated value of difference in efficiencies of module, bt.Δη, research status, challenges and recommendations. Renew Sustain Energy
regression coefficient (t on Δη); bΔη.t: regression coefficient (Δη on t); Rev 14(9), 3124–3131 (2010)
t: arithmetic average of the difference in efficiencies of photovoltaic module.
doi:10.1186/2251-6832-3-26
Competing interests Cite this article as: Siddiqui and Bajpai: Correlation between thicknesses
The authors declare that they have no competing interests. of dust collected on photovoltaic module and difference in efficiencies
in composite climate. International Journal of Energy and Environmental
Engineering 2012 3:26.
Authors' contributions
UB supervised the whole work, drafted the manuscript, and corrected the
manuscript. RS conducted the experimental analysis, studying the
performance of solar photovoltaic module under different climatic
parameters, and conducted the calculation according to the instruction
provided by the supervisor. All authors read and approved the final
manuscript.

Acknowledgments
The authors would like to thank the Ministry of New and Renewable Energy
(MNRE), New Delhi, India for providing financial support for current research
work.

Received: 13 June 2012 Accepted: 20 September 2012

Published: 3 October 2012

References
1. Omubo-Pepple, V.B., Israel-Cookey, C., Alaminokuma, G.I.: Effects of
temperature, solarflux and relative humidity on the efficient conversion of
solar energy to electricity. European J Sci Res 35(2), 173–180 (2009)
2. Del Cueto, J.A.: Comparison of energy production and performance from
flat plate photovoltaic module technologies deployed at fixed tilt.
Proceedings of the 29th IEEE Photovoltaic Specialists Conference,
Piscataway (2002)
3. Meyer, E.L., Van Dyk, E.E.: Degradation analysis of silicon photovoltaic
modules. Proceedings of the 16th European Photovoltaic Solar Energy
Conference, Glasgow (2000)
4. Topic, M., Breel, K., Sites, J.: Effective efficiency of PV modules under field
conditions. Progress in Photovoltaics: Res Appl 15, 19–26 (2007)
5. Carlson, D.E., Romero, R., Willing, F., Meakin, D.H., Gonzalez, L., Murphy, R.:
Corrosion effects in thin film photovoltaic modules. Progress in
Photovoltaics: Res Appl 11, 377 (2003)
6. Al-Hasan, A.Y.: A new correlation for direct beam solar radiation received by
photovoltaic panel with sand dust accumulated on its surface. Solar Energy
63(5), 323–333 (1998)
7. Sukhatme, S.P.: Solar Energy: Principles of Thermal Collection and Storage.
Tata McGraw-Hill, New Delhi (2003)
8. Armstrong, S., Hurley, W.G.: A new methodology to optimize solar energy
extraction under cloudy conditions. Renew Energy 35, 780–787 (2010)
9. Abentin, J., Cueli, A.B., Diaz, J., Moracho, J., Lagunas, A.R.: New accredited
laboratory for PV module testing. Proceedings of the 20th European
Photovoltaic Solar Energy Conference, Barcelona (2005) Submit your manuscript to a
10. Akpabio, L.E., Udoimuk, A.B.: Characteristics distribution of total, diffuse and
direct solar radiation at Calabar. Glob J Pure Appl Sci 9(1), 45–49 (2003) journal and benefit from:
11. Osterwald, C.R., McMohan, T.J.: History of accelerated and qualification
testing of terrestrial photovoltaic modules: a literature review. Progress in 7 Convenient online submission
Photovoltaics: Res Appl 2017(1), 11–33 (2008) 7 Rigorous peer review
12. Malik, A.Q., Lim Chee, M., Tan Kha, S., Blundell, M.: Influence of temperature 7 Immediate publication on acceptance
on the performance of photovoltaic polycrystalline silicon module in the 7 Open access: articles freely available online
Bruneian climate. AJSTD 26(2), 61–72 (2010)
7 High visibility within the field
13. Arya, R.R., Carlson, D.E.: Amorphous silicon PV module manufacturing at BP
solar. Progress in Photovoltaics: Res Appl 10(2), 69–76 (2002) 7 Retaining the copyright to your article
14. Wohlgemuth, J.H., Conway, M., Meakin, D.H.: Reliability and performance
testing of photovoltaic modules, pp. 15–22. 28th IEEE PVSC, Egan
Submit your next manuscript at 7 springeropen.com
Convention Center, Anchorage (2000)