See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/224605112

Simplified model of a photovoltaic module

Conference Paper · October 2009

Source: IEEE Xplore

CITATIONS READS

119 5,707

4 authors, including:

A. Bellini Cristina Cornaro

University of Rome Tor Vergata University of Rome Tor Vergata
32 PUBLICATIONS 499 CITATIONS 105 PUBLICATIONS 2,636 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Cristina Cornaro on 22 May 2014.

The user has requested enhancement of the downloaded file.

 Simplified Model of a Photovoltaic Module
Armando Bellini1, Stefano Bifaretti1, Vincenzo Iacovone1, Cristina Cornaro2
1
Dept. of Electronic Engineering, University of Rome Tor Vergata
[email protected]
2
Dept. of Enterprise Engineering, University of Rome Tor Vergata
[email protected]

Abstract – The cost and the performance of PV and maximum power point conditions. Finally a
plants strongly depends on modules. However, the model with all parameters achievable from
electrical parameters of the modules can be manufactures datasheet is proposed in [5]. The
different from those provided by the parameters determination of such model requires the
manufacturer; moreover, such parameters can use of numerical methods.
change as the module is getting older. Therefore, The paper proposes an improved model of a PV
the behavior of the mathematical model of a PV module that make use only parameters provided by
module can’t match the real operating conditions. manufacturers datasheets and, moreover, doesn’t
The paper proposes an improved model of a PV require any numerical methods. In the paper after the
module that make use only parameters provided detailed description of the proposed model, different
by manufacturers datasheets without requiring the simulation results are pointed out. Finally an
use of any numerical methods. In the paper after experimental validation of the model is presented
the detailed description of the proposed model, together with a suitable procedure that takes into
different simulation results are pointed out. Finally account the real operating parameters in the PV
a suitable procedure that permits to account the model.
real operating parameters in the PV model and an
experimental validation is presented.

I. INTRODUCTION

Renewable energy generation systems, based on

photovoltaic (PV) modules, nowadays represent the
most suitable solution, in particular for domestic
power levels, to reduce CO2 emissions and the energy
consumption produced by oil and gas. Moreover in
different European countries, electricity companies
are providing money incentives for the energy
produced by renewable sources and injected into the Fig. 1 – Single-diode equivalent circuit of a PV module.
utility grid. The cost and the performance of PV
plants strongly depends on modules. However, the II. MODEL OF A PHOTOVOLTAIC MODULE
electrical parameters of the modules, i.e. open circuit
voltage and short circuit current, can be different than The current of a PV module can be expressed, as
those provided by the manufacturer; moreover, such function of voltage, by the simplified expression
parameters can change as the module is getting older. derived from [4]:
Therefore, the behavior of the mathematical model of
a PV module can’t match the real operating
conditions. · 1 · ·
1 (1)
Different model of solar cells were presented in
literature [1-5]. The most accurate model, denoted as where
double-diode model [1], uses an equivalent circuit 1 · ·
(2)
with two diodes but it is quite complex due to the
presence of a double exponential and six parameters
to assign. A different model, based on a single diode
circuit, was then proposed in [2, 3]. In both cases, the (3)
mathematical models require the knowledge
respectively of six and five parameters that are not
Coefficients C1 and C2 depends on the following
directly available on manufactures datasheets. A
module parameters:
simplified one-diode model, shown in Fig.1, using
• short circuit current ISC;
only four parameters was proposed in [4]. However,
in such model the voltage is independent from solar • open circuit voltage VOC;
irradiance. As consequence, a significant voltage error • maximum power point voltage VMPP;
in the I-V curves is present specially at open circuit • maximum power point current IMPP.
Such parameters can be expressed as follow: · · (11)

, 1 α (4)
III. SIMULATION RESULTS
β (5)
The behavior of the proposed PV model has been
, 1 α (6) simulated, in Matlab-Simulink environment,
considering the parameters, at STC, of Photowatt
β (7) PW1650 – 24V panel listed below:

where parameters ISCS, VOCS, Impps and Vmpps are • ISC = 5.3 A;
defined at standard conditions, STC (Gs=1000W/m2 • VOC = 44.6 V;
and Ts=25°C) and α and β are respectively the current • VMPP = 35.4 V;
and the voltage temperature coefficient; all the above • IMPP = 4.95 A;
parameters are provided by manufacturers on module • α = 1.46 mA/°C;
datasheet. It is possible to note that the parameters • β = −158 mV/°C.
referred to currents depends on module solar
irradiance G and temperature T, while the voltage In order to verify the model under different operating
ones depends only on temperature. conditions, several tests have been performed at
To improve the accuracy of the model it is various temperature and solar irradiance values. Fig. 2
convenient to modify expressions (5) and (7) inserting shows I-V curves for various solar irradiance (G
a correction term, ΔV(G), taking into account voltage varies from 100W/m2 to 1000 W/m2), while Fig. 3
variation as function of solar irradiance: illustrates I-V curves for different temperature values
(T varies from -10°C to 65°C).
, β ∆ (8)
6
, β ∆ (9)
5.5 G = 1000
5
Correction term ΔV(G) is obtained by the following 4.5 G = 800
relationship: 4
Current (A)

3.5 G = 600
∆ 3 G = 500
2.5
G = 400
where voltage VOCm represents the open circuit 2
voltage of the IV curve translated from STC to the 1.5
G = 200
considered irradiance G and it defined as: 1
0.5 G = 100
0
1 0 10 20 30 40 50
· · ln 1 . Voltage (V)
Fig. 2 – I-V curves obtained by simulation for various
irradiation levels.

6
It is the short circuit current at irradiance G and 5.5
can be written as: 5
4.5 T = 65°C T = -10°C
. 4
Current (A)

3.5
3
In order to determine the value of series RS, as 2.5
function of panel parameters, it is convenient to 2
express the module voltage as function of current by 1.5
inverting eq. (1): 1
0.5
0
0 10 20 30 40 50
· · ln 1 (10) Voltage (V)
Fig. 3 - I-V curves obtained by simulation for various temperature
values.
The value of series resistance RS can be calculated
by deriving eq. (10) with the current calculated for
IP=0:
 2
G=800 W/m T=45°C 10 minutes a complete IV curve is traced for each
5 module under test. In plane irradiance is measured by
Proposed
an EKO pyranometer with an uncertainty of 3% at
[3]
4 1000 W/m2 and by a polycrystalline reference cell
(ISET sensor) with an uncertainty of 4% in the whole
range. Module temperature is measured by a Pt100
Current (A)

3
attached on the back of the module. Temperature
uncertainty has been estimated to be +/- 0.16°C.
2
IV curves collected with the described system as
been compared with others measured using an
1 acquisition system built in our laboratory, based on a
switched capacitor [7] and shown in Fig. 6. Such
0 approach permits to obtain similar performance on I-
0 10 20 30 40 50
Voltage (V) V curve tracking than traditional ones but simple
implementation, lower costs and fast acquisition time.
Fig. 4 – Comparison of I-V curves obtained by simulation.
2
G=400 W/m T=30°C
3
[3]
2.5 Proposed

2
Current (A)

1.5

0.5

0
0 10 20 30 40 50
Voltage (V) Fig. 6 – Acquisition system used for experimental tests.
Fig. 5 - Comparison of I-V curves obtained by simulation. A first test has been performed on the considered
panel using both measurement systems to validate the
In order to evaluate the performance of the prototype built in our laboratory. Since the
proposed model a comparison with model presented experimental measures have been achieved at
in [4] was performed by different simulation tests. different temperature and solar irradiance, therefore it
Figs. 4 and 5 show the comparison of the IV has been necessary to translate both curves to STC.
curves obtained respectively for G = 800 W/m2 The obtained results have been compared as shown in
T=45°C and G = 400 W/m2 T=30°C. In the figures Fig. 7.
2
the continuous line represents the curve obtained with G=1000 W/m T=25°C
proposed model while, the dotted line, the curve 6 MPPT 3000
achieved with the model described in [3]. It is Capacitor method
5
possible to notice that, in both cases, the two curves
present a similar shape without any significant 4
Current (A)

discrepancy; therefore, in the successive comparisons

the proposed model will be used. 3

IV. EXPERIMENTAL RESULTS 2

Several tests have been carried out on a Photowatt 1

PW1650 – 24V polycrystalline photovoltaic module
in the outdoor ESTER facility located on the roof top 0
0 10 20 30 40 50
of the Engineering building of the University of Rome Voltage (V)
Tor Vergata [6]. The module has been exposed to the Fig. 7 – Comparison of I-V curves obtained by experimental test.
real environment since October 2007. It has been
mounted on a south oriented frame with variable tilt
angle. The module has been constantly monitored The translation has been performed according to
keeping it at maximum power point using an [8] that describes the procedures for temperature and
electronic device (MPPT3K) provided by ISAAC – irradiance corrections to the I-V characteristics of a
SUPSI, Lugano, Switzerland. Every minute module crystalline silicon PV module. Using this standard,
maximum power, module temperature, in plane each current-voltage pair (IM, VM) on the measured I-
irradiance, air temperature and wind speed and V curve is transformed into a corresponding pair at
velocity are collected and stored in a database. Every STC (ISTC, VSTC) by the following equations:
 the experimental curve is due to the different
· parameters values of the tested panel respect to the
theoretical ones furnished on the datasheet.
· In order to consider those variations, authors
propose to modify the PV model introducing two
where: additional resistors, RSX and RPX, as shown in Fig. 10.

· ln · .

The coefficients a and b are calculated by a least

square procedure applied on the expression shown
below:

, , · ln · .
Fig. 10 – Modified model of PV module.

It is possible to note, from Fig. 7, that the I-V

curves don’t present significant discrepancy; Resistors RSX and RPX can be identified using the
therefore, the acquisition system of Fig. 6 is used to following procedure.
perform the following tests.
The value of parallel resistance RPX can be
G=993 W/m
2
T=44.8°C determined by:
6 Experimental
Model ∆
5 ∆

4
where ΔVE and ΔIE are taken from the experimental
Current (A)

3 data nearby the short circuit current.

The value of series resistance RSX can be
2 determined as:

1 ∆
∆
0
0 10 20 30 40 50
Voltage (V) where ΔVE and ΔIE are taken from the
Fig. 8 – Comparison of I-V curves obtained by simulation and experimental data nearby the open circuit voltage
experimental test. and RS is calculated using (11).
Finally, the output current I of the new model can
G=1000 W/m2 T=25°C
6
be expressed as:
Experimental
Model
5

4
Current (A)

and the output voltage V of the new model is:

3
·
2 G=993 W/m
2
T=44.8°C
6 Modified model
1
Experimental
5
0
0 10 20 30 40 50
Voltage (V) 4
Current (A)

Fig. 9 – Comparison of I-V curves obtained by simulation and 3

experimental test at STC.
2
Fig. 8 shows the comparison among the simulation
result obtained with the proposed model (continuous 1
line) and the experimental curve (dotted line)
0
measured at G = 993 W/m2 and T=44.8°C. Fig. 9 0 10 20 30 40 50
illustrates the IV curves, referred to the same test, Voltage (V)
translated to STC. Fig. 11 – Comparison of I-V curves obtained by simulation and
The difference between the simulation model and experimental test.
 G=1000 W/m2 T=25°C
6 However, the electrical parameters of the modules,
i.e. open circuit voltage and short circuit current, can
5
be different than those provided by the manufacturer;
4 moreover, such parameters can change as the module
Current (A)

is getting older. Therefore, the behavior of the

3 Experimental mathematical model of a PV module can’t match the
Modified model real operating conditions. In such framework, authors
2 Model
propose a suitable modification to the model to take
into consideration the real operating parameters in the
1 PV model.
Several simulation and an experimental results
0
0 10 20 30 40 50 have been pointed out in order to validate the
Voltage (V) mathematical model. The results show that the
Fig. 12 – Comparison of I-V curves obtained by simulation and proposed model demonstrates a good agreement with
experimental test at STC. the experimental data measured by an acquisition
system specific for I-V characteristics measurements.
G=1000 W/m2 T=25°C
ACKNOWLEDGMENT
Model
Modified model
150 Experimental This work was supported by Polo Fotovoltaico
Organico della Regione Lazio (CHOSE - Center For
Hybrid and Organic Solar Energy).
Power (W)

100
REFERENCES

50 [1] Gow, J.A.; Manning, C.D.: Development of a

photovoltaic array model for use in power-
electronics simulation studies, IEE Proc. Electr.
0 Power Appl., Vol. 146, No.2, March 1999,
0 10 20 30 40 50
Voltage (V) pp.193-200.
[2] Liu, S.; Dougal, R.A.: Dynamic Multiphysics
Fig. 13 – Comparison of I-V curves obtained by simulation and
experimental test at STC. Model for Solar Array, IEEE Trans. on Energy
Conversion, Vol. 17, No. 2, June 2002, pp. 285-
294.
To verify the proposed procedure a comparison [3] Walker, G.: Evaluating MPPT Converter
among the experimental curve, obtained in the same Topologies using a Matlab PV Model, Journal of
operating condition of Fig. 8, the curves achieved by electrical and electronics engineering, Australia,
simulation with the model described in section II and Vol. 21, pp. 49–55, 2001.
the modified model has been performed. The results [4] Xiao, W.; Dunford, W.G.; Capel, A.: A novel
of the comparison of the IV curves are shown in Figs. modeling method for photovoltaic cells, IEEE
11 and 12 respectively for real operating conditions Power Electronics Specialists Conference,
and STC, while in Fig. 13 is represented the Aachen, Germany, 2004.
comparison among the power waveforms at STC. [5] Sera, S.; Teodorescu, R.; Rodriguez, P.: PV panel
As it possible to notice from the figures above, the model based on datasheet values, IEEE
experimental curve (dotted line) and the modified International Symposium on Industrial
model curve (continuous line) are very similar, while Electronics, ISIE 2007, 4-7 June 2007, pp. 2392 –
the curve of the initial model (dashed line) presents 2396.
significant differences from the others. Fig. 13 [6] Spena, A.; Cornaro, C.; Serafini, S.: Outdoor
highlights that the maximum peak power of the ESTER Test Facility for Advanced Technologies
module provided by the manufacturer datasheet is not PV Modules, Proc. of the 33rd IEEE PV specialists
correspondent to the experimental one. Conference, San Diego (CA), May, 11-16 2008.
[7] Mahmoud, M.M.: Transient analysis of a PV
V. CONCLUSION power generator charging a capacitor for
measurement of the I-V characteristics,
The paper presents an improved mathematical Renewable Energy, Vol. 31, Elsevier, 2006, pp.
model for photovoltaic modules that employs only 2198-2206.
parameters provided by manufacturers datasheets [8] Blaesser G. : On site power measurement on large
without requiring the use of any numerical methods. PV array, 10th EUPVSEC, Lisbon, 1991.
The model is derived from [4] by applying several
improvements, such as an opportune correction term
that permit to account the voltage variation as
function of the solar irradiance.

View publication stats