Turkish Journal of Electrical Engineering and Computer Sciences

Volume 27 Number 4 Article 46

1-1-2019

A new approach for parameter estimation of the single-diode

model forphotovoltaic cells/modules
BİLGE KAAN ATAY

ULAŞ EMİNOĞLU

Follow this and additional works at: https://journals.tubitak.gov.tr/elektrik

Part of the Computer Engineering Commons, Computer Sciences Commons, and the Electrical and
Computer Engineering Commons

Recommended Citation
ATAY, BİLGE KAAN and EMİNOĞLU, ULAŞ (2019) "A new approach for parameter estimation of the single-
diode model forphotovoltaic cells/modules," Turkish Journal of Electrical Engineering and Computer
Sciences: Vol. 27: No. 4, Article 46. https://doi.org/10.3906/elk-1805-161
Available at: https://journals.tubitak.gov.tr/elektrik/vol27/iss4/46

This Article is brought to you for free and open access by TÜBİTAK Academic Journals. It has been accepted for
inclusion in Turkish Journal of Electrical Engineering and Computer Sciences by an authorized editor of TÜBİTAK
Academic Journals. For more information, please contact [email protected].
 Turkish Journal of Electrical Engineering & Computer Sciences
Turk J Elec Eng & Comp Sci
http://journals.tubitak.gov.tr/elektrik/ (2019) 27: 3026 – 3039
© TÜBİTAK
Research Article doi:10.3906/elk-1805-161

A new approach for parameter estimation of the single-diode model for

photovoltaic cells/modules

Bilge Kaan ATAY1 ,, Ulaş EMİNOĞLU2,∗,

1
Graduate School of Natural and Applied Sciences, Gaziosmanpaşa University, Tokat, Turkey
2
Department of Electrical and Electronics Engineering, Gaziosmanpaşa University, Tokat, Turkey

Received: 23.05.2018 • Accepted/Published Online: 15.04.2019 • Final Version: 26.07.2019

Abstract: Solar energy has become a popular renewable energy source, leading to wide use of photovoltaic (PV)
cells/modules in energy production. For this reason, realistic modeling of PVs and determining the equivalent circuit
parameters is of great importance in terms of planning and operation. Hence, in this study, an analytical model for
identifying the single-diode equivalent circuit parameters; series resistance ( Rs ), shunt resistance ( Rp ), diode ideality
factor ( a ), diode reverse-saturation current ( Io ), and photon current ( Ipv ) for PV cells/modules is developed without
neglecting any term. In order to test the accuracy of the model, a number of PV modules from different manufacturers
are taken into account and the results are compared with those obtained by using such analytical models given in
the literature. Current-voltage (I-V ) characteristics of the PV modules, which are studied here, are also simulated
by comparing with the experimental I-V curves provided by the manufacturers. Results show that the values of the
parameters obtained for the PV modules are consistent with those extracted by using other analytical models. In
addition, I-V curves created by using the obtained parameters are in full agreement with the experimental data. The
curves also show a high degree of compatibility with the ones created by using the optimal parameters of the two-diode
models given in the literature. Moreover, the proposed model provides a great advantage in estimating equivalent circuit
parameters in terms of ease of use, requirements for input data, dependency on initial conditions as well as considering
the parameters which are neglected in such methods given in the literature.

Key words: Photovoltaic cells/modules, mathematical modeling, parameter extraction, single-diode model

1. Introduction
The ever-decreasing production costs of solar panels have made PV modules a preferred energy source to be used
in an off-grid or on-grid system. In this context, accurate modeling of PV modules plays quite an important
role in the effective assessment and quality control. In addition, the characterization of the output current and
power (I-V and P-V ) curves of the system, based on the equivalent circuit parameters, is of great importance
in terms of design and sizing [1]. Therefore, numerous studies have been conducted in the literature, especially
on determining the equivalent circuit parameters for PVs. In the literature, we see that particularly two types
of models, i.e. the single-diode five-parameter model and the two-diode seven parameters model, are used.
For instance, in a study by Chan and Phang [2], a PV cell has been presented with the single-diode and the
two-diode models using analytical techniques based on experimental data. For both models, there are analytical
and numerical solutions, in which various methods can be applied. In another study, Shongwe and Hanif have
∗ Correspondence: [email protected]

3026
This work is licensed under a Creative Commons Attribution 4.0 International License.
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

compared different methods developed for the extraction of the parameters of PV modules using the single-diode
five-parameter circuit model on two commercial PV modules [3]. The single-diode model has been improved
by adding a shunt resistor and a resistor in series. In a similar study [4], the authors have created a model
with a shunt and serial resistors at the standard test conditions (STC) by taking into account the temperature-
dependent saturation current; however, they have neglected some terms in the set of equations. Due to the
nonlinearity of the equations and the number of unknown variables, it is necessary to use iterative solutions for
these types of systems. Accordingly, Reis et al. formulated the parameters for the single-diode five-parameter
model [5]. The set of equations was solved using the Newton–Raphson method, and it was demonstrated that
the extraction of the five unknown parameters can easily be determined by this method. A new set of equations
for the determination of the equivalent circuit parameters based on Kirchhoff’s laws at STC was proposed
in [6], and the validity of the model was evaluated on some commercial PV modules using the Gauss–Seidel
method. In this study, the initial values of the unknown parameters were also determined by using approximate
analytical models since the performance of the Gauss method is significantly affected by the initial conditions.
For the single-diode model, however, a set of equations were created and solved for different commercial PV
cells/modules [7]; the shunt resistor, which is one of the main parameters of the equivalent circuit, was selected
as an input parameter. In other words, it is claimed that the value of the shunt resistor can be determined by
experimental measurements and hence the parameter is assumed to be known. Similarly, an analytical model
based on the equivalent circuit was developed by Javier et al. in [8] and tested on two well-known commercial
PV modules comparing the results with those available in the literature. Despite the reasonable results obtained
using this model, one main parameter in the equation set, namely the diode ideality factor was defined as an
input parameter and its range was kept between 1 and 1.5. As is well known, the Newton–Raphson method
has good accuracy, but it becomes disadvantageous as the number of iterations increase. Accordingly, Patel,
who created a new method by synthesizing an analytical and numerical solution, used Lambert W-function
to reduce the execution time [9]. In addition to these models, it is also possible to determine the parameters
using temperature-dependent models; however, manufacturer’s data, I-V and P-V characteristics, and solar
temperature become very important in this type of approach [10].
The main difference between the two main models in the literature is the accuracy and the number of
unknown parameters. A comparative study on both models was conducted by Shannan et al. [11]. In their
study, the differences between the models were analyzed in terms of accuracy and execution time, and it has
been concluded that the two-diode model has better accuracy, but it also has a longer execution time. Similarly,
in another study [12], the authors solved both models using a numerical solution on a PV module simulated
in a hybrid system. In addition, the authors stated that the single-diode model has fewer parameters so that
the unknowns can be calculated in a shorter time, and that the two-diode model is more consistent in terms
of accuracy. Since the determination of initial values in numerical studies is quite difficult, the approximate
analytical solutions have been proposed by Hejri et al. to determine the values needed for the numerical solution
of the two-diode model, and this method was verified with experimental data of different commercial PV models
[13]. The authors have also analyzed the single-diode five-parameter model [1]. In their study, some terms have
been neglected and the saturation current of the diode has been eliminated to decrease the number of unknown
parameters in the set of equations.
Metaheuristic methods are widely used to obtain the optimum solution for multiparametric complex
systems. Consequently, there are studies in the literature [14–17] conducted to obtain the parameters of PV
module using such optimization techniques namely genetic algorithm (GA) [18], particle swarm optimization

3027
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

(PSO) algorithm [19], and differential evaluation (DE) method [20]. In order to calculate the unknown
parameters on a single-diode model, Zagroube et al. [14] used GA, and the method was quite effective for
such a problem. In another study [15], the authors used the PSO algorithm for the first time to determine
the value of the parameters for the double-diode models by using the data for practical PV modules provided
by the manufacturers. A similar approach has also been used in a study for parameter extraction of the
single-diode model [16]. Another technique, called differential evaluation algorithm, was used by Ishaque and
Salam for identifying the parameters of the single-diode model [17]. The proposed model has been evaluated
on the experimental data and the results obtained have been found to be more reliable as compared with
the single diode Rs -model. In addition, parameter estimation by using different approaches such as multiple
optimization-based method [21], search algorithm based on the reduced form of the equations set [22], hybrid
form of the Nelder–Mead and modified PSO algorithm [23], partly search-based method using diode ideality
factor ( a) for the single-diode model [24] and series resistance ( Rs ) for the double-diode model [25] as an
independent variable, tool/software solution-based method [26] are also available in the literature. In [21], the
authors proposed an algorithm based on the hybrid GA-PSO to extract the parameters of the single-diode
model for PV modules, and hybrid approach was shown to be more advantageous than the method based on
the GA. Bencherif and Benouaz reduced the number of parameters to be calculated only to two and reached
a solution using the graphical method based on the search algorithm [22]. A hybrid method based on the
combination of the Nelder–Mead and modified PSO algorithm was developed for extracting the double-diode
equivalent circuit parameters for PV modules [23]. The proposed model was validated on a commercial solar
cell comparing the results with those obtained by using such metaheuristic approaches. In order to extract
the parameters of the single-diode equivalent circuit model, the authors in [24] developed a partly search-based
algorithm in which the diode ideality factor ( a) was defined as an independent variable. In the method, the
values of the remaining parameters are obtained analytically at each step by using the numerical solution of
the equation reduced only to the series resistance ( Rs ). The algorithm is terminated based on the error value
defined between the calculated output current and empirical data of I-V curves. A similar approach was also
implemented for a double-diode model of PV modules in [25]. The series resistance ( Rs ) was used as an
independent parameter and the remaining ones were determined based on the selected Rs in each iteration.
In this method, the procedure is reiterated until the selected value of Rs coincides closely with its calculated
value. In [26], the authors created five basic nonlinear equations from the equivalent circuit and solved by
using a solution technique of MATLAB for one practical PV module. The reliability of the model was validated
on an experimental system and the results were also compared with those obtained by two different methods
available in the literature. Detailed information about the methods developed for extracting equivalent circuit
parameters of PV cells/modules based on the analytical and/or numerical solution techniques are given in [27].
The single-diode five-parameter model is widely studied in the literature since it is an adequate method
for modeling the equivalent circuit of PV cells/modules. Accordingly, in this study, a mathematical model is
proposed for estimating the parameters of the single-diode model for PV cells/modules. The model is validated
on several commercial PV cells/modules at different temperature and solar radiation values including standard
test conditions. A number of analytical methods available in the literature are also taken into account for
comparison purposes. Results show that our model, which has been developed without neglecting any terms,
extracts the equivalent circuit parameters of PV modules in a more accurate manner. In addition, it is not
sensitive to the initial conditions as compared to such methods available in the literature. The rest of this study

3028
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

is structured as follows: The formulation of the proposed model is presented in Section 2. In Section 3, the set
of equations developed are validated on some practical PV cells/modules by comparing with those obtained by
using such models and the results of the studies given in the literature. Finally, the advantages of the proposed
model are emphasized in Section 4.

2. Mathematical formulation of the single-diode model of PV cells/modules

Since they have no moving parts, PV cells/modules are one of the most convenient systems in terms of
maintenance in energy production. In addition, their life cycle is longer than that of other renewable energy
conversation systems [6, 16]. In their electrical modeling, the single-diode five-parameter model shown in
Figure 1 is commonly used [1, 2].

Figure 1. Single-diode five-parameter model.

As shown in Figure 1, the photon current generated by a PV cell is modeled with Ipv with a parallel
diode. Rp and Rs denote shunt and serial resistor, respectively. Vt refers to the thermal voltage on the diode
(Vt = kNs T /q) , where Ns denotes the number of PV cells in series, k is the Boltzmann’s constant, which
is 1.38x10−23 J/K , and q is the electron charge (1.6x10−19 C). The last term, T , is the cell temperature in
Kelvin. Applying Kirchhoff’s current law to the system, we can get the output current as:

I = Ipv − Id − Ip , (1)

V +IRs V + IRs
I = Ipv − Io (e aVt − 1) − . (2)
Rp

At (Isc , 0),
Isc Rs Isc Rs
Isc = Ipv − Io (e aVt − 1) − . (3)
Rp

At (0, Voc ),
Voc Voc
0 = Ipv − Io (e aVt − 1) − . (4)
Rp

At (Impp , Vmpp ) ,
Vmpp +Impp Rs Vmpp + Impp Rs
Impp = Ipv − Io (e aVt − 1) − . (5)
Rp

3029
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

The generated power by the PV module is given by:

P = IV. (6)

where V and I refer to the terminal voltage and current, respectively. If the power is differentiated with respect
to the output voltage at its maximum point ( Impp , Vmpp ), Eq. (9) is obtained as follows:

dP dI
=I+ V, (7)
dV dV

dI
0 = Impp + Vmpp , (8)
dV

dI Impp
=− . (9)
dV Vmpp

and then Eq. (2) can be differentiated with respect to output voltage:

dI 1 Rs dI V +IR s 1 Rs dI
= −Io [( + )e aVt ] − − , (10)
dV aVt aVt dV Rp Rp dV

dI Io dI V +IRs 1 dI
=− (1 + Rs )e aV t − (1 + Rs ). (11)
dV aVt dV Rp dV

Substituting Eq. (9) into Eq. (11), we get the ratio of output current to the voltage at maximum power point
as given in the following equations.

Impp Io Impp Vmpp +Impp Rs 1 Impp

− =− (1 − Rs )e aV t − (1 − Rs ), (12)
Vmpp aVt Vmpp Rp Vmpp

Impp Io Impp Vmpp +Impp Rs 1 Impp

= (1 − Rs )e aV t + (1 − Rs ). (13)
Vmpp aVt Vmpp Rp Vmpp

Rearranging Eq. (3) by taking into account Eq. (4), we get the short circuit current equation as follows:

Voc Isc Rs Voc − Isc Rs

Isc = Io (e aVt − e aVt )+ . (14)
Rp

Similarly, by substituting Eq. (4) into Eq. (5), we get the following equation:

Rs Voc Vmpp +Impp Rs Voc − Vmpp

Impp (1 + ) = Io (e aVt − e aVt )+ . (15)
Rp Rp

It is seen in Eqs. (13)–(15) that we have four unknown parameters ( Rs , Rp , a, and Io ) but three equations.
For this reason, a fourth equation is needed to determine these parameters. On the other hand, at short circuit
situation (Isc , 0), we can write the following equality [1]:

dI 1
=− . (16)
dV Rp0

3030
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

By substituting the last equation into Eq. (11), for the short circuit case, we get the equation for Rp0 as given
in the following equations:

1 Io Rs Isc Rs 1 Rs
− =− (1 − )e aVt − (1 − ), (17)
Rp0 aVt Rp0 Rp Rp0

Io ( Isc Rs Rs
aVt Rs e aVt ) + Rp + 1
Rp0 = 1 Io ( Isc Rs
. (18)
Rp + aVt e aVt )

It can be inferred from [1] and [2] that Rp0 = Rp and therefore Eq. (18) can be rewritten as follows:

Io ( Isc Rs Rs
aVt Rs e aVt ) + Rp + 1
Rp = 1 Io ( Isc Rs
. (19)
Rp + aVt e aVt )

Now, the set of equations given in Eqs. (13)–(15) and (19) can be solved by applying a numerical solution
method, such as the Newton–Raphson method, to analytically calculate the values of unknown parameters for
the PV cells/modules. However, as stated in [1], convergence problem may occur when such numerical solution
methods are used. This is due to the fact that the saturation current of the diode ( Io ) has a very small value,
and as a result, the Jacobian Matrix converges to singularity. In order to overcome this problem, a new set of
equations can be obtained by eliminating Io as shown in the following procedures.
Eq. (14) can be rewritten as follows:
Voc −Isc Rs
Isc − Rp
Io = Voc Isc Rs . (20)
e aVt − e aVt

Then, substituting the last equation into Eqs. (13), (15), and (19), we obtain a new set of equations, only
dependent on the shunt, series resistors, and the ideality factor of the diode ( Rp , Rs and a) as given in the
following equations:

Voc −Isc Rs
Isc − Rp Impp Vmpp +I mpp Rs 1 Impp Impp
0= Voc Isc Rs (1 − Rs )e aVt + (1 − Rs )− , (21)
aVt (e aVt −e aVt ) Vmpp Rp Vmpp Vmpp

Voc −Isc Rs
Isc − Rp Voc Vmpp +Impp Rs Voc − Vmpp Rs
0= Voc Isc Rs (e aVt − e aVt )+ − Impp (1 + ), (22)
e aVt −e aVt Rp Rp

Isc − Voc −I
Rp
sc Rs Isc Rs
Rs
Voc Isc Rs Rs e aVt + Rp +1
aVt (e aVt −e aVt )
0= − Rp . (23)
Isc − Voc −I sc Rs Isc Rs
1 Rp
Rp + Voc Isc Rs e aVt

aVt (e aVt −e aVt )

3. Test cases
The last three independent equations allow us to extract the value of three unknown parameters; Rp , Rs , and
a for PVs by taking into account the technical data given by the manufacturers or obtained experimentally.
Technical data of the PV modules to be used for the validation of the proposed model are presented in Table 1. In

3031
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

order to solve this set of equations, the Newton–Raphson method widely used for solving this type of nonlinear
equations [2, 5, 13, 22] is utilized in MATLAB environment. The numerical solutions are found for varying
initial conditions and the results obtained for the BP-MSX120 PV module are listed in Table 2. The results
obtained by using the model developed in [1] and [4, 6–8, 26] are also presented in the table for comparison
purposes. In this context, first of all, it should be noted that the initial conditions do not have an effect on the
parameter values as one expects. In addition, the relative error used for terminating the numerical solution is
defined as follows:
Error = |xi − xi+1 | < 10−6 (24)

where xi shows the value of the parameter in the previous iteration. The remaining two parameters ( Io and Ipv )
are then computed using Eqs. (20) and (4), respectively. As shown in the table, the values of the parameters
for PV module are in close agreement with the solutions given in [1] and those determined by the models
proposed in [4, 6–8, 26]. It is clear that the proposed model allows getting more realistic results without any
prior knowledge except technical data. A similar situation is also observed for the model given in [6]. Even
though the models proposed in references [4, 7, 8, 26] give consistent results, these models have some handicaps.
For example, in the model proposed in [7], the parallel resistor is considered to be known and taken as an input
parameter. Similarly, in the model developed in [8], the diode ideality factor is taken as an input parameter.
For both models, the values of these parameters obtained by using the proposed method, which is marked in the
table, were used to find the solution. In addition, the performance of the models given in [4] and especially given
in [26] is highly dependent on the initial conditions. In other words, realistic results can only be obtained as
the initial conditions converge to the actual values. The model in [26] also requires a powerful tool/software for
the iterative solution of its nonlinear equations. In this study, Isqnonlin command based solution technique in
MATLAB program is used and the objective function is chosen as given in [26]. Therefore, it is quite possible to
say that these methods are not favorable for parameter extraction of PV cells/modules in terms of application.
Moreover, the results given in [1] cannot be obtained when the Newton–Raphson method applied to solve the
set of equations given in [1]. The model gives unrealistic results. For this reason, an accuracy analysis was
performed for the set of equations due to some noteworthy terms that should not be contained in the model.
Table 3 shows the values of equations calculated by substituting the solutions given in Table 2 into the set of
equations for both models. It is quite clear from the table that the solutions (roots) do not verify the equations
formulated by Hejri et al. in [1], at any degree.
It is well known that metaheuristic methods have been successfully applied to solve various types of
large-scale optimization problems. Therefore, in [21], the authors extracted the equivalent circuit parameters

Table 1. Manufacturers’ data for PV modules at T = 25 ◦ C .

PV module\ MSX60 KL070 BP-MSX120 BP-SX150 KC200GT SW255
Parameters [10] [21] [1] [24] [8] [7]
Voc (V) 21.1 21.5 42.1 43.5 32.9 38
Isc (A) 3.8 4.59 3.87 4.75 8.21 8.88
Vmpp (V) 17.1 17.1 33.7 34.5 26.3 30.9
Impp (A) 3.5 4.1 3.56 4.35 7.61 8.32
Ns 36 36 72 72 54 60
Pm (W) 60 70 120 150 200 255

3032
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

Table 2. Equivalent circuit parameters obtained for BP-MSX120 PV module.

BP-MSX120 (BP Solar Global Marketing)

PV module\
Ref. The proposed Model in Model in Model in Model in Model in Model
parameters
[1] model [1] [4] [6] [7] [8] [26]
Rs (Ω) 0.473 0.472 Na 0.473 0.472 0.472 0.472 0.472
Rp (Ω) 1365 1365 Na 1366 1365 1365 1364 1360
a 1.396 1.398 Na 1.396 1.396 1.396 1.398 1.398
Io (µA) 0.323 0.322 Na 0.323 0.322 0.322 0.322 0.325
Ipv (A) 3.871 3.871 Na 3.871 3.871 3.871 3.871 3.871

Table 3. The values of the set of equations for BP-MSX120 PV module after the solutions.

PV module\ BP-MSX120 (BP Solar Global Marketing)

related parameters Set of Eqs. (The proposed model) Sets of Eqs. (Given in [1])
Rs Eq. (22) = 8.8x10−16 Eq. (20) = −1.02x10−3
Rp Eq. (23) = 0 Eq. (21) = 529.17
−16
a Eq. (21) = −2.6x10 Eq. (19) = 2.2x−4
−17
Io Eq. (20) = −2.1x10 Eq. (18) = −8.1x10−9
Ipv Eq. (4) = 2.5x10−10 Eq. (12) = 9.93x10−2

of the KL070 PV module by applying metaheuristic methods, namely GA and hybrid GA-PSO algorithms.
Thus, the reliability of the proposed method is tested in this section by comparing the values with the ones
obtained by such applications. For comparison purposes, other methods are also taken into consideration, and
the values of the parameters extracted are presented in Table 4. One can see that the results obtained by
the proposed model and the metaheuristic methods slightly differ. This is due to the diode saturation current
neglected in the study cited in [21]. In other words, the saturation current has not been taken into account
in the equivalent circuit parameters and the analysis has been performed for four parameters. On the other
hand, the parameters extracted by using the proposed model are closely in agreement with those obtained by
using the existing methods. However, the significant dependency on initial conditions [4, 26], requirement of
a functional tool/software to solve the set of its nonlinear equations [26], and the use of the parallel resistor
[7] or diode ideality factor [8] as an input parameter are still among the limitations of the existing methods.
Additionally, the result obtained by using the equation given in [1] is unrealistic.

Table 4. Equivalent circuit parameters of KL070 PV module.

KL070 (KL solar PV module)
PV module\ Ref. Ref.
The proposed Model in Model in Model in Model in Model in Model in
parameters [21] [21]
model [1] [4] [6] [7] [8] [26]
(GA) (Hybrid GA-PSO)
Rs (Ω) 0.10 0.28 0.124 Na 0.123 0.123 0.123 0.124 0.124
Rp (Ω) 108.6 97.88 156.2 Na 156.2 156.2 156.2 156.1 156.2
a 1.242 1.292 1.712 Na 1.712 1.712 1.712 1.712 1.712
Io (µA) - - 5.61 Na 5.61 5.61 5.60 5.59 5.60
Ipv (A) 4.6 4.6 4.593 Na 4.593 4.593 4.593 4.593 4.594

3033
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

In this section, as an example for the application of the proposed model, the equivalent circuit parameters of
four commercial PV modules; MSX60, KC200GT, SW255, and BP SX-150 were extracted at standard test
conditions, i.e. air temperature of 25 ◦ C and solar radiation of 1000 W /m2 . Results are given in Tables 5 to 8
together with those obtained using existing methods. Although there are no satisfactory agreements between
the parameters extracted for MSX60 PV module by using the proposed model and quantum particle swarm
optimization algorithm [16], in Table 5, the results are close to each other. However, this is not the case for
the KC200GT parameters given in Table 6. This is due to the fact that the solutions were obtained for a fixed
value of the diode ideality factor ( a = 1.3 ) in [8]. The similar situation is also seen in the results obtained
for SW255 PV module given in Table 7. This is because the value of parallel resistor ( Rp ) has been assumed
to be in the range of 6.3 − 7.7 kΩ in the study conducted in [7]. However, the proposed model and the ones
given in [4, 6–8, 26] indicate that this value is around 2.5 kΩ . In Table 8, the results identified for BP SX-150
PV module are presented with those given in [24]. For the PV module, the experimental data at standard test
conditions were taken from [24]. Considering the first two columns in the table, it is clear that the values of
parameters do not match each other. In this case, first of all, it should be noted that the results (solutions)
given in [24] do not provide their own equations. This is due to some noteworthy terms that should not be
contained in the formula based on Rs . For this reason, the results obtained by rearranging the equation related
to Rs are given in the third column in the table and they are close to those determined by the proposed method
to a great extent. Hence, it is possible to say that both models are reliable but the requirement of experimental
I-V curves for PV modules creates a great handicap for the model in [24] in terms of implementation.

Table 5. Equivalent circuit parameters extracted for MSX60 PV module.

MSX60 (Solarex)
PV module\
Ref. The proposed Model in Model in Model in Model in Model in Model in
parameters
[16] model [1] [4] [6] [7] [8] [26]
Rs (Ω) 0.146 0.169 Na 0.169 0.169 0.169 0.169 0.178
Rp (Ω) 561.6 637.5 Na 637.8 637.6 637.5 637.8 649.9
a 1.331 1.404 Na 1.405 1.404 1.404 1.404 1.386
Io (µA) 0.150 0.329 Na 0.329 0.329 0.329 0.330 0.263
Ipv (A) 3.802 3.801 Na 3.801 3.801 3.801 3.801 3.800

Table 6. Equivalent circuit parameters extracted for KC200GT PV module.

KC200GT (Kyocera Solar)

PV module\
Ref. The proposed Model in Model in Model in Model in Model in Model in
parameters
[8] Model [1] [4] [6] [7] [8] [26]
Rs (Ω) 0.231 0.217 Na 0.217 0.217 0.217 0.217 0.217
Rp (Ω) 594.85 951.92 Na 951.93 951.93 951.93 948.93 899.90
a 1.3 1.342 Na 1.342 1.342 1.342 1.342 1.341
Io (µA) 0.096 0.171 Na 0.171 0.171 0.171 0.170 0.168
Ipv (A) 8.213 8.211 Na 8.211 8.211 8.211 8.211 8.212

In order to further investigate the reliability of the proposed model, the output current-voltage (I-V ) charac-
teristics of two PV modules and a solar cell are examined and the results are compared with the experimental

3034
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

Table 7. Equivalent circuit parameters extracted for SW255 PV module.

SW255 (SolarWorld)
PV module\
Ref. The proposed Model in Model in Model in Model in Model in Model in
parameters
[7] model [1] [4] [6] [7] [8] [26]
Rs (Ω) 0.2039 0.21 Na 0.21 0.21 0.21 0.21 0.21
Rp (Ω) 6300 2570.3 Na 2570.6 2570.3 2570.3 2568.5 2499.9
a 1.2647 1.2484 Na 1.2484 1.2484 1.2484 1.2484 1.2457
Io (nA) 0.3041 23.176 Na 23.177 23.176 23.176 23.168 22.204
Ipv (A) 8.88 8.8807 Na 8.8807 8.8807 8.8807 8.8807 8.8803

Table 8. Equivalent circuit parameters extracted for BP SX-150 PV module.

BP SX-150 (ABC Solar)

PV module\
Ref. The proposed The corrected model
parameters
[24] model in [24]
a 1.64 1.4851 1.4851 (fixed)
Rp (Ω) 1799.3 960.06 782.52
Rs (Ω) 0.3125 0.4543 0.4415
Io (µA) 2.8016 0.6166 0.6155
Ipv (A) 4.7508 4.7522 4.7527

data. Figure 2 demonstrates the consistency between the experimental data and predicted data for Solarex
MSX-60 PV module. Experimental data were taken from [16] and the I-V curve was plotted using the results
of the model given in Table 5. I-V characteristics predicted using the results given in [16] are also used for
comparison. It is clear from the figure that the calculated values are in good agreement with the experimental
data provided by the manufacturer. On the other hand, the data predicted using the value of equivalent circuit
parameters given in [16] do not exactly match the experimental data. Similarly, Figure 3 shows the I-V curves
for BP Solar MSX120 PV module, plotted by using the equivalent circuit parameters which were determined by
◦
the proposed model and those given in [22] at an air temperature of 40 C and solar radiation of 300 W /m2 .
The experimental data were taken from [22]. The parameters determined by the proposed model and the ones
given in [22] are presented in Table 9. Results show that the value of equivalent circuit parameters do not match
each other, but it is quite clear in Figure 3 that the I-V curve based on the proposed model exactly match the
experimental data at all points. Hence, it is possible to say that the parameters extracted by using the proposed
model are more realistic than those given in [22]. Figure 4 shows I-V curves of a 57 mm-diameter commercial
(R.T.C. France) silicon solar cell plotted using the optimal values of single-diode equivalent circuit parameters
identified by the proposed model and the optimal values of two-diode equivalent circuit parameters given in [23].
The experimental I-V characteristic is also presented in the figure to make a comparison. For the solar cell,
the experimental data were taken from [23] and the identified parameters are given in Table 9. As can be seen
in the figure, I-V curves constructed with the optimal parameters determined by the model and the two-diode
hybrid model (model-1) in [23] are exactly coincident with each other. Although there is a slight difference as
the output voltage increases, the curves created using the results of both methods are also consistent with the

3035
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

experimental data. On the other hand, it is clear in the figure that I-V characteristics created with the results in
[23] which were determined by using the pattern search algorithm (model-II) do not coincide with experimental
data at a satisfactory level. Consequently, all three comparisons suggest that the model extracts equivalent
circuit parameters in a more accurate manner, and as a result, it is well-suited for characterizing I-V curves.

Table 9. Extracted parameters for BP-MSX120 PV module and R.T.C France solar cell.

MSX120 (BP Solar Global Marketing) Solar cell (R.T.C France)

PV module\
The proposed Model - I Model - II The proposed
parameters Ref. [22]
model Ref. [23] Ref. [23] model
Rs (Ω) 1.769 1.4922 0.0367 0.032 0.0389
Rp (Ω) 2438.65 4288.2 55.529 81.3 52.252
a 1.1592 1.2472 1.45-1.99 1.6-1.19 1.4582
Io (nA) 28.164 100.14 224.7-755.2 988.9-1 253.83
Ipv (A) 1.1713 1.1709 0.7607 0.7602 0.7611

3.5

3 4
Output current (A)

3.5
2.5 3
2 2.5
2
1.5 1.5
1
1 Proposed Model 0.5
Model in [16]
0.5 Experimental data 0
17 17.5 18 18.5 19 19.5 20 20.5 21 21.5
0
0 2 4 6 8 10 12 14 16 18 20 22
Terminal voltage (V)

Figure 2. I-V curves for Solarex MSX60 PV module ( T = 25 ◦ C )

1.2

1 1
Output current (A)

0.8 0.8

0.6
0.6
Proposed Model 0.4
0.4 Experimental data
Model in [22] 0.2
0.2 0
34 35 36 37 38 39
0
0 5 10 15 20 25 30 35 40
Terminal voltage (V)

Figure 3. I-V curves for BP Solar MSX120 PV module ( T = 40 ◦ C )

3036
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

0.8
0.7
0.7

Output current (A)

0.6
0.6
0.5 0.5
0.4 0.4
0.3
0.3
Proposed Model 0.2
0.2 Two-diode model-I in [23] 0.1
Two-diode model-II in [23] 0
0.1 Experimental data 0.46 0.48 0.5 0.52 0.54 0.56 0.58
0
0 0.1 0.2 0.3 0.4 0.5 0.6
Terminal voltage (V)
Figure 4. I-V curves for R.T.C France solar cell ( T = 33 ◦ C )

4. Conclusions
In this study, an analytical model for extracting unknown parameters of PV cells/modules using the single-diode
model was developed without neglecting any of the valuable terms. It was tested on six well-known commercial
PV modules by comparing the results with those obtained by using such analytical methods presented in [1, 4, 6–
8, 26]. The results given in [16, 21–24] were also taken into account for comparison. In addition, the accuracy of
the extracted parameters was verified by experimental I-V curves and also those created by using the parameters
of the two-diode models given in [23]. All applications suggest that the proposed method allows to determine
all unknown parameters of the single-diode equivalent circuit in a more accurate manner. It requires only the
main data points of the I-V characteristic; open circuit, short circuit, and the maximum power point as input
data which make it a fast and low-cost parameter identification approach. Although the computation time was
not undertaken comprehensively, the results from MATLAB environment indicate that the model has the same
execution time as the other analytical models given in [1, 4, 6–8, 26] and it is negligible. On the other hand,
it is quite clear that the models given in [1] and [24] were developed incorrectly. Additionally, other models
considered for comparison have several drawbacks. For instance, the parallel resistance ( Rp ) or diode ideality
factor ( a), which are of great importance in the realization of I-V curves for PV modules, are considered to be
known and used as an input parameter in [7] and [8], respectively. This makes these models dependent on the
empirical data from the literature, except for the main data points of the I-V characteristics. Moreover, the
solution of their equations is highly dependent on the initial values of the parameters [4, 26]. The model proposed
in [26] also requires a functional tool/software for solving the set of its nonlinear equations. Additionally, the
model developed in [24] is a partly numerical-based method, and as a result, it requires experimental I-V curves
of PV modules which are not provided by most manufacturers.

5. Future work
The proposed method will be improved to extract equivalent circuit parameters of two-diode seven-parameter
model of PV cells/modules.

References

[1] Hejri M, Mokhtari H, Azizian MR, Söder L. An analytical-numerical approach for parameter determination of a
five-parameter single-diode model of photovoltaic cells and modules. International Journal of Sustainable Energy
2016; 35 (4): 396-410. doi: 10.1080/14786451.2013.863886

3037
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

[2] Chan DSH, Phang JCH. Analytical methods for the extraction of solar-cell single-and double-diode model pa-
rameters from I-V characteristics. IEEE Transactions on Electron Devices 1987; 34 (2): 286-293. doi: 10.1109/T-
ED.1987.22920
[3] Shongwe S, Hanif M. Comparative analysis of different single-diode PV modeling methods. IEEE Journal of
Photovoltaics 2015; 5 (3): 938-946. doi: 10.1109/JPHOTOV.2015.2395137
[4] Sera D, Teodorescu R, Rodriguez P. PV panel model based on datasheet values. In: IEEE International Symposium
on Industrial Electronics; Vigo, Spain; 2007. pp. 2392-2396. doi: 10.1109/ISIE.2007.4374981
[5] Reis LRD, Camacho JR, Novacki DF. The Newton Raphson method in the extraction of parameters of PV modules.
In: International Conference on Renewable Energies and Power Quality (ICREPQ’17); Malaga, Spain; 2017. pp.
634-639. doi: 10.24084/repqj15.416
[6] Kareem MSA, Saravanan M. A new method for accurate estimation of PV module parameters and extraction
of maximum power point under varying environmental conditions. Turkish Journal of Electrical Engineering &
Computer Sciences 2016; 24 (4): 2028-2041. doi: 10.3906/elk-1312-268
[7] Chaibi Y, Salhi M, El-jouni A, Essadki A. A new method to extract the equivalent circuit parameters of a
photovoltaic panel. Solar Energy 2018; 163: 376-386. doi: 10.1016/j.solener.2018.02.017
[8] Cubas J, Pindado S, Farrahi A. New method for analytical photovoltaic parameter extraction. In: International
Conference on Renewable Energy Research and Applications (ICRERA); Madrid, Spain; 2013. pp. 873-877. doi:
10.1109/ICRERA.2013.6749874
[9] Patel S. The single diode model of I-V and P-V characteristics using the lambert w function. Interna-
tional Journal of Innovative Research in Science Engineering and Technology 2016; 5 (5): 7034-7039. doi:
10.15680/IJIRSET.2016.0505058
[10] Farivar G, Asaei B. Photovoltaic module single diode model parameters extraction based on manufacturer datasheet
parameters. In: IEEE International Conference on Power and Energy (PECon); Kuala Lumpur, Malaysia; 2010.
pp. 929-934. doi: 10.1109/PECON.2010.5697712
[11] Shannan NMAA, Yahaya NZ, Singh B. Single-diode model and two-diode model of PV modules: a comparison.
In: IEEE International Conference on Control System, Computing and Engineering (ICCSCE); Mindeb, Malaysia;
2013. pp. 210-214. doi: 10.1109/ICCSCE.2013.6719960
[12] Yahfdhou A, Mahmoud AK, Youm I. Evaluation and determination of seven and five parameters of a photovoltaic
generator by an iterative method. Smart Grid and Renewable Energy 2016; 7: 247-260. doi: 10.4236/sgre.2016.79019
[13] Hejri M, Mokhtari H, Azizian MR, Ghandhari M, Söder L. On the parameter extraction of a five-parameter
double-diode model of photovoltaic cells and modules. IEEE Journal of Photovoltaics 2014; 4 (3): 915-923. doi:
10.1109/JPHOTOV.2014.2307161
[14] Zagrouba M, Sellami A, Bouaicha M, Ksouuri M. Identification of PV solar cells and modules parameters using
the genetic algorithms: Application to maximum power extraction. Solar Energy 2010; 84 (5): 860-866. doi:
10.1016/j.solener.2010.02.012
[15] Sandrolini L, Artioli M, Reggiani U. Numerical method for the extraction of photovoltaic module double-diode model
parameters through cluster analysis. Applied Energy 2010; 87 (2): 442-451. doi: 10.1016/j.apenergy.2009.07.022
[16] Muralidharan R. Parameter extraction of solar photovoltaic cells and modules using current-voltage characteristics.
International Journal of Ambient Energy 2017; 38 (5): 509-513. doi: 10.1080/01430750.2016.1144525
[17] Ishaque K, Salam Z. An improved modeling method to determine the model parameters of photovoltaic (PV)
modules using differential evolution (DE). Solar Energy 2011; 85 (9): 2349-2359. doi: 10.1016/j.solener.2011.06.025
[18] Goldberg DE. Genetic algorithms in search optimization and machine learning. Boston, MA, USA: Addison Wesley
Publishing Company, 1989.

3038
 ATAY and EMİNOĞLU/Turk J Elec Eng & Comp Sci

[19] Kennedy J, Eberhart RC. Particle swarm optimization. In: IEEE International Conference on Evolutionary Com-
putation (ICNN’95); Perth, WA, Australia; 1995. pp. 1942-1948. doi: 10.1109/ICNN.1995.488968
[20] Storn R, Price K. Differential Evolution–a simple and efficient heuristic for global optimization over continuous
spaces. Journal of Global Optimization 1997; 11 (4): 341-359. doi: 10.1023/A:1008202821328
[21] Saravanan C, Panneerselvam MA. A comprehensive analysis for extracting single diode PV model parameters by
hybrid GA-PSO algorithm. International Journal of Computer Applications 2013; 78 (8): 16-19. doi: 10.5120/13509-
1265
[22] Bencherif M, Benouaz T. Parameter extraction of solar panels using the graphical method. International Journal
of Ambient Energy 2018; 1-18. doi: 10.1080/01430750.2018.1501735
[23] Hamid NFA, Rahim NA, Selvaraj J. Solar cell parameters identification using hybrid Nelder-Mead and modified
particle swarm optimization. Journal of Renewable and Sustainable Energy 2016; 8: 1-21. doi: 10.1063/1.4941791
[24] Hussein A. A simple approach to extract the unknown parameters of PV modules. Turkish Journal of Electrical
Engineering & Computer Sciences 2017; 25: 4431-4444. doi:10.3906/elk-1703-14
[25] Amin YK, Shahhoseini A. A fast modeling of the double-diode model for PV modules using combined analytical
and numerical approach. Solar Energy 2018; 162: 403-409.
[26] Mohapatra A, Nayak B, Mohanty KB. Parameter extraction of PV modules using NLS algorithm with Ex-
perimental Validation. International Journal of Electrical and Computer Engineering 2017; 7 (5): 23922400.
doi:10.11591/ijece.v7i5.pp2392-2400
[27] Humada AM, Hojabri M, Mekhilef S, Hamada HM. Solar cell parameters extraction based on single and double-diode
models: a review. Renewable and Sustainable Energy Reviews 2016; 56: 494-509. doi: 10.1016/j.rser.2015.11.051

3039