Professional Documents
Culture Documents
EUB_7_133
Photovoltaics
2012/13
Greek
λ m Wavelength
Constants
c 299,792,458 m·s-1 Speed of light
h 6.63×10-34 J·s Planck’s Constant
e 1.602×10-19 C Elementary Charge
q 1.602×10-19 C Charge of the electron
k 1.381×10-23 J∙K-1 Boltzmann’s constant
3. Basic principles
Although there are an increasing number of materials being used to construct PV cells we
shall examine the principles using the most common, silicon, which is used in about 80% of
all installations.
Each shell corresponds to a particular energy level and there are no intermediate levels. If an
electron receives energy, by absorbing a photon, it can move up to the next shell. If it loses
energy by emitting a photon then it will drop back to a lower shell.
The outer shell, known as the valence shell, determines the element’s chemical properties –
how it combines with other atoms. If the valence shell is full, the element will be inert; if it is
not full, then it can combine with other atoms by sharing electrons.
The valence shell of pure silicon is not full and in its natural state will
share an electron from four adjacent silicon atoms forming a cubic
crystal lattice as shown in Figure 3. Above the valence band is the
conductance band which is empty at 0 oK, but if electrons in the valence
band absorb sufficient energy they can jump into the conductance band.
Figure 3 Structure of
In metals electrons move freely into the conductance band creating an monocrystalline silicon
electron cloud that can move through the solid transporting both
electrical charge and heat. Conversely, insulators have electrons locked in the valence shell
unable to move freely and therefore insulators do not conduct heat or electrical charge.
3.2. Doping
The ability of semiconductors to conduct charge, as you might expect, lies somewhere
between conductors and insulators. Natural semiconductors are to be found in Group IV of
the periodic table and include silicon and germanium. Others may be formed by binary
compounds of elements from adjoining groups. E.g. gallium arsenide (III-V), cadmium
telluride (II-VI).
The electrical properties of pure semiconductors may be modified by the process of doping.
Doping consists of introducing small quantities of another element into the crystalline
structure of the silicon. Doping silicon with group V elements such as phosphorus, which has
five electrons in its outer shell, introduces an extra valence electron in the lattice which can
become unbounded creating an electrically conductive n-type semiconductor.
Doping with a group III elements such as boron, which only has three electrons in the outer
shell, creates a broken bond in the lattice which is free to move and forms an electrically
conducting p-type semiconductor (Figure 4).
If an n-type and p-type are placed together it will create a p-n junction. Electrons will diffuse
across the junction from the n-type material and fill the holes in the p-type material. As the
electrons migrate they will leave the phosphorus atoms in the n-type material with a positive
charge (+ve ion) and as they fill the holes the p-type material becomes negatively charged (-
ve ion). This build-up of charge creates a region depleted of charge carriers across which an
electric field opposes the flow of electrons and the flow will eventual stop.
c= λf 3-1
When dealing with photons it is usual to use units of electron volt (eV) rather than Joules.
Dividing by the elementary charge, e gives E in eV
hc
E= 3-4
λe
We can use Equation 3 -4 to find the and wavelength corresponding to a given band gap, b .
ch 1
λ= ∙ 3-5
e b
(Note: the wavelength is in metres.)
The band gap and corresponding wavelengths for various PV materials are shown in Figure 6.
Example
If a load is connected, a current will flow with electrons leaving the cell, passing through the
load, and returning to the other terminal of the cell where they recombine with the waiting
holes.
Not all the electrons make it to the terminal. When they leave the depleted region and
diffuse through the n-region they may meet a boron atom with a missing electron (a hole)
and recombine. The electron will no longer generate a current and recombination reduces
the efficiency of the cell.
In Solar Resource we saw that solar radiation is distributed over a wide spectrum. Any
radiation at a wavelength having less energy than the band gap will not move electrons into
the conduction band and therefore will not generate electricity. In the example calculation
above radiation of a wavelength less than 1,117nm will not be converted to electricity.
Radiation of a shorter wavelength having energy greater than the band width will dislodge
electrons, but energy possessed by photons in excess to
the band with (1.11 eV for silicon) will simply heat the
cell.
a. Monocrystalline Silicon
A large, pure crystal is grown, cut into thin wafers and doped. The process is expensive
but the technology is well established.
b. Polycrystalline
Researchers are working in many different areas to overcome the Shockley–Queisser limit of
33% power efficiency for single band gap solar cells as discussed in Section 3.4. Most
attempts remain in the laboratory but two techniques are close to commercial use.
4.4. Summary
Silicon Thin Film
Panel Type
Monocrystalline Polycrystalline Amorphous Silicon CIGS CdTe
Peak Power W/m2 160 140 120 100 50
Typical Efficiency/% 16 14 12 9.5 5
Area for 1 kWp 6.25 7.1 8.3 10 20
Table 2 Summary of PV performance data
2
A typical silicon cell has a voltage around 0.5 – 0.6 V and so 36 cells in series produces a
voltage of 18 – 21.6 V which is just right for charging 12 V lead acid batteries.
The test data provided by manufacturers is shown in Table 4. Manufacturers will provide at
plot of current versus voltage (an I-V curve) measured under STC.
The short-circuit current, I SC occurs when the module terminals are shorted and there is no
load resistance.
If they are connected in series they form what is known as a string. Each module will share
the same current but the voltage across the string will be the sum of the voltages across
each module. If they are connected in parallel the total current will be the sum of the
module currents and the voltage will be the same as a module.
The published temperature coefficients (Table 4) indicate the degree of dependence which is
also shown in Figure 14. Notice that the short circuit current is the least affected, reducing
slightly with an increase in cell temperature. For practical purposes this dependence is often
neglected. Open circuit voltage is more strongly affected, increasing with an increase in
temperature.
Maximum Voltage
This occurs when the modules are open
circuit and depends upon the cell
temperature. The sensitivity to
temperature change, the temperature
coefficient of V oc , is part of the
standard test and provided by
manufactures.
V oc =n × ( V oc + ΔV oc × ΔT c )=¿
This compares with a V oc of 21.6 V under standard test conditions – an increase of nearly
9%.
Maximum Current
The maximum current produced by a cell
occurs under short circuit conditions and is
affected slightly by temperature, but more
significantly by the irradiance. Even in the UK
the irradiance can exceed the STC of 1000
W·m-2 if the sun is high in the sky, the sun is
not obstructed, and there is significant diffuse
radiation from surrounding cloud.
A design multiplication factor of 1.25 is
normally used for I sc .
The above factors do not include the initial high output during the first few weeks of
operation, nor does it include allowances for fouling (10-15%).
The above factors are for monocrystalline cells, other types of cell have different
characteristics. Peak values for V oc and I sc should be determined for individual
cases.
T c =T a+ ( NOCT−20
0.8 )G
This equation may be used to predict the cell temperature when operating under non-
standard conditions. This is a very simplified approach and subject of predicting cell
temperature is discussed further in Section 8.
Module Data
Number of cells 34
Cell size 125mm x125 mm
I sc 3 A
∆ I sc negligible
V oc 20.4
∆ V oc -2.3 mV·K-1
Pmax 45.9 W
NOCT 43oC
Solution
Short-circuit current
Ignoring the temperature coefficient for I sc which is small:
G
I sc , act =I sc, STC ×
G STC
Where G = actual irradiance, kW·m-2
0.7
I sc , act =3.0× =2.10 A
1.0
Open-circuit voltage
Adjusting for the temperature dependence of V oc
V oc =20.4−2.28=18.1V
Peak Power
Calculate the module Fill Factor under Standard Test Conditions:
Pmax 45.9
FF= = =0.75
I sc V oc 3.0 × 20.4
Assuming that the Fill Factor is independent of ambient temperature, T a and irradiance, G ,
the power output under the actual operating conditions is:
Efficiency
Cell efficiency is given by:
Pmax
η=
Ac G
45.9
η= =0.086
( 34 × 0.125 ×0.125 ) ×1000
6.1. Shading
When deciding on the location and orientation of PV arrays careful consideration must be
given to shading from distant and nearby objects[ CITATION DGS08 \l 2057 ]. Analysis
requires the use of software such as EcoTect of IES<VE>3 to model potential obstructions
throughout the year.
A shadow not only reduces the area of module in direct sunshine but turns shaded cells into
resistors.
6.2. Soiling
Soiling (or fouling) is the build-up of dirt on the glazed face of the collector. Increased
collector tilt improves the self-cleaning properties of a module (providing it rains) and there
is research into surface treatments to enhance the self-cleaning properties.
In some climate zones snow may cover PV modules for extended periods, even when
significant solar radiation is available.
3
LSBU students can download Ecotect for free and buy a student licence for IES<VE> for £50
per annum.
Module performance data is based on a zero angle of incidence (i.e. radiation normal to the
POA). Unless the PV array is mounted on a 2-axis tracker, the angle of incidence will vary
throughout the day as the sun moves through the sky. As the angle of incidence increases,
the radiation reaching the PV will decrease because of increased reflection from the glass
cover. A correction is normally only made to the beam radiation, but there is also an effect
on the diffuse radiation. An Incident Angle Modifier for beam radiation, IAM B may be
defined as:
τ (θ )
IAM B = 6-6
τ (0 )
n2 v 1 sin θ1
= =
n1 v 2 sin θ2
1-0.5*(A+B)
The refractive index of air, n1 =1.0003, which can be taken as 1. The angle of refraction, θ2
is therefore
−1
θ2=sin ( n2 sinθ 1 ) 6-7
2 2
− ( KL/ cos θ2)
τ ( θ ) =e
[ (
1−
1 sin ( θ2−θ 1 ) tan ( θ2 −θ1 )
+
2 sin 2 ( θ2 +θ 1) tan 2 ( θ 2+θ 1 ) )] 6-8
Where
θ = incidence angle on the plane of the array.
b 0 is normally taken as 0.1 for a thermal collector, but PVSYST recommends 0.05 for PVs.
Notice that there is a discontinuity at θ=90o and some authors recommend that the
expression should not be used for θ>80 o. This not a serious problem as at such low angles
little electricity is being produced.
It should be noted that the incident angle modifier is affected by the degree of soiling.
7. Model of a PV Cell
In section 5.5 we calculated the new maximum power and efficiency of a PV operating under
non-standard conditions. This is often sufficient for our modelling needs, however there are
times when we would like to generate the complete I-V curve under non-standard
conditions. Such a model is required for the software in a maximum power point tracker.
A variety of models are available based on equivalent circuits[ CITATION Cas02 \l 2057 ],
[ CITATION Gra11 \l 2057 ] and the simplest is present below.
4
Also known as the Beer–Lambert–Bouguer law, or various combinations thereof.
The shunt resistor R sh models shunts at the emitter layer or cell border. Recombination
losses may be modelled by the addition of a second diode, but are not considered here.
WhereV + I ∙ Rs is the voltage across the diode, n is the diode ideality factor (dimensionless
and usually between 1 and 2 for a single junction cell), I 0 is the saturation current, and V T is
the thermal voltage given by:
V + I R s + I sh Rsh
Rearranging
V + I Rs
I sh = 7-12
R sh
Substituting equations 7 -11 and 7 -12 into equation 7 -10 gives the complete governing
equation for the single diode model
V + I ∙ Rs V + I ∙ Rs
[
I =I ph−I 0 exp
( nVT ) ]
−1 −
R sh
7-13
To use the equation five parameters must be known, all of which may be a function of cell
temperature:
Model parameters can be estimated from manufacturers’ data measured a standard test
conditions.[ CITATION Duf91 \l 2057 ][ CITATION Sch07 \l 2057 ].
To determine five parameters we need to know the current and voltage at five measured
points. However, manufacturers normally only publish three: open circuit, short circuit and
maximum power point.
PVSyst is a sophisticated tool for modelling the performance of solar systems. It uses the
single diode model and the manual[ CITATION PVs13 \l 2057 ] explains how the equations
are solved.
8. Module Temperature
We have seen that a PV module does not convert all the incident radiation into electricity.
The radiation that is not converted will heat the module, increasing its temperature until the
long-wave radiative and convective heat losses are equal to the radiant heat input. At this
point the module will reach an equilibrium temperature.
Factors that determine the module temperature include incident solar radiation, air
temperature, wind speed and direction, mean radiant temperature of the surroundings, the
geometry of the module installation, and the construction of the module itself.
It should be noted that the module rarely reaches an equilibrium temperature as the climate
variables are constantly changing and change is not instantaneous because of the thermal
capacity of the module. Other real-world complexities which are rarely considered:
Faiman published a module temperature model that has been widely adopted.[CITATION
Fai08 \l 2057 ] The model uses a modified form of the Hottel-Whillier-Bliss equation used to
analyse flat-plate solar collectors.
G
T c =T a+ 8-14
U +U '1 v
'
0
Values of U '0 and U '1 are found by fitting data. Faiman found the values to be fairly constant
for the seven collector types he studied and proposed the following values:
Example
What will be the temperature of a PV module when the irradiance is 700
W·m-2, the ambient temperature is 26oc and the wind speed is 4 m·s-1?
G 700
T c =T a+ =26+ =39.2o C
' '
U +U 1 v
0
24.9+7 × 4
10. References
Alsema, E. A., & de Wild-Scholten, M. J. (2006). Environmental impacts of PV electricity
generation - a critical comparison of energy supply options. 21st European
Photovoltaic Solar Energy Conference,. Dresden.
Castañer, L., & Silvestre, S. (2002). Modelling photovoltaic systems using PSpice. Chichester:
Wiley.
CIBSE. (2000). TM 25 Understanding building integrated photovoltaics. London: Chartered
Institution of Building Services Engineers.
De Soto, W., Klein, S. A., & Beckman, W. A. (2006). Improvement in the validation of a model
for photovoltaic array performance. Solar Energy 80, 78-88.
DGS. (2008). Site surveys and shading analysis. In DGS, Planning and installing photovoltaic
systems (pp. 129-150). London: Earthscan.
Duffie, J. A., & Beckman, W. A. (1991). Solar Engineering of Thermal Processes (2nd ed.). New
York: Wiley.
Faiman, D. (2008). Assessing the outdoor operating temperature of photovoltaic modules.
Progress in Photovoltaics: Research and Application, 16(4), 307-315.
Gray, J. L. (2011). The Physics of the Solar Cell. In S. Hegedus, & A. Luque, Handbook of
Photovoltaic Science and Engineering. John Wiley.
IEC. (2005). 61215 Crystalline silicon terrestrial photovoltaic (PV) modules – Design
qualification and type approval. Geneva: International Electrotechnical Commission,.
Kazuhiko, K. (2001, March). A life-cycle analysis on thin-film CdS/CdTe PV modules. Solar
Energy Materials and Solar Cells, 67(1-4), 279–287.
Martin, N., & Ruiz, J. M. (2001). Calculation of the PV modules angular losses under field
conditions by means of an analytical model. Solar Energy Materials & Solar Cells 70,
25-38.
Martin, N., & Ruiz, J. M. (2002). A New Model for PV Modules Angular Losses Under Field
Conditions. International Journal of Solar Energy , 22(1) 19-31.
Martin, N., & Ruiz, J. M. (2005). Annual Angular Reflection Losses in PV Modules. Progress in
Photovoltaics: Research and Applications, 13 75-84.
Photon. (2013, September 30). Solar Module Database. Retrieved 9 30, 2013, from
Photon.info: http://www.photon.info/photon_site_db_solarmodule_en.photon
PVsyst. (2013, 9 30). PV Module - model description. Retrieved 9 30, 2013, from PVsyst 6
Help: http://files.pvsyst.com/help/index.html
Sandia National Laboratories. (2013). Retrieved 5 30, 2013, from PV Performance Modeling
Collaborative: http://pvpmc.org/
Sandia National Laboratories. (2013). Single Diode Equivalent Circuit Models. Retrieved 5 28,
2013, from PV Performance Modelling Collaborative: http://pvpmc.org/modeling-
steps/module-iv-curve/diode-equivalent-circuit-models/
Schumacher, J., Eicke, U., Pietruschka, D., & Catani, A. (2007). Exact Analytical Calculation of
the One-diode Model Parameters from PV Module Data Sheet Information. 22nd
European Photovoltaic Solar Energy Conference. Milan.
Souka, A. F., & Safwat, H. H. (1966). Determination of the optimum orientations for the
double-exposure, flat-plate collector and its reflectors. Solar Energy, 10, 170-174.
(h) DC to AC Conversion
Most modern inverters operate with a conversion efficiency of 95%, but this does vary with
DC voltage, AC output power and sometimes temperature.
(i) AC Losses
V + I ∙ Rs V + I ∙ Rs
I =I ph−I 0 exp[ ( a )
−1 − ]
R sh
12-15
I ph A Photo current
Io A Diode reverse saturation current
Rs Ω Series resistance
R sh Ω Shunt resistance
a Modified ideality factor
sh
s
12-16
[
0=I ph−I 0 exp ( Va )−1]− VR
oc oc
sh
12-17
[
I mp=I ph−I o exp (V mp + I mp ∙ R s
a ) V +I R
]
−1 − mp mp s
R sh
12-18
d ( IV ) dI
dV |
mp
=I mp −V mp
dV | =0mp
12-19
Where
a R sh
Fifth known condition: derivative of the current at I sc
−I o I Rs
1
sc
a
e −
dI a Rsh 1
dV |I= I sc
=
1+
I o Rs
e
I sc R s
a
+
Rs
=
Rsh
a R sh
12.3. Example
A PV module with 36 cells in series and an area of 0.427 m 2 has the following measured
characteristics at reference conditions:
I sc 2.9 A
V oc 20 V
I mp 2.67 A
V mp 16.5 V
μ I ,sc 1.325x10-3 A∙K-1
μV , oc -0.0775 V∙ K-1