Identification study of solar cell/module using recent optimization techniques

International Journal of Electrical and Computer Engineering (IJECE)
Vol. 12, No. 2, April 2022, pp. 1189~1198
ISSN: 2088-8708, DOI: 10.11591/ijece.v12i2.pp1189-1198  1189
Journal homepage: http://ijece.iaescore.com
Identification study of solar cell/module using recent
optimization techniques
Mahmoud Abbas El-Dabah1
, Ragab Abdelaziz El-Sehiemy2
, Mohamed Ahmed Ebrahim3
,
Zuhair Alaas4
, Mohamed Mostafa Ramadan4,5
1
Department of Electrical Engineering, Faculty of Engineering, Al-Azhar University, Cairo, Egypt
2
Department of Electrical Engineering, Faculty of Engineering, Kafr El-Sheikh University, Kafr El Sheikh, Egypt
3
Department of Electrical Engineering, Faculty of Engineering at Shoubra, Benha University, Banha, Egypt
4
Department of Electrical Engineering, Faculty of Engineering, Jazan University, Jazan, Saudi Arabia
5
Department of Electrical Engineering, Faculty of Engineering, Helwan University, Helwan, Egypt
Article Info ABSTRACT
Article history:
Received Mar 22, 2021
Revised Nov 1, 2021
Accepted Nov 14, 2021
This paper proposes the application of a novel metaphor-free population
optimization based on the mathematics of the Runge Kutta method (RUN)
for parameter extraction of a double-diode model of the unknown solar cell
and photovoltaic (PV) module parameters. The RUN optimizer is employed
to determine the seven unknown parameters of the two-diode model. Fitting
the experimental data is the main objective of the extracted unknown
parameters to develop a generic PV model. Consequently, the root means
squared error (RMSE) between the measured and estimated data is
considered as the primary objective function. The suggested objective
function achieves the closeness degree between the estimated and
experimental data. For getting the generic model, applications of the
proposed RUN are carried out on two different commercial PV cells. To
assess the proposed algorithm, a comprehensive comparison study is
employed and compared with several well-matured optimization algorithms
reported in the literature. Numerical simulations prove the high precision
and fast response of the proposed RUN algorithm for solving multiple PV
models. Added to that, the RUN can be considered as a good alternative
optimization method for solving power systems optimization problems.
Keywords:
Double diode model
Optimization algorithms
Parameter extraction
Runge Kutta optimizer
This is an open access article under the CC BY-SA license.
Corresponding Author:
Mahmoud Abbas El-Dabah
Department of Electrical Engineering, Al-Azhar University
El Mokhayam El Daam Street, Nasr City, Cairo, Egypt
Email: Dr_mdabah@azhar.edu.eg
1. INTRODUCTION
The increased demand for electrical energy while decreasing CO2 emission associated with
electricity generation requires green renewable energy sources (GRERs). As a source of green energy,
photovoltaic (PV) can generate electricity from sunlight using semiconductor materials. Globally, there is a
massive increase in solar photovoltaic capacities in comparison with those of wind energy as per the REN21
latest published report in 2021 [1]. The Egyptian government pays a concentration towards the exploitation
of available GRERs and PV especially. One of the most significant constructed PV projects in the world was
at Benban, Aswan, Egypt, which has an installed capacity of 1.8 GW [2].
Due to this high potential spread, accurate modeling of PVs has a high impact on their performance
under varied environmental conditions and shading. To express the non-linearity relationship between
current-voltage (I-V) and power-voltage (P-V) of the PV cell, equivalent circuit models will be required. As

 ISSN: 2088-8708
Int J Elec & Comp Eng, Vol. 12, No. 2, April 2022: 1189-1198
1190
per the literature, the single and double diode equivalent circuit models are used in parameter extraction of
PV cells, which can be reviewed in [3]. However, its simplicity and reduced number of estimated parameters
of the single diode equivalent circuit model, its precision diminishes at low irradiation levels with
temperature variation [4], [5]. The estimated parameters in the single diode model are the photogenerated
current, the reverse saturation currents, the series and shunt resistances, and the ideality factor. The double
diode equivalent circuit model is composed of seven parameters, five of them the same as five parameters of
the single diode in addition to two parameters, which are the reverse saturation current and the ideality factor
of the additional diode.
As per the literature review, the parameter extraction techniques of the PV equivalent circuit model
can be classified into three categories, namely analytical, numerical (deterministic and stochastic), and hybrid
techniques [6], [7]. The analytical methods utilize the power-voltage and current-voltage data curves with the
aids of selected data points from the manufacturer datasheets (the open-circuit voltage V_oc, the maximum
output power P_m, the short circuit current I_sc, the maximum output current I_m, the maximum output
voltage V_m) to constitute the mathematical parameter estimation problem [8]–[13]. These methods require
some mathematical approximations (simplification) to reduce the number of extracted parameters. Although
the used approximation provides ease of implementation and less computational effort, however, it has a
significant impact on the solution accuracy. In study [14], it was approved that these approaches are less
accurate than numerical approaches. To overcome modeling imprecision that may arise if there are inaccurate
selected data from datasheets, the measured data of I-V are the same as used in numerical (iterative)
techniques. The parameter extraction techniques can be categorized into two categories: deterministic and
stochastic (heuristic and meta-heuristic) techniques. The deterministic techniques such as Newton-Raphson
[15], linear identification [16], or the Levenberg-Marquardt (LM) algorithm [17] can be used for parameter
extraction of PV equivalent circuit parameters. The main drawback of these conventional methods is local
optimum trapping due to its sensitivity to the initial solution. As an alternative to the deterministic
techniques, the naturally inspired (meta-heuristic) are extensively used in the last decade. In literatures, there
are a lot of these methods such as, biogeography-based heterogeneous cuckoo search (BHCS) [18], pattern
search (PS) [19], firefly algorithm (FA) [20], ant lion optimizer (ALO) [21], Jaya algorithm [22], salp swarm
algorithm (SSA) [23], elephant herd optimizer [24], enhanced sine cosine algorithm (ISCA) [25], hybridized
interior search algorithm (HISA) [26], an artificial bee colony-differential evolution (ABC-DE) [27],
improved adaptive Nelder-Mead simplex (NMS) hybridized with the artificial bee colony (ABC)
metaheuristic, algorithm of hybrid adaptive and Nelder-Mead simplex (EHA-NMS) [28], mutative-scale
parallel chaos optimization algorithm (MPCOA) [29], classified perturbation mutation based particle swarm
optimization (CPMPSO) [30], heterogeneous comprehensive learning particle swarm optimizer (HCLPSO)
[31], and improved shuffled complex evolution (ISCE) [32]. A forensic based optimization algorithm was
developed in [33] for finding the optimal parameters of solar cell modules. Another optimizer called
turbulent flow of water optimizer in [34] was developed to optimize the parameters of three solar cell models.
An assessment study based on the elephant herd optimization, which developed with different versions [35],
is compared with closed loop particle swarm optimizer in [36]. The An interval branch and bound global
optimization algorithm (IBBGO) that is referred to, interval branch/bound global optimizer, was integrated to
find the optimal parameters of three PV models [37].
All these meta-heuristic optimization algorithms aim at minimizing the objective/cost function while
balancing between exploration and exploitation phases. At the same time, meta-heuristic optimization
algorithms cannot provide robust search capability towards optimal solutions [38]. To overcome these
challenges, a novel metaphor-free population optimization based on the mathematics of the Runge Kutta
method (RUN) is presented [39]. RUN optimizer balances between exploration and exploitation phases in
dynamic behavior. Moreover, the RUN optimization algorithm has a competitive convergence speed as well
as an enhanced solution quality to avoid local optimal solutions.
It is cleared that many meta-heuristic optimization algorithms are used for parameter extraction of
the equivalent circuit of the PV model. Nevertheless, most of these algorithms have several control
parameters that require tuning to achieve better performance for an optimization problem. The RUN
algorithm has fewer control parameters due to its specific property which depends on the main logic of the
Runge Kutta technique. This advantage motivates the authors of the presented article to use it in parameter
estimation of the double diode model of PV cell/module. The main contribution of this research can be
summarized as: i) develop RUN optimization algorithm for PV parameter extraction based on double diode
model, ii) assessment of the extracted parameter using RUN with the recent optimization algorithms, iii)
recommending the best method and best equivalent circuit model to be used for PV cell/module, and iv) the
proposed RUN has the best convergence rates compared with the competitive methods.
This paper will be organized as follows: After this section, problem formulation will be introduced
in section 2. Section 3 will develop the RUN optimization algorithm for use in parameter extraction of a PV

Int J Elec & Comp Eng ISSN: 2088-8708 
Identification study of solar cell/module using recent optimization techniques (Mahmoud Abbas El-Dabah)
1191
cell/module. Application of the proposed algorithm to different commercial cell/module and test results will
be emphasized in section 4. Section 5 will conclude the results drawn from this research.
2. PROBLEM FORMULATION
2.1. Electrical models
Mathematical modeling of PV cell/module equivalent circuits will be illustrated. The equivalent
circuit models emulate the non-linear (I-V) and (P-V) relationships of the PV cell. This model is considered
as the simplest equivalent circuit model. It has five parameters which are the photogenerated current 𝐼𝑝ℎ,
series and shunt resistances 𝑅𝑠, 𝑅𝑠ℎ, the ideality factor 𝑛 and the reverse saturation current 𝐼𝑅𝑆 as shown in
Figure 1(a). The load current I of solar cells can be represented by (1):
𝐼 = 𝐼𝑝ℎ − 𝐼𝑅𝑆−1 [exp (
𝑉 + 𝐼𝑅𝑠
𝑛 𝑉𝑡
− 1)] −
(𝑉 + 𝐼𝑅𝑠)
𝑅𝑠ℎ (1)
In this model, the recombination of generated charge carriers was neglected. Moreover, its precision
diminishes at low irradiation levels with temperature variation [4]–[6]. The recombination effect was
modeled in the double diode model that is presented in Figure 1(b), the load current 𝐼of solar cells can be
represented by (2):
𝐼 = 𝐼𝑝ℎ − 𝐼𝑅𝑆−1 [𝑒𝑥𝑝 (
𝑉 + 𝐼𝑅𝑠
𝑛1𝑉𝑡
− 1)] − 𝐼𝑅𝑆−2 [𝑒𝑥𝑝 (
𝑉 + 𝐼𝑅𝑠
𝑛2𝑉𝑡
− 1)] −
(𝑉 + 𝐼𝑅𝑠)
𝑅𝑠ℎ (2)
(a) (b)
Figure 1. Diode equivalent circuit models, (a) single and (b) double
2.2. Mathematical representation of parameter extraction problem
The parameter extraction problem is considered an optimization problem. The main objective of the
estimation problem is to minimize the variance between the experimental data and simulated results so that
the optimal values of these unknown model parameters can be extracted. The objective function is defined as
the overall root mean square error (RMSE). For N-measurements, the objective function is formulated as [38]
as in (3):
𝑅𝑀𝑆𝐸(𝑥) = √
1
𝑁
∑[𝑓(𝑉𝑘, 𝐼𝑘, 𝑥) − 𝐼𝑘]2
𝑁
𝑘=1 (3)
where, 𝑥 = {𝐼𝑝ℎ, 𝐼𝑅𝑆−1, 𝐼𝑅𝑆−2, 𝑅𝑆, 𝑅𝑠ℎ, 𝑛1, 𝑛2} and the 𝑓(𝑉𝑘, 𝐼𝑘, 𝑥) is used for current calculation from (2).
3. RUNGE KUTTA OPTIMIZER (RUN)
The Runge Kutta-based optimization (RUN) is a novel optimization algorithm proposed in [39]. The
RUN is developed based on the mathematics of the Runge Kutta method. It is inspired by the logic of slope
variations computed by the Runge Kutta method as a promising and logical searching mechanism for global
optimization. In RUN, three distinctive phases are simulated. These phases are active exploration and
exploitation phases for exploring the promising regions in the feature space and constructive toward the
global best solution. Moreover, in the last phase, an enhanced solution quality mechanism is employed to

 ISSN: 2088-8708
1192
avoid the local optimal solutions and increase convergence speed. Figure 2 demonstrates the phases of the
proposed RUN algorithm. The phases of the suggested RUN optimizer are discussed in detail in the
following sub-sections.
Algorithm 1. The pseudo-code of RUN
Stage 1. Initialization
Initialize 𝑎, 𝑏
Generate the RUN population 𝑋𝑛(𝑛 = 1, 2, …, 𝑁)
Calculate the objective function of each member of the population
Determine the solutions 𝑥𝑤, 𝑥𝑏, and 𝑥𝑏𝑒𝑠𝑡
Stage 2. RUN operators
for i= 1: Maxi
for n = 1 : N
for l = 1 : D
Calculate position 𝑥𝑛+1,𝑙 using Eq. 4
end for
Enhance the solution quality
if 𝑟𝑎𝑛𝑑 < 0.5
Calculate position 𝑥𝑛𝑒𝑤2 using Eq. 7
if 𝑓(𝑥𝑛) < 𝑓(𝑥𝑛𝑒𝑤2)
if rand< 𝑤
Calculate position 𝑥𝑛𝑒𝑤3 using Eq. 9
end
end
end
Update positions 𝑥𝑤 and 𝑥𝑏
end for
Update position 𝑥𝑏𝑒𝑠𝑡
i=i+1
end
Stage 3. return 𝑥𝑏𝑒𝑠𝑡
Figure 2. Pseudo-code of the proposed algorithm (RUN)
3.1. Updating solutions stage
In the updating solutions stage, the RUN uses a search mechanism (SM) based on the runge kutta
method to update the position of the current solution at each iteration, which is defined as (4):
𝑖𝑓 𝑟𝑎𝑛𝑑 < 0.5
(Exploration phase)
𝑥𝑛+1 = (𝑥𝑐 + 𝑟. 𝑆𝐹. 𝑔. 𝑥𝑐) + 𝑆𝐹. 𝑆𝑀 + 𝜇. 𝑟𝑎𝑛𝑑𝑛.(𝑥𝑚 − 𝑥𝑐)
𝑒𝑙𝑠𝑒
(Exploitation phase)
𝑥𝑛+1 = (𝑥𝑚 + 𝑟. 𝑆𝐹. 𝑔. 𝑥𝑚) + 𝑆𝐹. 𝑆𝑀 + 𝜇. 𝑟𝑎𝑛𝑑𝑛. (𝑥𝑟1 − 𝑥𝑟2)
𝑒𝑛𝑑 (4)
where, 𝑟 is an integer number, which is 1 or-1. 𝑔 is a random number in the range [0, 2]. 𝑆𝐹 is an adaptive
factor. Where, 𝜇 is a random number. The formula of 𝑆𝑀 is defined in [34]. The formula of 𝑆𝐹 is as (5):
𝑆𝐹 = 2.(0.5 − 𝑟𝑎𝑛𝑑)×𝑓 (5)
where, 𝑓 = 𝑎 × 𝑒𝑥 𝑝 (−𝑏 × 𝑟𝑎𝑛𝑑 × (
𝑖
𝑀𝑎𝑥𝑖
)), Maxi stands for the largest number of iterations. The formula
of xc and xm are as (6):
𝑥𝑐 = 𝜑 × 𝑥𝑛 + (1 − 𝜑) × 𝑥𝑟1
𝑥𝑚 = 𝜑 × 𝑥𝑏𝑒𝑠𝑡 + (1 − 𝜑) × 𝑥𝑙𝑏𝑒𝑠𝑡 (6)
where, φ is a random number in the range of (0,1). xbest is the best-so-far solution. xlbest is the best position
obtained at each iteration.
3.2. Enhanced solution quality stage
In the RUN algorithm, enhanced solution quality (ESQ) is employed to increase the quality of
solutions and to avoid local optima in each iteration. The following scheme is executed to create the solution
(𝑥𝑛𝑒𝑤2) by using the ESQ (7), (8):

1193
𝑖𝑓 𝑟𝑎𝑛𝑑 < 0.5
𝑖𝑓 𝑤 < 1
𝑥𝑛𝑒𝑤2 = 𝑥𝑛𝑒𝑤1 + 𝑟.𝑤.|(𝑥𝑛𝑒𝑤1 − 𝑥𝑎𝑣𝑔) + 𝑟𝑎𝑛𝑑𝑛|
𝑒𝑙𝑠𝑒
𝑥𝑛𝑒𝑤2 = (𝑥𝑛𝑒𝑤1 − 𝑥𝑎𝑣𝑔) + 𝑟.𝑤.|(𝑢.𝑥𝑛𝑒𝑤1 − 𝑥𝑎𝑣𝑔) + 𝑟𝑎𝑛𝑑𝑛|
𝑒𝑛𝑑 (7)
𝑤 = 𝑟𝑎𝑛𝑑(0, 2).𝑒𝑥 𝑝 (−𝑐 (
𝑖
𝑀𝑎𝑥𝑖
)),𝑥𝑎𝑣𝑔
=
𝑥𝑟1+𝑥𝑟2+𝑥𝑟3
3
,
𝑥𝑛𝑒𝑤1 = 𝛽 × 𝑥𝑎𝑣𝑔 + (1 − 𝛽) × 𝑥𝑏𝑒𝑠𝑡 (8)
where, 𝛽 is a random number in the range of [0, 1]. 𝑐 is a random number, which is equal to 5× 𝑟𝑎𝑛𝑑 in this
study. 𝑟 is an integer number, which is 1, 0, or-1. 𝑥𝑏𝑒𝑠𝑡 is the best solution explored so far. The solution
calculated in this part (𝑥𝑛𝑒𝑤2) may not have better fitness than that of the current solution (i.e., 𝑓(𝑥𝑛𝑒𝑤2) >
𝑓(𝑥𝑛)). To have another chance for creating a good solution, another new solution (𝑥𝑛𝑒𝑤3) is generated,
which is defined as in (9),
if rand < 𝑤
𝑥𝑛𝑒𝑤3 = (𝑥𝑛𝑒𝑤2 − 𝑟𝑎𝑛𝑑.𝑥𝑛𝑒𝑤2) + 𝑆𝐹.(𝑟𝑎𝑛𝑑.𝑥𝑅𝐾 + (𝑣.𝑥𝑏 − 𝑥𝑛𝑒𝑤2))
𝑒𝑛𝑑 (9)
where 𝑣 is a random number with a value of 2 × 𝑟𝑎𝑛𝑑.
4. SIMULATION AND RESULT
To assess the proposed algorithm, a comprehensive study is employed compared with several
previous optimization algorithms. Numerical simulations on different commercial PV cells/modules will be
illustrated below. To evaluate the performance of the proposed algorithm, A set of standard data for a
commercial silicon solar cell (made by R.T.C. company from France) with a diameter of 57 mm, at a
temperature of 33 °C, and 1 sun (1,000 W/m2
) [40]. The PhotowattPWP201 PV commercial module with 36
polycrystalline silicon cells connected in series, operating under an irradiance of 1,000 W/m2
and temperature
of 45 °C [41]. Table 1 summarizes the datasheets of the selected commercial PV cell/module to be tested in
this work. The boundary constraints of the extracted parameters are given in Table 2. The proposed RUN-
based model and the other comparative algorithms are executed by the authors via the matrix laboratory
(MATLAB) 2017a platform using an Intel ® core TM i5-7200U CPU, 2.50 GHz, 8 GB RAM Laptop.
Accuracy examination of the proposed optimization RUN algorithm for parameters identification will be
additionally accomplished by the current calculation based on the values estimated for the two models
considered for comparison with that taken from the experimental measurements. The error concerning each
of the measured values was evaluated by relative error (RE) and individual absolute error (IAE), calculated
as given in (10) and (11), respectively.
𝑅𝐸 = (𝐼𝑀𝑒𝑎𝑠𝑢𝑟𝑒𝑑 − 𝐼𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑒𝑑)/𝐼𝑀𝑒𝑎𝑠𝑢𝑟𝑒𝑑 (10)
𝐼𝐴𝐸 = |𝐼𝑀𝑒𝑎𝑠𝑢𝑟𝑒𝑑 − 𝐼𝐸𝑠𝑡𝑖𝑚𝑎𝑡𝑒𝑑| (11)
Table 1. Datasheets of selected cell/module for study
Company Cell type Voc [V] Isc [A] Vm [V] Im [A] Pm [W] Reference temp. [º
C] # of cells / module
France solar NA 0.5728 0.76 0.45 0.691 11.315 33 1
Photowatt Polycrystalline 16.778 1.03 12.60 0.898 0.311 45 36
Table 2. Boundary constraints of the extracted parameters
Model Parameters
France solar 𝐼𝑝ℎ [A] 𝐼𝑅𝑆1,2[𝜇𝐴] 𝑅𝑆 [Ω] 𝑅𝑠ℎ [Ω] 𝑛1,2
Min 0 0 0 0 1
Max 1 1 0.5 100 2
Photowatt 201 Min 0 0 0 0 1
Max 2 10 0.5 3000 2

 ISSN: 2088-8708
1194
4.1. RTC. france solar cell
For RTC solar cell model, the optimal parameters attained using different optimization techniques
are presented in Table 3. The solar cell parameters estimated using the RUN algorithm have the lowest
RMSE value. Figure 3 demonstrates that the estimated I-V and P-V characteristics using RUN are identical
compared to the measurement as well as the datasheet. Table 4 spots the light on the coincidence between
both the measured and estimated values for voltage and current. The results of MPP are highlighted in green
in Table 4 and illustrated in Figure 3. As can be concluded from Figure 4, RUN has the lowest RMSE value
(0.0009829) in comparison with sunflower optimization (SFO), which has the highest value (0.002). On the
other hand, RUN has the lowest execution time of (9.95 s) in comparison with SFO, which has the highest
execution time of (44.402 s).
Table 3. Optimum parameter settings using different optimization techniques for RTC France solar cell
RCGA CSA SSA PSO SFO GW-CS RUN
𝑰𝒑𝒉 (𝑨) 0.7606 0.7608 0.7604 0.76077 0.76275 0.76002 0.76077
𝑰𝑹𝑺−𝟏(𝝁𝑨) 0.3743 0.6489 0.3165 0.6680 0.2452 0.2272 0.22266
𝑰𝑹𝑺−𝟐(𝝁𝑨) 0.0731 0.2199 0.1728 0.2389 0.8324 0.2553 0.67737
𝑹𝑺 (𝛀) 0.0357 0.0367 0.0369 0.0366 0.0339 0.0359 0.03674
𝑹𝒔𝒉(𝛀) 60.823 54.946 59.7711 55.691 51.165 67.0574 55.4973
𝒏𝟏 1.4968 1.9454 1.69629 2.000 1.4768 1.4632 1.45006
𝒏𝟐 1.9607 1.4493 1.43602 1.4559 1.7895 1.6630 1.96604
RMSE 0.0010 0.000985 0.00103 0.00098 0.002 0.00110 0.0009829
Duration [s] 13.655 11.524 10.1512 10.567 44.402 14.131 9.95
Figure 3. The coincidence between measured and estimated results in V-I and V-P curves of RTC-solar cell
Table 4. The relative error for each measurement using a RUN-based two diode model
Measured
Voltage
Measured
current
Estimated
current
IAE RE Measured
Voltage
Measured
current
Estimate
d current
IAE RE
-0.2057 0.764 0.7640 0.0000 0.0000 0.4373 0.728 0.7399 0.0014 0.0019
-0.1291 0.762 0.7626 0.0006 0.0008 0.4590 0.7065 0.7271 0.0009 0.0012
-0.0588 0.7605 0.7613 0.0008 0.0011 0.4137 0.6755 0.7066 0.0001 0.0001
0.0057 0.7605 0.7602 0.0003 0.0004 0.4784 0.632 0.6748 0.0007 0.0010
0.0646 0.7600 0.7591 0.0009 0.0012 0.4960 0.573 0.6302 0.0018 0.0029
0.1185 0.7590 0.7581 0.0009 0.0012 0.5119 0.499 0.5711 0.0019 0.0033
0.1678 0.7570 0.7572 0.0002 0.0002 0.5265 0.413 0.4982 0.0008 0.0017
0.2132 0.7570 0.7562 0.0008 0.0010 0.5398 0.3165 0.4118 0.0012 0.0030
0.2545 0.7555 0.7552 0.0003 0.0004 0.5521 0.212 0.3150 0.0015 0.0046
0.2924 0.7540 0.7537 0.0003 0.0004 0.5633 0.1035 0.2095 0.0025 0.0119
0.3269 0.7505 0.7514 0.0009 0.0012 0.5736 -0.01 0.0997 0.0038 0.0368
0.3585 0.7465 0.7473 0.0008 0.0010 0.5833 -0.123 -0.0126 0.0026 -0.2618
0.3873 0.7385 0.7640 0.0000 0.0000 0.5900 -0.21 -0.1280 0.0050 -0.0408

1195
Figure 4. The objective function of the competitive estimation algorithms
4.2. Photo watt PWP201
In the Photowatt PWP201 module, from the obtained results in Tables 5 and 6, the RUN has the
lowest RMSE value (0.003139) among the competitive optimization algorithms. The maximum power point
(MPP) is highlighted in green in Table 6 and illustrated clearly in Figure 5. The RMSE for all optimization
techniques are plotted in Figure 6. Generally speaking, RUN has the lowest RMSE value (0.003139)
compared to all the applied optimization techniques. The crow search algorithm (CSA) comes in the first
class in terms of execution time with (10.2910 s) in comparison with the SFO with (46.7350 s), which has the
highest value of elapsed time.
Table 5. Optimum parameter settings using competitive optimization algorithms for Photowatt-PWP201
RCGA CSA SSA PSO SFO GW-CS RUN
𝑰𝒑𝒉 (𝑨) 1.0252 1.0258 1.02867 1.0258 1.01883 1.02574 1.0270
𝑰𝑹𝑺−𝟏(𝝁𝑨) 5.1463 9.8979 7.59250 3.8226 0.2769 0.2527 1.4074
𝑰𝑹𝑺−𝟐(𝝁𝑨) 2.4599 3.8944 7.53697 5.0705 0.9499 6.01257 5.5321
𝑹𝑺 (𝛀) 0.0321 0.0322 0.02864 0.03175 0.03724 0.03191 0.0321
𝑹𝒔𝒉(𝛀) 1000 1001 629.437 1000 351.092 248.962 100.3905
𝒏𝟏 1.3982 1.877 1.48011 1.7271 1.32756 1.89869 1.9920
𝒏𝟐 1.767 1.3723 1.61563 1.4007 1.23380 1.41230 1.4034
RMSE 0.0035 0.0035 0.00491 0.00349 0.00734 0.00347 0.003139
Duration[s] 12.0730 10.2910 19.5530 10.4660 46.7350 11.4380 11.25
Note: RCGA= real coded genatic algorithm, SSA = introduced in the introduction, PSO is particle swarm optimization,
GW-CS = gray wolf cuckoo search algorithm
Table 6. The relative error for each measurement using a RUN-based two diode model
Measured
voltage
Measured
current
Estimated
current
IAE RE
Measured
Voltage
Measured
current
Estimated
current
IAE RE
1.9426 1.0345 1.0272 0.0073 0.0070 12.6490 0.9120 0.9227 0.0028 0.0030
0.1248 1.0315 1.0266 0.0049 0.0047 13.1231 0.8725 0.9117 0.0003 0.0003
1.8093 1.0300 1.0261 0.0039 0.0038 14.2221 0.7265 0.8715 0.0010 0.0012
3.3511 1.0260 1.0256 0.0004 0.0004 14.6995 0.6345 0.7264 0.0001 0.0002
4.7622 1.0220 1.0249 0.0029 0.0029 15.1346 0.5345 0.6352 0.0007 0.0010
6.0538 1.0180 1.0239 0.0059 0.0058 15.5311 0.4275 0.5349 0.0004 0.0007
7.2364 1.0155 1.0222 0.0067 0.0066 15.8929 0.3185 0.4285 0.0010 0.0024
8.3189 1.0140 1.0189 0.0049 0.0049 16.2229 0.2085 0.3188 0.0003 0.0010
9.3097 1.0100 1.0131 0.0031 0.0031 16.5241 0.1010 0.2083 0.0002 0.0009
10.2163 1.0035 1.0029 0.0006 0.0006 16.7987 -0.0080 0.0989 0.0021 0.0204
11.0449 0.9880 0.9862 0.0018 0.0018 17.0499 -0.1110 -0.0076 0.0004 -0.0472
11.8018 0.9630 0.9603 0.0027 0.0028 17.2793 -0.2090 -0.1106 0.0004 -0.0034
12.4929 0.9255 1.0272 0.0073 0.0070 17.4885 -0.3030 -0.2091 0.0001 -0.0005

 ISSN: 2088-8708
1196
Figure 5. The coincidence between measured and estimated results in V-I and V-P curves of PWP201
Figure 6. The convergence rates of objective function decline using competitive algorithms
5. CONCLUSION
This manuscript presents a new physical implementation of modern inspired optimization
algorithms called RUN for PV cells/modules parameter estimation. To validate and verify the main findings,
the proposed algorithm was implemented and tested on two commercial cells/modules. Moreover, a
comprehension comparison of the latest meta-heuristic optimization algorithms was illustrated. The attained
theoretical and experimental results are coincident, which proves the superiority of RUN in the field of
parameter estimation for PV cells/modules. The results signify the effectiveness and reliability of the
proposed RUN in estimating the accurate double diode model of two practical PV cells/modules. The RUN
realizes steady convergence rates than other competitive algorithms. Finally, the main findings of this article
will pave the way for the authors as well as the researchers to evaluate the impact of the estimated parameters
in emulating PV equivalent circuits under various real operating conditions.
REFERENCES
[1] IRENA, Renewable capacity statistics. International Renewable Energy Agency (IRENA), 2021.
SFO
SSA
GW-CS
RCGA
RUN

1197
[2] NREA, “Strategic environmental and social assessment of wind energy projects in the East Nile Region (Arab Republic of
Egypt),” in "Non of the technical summary final draft (NTS) SESA wind report," New and Renewable Energy Authority (NREA)
Ministry of Electricity and Renewable Energy, Egypt, 2018.
[3] A. R. Jordehi, “Parameter estimation of solar photovoltaic (PV) cells: A review,” Renewable and Sustainable Energy Reviews,
vol. 61, pp. 354–371, Aug. 2016, doi: 10.1016/j.rser.2016.03.049.
[4] Z. Salam, K. Ishaque, and H. Taheri, “An improved two-diode photovoltaic (PV) model for PV system,” in 2010 Joint
International Conference on Power Electronics, Drives and Energy Systems & 2010 Power India, Dec. 2010, pp. 1–5, doi:
10.1109/PEDES.2010.5712374.
[5] K. Ishaque, Z. Salam, and H. Taheri, “Accurate MATLAB simulink PV system simulator based on a two-diode model,” Journal
of Power Electronics, vol. 11, no. 2, pp. 179–187, Mar. 2011, doi: 10.6113/JPE.2011.11.2.179.
[6] M. Ćalasan, S. H. E. A. Aleem, and A. F. Zobaa, “On the root mean square error (RMSE) calculation for parameter estimation of
photovoltaic models: A novel exact analytical solution based on Lambert W function,” Energy Conversion and Management,
vol. 210, Apr. 2020, Art. no. 112716, doi: 10.1016/j.enconman.2020.112716.
[7] F. Ayala-Mató, D. Seuret-Jiménez, J. J. Escobedo-Alatorre, O. Vigil-Galán, and M. Courel, “A hybrid method for solar cell
parameter estimation,” Journal of Renewable and Sustainable Energy, vol. 9, no. 6, Nov. 2017, Art. no. 063504, doi:
10.1063/1.4985953.
[8] Y. Chaibi, M. Salhi, A. El-jouni, and A. Essadki, “A new method to extract the equivalent circuit parameters of a photovoltaic
panel,” Solar Energy, vol. 163, pp. 376–386, Mar. 2018, doi: 10.1016/j.solener.2018.02.017.
[9] A. Şentürk, “New method for computing single diode model parameters of photovoltaic modules,” Renewable Energy, vol. 128,
pp. 30–36, Dec. 2018, doi: 10.1016/j.renene.2018.05.065.
[10] A. Ayang et al., “Maximum likelihood parameters estimation of single-diode model of photovoltaic generator,” Renewable
Energy, vol. 130, pp. 111–121, Jan. 2019, doi: 10.1016/j.renene.2018.06.039.
[11] F. J. Toledo, J. M. Blanes, and V. Galiano, “Two-step linear least-squares method for photovoltaic single-diode model parameters
extraction,” IEEE Transactions on Industrial Electronics, vol. 65, no. 8, pp. 6301–6308, Aug. 2018, doi:
10.1109/TIE.2018.2793216.
[12] H. El Achouby, M. Zaimi, A. Ibral, and E. M. Assaid, “New analytical approach for modelling effects of temperature and
irradiance on physical parameters of photovoltaic solar module,” Energy Conversion and Management, vol. 177, pp. 258–271,
Dec. 2018, doi: 10.1016/j.enconman.2018.09.054.
[13] E. Cuce, P. M. Cuce, I. H. Karakas, and T. Bali, “An accurate model for photovoltaic (PV) modules to determine electrical
characteristics and thermodynamic performance parameters,” Energy Conversion and Management, vol. 146, pp. 205–216, Aug.
2017, doi: 10.1016/j.enconman.2017.05.022.
[14] D. Allam, D. A. Yousri, and M. B. Eteiba, “Parameters extraction of the three diode model for the multi-crystalline solar
cell/module using Moth-Flame Optimization Algorithm,” Energy Conversion and Management, vol. 123, pp. 535–548, Sep.
2016, doi: 10.1016/j.enconman.2016.06.052.
[15] M. Louzazni, A. Khouya, K. Amechnoue, M. Mussetta, and A. Crăciunescu, “Comparison and evaluation of statistical criteria in
the solar cell and photovoltaic module parameters’ extraction,” International Journal of Ambient Energy, vol. 41, no. 13, pp.
1482–1494, Nov. 2020, doi: 10.1080/01430750.2018.1517678.
[16] H. M. Ridha, C. Gomes, and H. Hizam, “Estimation of photovoltaic module model’s parameters using an improved
electromagnetic-like algorithm,” Neural Computing and Applications, vol. 32, no. 16, pp. 12627–12642, Aug. 2020, doi:
10.1007/s00521-020-04714-z.
[17] F. Dkhichi, B. Oukarfi, A. Fakkar, and N. Belbounaguia, “Parameter identification of solar cell model using Levenberg–
Marquardt algorithm combined with simulated annealing,” Solar Energy, vol. 110, pp. 781–788, Dec. 2014, doi:
10.1016/j.solener.2014.09.033.
[18] X. Chen and K. Yu, “Hybridizing cuckoo search algorithm with biogeography-based optimization for estimating photovoltaic
model parameters,” Solar Energy, vol. 180, pp. 192–206, Mar. 2019, doi: 10.1016/j.solener.2019.01.025.
[19] A. M. Beigi and A. Maroosi, “Parameter identification for solar cells and module using a hybrid firefly and pattern search
algorithms,” Solar Energy, vol. 171, pp. 435–446, Sep. 2018, doi: 10.1016/j.solener.2018.06.092.
[20] R. A. P. Franco, G. L. Filho, and F. H. T. Vieira, “Firefly algorithm applied to the estimation of the parameters of a photovoltaic
panel model,” Springer International Publishing, 2019, pp. 101–134.
[21] K. G. and Harish Kumar, “Modeling of solar cell under different conditions by Ant Lion optimizer with LambertW function,”
Applied Soft Computing, vol. 71, pp. 141–151, Oct. 2018, doi: 10.1016/j.asoc.2018.06.025.
[22] K. Yu, B. Qu, C. Yue, S. Ge, X. Chen, and J. Liang, “A performance-guided JAYA algorithm for parameters identification of
photovoltaic cell and module,” Applied Energy, vol. 237, pp. 241–257, Mar. 2019, doi: 10.1016/j.apenergy.2019.01.008.
[23] R. Abbassi, A. Abbassi, A. A. Heidari, and S. Mirjalili, “An efficient salp swarm-inspired algorithm for parameters identification
of photovoltaic cell models,” Energy Conversion and Management, vol. 179, pp. 362–372, Jan. 2019, doi:
10.1016/j.enconman.2018.10.069.
[24] A. A. Zaky et al., “Optimal performance emulation of PSCs using the elephant herd algorithm associated with experimental
validation,” ECS Journal of Solid State Science and Technology, vol. 8, no. 12, pp. Q249–Q255, Dec. 2019, doi:
10.1149/2.0271912jss.
[25] H. Chen, S. Jiao, A. A. Heidari, M. Wang, X. Chen, and X. Zhao, “An opposition-based sine cosine approach with local search
for parameter estimation of photovoltaic models,” Energy Conversion and Management, vol. 195, pp. 927–942, Sep. 2019, doi:
10.1016/j.enconman.2019.05.057.
[26] D. Kler, Y. Goswami, K. P. S. Rana, and V. Kumar, “A novel approach to parameter estimation of photovoltaic systems using
hybridized optimizer,” Energy Conversion and Management, vol. 187, pp. 486–511, May 2019, doi:
10.1016/j.enconman.2019.01.102.
[27] O. Hachana, K. E. Hemsas, G. M. Tina, and C. Ventura, “Comparison of different metaheuristic algorithms for parameter
identification of photovoltaic cell/module,” Journal of Renewable and Sustainable Energy, vol. 5, no. 5, Sep. 2013, Art. no.
053122, doi: 10.1063/1.4822054.
[28] Z. Chen, L. Wu, P. Lin, Y. Wu, and S. Cheng, “Parameters identification of photovoltaic models using hybrid adaptive Nelder-
Mead simplex algorithm based on eagle strategy,” Applied Energy, vol. 182, pp. 47–57, Nov. 2016, doi:
10.1016/j.apenergy.2016.08.083.
[29] X. Yuan, Y. Xiang, and Y. He, “Parameter extraction of solar cell models using mutative-scale parallel chaos optimization
algorithm,” Solar Energy, vol. 108, pp. 238–251, Oct. 2014, doi: 10.1016/j.solener.2014.07.013.
[30] J. Liang et al., “Classified perturbation mutation based particle swarm optimization algorithm for parameters extraction of
photovoltaic models,” Energy Conversion and Management, vol. 203, Jan. 2020, Art. no. 112138, doi:

 ISSN: 2088-8708
1198
10.1016/j.enconman.2019.112138.
[31] D. Yousri, D. Allam, M. B. Eteiba, and P. N. Suganthan, “Static and dynamic photovoltaic models’ parameters identification
using chaotic heterogeneous comprehensive learning particle swarm optimizer variants,” Energy Conversion and Management,
vol. 182, pp. 546–563, Feb. 2019, doi: 10.1016/j.enconman.2018.12.022.
[32] X. Gao et al., “Parameter extraction of solar cell models using improved shuffled complex evolution algorithm,” Energy
Conversion and Management, vol. 157, pp. 460–479, Feb. 2018, doi: 10.1016/j.enconman.2017.12.033.
[33] A. M. Shaheen, A. R. Ginidi, R. A. El-Sehiemy, and S. S. M. Ghoneim, “A forensic-based investigation algorithm for parameter
extraction of solar cell models,” IEEE Access, vol. 9, pp. 1–20, 2021, doi: 10.1109/ACCESS.2020.3046536.
[34] M. Said et al., “Estimating parameters of photovoltaic models using accurate turbulent flow of water optimizer,” Processes,
vol. 9, no. 4, Apr. 2021, Art. no. 627, doi: 10.3390/pr9040627.
[35] M. A. Elhosseini, R. A. El Sehiemy, Y. I. Rashwan, and X. Z. Gao, “On the performance improvement of elephant herding
optimization algorithm,” Knowledge-Based Systems, vol. 166, pp. 58–70, Feb. 2019, doi: 10.1016/j.knosys.2018.12.012.
[36] A. S. Bayoumi, R. A. El-Sehiemy, K. Mahmoud, M. Lehtonen, and M. M. F. Darwish, “Assessment of an improved three-diode
against modified two-diode patterns of MCS solar cells associated with soft parameter estimation paradigms,” Applied Sciences,
vol. 11, no. 3, Jan. 2021, Art. no. 1055, doi: 10.3390/app11031055.
[37] R. Chenouard and R. A. El-Sehiemy, “An interval branch and bound global optimization algorithm for parameter estimation of
three photovoltaic models,” Energy Conversion and Management, vol. 205, Feb. 2020, Art. no. 112400, doi:
10.1016/j.enconman.2019.112400.
[38] P. J. Gnetchejo, S. Ndjakomo Essiane, P. Ele, R. Wamkeue, D. M. Wapet, and S. P. Ngoffe, “Important notes on parameter
estimation of solar photovoltaic cell,” Energy Conversion and Management, vol. 197, Oct. 2019, Art. no. 111870, doi:
10.1016/j.enconman.2019.111870.
[39] I. Ahmadianfar, A. A. Heidari, A. H. Gandomi, X. Chu, and H. Chen, “RUN beyond the metaphor: An efficient optimization
algorithm based on Runge Kutta method,” Expert Systems with Applications, vol. 181, Nov. 2021, Art. no. 115079, doi:
10.1016/j.eswa.2021.115079.
[40] T. Easwarakhanthan, J. Bottin, I. Bouhouch, and C. Boutrit, “Nonlinear minimization algorithm for determining the solar cell
parameters with microcomputers,” International Journal of Solar Energy, vol. 4, no. 1, pp. 1–12, Jan. 1986, doi:
10.1080/01425918608909835.
[41] K. Nishioka, N. Sakitani, Y. Uraoka, and T. Fuyuki, “Analysis of multicrystalline silicon solar cells by modified 3-diode
equivalent circuit model taking leakage current through periphery into consideration,” Solar Energy Materials and Solar Cells,
vol. 91, no. 13, pp. 1222–1227, Aug. 2007, doi: 10.1016/j.solmat.2007.04.009.

Identification study of solar cell/module using recent optimization techniques

More Related Content

Similar to Identification study of solar cell/module using recent optimization techniques (20)

More from IJECEIAES (20)

Recently uploaded (20)

Identification study of solar cell/module using recent optimization techniques