Test bench for Photovoltaic Modules

EFEEA’10 International Symposium on Environment Friendly Energies in Electrical Applications

2-4 November 2010, Ghardaïa, Algeria

Test bench for Photovoltaic Modules A. Mahrane#1 , A. Guenounou#1, Z. Smara#1, M. Chikh#1, M. Lakehal#1 #1

Renewable Energies Equipment division (EER) Solar Equipment Development Unit (UDES) Route Nationale n°11, BP 386, 42415, Bou Ismail, Tipaza, Algeria [email protected], [email protected], [email protected], [email protected], [email protected]

standard conditions (STC) known as an illumination of 1000W/m², a module temperature of 25 ° C and an Air Mass AM = 1.5 [2]. However, the usual conditions of use of PV modules in a natural environment are different. For this purpose we propose to realize a test platform that will allow us to characterize and monitor the performance of photovoltaic modules under natural conditions. As a first phase, we set up an I -V measurement bench for PV modules that we describe in this article. Thus, in paragraph II, we discuss the modeling of PV module. In section III we describe the electronic load we made and present its operating principle. In section IV we tackle the simulation of the I-V module. After having presented the I-V test bench in section V, we treat, in section VI, an example of characterization of a PV module and we analyze in section VI the results obtained by simulation and measurements.

Abstract— The photovoltaic modules performances are evaluated from their I-V characteristic. We present in this article the I-V measurement bench that we have set up at UDES in order to characterize the photovoltaic modules under natural conditions. We present, first, the method used for modeling PV module that allows us to simulate the I-V curve under a given conditions of illumination and temperature. We then describe the bench measurement focusing on the power MOSFETs electronic load that we realized. As an example, the plot of an I-V characteristic of a c-Si photovoltaic module got with our testing bench at a given operating conditions and translated to STC conditions are in good agreement with the one supplied by the manufacturer with a few correction made. KeywordsPhotovoltaic characteristics, electronic load

I.

modules,

current-voltage

INTRODUCTION

The combined effects of the gradual depletion of fossil resources and greenhouse gas emissions have generated awareness at the global level to adopt new behavior of energy consumption by focusing on the one hand, on the energy saving and, on the other hand, trying to have diversified energy mix. In this favorable context, solar photovoltaic’s grown steadily in recent years so that the world market has reached an annual size of the order of several gigawatts [1]. As a matter of fact, the photovoltaic module is the object of all issues so all efforts are made in order to achieve a viable price of peak watts of solar electricity as compared to the conventional energy which will result to increase the actual contribution of green energy in the global energy balance. The indicator of performance of a photovoltaic module is its I-V characteristic. The shape of the curve tells us about the quality and the data allow us to record or extract all the relevant parameters such as the short circuit current ISC, the open circuit voltage VOC, the current and the voltage respectively at the maximum power point Imp, Vmp, the power at the maximum power point Pmp, the field factor FF, the series resistance RS and the shunt resistance Rsh. In production, the I–V characteristic, provided by the manufacturers, is obtained using a solar simulator under

II.

MODELISATION OF PHOTOVOLTAIC MODULE

Rsh

Figure 1 One diode model for PV module

The study of the performance of a photovoltaic (PV) module is done through its I–V characteristic. The simulation of this feature for the analysis of the module’s behavior in terms of operating conditions requires the modeling of PV module. This last, which is an arrangement of a series of several solar cells, is based on the model of a cell, which in the case of a one diode model that we use, is represented by a diode, a current source, a series resistance and a shunt resistance. The current source generates a photocurrent which is a function of solar irradiance and temperature of the cell. The diode represents the PN

1



R1

junction solar cell. The I-V characteristic of a module has the following expression [3]:  V + IR s  q ( V + IR s ) . (1) ) − 1 − I = I ph − I 0  exp( R sh mNs kT C  

Control Stage

T

Power Stage

2

U 2m 1 R3

R6

M4

M1

1 IRF150

IRF150

100

0

Where I is the current output of the module, Iph is the photo current, Io the diode saturation current, NS is the number of the solar cells in series in the module, V is the voltage output of the module, q is the electric charge (1.6x10-19 C), k is the Boltzmann constant (1.38x10-23J/K), TC the cell temperature in degrees Kelvin and m is the ideality factor comprised between 1 and 2.

W

102

V

R R7

C 2200u CMAX

R2

R10

PARAMETERS: v ar = 25

V7 Iph

Req 0

Figure 3: Representation of the entire PV Module – Electronic load in the Spice environment.

The starting point of the modeling of a given type module is the I-V characteristic provided by the manufacturer. If the electrical parameters such as Voc, Pmp, Imp, Vmp and the FF can be deduced from the I-V characteristic, the parameters needed for modeling such as (Iph, I0, m, Rs, Rsh) require special extraction methods. In our case, we used an iterative method which consists in solving numerically a system of five nonlinear equations obtained by applying equation 1 to within five points of a typical experimental I-V. These five points should be as defined in (2). An interactive program that we have developed allows us to perform this task [4]:

In our case, the goal was to have an electronic load which allows us to get the I-V characteristics of the most type of modules available on the market. We then decided to realize an electronic load based on MOSFET which has the advantage of being simple, inexpensive, fast plot and by its design allows us to increase its power just by adding as many power units as necessary. The electronic load that we build (3) consists of two stages: A power stage connected directly to the photovoltaic module, which acts as a variable impedance modifying the operating point of the module and so collect the two quantities I and V. This part is composed of four MOSTEFs (IRF 150) in parallel to increase the absorption capacity of our load. To increase the power of the load we simply add more power stages or use more robust MOSFETs. The power resistors R7-R10 have a role to reduce the power dissipated by the MOSFETs. A control stage which controls the power stage, in other words its impedance, is composed of capacitor C, a pushbutton T, a potentiometer R and the resistors R1 to R6. The operating principle of this stage is based on the charging and discharging of the capacitor. Indeed, as soon as we activate the push button T, the capacitor C will charge through resistor R1. When we release it, it will be discharging through the resistance Req (potentiometer R and the resistance R2) resulting in the variation of the gate voltage VG of MOSFET transistors and thus change their drain current. We have used for the discharge circuit a potentiometer whose value is adjusted depending to the acquisition system utilized. The lower is its value, the faster is the I-V plot. The resistors R3 to R6 are necessary for the protection of MOSFET transistors.

Figure 2: The five points of the I-V curve which have been used in the extraction method

III.

3

103 2

ELECTRONIC LOAD

The plot of the I-V module characteristic involves the use of a load. This can be done with a simple rheostat. By varying its value from zero to its maximum value we can get different points of the I-V curve. This method is only applicable to low-power modules because high power resistors are not available. [5]. We can also use as a load a capacitor, but getting a reliable I-V curve by this method requires a high quality capacitor (low equivalent series resistor) with low loss [6]. Transistors (typically IGBTs or MOSFETs) can also be used as electronic load for testing PV modules. The potential advantage of this method is the rapid variation of the equivalent load resistance [7].

IV.

SIMULATION OF THE I-V CHARACTERISTIC OF A PV MODULE

In order to analyze the behavior of a photovoltaic module according to the irradiance and temperature, we have implemented the PV module model described in paragraph I

2


and the circuit of electronic charge in the PSPICE environment [8]. We have treated as an example the case of a Philadelphia Solar module type MS36 (SN 120-0-21-030110). The data for Standard Tests Conditions (STC) given by the manufacturer are as follow: Isc = 8.37 A, Voc = 22.86 V, Pmp = 140.7 W, Imp = 7.8 A, Vmp = 18.04 V, FF = 0.736, η = 14.1%, αIsc = 10 µA/cm²/°C, βVoc = -2.1 mV/Cell/°C. Where, η is the yield of the module, αIsc the temperature coefficient of the short circuit current, β Voc the temperature coefficient of the open circuit voltage. The parameters extracted from the experimental curve are: Iph = 8.386 A, Io = 6.124.10-10 A, m = 1.06, Rs = 0.22 Ω, Rsh = 112.30 Ω. These are then injected in the model of the module developed under SPICE.

Oscilloscope

Computer PV Module

Electronic load Figure 5: ‘Outdoor’ I-V test bench for PV modules.

VI.

The plot of the I-V characteristic of a PV module under natural conditions of illumination and temperature is done as follow. After the acquisition of I and V raw data, the treatment of these data is necessary. It is done to eliminate noise, to make adjustments, interpolations and extrapolations of the curve. Then, the electrical parameters are extracted. All these operations are done through the program developed by A. Guenounou [4]. As an example, we present the measurement of the I-V characteristic of a monocrystalline silicon module (MS36 SN-120-0-21-030110). The measurement conditions are: G= 900.9 W/m2, T= 46.65°C. The curve obtained after the treatment of the rough data is represented on (6).

AT STC conditions

IV of manufacturer IV calculated

Current [A]

8 6 Iph=8.386 A Io= 6.124E-10 A m=1.06 Rs=0.22 Ohm Rsh=112.3Ohm

4 2 0

0

5

15 10 Voltage [V]

20

EXAMPLE OF MEASUREMENT OF CHARACTERISTIC OF PHOTOVOLTAIC MODULE

Figure 4 shows simulated and measured curves obtained for the module (MS36 SN-120-0-21-030110) considered. We note that there is a good agreement between the two curves which allows us to validate the simulation tool developed under SPICE of our I-V test bench for the module type considered. 10


25

Figure 4: Simulated and measured I-V Characteristic obtained for the module (MS36 SN–120-0-21-030110) under standard conditions

V.

BENCH MEASUREMENT OF THE I-V CHARACTERISTIC OF A PHOTOVOLTAIC MODULE

Figure 6: Representation of I-V curve obtained after the treatment of the rough data for the module (MS36 –120-0-21-030110).

The bench 'outdoor' of the I-V characteristic of a PV module (Figure 5) that we have set put at UDES is composed of the electronic charge whose terminals are connected to the PV module, a digital oscilloscope TECKTRONIX HRT-710 for the acquisition and display of the IV characteristic and a computer to which data are transferred via RS 232 and where they are processed. The solar radiation received by the PV module under test is measured using a pyranometer and the module temperature is measured using a thermocouple fixed to the rear of the module PV.

To compare the measured I-V characteristic of the module (MS36 SN-120-0-21-030110) with those provided by the manufacturer, the curve measured under natural conditions should be translated to standard conditions. To do this we use the following equations of translation [9], [10]: Iph

3

STC

=

G STC G meas

(

+ α Isc TCSTC − TC . Iph meas  meas

). (2)


3    TCSTC   exp q.Eg  1 − 1  Io = Io STC meas  T   m.K  TCmeas TCSTC  Cmeas   

R sh STC G = meas . G STC R sh


factor for each type of PV module. In our case, we tried to define a value of this coefficient around the temperature coefficient of the open circuit voltage. Moreover, in the region where the module behaves as a current generator, the discrepancy is probably due to measurement error of the illumination.

 . (3)  

(4)

meas

We recorded in Table 1, in standard conditions, the values of the electrical parameters provided by the manufacturer and those translated from the measurements and we assessed the differences between the values obtained.

Where G and TC are the illumination and temperature. The indices ‘STC’ and ‘meas’ are respectively for Standard Test Conditions and measurements conditions. Eg is the energy gap. Two other parameters must be translated: the series resistance RS and the ideality factor m. The series resistance affects the slope of the I-V characteristic in the region where the module functions as a generator of tension, but this variation with temperature is not significant to induce an error on the maximum power point so we assume that it remains constant. We also think that the ideality factor m remains constant [9]. In Figure 7 we have shown the I-V characteristic of the module (MS36 SN-120-0-21-030110) at standard conditions translated from the curve measured under natural conditions and that provided by the manufacturer under STC.

10

Current [A]

8

6

4

Manufacturer's I-V caracteristic I-V caracteristic calculated at STC from G=900.9 W/m²,Tc=46.65°C, with correction in ideality factor (alpham = -0.0032).

2

0

0

5

15 10 Voltage [V]

20

25

Figure 8: Representation of the I-V curves for the module (MS36 SN–120-021-030110) at STC conditions. The first one given by the manufacturer and the other deduced from that measured at the conditions G=900.9W/m2, Tc=46.65C with correction of the ideality factor m.

TABLE I.

VALUES, AT STANDARD CONDITIONS, OF THE MAIN

ELECTRICAL PARAMETERS FOR THE PV MODULE (MS36 SN-120-0-21-030110) PROVIDED BY THE MANUFACTURER AND THOSE CALCULATED AFTER TRANSLATION FROM THE MEASUREMENTS AT OPERATING CONDITIONS TAKING INTO ACCOUNT THE IDEALITY FACTOR CORRECTION.

Parameters Figure 7: Representation of the I-V curves for the module (MS36 SN–120-021-030110) at STC conditions. The first one is provided by the manufacturer and the other is translated from the curve obtained at the conditions G=900.9W/m2, Tc= 46.65C.

Isc (A) Voc (V) Pmp (W) Vmp (V) Imp (A) FF η (%)

We note that there is a gap between the two curves. In the region where the module acts as a voltage generator we think that the gap is due to the fact that we have neglected the variation of the ideality factor as a function of temperature. Indeed, this gap can be reduced (8) by introducing a correction factor [10] for the ideality factor whose expression is:

mSTC = m meas[1 + α m (TC − TCmeas )] .

Manufacturer Data (STC) 8.37 22.86 140.7 18.04 7.8 0.736 14.1

Calculated Data (STC)

Gap (%)

8.7 22.98 142.4 18.23 7.81 0.713 16.26

3.94 0.52 1.21 1.05 0.13 -3.13 0.15

VII. CONCLUSION This paper presents a semi automatic I-V test bench for PV module characterization set up at UDES. Currently, we can get the I-V plot of any type of PV modules at outdoor conditions and we can also translate them into any desired operating conditions. However, this testing bench can be improved to reach full automation with more precise measurements. Our goal is to approach fully automatic start up of I-V scan.

(5)

It should be noted, nonetheless, that this method requires further study to be validated. Because, currently, there is no accurate method to define a temperature coefficient of ideality

4


[4]

A. Guenounou, “ Mise au point de nouveaux procédés de mesure des caractéristiques directe et inverse des modules photovoltaïques de différentes technologies,” Mémoire de magister, Ecole Doctorale “Energies Renouvelables” (CDER / Université de Tlemcen), Janvier 2010. [5] Duran, E. Piliougine, M. Sidrach-de-Cardona, « Different methods to obtain the I-V curve of PV modules: A review», Photovoltaic Specialists Conference, PVSC '08. 33 rd IEEE, 2008. USA. [6] E. Caamano, E. Lorenzo, R. Zilles, « Quality Control of Wide Collections of PV Modules: Lessons Learned from the IES Experience », Progress in Photovoltaics: Research and Applications. 7, 137-149, 1999. [7] Yingying Kuai, S. Yuvarajan, « An electronic load for testing photovoltaic panels», Journal of Power Sources 154, pp 308–313, 2006. [8] http//www.cadencepcb.com [9] W. De Soto, S.A. Klein, W.A. Beckman, “Improvement and validation of a model for photovoltaic array performance,” Solar Energy 80 (2006) 78–88. [10] A. Mermoud, “Conception et Dimensionnement de Systèmes Photovoltaïques¨. PVSYST, Rapport final, Université de Genève, Mai 2005.

Furthermore, we intend to improve the electronic load that will allow us to acquire data close to ISC and VOC. Currently, our equipment gives us a very large number of data points that are not practical for analysis. Consequently we intend to devise a technique that will allow us to optimize the number of data points to obtain a representative trend of the I-V characteristic. ACKNOWLEDGMENT We would like to thank Mr S. Elmetnani (UDES) for his valuable advices on the writings of this article. REFERENCES [1] [2] [3]


http://www.solarbuzz.com CEI 60904-1:1987, Dispositifs photovoltaïques – Partie 1 : Mesures des caractéristiques courant-tension des dispositifs photovoltaïques. Engin Karatepe *, Mutlu Boztepe, Metin C¸ olak, «Development of a suitable model for characterizing photovoltaic arrays with shaded solar cells», Sol. Energy 2007, doi:10.1016/j.solener.2006.12.001.

5