Available online at www.sciencedirect.com
Energy Procedia 00 (2014) 000–000
www.elsevier.com/locate/procedia
SHC 2013, International Conference on Solar Heating and Cooling for Buildings and Industry
September 23-25, 2013, Freiburg, Germany
BiPV system performance and efficiency drops: overview on PV
module temperature conditions of different module types
Laura Maturia*, Giorgio Belluardoa, David Mosera, Matteo Del Buonoa
a
EURAC, Viale Druso 1, Bolzano 39100, Italy
*tel.: +39-0471-055633; fax: +39-0471-055699; e-mail: [email protected]
Abstract
A typical problem of BiPV systems (Building Integrated Photovoltaic) is the power loss due to temperature increase, because
modules often operate close to the building envelope with low ventilation.
It is thus essential to properly evaluate and compare the PV temperature conditions of different PV module categories (in terms
of PV technology and material type). Several explicit correlations exist for the evaluation of the PV module temperature, among
which the simplest and most handy is a linear expression (i.e. Tmod=Tamb+k G) which links Tmod with the ambient temperature
(Tamb) and the incident solar radiation flux (G).
Within this expression the value of the dimensional parameter k, known as the Ross coefficient, depends on several aspects (i.e.
module type, wind velocity and integration characteristics).
However, dispersed values for this parameter can be found in literature (in the range of 0.02-0.06 K m2/W) according to different
module types, while more information are provided regarding different integration characteristics.
This paper aims at giving an overview of the value of k for different module types according to monitored data registered over
one year time period at the ABD-PV plant in Bolzano (Italy).
The highest values of k, which means the highest module temperature at a certain G, are registered for the three glass-glass (G-G)
module types (kTmod,c from 0.033 K m2/W to 0.037 K m2/W). The glass-tedlar (G-T) module types operate at slight lower
temperature values (kTmod,c ranging from 0.029 K m2/W to 0.032 K m2/W).
© 2014 The Authors. Published by Elsevier Ltd.
Selection and peer review by the scientific conference committee of SHC 2013 under responsibility of PSE AG.
Building Integrated Photovoltaic; BiPV; PV temperature; PV modules; PV outdoor monitoring
1876-6102 © 2014 The Authors. Published by Elsevier Ltd.
Selection and peer review by the scientific conference committee of SHC 2013 under responsibility of PSE AG.
Author name / Energy Procedia 00 (2014) 000–000
1. Introduction
A typical problem of BiPV systems (Building Integrated Photovoltaic) is the power loss due to temperature
increase, because modules often operate close to the building envelope with low ventilation (Nordmann and
Clavadetscher, 2003 [1]) (Trinuruk et al., 2009 [2]) (Maturi et al., 2012 [3]). The operating temperature in fact plays
a central role in the photovoltaic conversion process, as well as on the electrical efficiency and, hence, the power
output of a PV module depends practically linearly but rather strongly on Tmod, decreasing with it [4]. It is thus
essential to properly evaluate the module temperature profile for different PV working conditions.
Several explicit correlations exist for the evaluation of PV module temperature [2,4,5,6,7,8], among which the
simplest is a linear expression (i.e. Tmod=Tamb+k G) which links Tmod with the ambient temperature (Tamb) and the
incident solar radiation (G).
This paper aims at giving an overview of the values of k for different module categories according to monitored
data registered over one year time period (August 2012 – July 2013) at the ABD-PV plant in Bolzano (Italy)
allowing to evaluate and compare PV module temperatures of different module types.
Nomenclature
m-Si
a-Si
IR
ΔTmax
Tmod
Tmod,avg
Tmod,c
Tamb
k
R2
RMSE
G-G
G-Gbs
G-T
WF
NF
vwind
G
GSTC
Pn
PR
mono crystalline
amorphous silicon
infrared
maximum temperature gradient within the PV module (among four points)
PV module temperature
back of module temperature average of four points of the module back side
back of module temperature in the center of the module back side
ambient temperature
Ross coefficient
coefficient of determination
root-mean-square error
glass-glass PV module type
glass-glass PV module type, with a black sheet in the back side
glass-tedlar PV module type
PV module with frame
PV module without frame
wind velocity
irradiance
irradiance under Standard Test Conditions
array nominal power
array performance ratio
2. Experimental set-up
The modules investigated are listed on an anonymous basis in Table 1.
The six monitored modules are installed on a rack with a fixed inclination of 30° (grid connected system) almost
facing South, located in the ABD-PV plant of the airport “Aeroporto Bolzano Dolomiti” in Bolzano (North of Italy),
which is in operation since August 2010. The European Academy of Bozen/Bolzano (EURAC) is the scientific
responsible for monitoring the performance of the ABD-PV plant (Colli et al. [10]). Electric and production data
regarding PV modules are recorded automatically by the inverters with a frequency of 15 minutes, while the PV
back of module temperature is recorded every minute by means of temperature sensors (Pt100) attached on the back
Author name / Energyy Procedia 000 (2014) 0000–000
side of thhe modulles. The w
weather ddata are rrecorded eevery minnute withh the suppport of a meteo sttation whhich is
located inn the ABD
D-PV plannt area annd averageed over a 15-minuutes time iinterval. IIt is equippped with sensors ffor the
measurem
ment of irrradiance on horizoontal andd module plane, am
mbient tem
mperaturee, wind sppeed and directionn. The
acquisitioon system
m and meassurement errors aree describeed in detaiils by Fannni et al [111].
Table 1. Liist of investiigated moduules
nrr.
name
technollogy
Stratigrap
aphy
Frame
1
CIGS3
CIGS
glass-glaass (G-G)
WF
2
mc-Si44
m-Si-baack contact
glass-teddlar (G-T)
WF
3
mc-Si3
m-Si
glass-teddlar (G-T)
WF
4
mc-Si1
m-Si
glass-glaass (G-G)
NF
5
mc-Si22
m-Si
glass-glaass-black sheeet (G-Gbs)
NF
6
1j-a-Si22
a-Si
glass-teddlar (G-T)
WF
V arrays: aall modulees consideered for thhe analysis in this paper aree installedd close
Figuree 1 shows the moniitored PV
to each otther, in iddentical coonditions (i.e. tilt anngle 30°, azimuth 88.5° Westt of Southh, rack moounted).
The baack of moodule tem
mperature values arre acquired with a Pt100 attaached in the centerr of the bback side of the
module loocated in the midddle of eachh array. T
The globaal irradiannce is meaasured witth a pyrannometer ((Kipp & Z
Zonen
CMP) at tilt of 30°° (same pplane as thhe modulees). The aambient teemperaturre (Tamb) iis acquireed with a Pt100 whhich is
covered bby a weatther and rradiation protection shield. The windd velocityy and direection aree measureed with a sonic
anemomeeter. All w
weather daata are loggged withh one minuute time resolution and averaage on 155-minutes time interrvals.
a
b
d
c
e
f
Figure 1: aarrays of the six monitorred PV moddules. (a) CIG
GS3, (b) mcc-Si4, (c) mcc-Si3, (d) m
mc-Si1, (e) m
mc-Si2, (f) 1jj-a-Si2
3. Methoodology
3.1. Therm
mographyy measureements
The thhermograpphy technnique is uused for a preliminnary evaluuation off the backk-side tem
mperature distributiion of
each moddule.
Therm
mography measurem
ments werre perform
med at a ssunny clooudless daay, with thhe modulee workingg at mpptt, with
an angle vview highher than 60° to the m
module glass and ssetting thee right em
missivity ddepending on the m
module typpe.
The IR
R-image oof the mc-Si2 moddule in Figgure 2 (a)) shows a temperaature gradient withiin the pannel up to 7.1°C
along linee 1 as displayed inn Figure 2 (b). In thhis case, tthe tempeerature peak is reacched in thhe middle of the moodule:
the lateraal parts arre significcantly coooler, due tto the abssence of thhe metal frame. Thhe same ttemperatuure distribbution,
i.e. with a significant tempeerature inncrease in the midddle, is obsserved forr the otheer module without frame (i.ee. mcSi1 moduule). The other moddules, whhich have a metal fframe (i.e. CIGS3, mc-Si4, m
mc-Si3, 11j-a-Si2), present a more
homogenneous tempperature ddistributioon within the moduule. Figurre 3 show
ws the typpical tempperature ddistributionn of a
module w
with framee: in this ccase, the ttemperatuure gradiennt along line 1 is 2..5°C.
a
b
Author naame / Energgy Procedia 00 (2014) 0000–000
Temperature °C
50,3
Line 1 Max: 50
0,0°C Min: 42,9°C
C Avr: 47,6°C
48
46
44
42,6
1
5
10
15
20
25
Pixels
30
35
40
45
50
Figure 2: (a) IR-imagge of the mc-Si2 modulee, (b) tempeerature distriibution referrred to line 1 of the IR-iimage show
wn in (a)
b
a
Temperature °C
45,31
Line 1
Line 1 Max: 45,2°C
4
Min: 42,7°C Avr: 44,5°C
44,5
44
43,5
43
42,55
1
5
10
15
20
25
30
35
40
Pixels
45
50
55
60
65
70
0
76
Figure 3: (a) IR-imagge of the mc-Si4 modulee, (b) tempeerature distriibution referrred to line 1 of the IR-iimage show
wn in (a)
mperaturee measureements
3.2. Addiitional tem
In ordder to bettter evaluaate the am
mplitude oof the tem
mperature gradient within thhe panel oover diffeerent condditions,
additionaal measurrements arre carried out by m
means of a portable thermomeeter with four therm
mocouplees.
The bback side ttemperatuure is thuss measured for at leeast three sunny dayys, in fouur points oof each moodule (i.e. in the
center, inn the cornner and in the two laateral sidees, as indiicated in tthe IEC 60891 [12]]).
The m
maximum
m registereed temperrature graddient (ΔT
Tmax) is repported in Table 2 ffor each m
module. T
The tempeerature
gradient ΔTmax is ddefined ass the diffeerence bettween the max and min values amongg the four measuredd points.
Valuees of ΔTmaax range ffrom 4.5°°C to 6.3°°C for moodules witth frame. Similar vvalues are reported in other sstudies
(i.e. [13,14]) that rreport tem
mperature gradient within fraamed moddules of aabout 3°C--5°C.
me (i.e. m
mc-Si1 annd mc-si2)), measurred valuess of ΔTmaxx exceed 88°C. Thesse data
For thhe two moodules wiithout fram
confirm the trend shown inn the therrmographyy measureements, w
where the highest Δ
ΔTmax is ddisplayed for the m
modules
without fframe.
T
Table 2. valuues of ΔTmax for each moodule (i.e. m
maximum tem
mperature ggradient withhin the PV m
module amoong four meaasured
pooints). Fram
med moduless in bold
nr.
Δ
ΔTmax [°C]
1
4.5
2
5.7
3
6.1
4
8.0
5
88.8
6
6.3
3.3. Temp
mperature data
The P
Pt100 senssor is attaached in thhe center of the moodules back side, aand thus a completee data-sett on tempeerature
values ovver a whoole year iss availablee just for tthe modulles centraal point.
On thhe other hhand, paraagraphs 3.1and 3.22 show thhat there ccould be a significcant temperature grradient w
within a
single m
module, especially iif it is witthout fram
me. Conseequently, the tempperature m
measured iin one sinngle pointt could
not be reepresentative of thee module ttemperatuure.
Author name / Energy Procedia 00 (2014) 000–000
For this reason we also used the additional data measured in four points (as specified in paragraph 3.2) to
calibrate the acquired temperature values.
The spatial average temperature measured in the four points (Tmod,avg) is plotted against the temperature measured
in the center of each module Tmod,c and the calibration coefficients are evaluated through a linear regression,
according to the following equation:
∗
,
,
(1)
The resulting coefficients a and b are reported in Table 3 for each module.
Table 3. calibration coefficients a and b referred to equation (1) and coefficient of determination related to each linear
regression. Framed modules in bold
nr.
a
b
R2
1
0.583
0.983
0.998
2
0.507
0.976
0.998
4
1.318
0.943
0.998
3
0.247
1.004
0.998
5
1.350
0.948
0.999
6
-0.042
1.011
0.997
3.4. Filtering procedure
A filtering is performed on the 15-minutes data before calculating the Ross coefficient k. The first filter exclude
points corresponding to a ratio of diffuse and global irradiance higher than 0.20 in order to consider just clear sky
conditions, i.e. conditions in which the direct irradiance component is predominant on the diffuse one.
A second filter considers points as valid only within the range PRavg±σ, where:
 PRavg is the average value of performance ratio of the PV array for the considered period, defined by the
international standard IEC 61724 [15] as:

where E is the energy production (Wh), Pn the array nominal power (W) and GSTC the irradiance under STC
conditions (1000 W/m2)
σ is the standard deviation
In this way, outlier points are excluded which correspond to situations in which the PR is too high or too low,
respectively due to irradiance on the module larger than the one measured by pyranometer, and irradiance on the
module lower than the one measured by pyranometer. This condition may occur with a not-uniform cloud cover or
shadowing due to obstacles, such as mountains, and is more evident the longer is the distance between module and
irradiance sensor.
Finally, points corresponding to morning time are filtered out. The reason for this can be seen in Figure 4, which
represents Tmod,avg-Tamb against irradiance after applying the two previously described filters. Points related to glassglass modules (as for example mc-Si2) show a different trend depending on the time of the day. A lower curve is
clearly visible, which correspond to points measured in the morning (approximately before 1 PM). During this
period of the day, the glass-glass module seem to be affected by thermal inertia: the module heats up due to
increasing ambient temperature and irradiance, but the back of the module temperature does not increase
simultaneously, lagging up to 5 degrees behind the cell temperature until thermal equilibrium is reached . The same
effect occurs also during the afternoon when the module cools down but in a minor extent, and the linearity of
Tmod,avg-Tamb with G is less affected.
Thermal inertia has been observed by different authors [16,17], and it seems to occur also on glass-tedlar module
(see Figure 4a), but is less visible. Its effect on different module types will be deeper analyzed in a future study.
Author naame / Energgy Procedia 00 (2014) 0000–000
a
b
Figurre 4: Plots off Tmod,avg-Tammb against irrradiance for a) a glass-teedlar modulle type b) a gglass-glass m
module typee; ratio of diiffuse to gloobal
irradiance andd PRavg±σ fillters appliedd
4. Results
4.1. Rosss coefficieents overvview
The vvalues of Tmod,avg-Taamb are plootted agaiinst irradiiance for each moddule type (see Figuure 5). Thhe plotted values
refer to data filttered acccording too the proocedure ddescribedd in paraagraph 3.44, with tthe addittional filtter 0.9
m/s<vwindd<1.1m/s..
The lleast-squaares-fit linne througgh the sett of data is evaluaated with an addittional connstraint, ssuch as Tmod,avgTamb=0˚C
C when G
G=0 W/m2, for the physical meaning related too the therm
mal equillibrium, inn steady sstate condditions,
when no irradiancce is preseent. The R
Ross coeffficient k iss evaluateed as the sslope of thhat line.
wind veloocities of 1m/s (±0..1m/s).
Tablee 4 providdes an oveerview off the evaluuated k vvalues, whhich are vvalid for w
kTmod,c is defined aas the Ross coefficcient evaluuated plotting tempperature ddata meassured justt in the baack side m
module
center; kTmod,avg is defined as the Rooss coefficient evalluated plootting “caalibrated” temperatture data which takke into
account also periipheral values, byy applyingg the coeefficients presenteed in paraagraph 3.3. In thee last columns,
kTmod,avg,GGarc. is alsoo reportedd, which ccorresponnds to kTmood,avg calcuulated takiing into aaccount pooints filterred as desscribed
in paragrraph 3.4, with the only diff
fference of excludinng valuess of G<8000W/m2 iinstead off values m
measured in the
morning, accordinng to the oobservatioons of Aloonso Garccía and Baalenzateguui [17].
The ddifference between kTmod,c annd kTmod,avvg is negliigible for WF moddule typess, while itt becomess slightly higher
for NF m
module tyypes. Thiss finding iis in agreeement wiith resultss shown inn paragraaphs 3.1 aand 3.2, showing thhat the
highest Δ
ΔTmax occuurs for thhe NF moddules andd the highhest back-side tempperature vvalues are registered in the m
module
center. Inn fact, thee values oof kTmod,avgg takes innto considderation allso periphheral moddule tempeerature vaalues, whiich are
highly innfluencedd not onlyy by the sstratigraphhy of thee module, but also by the aabsence oof the fram
me (see m
module
temperatture distriibution off a NF moodule in F
Figure 2). In generral the vaalues of kTmod,avg,Garrc. are sligghtly loweer than
kTmod,avg, but resultt in the saame trend..
mpare thee PV backk of moduule tempeerature off differentt moduless stratigraaphy, the least-squaares-fit
In ordder to com
lines of eeach moduule type, aaccordingg to the vaalues of kTmod,c of T
Table 4 , aare plottedd in Figurre 5.
Author name / Energyy Procedia 000 (2014) 0000–000
Figuree 5: Plots off Tmod,avg-Tammb against irrradiance for each modulle type; all ffilters applieed
By consideringg the backk of moduule tempeerature Tmmod,c, the highest P
PV tempeeratures aat a certaiin irradiannce G
For these module ttypes the value of kTmod,c
corresponnd to the three glasss-glass (G-G) module typees (see Figgure 6). F
ranges froom 0.033 K m2/W to 0.037 K m2/W ((see red linnes in Figgure 5). The glass-ttedlar (G--T) modulle types opperate
at slight lower tem
mperaturee values, w
with kTmood,c ranginng from 00.029 K m2/W to 00.032 K m2/W (seee blue linnes in
mple, acccording too the evaluuated kTmood,c, in connditions oof irradiannce of 10000W/m2 aand air veelocity
Figure 5). For exam
mest moduule (i.e. C
CIGS3) opperates at a temperature whiich is 8˚C
C higher thhan the ccoolest onne (i.e.
of 1m/s, tthe warm
mc-Si4), possibly affectingg the PV power pproductionn up to 33% (e.g. considering a tem
mperaturee coefficieent of
C for CIGS
S3 and off 0.38%/°C
C for mc--Si4).
0.36%/°C
Taable 4. Ross coefficient k overview evaluated eexperimentallly
n
nr.
nam
me
tecchnolog
y
Stratigraphyy
Frame
kTmod,c
(K m2/W)
RMSEk
_Tmod,c
kTmod,avg
(K m2/W)
RMSEk
_Tmod,avg
kTmod,avg,GGarc.
(K m2/W)
RMS
SEk
_Tmod,avg,
Garc.
1
CIG
GS3
CIIGS
glass-glass (G-G)
WF
0.037
2.3
0.037
2.4
0.036
2.1
2
mc-S
Si4
m--Si-back
conntact
glass-tedlarr (G-T)
WF
0.029
2.0
0.028
2.1
0.028
1.9
3
mc-S
Si3
m--Si
glass-tedlarr (G-T)
WF
0.032
2.2
0.032
2.2
0.031
1.9
4
mc-S
Si1
m--Si
glass-glass (G-G)
NF
0.033
2.0
0.031
2.2
0.030
1.9
5
mc-S
Si2
m--Si
glass-glass--black
sheet (G-Gbbs)
NF
0.035
2.4
0.033
2.5
0.031
2.0
6
1j-a--Si2
a-S
Si
glass-tedlarr (G-T)
WF
0.031
1.7
0.032
1.6
0.031
1.5
Author naame / Energgy Procedia 00 (2014) 0000–000
Figure 66: Plots of thhe least-squaares-fit liness of each moodule type, aaccording too the kTmod,c vvalues
4.2. Windd dependeence of Rooss coefficcient
The rresults preesented abbove, refe
fer to a coonstant w
wind veloccity of 1m
m/s (±0.11m/s). Annyway, thhe wind velocity
could haave a stronng influennce on the k parameeter [4].
A corrrelation bbetween k and the wind vellocity is thhus evaluuated for each moddule type,, by plottiing the k values
(i.e. (Tmood,avg-Tamb))/G), againnst wind vvelocity uusing expeerimental data filteered accorrding to thhe proceduure discusssed in
paragrapph 3.4.
The eexperimenntal correllation fouund for eaach moduule type, iin agreem
ment with earlier sttudies [4,118,19], is in the
form:
∗ exp
p
∗
(2)
G), againsst wind veelocity forr each moodule typee are preseented.
In Figuree 7, the pllots of k (ii.e. (Tmod,aavg-Tamb)/G
Figgure 7: Plots of k (i.e. (T
Tmod,avg-Tamb)//G), against wind velociity for each module typee
Author name / Energy Procedia 00 (2014) 000–000
The experimental coefficients a, b and c referred to equation (2), are summarized in Table 5. As expected, results
obtained from this equation for vwind=1, are in good agreement with the ones presented in paragraph 4.1.
Table 5. experimental coefficients a, b and c for each module type, referred to the formula (2)
nr.
name
technology
Stratigraphy
Frame
a
b
c
RMSE
1
CIGS3
CIGS
glass-glass (G-G)
WF
2
mc-Si4
m-Si-back contact
glass-tedlar (G-T)
WF
0.011
0.011
0.016
0.013
0.015
0.008
0.042
0.036
0.030
0.031
0.033
0.034
0.466
0.748
0.634
0.539
0.530
0.373
0.0033
0.0035
0.0031
0.0032
0.0040
0.0026
3
mc-Si3
m-Si
glass-tedlar (G-T)
WF
4
mc-Si1
m-Si
glass-glass (G-G)
NF
5
mc-Si2
m-Si
glass-glass-black sheet (G-Gbs)
NF
6
1j-a-Si2
a-Si
glass-tedlar (G-T)
WF
5. Conclusions
In this paper an overview of k values evaluated experimentally for six module types is provided and a wind
dependence of k has been introduced.
The main conclusions follow:
 The highest values of k, which means the highest back of the module temperature at a certain G, is given by
the three glass-glass (G-G) module types (kTmod,c from 0.033 K m2/W to 0.037 K m2/W ). The glass-tedlar
(G-T) module types operate at slight lower temperature values (kTmod,c ranging from 0.029 K m2/W to 0.032
K m2/W).
 The thermography measurements show that the NF modules present a temperature gradient within the
panel (around 7.1°C) quite higher compared to the WF modules (around 2.5°C). This behavior is confirmed
by other measurements carried out with 4 temperature sensors placed in 4 points of the module back side,
which report a ΔTmax from 4.5°C to 6.3°C for WF modules and ΔTmax exceeding 8°C for NF modules. This
could be due to the fact that lateral parts of the NF modules are significantly cooler, due to the absence of
the metal frame. Consequently, temperature values measured in one central point of the module could be
considered as representative of the average operating temperature for WF modules while, for NF modules,
it could be significantly different.
 The role of the module thermal inertia should be taken into consideration when analyzing module back-side
temperature data. As the thermal inertia of G-G modules is higher than the one of G-T modules, the effect
of thermal inertia is more visible for these type of modules.
Acknowledgements
The authors would like to thank the fund EFRE-Provincia Autonoma di Bolzano Alto Adige (Fondo Europeo di
sviluppo regionale) for the financial contribution through the project 2-la-97 “PV-Initiative”, as well as the
European Union for funding the project n.310220 "On-the-fly alterable thin-film solar modules for design driven
applications" within the Seventh Framework Programme.
Reference
[1] Nordmann T, Clavadetscher L. Understanding temperature effects on PV system performance. Proceedings of 3rd World Conference on
Photovoltaic Energy Conversion, 2003, Vol. 3, p. 2243 –2246.
[2] Trinuruk P, Sorapipatana C, Chenvidhya D. Estimating operating cell temperature of BIPV modules in Thailand. Renewable Energy, 2009,
Vol. 34, 2515–2523.
Author name / Energy Procedia 00 (2014) 000–000
[3] Maturi L et al. BiPV façade systems: effectiveness of a passive strategy to improve the PV performance. Proceedings of the 27th European
Photovoltaic Solar Energy Conference, 2012, p. 4203 – 4206.
[4] Skoplaki E, Boudouvis AG, Palyvos JA. A simple correlation for the operating temperature of photovoltaic modules of arbitrary mounting.
Solar Energy Materials and Solar Cells. 2008 Nov, Vol. 92, p.1393–1402.
[5] Skoplaki E, Palyvos JA. Operating temperature of photovoltaic modules: A survey of pertinent correlations. Renewable Energy, 2009, Vol.
34, p. 23–29.
[6] Chenni R et al. A detailed modeling method for photovoltaic cells. Energy, 2007, Vol. 32, p.1724–1730.
[7] Risser VV, Fuentes MK. Linear regression analysis of flatplate photovoltaic system performance data. Proceedings of the 5th E.C.
photovoltaic solar energy conference, 1983, p. 623–627.
[8] Servant JM, 1985. Calculation of the cell temperature for photovoltaic modules from climatic data. Proceedings of the 9th biennial congress
of ISES – Intersol 85, Montreal, Canada, extended abstracts, p. 370.
[9] Ross RG, Smokler MI. Flat-plate solar array project final report-VolVI: Engineering Sciences and reliability. Report DOE/JLP-1012-125,
1986.
[10] Colli A et al., 2010. Performance monitoring of different PV technologies at a PV field in Northern Italy. Proceedings of the 25th European
Photovoltaic Solar Energy Conference and Exhibition / 5th World Conference on Photovoltaic Energy Conversion, 2010, p. 4344 – 4349.
[11] Fanni L et al. How accurate is a commercial monitoring system for photovoltaic plant?. Progress in Photovoltaics: Research and
Applications, 2012.
[12] International Standard IEC 60891, 2009-12. Photovoltaic devices – Procedures for temperature and irradiance corrections to measured I-V
characteristics.
[13] Buerhop C, Scheuerpflug H, Weissmann R. The role of infrared emissivity of glass on IR-Imaging of PV-plants. Proceedings of the 26th
European Photovoltaic Solar Energy Conference 2011; 3413 – 3416.
[14] Lee Y, Tay AAO. Finite element thermal analysis of a solar photovoltaic module. Energy Procedia 2012;15;413-420.
[15] International Standard IEC 61724, 1998. Photovoltaic System Performance monitoring – Guidelines for Measurement, Data Exchange, and
Analysis.
[16] King DL, Boyson WE, Kratochvill JA. Photovoltaic array performance model. Sandia Report SAND2004-3535, 2004.
[17] Alonso García MC, Balenzategui JL. Estimation of photovoltaic module yearly temperature and performance based on Nominal Operation
Cell Temperature calculations. Renewable Energy, Vol. 29, 2004, p. 1997-2010.
[18] Kurnik J, Jankovec M, Brecl K, Topic M. Outdoor testing of PV module temperature and performance under different mounting and
operational conditions. Solar Energy Materials and Solar Cells. 2011, Vol. 95, p. 373–376.
[19] Hart GW, Raghuraman P. Simulation of thermal aspects of residential photovoltaic systems. MIT Report DOE/ET/20279-202, 1982.