Next Article in Journal
Mechanism Analysis of Sympathetic Inrush in Traction Network Cascaded Transformers Based on Flux-Current Circuit Model
Next Article in Special Issue
A Novel MPPT Technique for Single Stage Grid-Connected PV Systems: T4S
Previous Article in Journal
Research on Overburden Movement Characteristics of Large Mining Height Working Face in Shallow Buried Thin Bedrock
Previous Article in Special Issue
Parameters Extraction of Photovoltaic Models Using an Improved Moth-Flame Optimization
Open AccessArticle

Estimation of Single-Diode and Two-Diode Solar Cell Parameters by Using a Chaotic Optimization Approach

1
Faculty of Electrical Engineering, University of Montenegro, Džordža Vašingtona bb, 81000 Podgorica, Montenegro
2
Montenegrin Electrical-energy distribution company—CEDIS, 81000 Podgorica, Montenegro
*
Author to whom correspondence should be addressed.
Energies 2019, 12(21), 4209; https://doi.org/10.3390/en12214209
Received: 3 October 2019 / Revised: 21 October 2019 / Accepted: 29 October 2019 / Published: 4 November 2019

Abstract

Estimation of single-diode and two-diode solar cell parameters by using chaotic optimization approach (COA) is addressed. The proposed approach is based on the use of experimentally determined current-voltage (I-V) characteristics. It outperforms a large number of other techniques in terms of average error between the measured and the estimated I-V values, as well as of time complexity. Implementation of the proposed approach on the I-V curves measured in laboratory environment for different values of solar irradiation and temperature prove its applicability in terms of accuracy, effectiveness and the ease of implementation for a wide range of practical environment conditions. The COA-based parameter estimation is, therefore, useful for PV power converter designers who require fast and accurate model for PV cell/module.
Keywords: Solar cell parameters; single-diode model; two-diode model; COA Solar cell parameters; single-diode model; two-diode model; COA

1. Introduction

The contribution of solar energy in total electric energy production is growing constantly. As the price of solar inverters and solar panels constantly decreases, most countries are basing their energy policy on higher use of solar energy. Studies on energy networks, and especially testing of the integration of solar energy sources into power networks, requires accurate calculation of the solar output power, as well as accurate modeling of solar cells. For that reason, modeling of solar cells (corresponding equivalent circuit and accurate parameters value) represents a very popular research field.
In the literature, two basic models of the equivalent circuits of solar cell can be found, namely the single-diode model (SDM) [1] and the double-diode model (DDM) [2]. DDM considers the composite effect of the neutral region of the junction, and, therefore, models the solar cells more accurately [3]. However, it is characterized by seven unknown parameters. Because of the complexity of DDM, some authors reduce the number of unknown parameters [3,4], which can greatly affect the model accuracy [5]. In this work, we focus on both SDM and DDM, without neglecting any of the model parameters of the solar cell.
For the estimation of solar cell parameters, two main sets of “input data” and corresponding estimations can be found, namely
(a)
estimation based on datasheet information [6,7] and
(b)
estimation based on experimental data [8].
The former uses the datasheet information (open circuit voltage, short circuit current, voltage and current value at maximum power point characteristics) provided by photovoltaic (PV) manufacturers under standard test conditions. However, recent research [7] on the usage of datasheet values for solar cell parameter estimation shows that current-voltage characteristic is not unique when designers focus on three datasheet points (open circuit, short circuit and maximum power). It is shown that by observing only three points, we can have multiple I-V characteristics, although in reality a solar cell has a single defined I-V characteristic corresponding to a specific set of cell parameters. To obtain unique and accurate I-V characteristic of PV cell, experimental data on more than three major points are necessary [7]. Research [7] has also implied that only approaches based on experimental data generate accurate models.
Evaluating the performance of solar cells (or PV panels) requires as accurate an estimation of the equivalent circuit solar cell parameters as possible. The approaches used for this purpose can be categorized as follows:
(a)
analytical techniques [9,10,11,12,13],
(b)
numerical extraction [13] and
(c)
meta-heuristic techniques [14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59].
Analytical techniques provide mathematical expressions for solving equivalent circuit parameters based on some input data (manufacturer data or/and data obtained from measurements). A review and comparative assessment of non-iterative methods for the extraction of the single-diode model parameters of photovoltaic modules is given in [11]. In general, analytical techniques provide rapid solution. On the other hand, these techniques give erroneous results when the estimated and measure solar cell output characteristics are compared [12].
Numerical techniques are based on curve fitting, usually via iterative methods. However, the application of curve fitting to nonlinear diode equations is quite complex, making numerical determination of solar cell parameters unpopular [14].
Recently, meta-heuristic algorithms for solar cell parameter estimation have been proposed [14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58]. They impose no restrictions on the problem formulation, they are excellent in dealing with nonlinear equations, and they can be applied for different numbers of unknown parameters.
Of all the proposed techniques, none excels in terms of accuracy and efficiency with respect to others. This was our main incentive for doing research in this field. We propose both accurate and efficient parameter optimization of solar cell SDM and DDM through chaotic optimization approach (COA).
Recently, COA has been used in solving various optimization problems: parameter identification of Jiles-Atherton hysteresis model [59], single-phase transformer parameter estimation [60], design of PID parameters for automatic voltage regulation of synchronous machine [61], antenna array radiation pattern synthesis [62,63]. The main advantages of COA over other optimization techniques are easy implementation and short execution time [64]. It should be noted that different versions of chaotic algorithm have also been used in solar cell parameters estimation, namely chaotic heterogeneous comprehensive learning particle swarm optimizer variants [16], chaotic asexual reproduction optimization [33], mutative-scale parallel chaos optimization algorithm [41], chaos-embedded gravitational search algorithm [65], chaotic improved artificial bee colony algorithm [66], improved chaotic whale optimization algorithm [67], etc. Unlike methods proposed in [16,33,41,65,66,67], this paper will use COA based on Logistic map for solar cell parameter estimation. Logistic map has a simple form with one variable and one control parameter, and it can produce chaotic behavior similar to more complex chaotic systems. [68]. The power of this optimization approach is demonstrated in [60], where it was used for the estimations of the transformer’s parameters.
The effectiveness of the proposed approach will be evaluated on different solar cells (different with respect to solar cell voltage and current level) found in the literature and the laboratory environment. Furthermore, COA-based parameter estimation will be compared with 50 various literature techniques for SDM, and with 12 various techniques for DDM. Also, we will compare parameters obtained by using the proposed method on the measured data with analytically and numerically obtained parameters. Finally, we will apply COA on the measured I-V characteristics using the laboratory environment.
The paper is organized as follows. SDM and DDM for solar cells are described in Section 2. In Section 3, COA and its implementation for solar cell parameters estimation are described. The comparison of solar cell parameters estimation accuracy obtained by using the COA-based and other methods, for one solar cell and one solar module, is presented in Section 4. The experimental setup for measuring I-V curves is presented in Section 5, along with the COA-based parameter estimation results. The concluding remarks are given in Section 6.

2. Mathematical Modeling of Single and Double Solar Cells

SDM is commonly used model for solar cell representation [1], and the equivalent circuit is shown in Figure 1a. The I-V relationship for this model can be described by the following equation:
I = I p v I o ( e V + I R s n V t h 1 ) V + I R s R p
where Ipv is the photo-generated current, Rs the series parasitic resistance, Rp the parallel parasitic resistance, I0 the saturation current, n is ideality factor and Vth = kBT/q is the thermal voltage (kB is Boltzmann constant equal to 1.38 × 10−23 J/K, T the temperature and q the electron charge equal to 1.602 × 10−19 C).
The equivalent circuit with DDM for the solar cell is shown in Figure 1b. Therefore, unlike the SDM model, the DDM model of the solar cell, in addition to the rectifying diode, includes one more diode to consider the space charge recombination current [48]. The I-V characteristic of DDM is given as
I = I p v I o 1 ( e V + I R s n 1 V t h 1 ) I o 2 ( e V + I R s n 2 V t h 1 ) V + I R s R p
where Io1 and Io2 are the diffusion and saturation currents, whereas n1 and n2 are the diffusion and recombination diode ideality factors [48]. The ideality factor is discussed in [69,70], whereas [71] presents a method for ideality factor calculation.

3. COA and Objective Function

COA is a very powerful optimization technique that has found numerous scientific applications [59,60,61,62,63]. This approach is based on the theory of chaos, which is, in a mathematical sense, described by ordinary differential equations or by an iterative map [64].
Different chaotic systems, including the logistic map, lozi map, tent map and Lorenz system, can be found in the literature. In this paper, we will base COA on the logistic map [59,60,61,62,63,64].
The task of COA is to estimate a set of unknown parameters X which minimizes the objective function (OF). In our case, for SDM, X = [Rs, Rp, Ipv, Io, n], and for DDM, X = [Rs, Rp, Ipv, Io1, Io2, n1, n2]. Therefore, in general, vector X = [x1, x2, ... xn] contains variables limited to the lower (LV) and upper (UV) permitted value, i.e, x i [ L i , U i ] . On the other side, the OF for SDM is
O F = t = 1 P ( I p v I o ( e V t + I t R s n V t h 1 ) V t + I t R s R p I t )
whereas for DDM it reads
O F = t = 1 P ( I p v I o 1 ( e V t + I t R s n 1 V t h 1 ) I o 2 ( e V t + I t R s n 2 V t h 1 ) V t + I t R s R p I t )
where P is the number of measured I-V pairs from the I-V characteristics, and Vt and It represent the voltage and current value of pair t.
Figure 2 presents the search procedure, i.e., the COA flowchart. The detailed description of COA flowchart can be found in [59].
In this paper, the following COA parameters were used: M = 1000, N = 50,000. The COA-based estimation is compared with other approaches through the root mean square error (RMSE), defined as follows:
R M S E = k = 1 P ( I e s t , k I m e a s , k ) P
where Iest,k and Imeas,k represent the estimated and the measured values of solar output current in point k, respectively.

4. Simulation Results

To evaluate COA for solar cell parameters estimation, we first applied the proposed method to an experimental current-voltage characteristic extracted from the manufacturer’s datasheets of a well-known R.T.C. France solar cell operating under standard test conditions.
The values of parameters obtained by using COA for the R.T.C. France solar cell are summarized, by year of publication, in Table 1, for SDM and Table 2 for DDM. These values are compared with the values of parameters published in recent papers (column Reference) for the same experimental data. During the estimation process, the parameter ranges for SDM estimation were R s ( Ω ) [ 0.02 ,   0.05 ] , I p v ( A ) [ 0.74 ,   0.78 ] , I o ( μ A ) [ 0.2 ,   0.4 ] , R p ( Ω ) [ 50 ,   55 ] and n [ 1.35 ,   1.6 ] , whereas for DDM, they were R s ( Ω ) [ 0.02 ,   0.04 ] , R p ( Ω ) [ 54 ,   58 ] , n 1 [ 1.4 ,   1.5 ] , n 2 [ 1.95 ,   2 ] , I p v ( A ) [ 0.75 ,   0.77 ] I o 1 ( μ A ) [ 0.2 ,   0.25 ] and I o 2 ( μ A ) [ 0.7 ,   0.8 ] .
Table 1 and Table 2 report parameters as they appear in the cited papers with no modification. However, in some papers in the Energy Conversion and Management journal (in Table 1 marked by *), inaccuracies occurred in parameter estimation of the PV cell using metaheuristic techniques. Namely, the results proposed in [15,16,17,18,21] do not correspond to the objective function [56].
The presented results, especially the value of RMSE, show that COA offers solar characteristics closer to the measured characteristics than the other existing methods, i.e., it outperforms other methods in terms of accuracy. In addition, by observing Table 1 and Table 2, it is also evident that DDM characterizes solar cells more accurately than SDM, which supports the conclusion regarding DDM accuracy noted in [3].
It can be seen that COA outperforms several other techniques, such as evaporation rate-based water cycle algorithm (ER-WPA) [19] and cat swarm optimization (CSO) [24] for SDM, and with the generalized opposition-flower pollination algorithm-nelder-mead simplex method (GOFPANM) [53], by a small margin. However, the implementation of COA is simpler than implementation of ER-WPA, CSO and GOFPANM. Furthermore, COA is computationally less demanding than CSO since CSO requires changing the operation mode during the estimation process [24]. On the other side, GOFPANM is a hybrid algorithm which combines local and global search as well as different algorithms during estimation [53]. In general, most evolutionary algorithms have the complexity of O((np + Cof p)Ni), where O is the big O notation, n is the dimension of the parameter space, p is the population size, Ni is the number of iterations and Cof is the complexity of the OF. The complexity of COA is O(QCof), where Q is the number of points in the parameter space in which the OF is calculated. Therefore, the proposed COA-based estimation has significantly lower computational complexity than evolutionary algorithms.
To show the additional advantage of COA over other techniques, we conducted a comparison in terms of required time for one iteration. In that sense, in MATLAB 2015 (MathWorks, Natick, MA, USA) we have implemented the following algorithms for solar cell parameter estimation: evaporation rate-based water cycle algorithm (ER-WCA) [19], cuckoo search (CS) [46] and harmony search (HS) [48]. ER-WCA algorithm has a very good accuracy, very close to that obtained by the proposed method (see Table 1). On the other hand, HS and CS also have a good accuracy (~10−4). All computer simulations were carried out on a PC with Intel(R) Core (TM) i3-7020U CPU @ 2.30 GHz and 4 GB RAM. The obtained results, i.e., the mean, maximal and minimal required time per one iteration, obtained over 20 runs, are presented in Table 3. Clearly, the COA-based algorithm is the most efficient method, as it is characterized by the lowest value of required time per iteration. Note, in order to draw a fair comparison between the considered algorithms, MATLAB implementation follows the same rules for each algorithm (e.g., avoiding loops and using array operations such as dot product and matrix product whenever possible).
The measured I-V and P-V characteristics and the corresponding simulated characteristics, for parameters obtained by using COA, are shown in Figure 3. Very good agreement can be seen between the measured and estimated curves. Also, the difference between the DDM and SDM simulated curves is small but consistent and always in favor of DDM. In addition, in Table 4, we presented the estimated value of the unknown DDM parameters of the BPSolar MSX-60 module. These parameters are obtained by using COA as well as by using analytical, numerical, iteration and Newton methods presented in the literature. In the COA-based estimation, the ranges of parameters were R s ( Ω ) [ 0.2   0.4 ] , R p ( Ω ) [ 150 ,   300 ] , n 1 [ 0.5 ,   1.5 ] , I p v ( A ) [ 3.5 ,   4 ] , I o 1 ( A ) [ 10 10 ,   10 6 ] , I o 2 ( A ) [ 10 10 ,   10 6 ] , and n 2 [ 1.5 ,   2 ] . From the presented results, it is clear that COA outperforms the considered non-metaheuristic methods for solar cell parameters determination in terms of accuracy.
The measured and corresponding simulated I-V and P-V characteristics, for parameters obtained by using COA and other methods, are shown in Figure 4 and Figure 5, respectively. It is evident that COA outperforms the other methods in terms of approaching the measured characteristics.
Based on all of the presented results, it can be concluded that COA can precisely estimate the solar cell/module circuit parameters, outperforming the other metaheuristics as well as analytical or numerical methods in terms of estimation accuracy.

5. Experimental Results and Analysis

To check the applicability and efficiency of COA for solar cell parameter estimation, we also observed solar cells from the Clean Energy Trainer setup. The main motivation to use these solar modules is that this setup enables adjustable solar insolation, USB data monitoring for PC-supported data acquisition and analysis, as well as highly advanced didactic software for system control and real-time data plotting.
The observed system contains of:
  • two solar modules and one module: 4 solar cells, 400 mW, 2 V, 0.5 A,
  • TES 1333R data logging Solar power meter—instrument with range of 2000 W/m2, high resolution (0.1 W/m2), and wide spectral resolution (400–1100 nm), etc.
  • lamp—special double spotlight lamp that simulates sunlight. It provides the optimal light spectrum for the solar module.
  • USB Data Monitor—used for data acquisition. Also, it is connected to the computer and software through the USB port.
  • load—simulates electric consumer load.
  • software—designed to facilitate system control, parameter monitoring, data acquisition and graphical representation of the collected data.
The experimental setup, installed in Laboratory for Automatics, at the Faculty of Electrical engineering, University of Montenegro, is presented in Figure 6.
Firstly, we measured the I-V characteristics for insolation of 1285 W/m2 and temperature of 42 °C. For the measured I-V pairs, we determined single and double diode solar cell parameters (see Table 5). The parameter ranges for SDM estimation were R s ( Ω ) [ 0.1 ,   0.4 ] , I p v ( A ) [ 0.2 ,   0.4 ] , I o ( A ) [ 5 × 10 8 ,   15 × 10 8 ] , R p ( Ω ) [ 200 ,   600 ] and n [ 0.2 ,   1 ] , whereas for DDM were R s ( Ω ) [ 0.1 ,   0.4 ] , R p ( Ω ) [ 600 ,   900 ] , n 1 [ 0.2 ,   1 ] , n 2 [ 1.95 ,   2 ] , I p v ( A ) [ 0.2 ,   0.4 ] , I o 1 ( A ) [ 5 × 10 8 ,   15 × 10 8 ] , and I o 2 ( A ) [ 5 × 10 8 ,   15 × 10 8 ] . Then we measured the I-V and P-V characteristics for different values of insolation and temperature. The corresponding simulated characteristics were determined by taking into account the change of parameters with insolation and temperature (see [13]). The measured and estimated I-V and P-V characteristics for different values of insolation and temperature are presented in Figure 7, Figure 8, Figure 9 and Figure 10. The agreement between the measured and estimated characteristics is evident (see zoomed parts in these figures). Finally, we repeated the estimation procedure on all measured I-V characteristics. The estimated values of parameters were in range of ±4% of the initially observed, which confirms that we can use any of the measured characteristics for parameter estimation. On the other hand, by observing the data provided in Table 5, even for this module, it is evident that DDM is more accurate than SDM.

6. Conclusions

Modeling of solar cells is a very popular research direction, which is supported by numerous recent contributions in the literature. This paper proposes COA as a very successful approach for this purpose.
The proposed method is verified using practical data from various manufacturers. Its accuracy is confirmed by comparing its RMSE with numerous metaheuristics and non-metaheuristics methods for different solar cells. Experimental testing of COA applicability for parameter estimation is also implemented in laboratory environment. In all considered scenarios, a high level of accuracy is demonstrated. Apart from this, excellent matching of the simulated I-V and P-V curves with the measured characteristics additionally confirms the COA accuracy and its applicability for parameter estimation.
In future work, our attention will be focused on the usage of COA for estimation of solar cell parameters when solar cell output current is represented through the Lambert W function.

Author Contributions

Conceptualization, M.Ć. and S.Đ.; Methodology, M.Ć.; software, D.J. and V.R.; validation, V.R and M.Ć., formal analysis, S.M.; investigation, M.Ć., V.R and S.Đ.; resources, S.M.; writing—original draft preparation, M.Ć. and S.Đ.; writing—review and editing, S.Đ.; visualization, D.J.; supervision, M.Ć.

Funding

This work has been supported through European Union’s Horizon 2020 research and innovation program under project CROSSBOW-CROSS BOrder management of variable renewable energies and storage units enabling a transnational Wholesale market (Grant No. 773430).

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Moshksar, E.; Ghanbari, T. Adaptive Estimation Approach for Parameter Identification of Photovoltaic Modules. IEEE J. Photovolt. 2017, 7, 614–623. [Google Scholar] [CrossRef]
  2. Ali, H.; Mojgan, H.; Saad, M.; Hussein, H. Solar cell parameters extraction based on single and double-diode models: A review. Renew. Sustain. Energy Rev. 2016, 56, 494–509. [Google Scholar]
  3. Kashif, I.; Zainal, S.; Hamed, T. Simple, fast and accurate two-diode model for photovoltaic modules. Sol. Energy Mater. Sol. Cells 2011, 95, 586–594. [Google Scholar]
  4. Chitti, B.B.; Suresh, G. A novel simplified two-diode model of photovoltaic (PV) module. IEEE J. Photovolt. 2014, 4, 1156–1161. [Google Scholar]
  5. Chen, Y.; Sun, Y.; Meng, Z. An improved explicit double-diode model of solar cells: Fitness verification and parameter extraction. Energy Convers. Manag. 2018, 169, 345–358. [Google Scholar] [CrossRef]
  6. Laudani, A.; Fulginei, F.R.; Salvini, A. Identification of the one-diode model for photovoltaic modules from datasheet values. Sol. Energy 2014, 108, 432–446. [Google Scholar] [CrossRef]
  7. Biswas, P.P.; Suganthan, P.N.; Wu, G.; Amaratunga, G.A.J. Parameter estimation of solar cells using datasheet information with the application of an adaptive differential evolution algorithm. Renew. Energy 2019, 132, 425–438. [Google Scholar] [CrossRef]
  8. Laudani, A.; Riganti, F.F.; Salvini, A. High performing extraction procedure for the one-diode model of a photovoltaic panel from experimental I-V curves by using reduced forms. Sol. Energy 2014, 103, 316–326. [Google Scholar] [CrossRef]
  9. Chatterjee, A.; Keyhani, A.; Kapoor, D. Identification of photovoltaic source models. IEEE Trans. Energy Convers. 2011, 24, 883–889. [Google Scholar] [CrossRef]
  10. Shongwe, S.; Hanif, M. Comparative analysis of different single-diode PV modeling methods. IEEE J. Photovolt. 2015, 5, 938–946. [Google Scholar] [CrossRef]
  11. Batzelis, E. Non-Iterative Methods for the Extraction of the Single-Diode Model Parameters of Photovoltaic Modules: A Review and Comparative Assessment. Energies 2019, 12, 358. [Google Scholar] [CrossRef]
  12. Jordehi, A.R. Parameter estimation of solar photovoltaic (PV) cells: A. review. Renew. Sustain. Energy Rev. 2016, 61, 354–371. [Google Scholar] [CrossRef]
  13. Kumar, M.; Kumar, A. An efficient parameters extraction technique of photovoltaic models for performance assessment. Sol. Energy 2017, 158, 192–206. [Google Scholar] [CrossRef]
  14. Kashif, I.; Salam, Z.; Mekhilef, S.; Shamsudin, A. Parameter extraction of solar photo voltaic modules using penalty based differential evolution. Appl. Energy 2012, 99, 297–308. [Google Scholar]
  15. Dhruv, K.; Goswami, Y.; Kumar, R.V. A novel approach to parameter estimation of photovoltaic systems using hybridized optimizer. Energy Convers. Manag. 2019, 187, 486–511. [Google Scholar]
  16. Dalia, Y.; Dalia, A.; Eteiba, M.B.; Ponnuthurai, N.S. Static and dynamic photovoltaic models’ parameters identification using Chaotic Heterogeneous Comprehensive Learning Particle Swarm Optimizer variants. Energy Convers. Manag. 2019, 182, 546–563. [Google Scholar]
  17. Abd, E.M.; Diego, O. Parameter estimation of solar cells diode models by an improved opposition-based whale optimization algorithm. Energy Convers. Manag. 2018, 171, 1843–1859. [Google Scholar]
  18. Manel, M.; Anis, S.; Faouzi, M.M. Particle swarm optimisation with adaptive mutation strategy for photovoltaic solar cell/module parameter extraction. Energy Convers. Manag. 2018, 175, 151–163. [Google Scholar]
  19. Kler, D.; Sharma, P.; Banerjee, A.; Rana, V.; Kumar, K.P.S. PV cell and module efficient parameters estimation using evaporation rate-based water cycle algorithm. Swarm Evolut. Comput. 2017, 35, 93–110. [Google Scholar] [CrossRef]
  20. Lin, P.; Cheng, S.; Yeh, W.; Chen, Y.; Wu, L. Parameters extraction of solar cell models using a modified simplified swarm optimization algorithm. Sol. Energy 2017, 144, 594–603. [Google Scholar] [CrossRef]
  21. Ram, J.P.; Babu, T.S.; Dragicevic, T.; Rajasekar, N. A new hybrid bee pollinator flower pollination algorithm for solar PV parameter estimation. Energy Convers. Manag. 2017, 135, 463–476. [Google Scholar] [CrossRef]
  22. Fathy, A.; Rezk, H. Parameter estimation of photovoltaic system using imperialist competitive algorithm. Renew. Energy 2017, 111, 307–320. [Google Scholar] [CrossRef]
  23. Chen, X.; Yu, K.; Du, W.; Zhao, W.; Liu, G. Parameters identification of solar cell models using generalized oppositional teaching learning-based optimization. Energy 2016, 99, 170–180. [Google Scholar] [CrossRef]
  24. Guo, L.; Meng, Z.; Sun, Y.; Wang, L. Parameter identification and sensitivity analysis of solar cell models with cat swarm optimization algorithm. Energy Convers. Manag. 2016, 108, 520–528. [Google Scholar] [CrossRef]
  25. Hamid, N.F.A.; Rahim, N.A.; Selvaraj, J. Solar cell parameters identification using hybrid Nelder-Mead and modified particle swarm optimization. J. Renew. Sustain. Energy 2016, 8, 1–10. [Google Scholar] [CrossRef]
  26. Zhang, Y.; Lin, P.; Chen, Z.; Cheng, S. A population classification evolution algorithm for the parameter extraction of solar cell models. Int. J. Photoenergy 2016, 2016, 2174573. [Google Scholar] [CrossRef]
  27. Tong, N.T.; Pora, W. A parameter extraction technique exploiting intrinsic properties of solar cells. Appl. Energy 2016, 176, 104–115. [Google Scholar] [CrossRef]
  28. Jamadi, M.; Merrikh-Bayat, F.; Bigdeli, M. Very accurate parameter estimation of single- and double-diode solar cell models using a modified artificial bee colony algorithm. Int. J. Energy Environ. Eng. 2016, 7, 13–25. [Google Scholar] [CrossRef]
  29. Ali, E.E.; El-Hameed, M.A.; El-Fergany, A.A.; El-Arini, M.M. Parameter extraction of photovoltaic generating units using multi-verse optimizer. Sustain. Energy Technol. Assess. 2016, 17, 68–76. [Google Scholar] [CrossRef]
  30. Chellaswamy, C.; Ramesh, R. Parameter extraction of solar cell models based on adaptive differential evolution algorithm. Renew. Energy 2016, 97, 823–837. [Google Scholar] [CrossRef]
  31. Jordehi, A.R. Time varying acceleration coefficients particle swarm optimisation (TVACPSO): A new optimisation algorithm for estimating parameters of PV cells and modules. Energy Convers. Manag. 2016, 129, 262–274. [Google Scholar] [CrossRef]
  32. Ma, J.; Man, K.L.; Guan, S.U.; Ting, T.O.; Wong, P.W. Parameter estimation of photovoltaic model via parallel particle swarm optimization algorithm. Int. J. Energy Res. 2016, 40, 343–352. [Google Scholar] [CrossRef]
  33. Yuan, X.; He, Y.; Liu, L. Parameter extraction of solar cell models using chaotic asexual reproduction optimization. Neural Comput. Appl. 2015, 26, 227–239. [Google Scholar] [CrossRef]
  34. Lim, L.H.I.; Ye, Z.; Ye, J.; Yang, D.; Du, H. A linear identification of diode models from single I-V characteristics of PV panels. IEEE Trans. Ind. Electron. 2015, 62, 4181–4193. [Google Scholar] [CrossRef]
  35. El-Fergany, A. Efficient tool to characterize photovoltaic generating systems using mine blast algorithm. Electr. Power Compon. Syst. 2015, 43, 890–901. [Google Scholar] [CrossRef]
  36. Alam, D.F.; Yousri, D.A.; Eteiba, M.B. Flower pollination algorithm based solar PV parameter estimation. Energy Convers. Manag. 2015, 101, 410–422. [Google Scholar] [CrossRef]
  37. Dkhichi, F.; Oukarfi, B.; Fakkar, A.; Belbounaguia, N. Parameter identification of solar cell model using Levenberg–Marquardt algorithm combined with simulated annealing. Sol. Energy 2014, 110, 781–788. [Google Scholar] [CrossRef]
  38. Niu, Q.; Zhang, L.; Li, K. A biogeography-based optimization algorithm with mutation strategies for model parameter estimation of solar and fuel cells. Energy Convers. Manag. 2014, 86, 1173–1185. [Google Scholar] [CrossRef]
  39. Niu, Q.; Zhang, H.; Li, K. An improved TLBO with elite strategy for parameters identification of PEM fuel cell and solar cell models. Int. J. Hydrogen Energy 2014, 39, 3837–3854. [Google Scholar] [CrossRef]
  40. Oliva, D.; Cuevas, E.; Pajares, G. Parameter identification of solar cells using artificial bee colony optimization. Energy 2014, 72, 93–102. [Google Scholar] [CrossRef]
  41. Yuan, X.; Xiang, Y.; He, Y. Parameter extraction of solar cell models using mutative-scale parallel chaos optimization algorithm. Sol. Energy 2014, 108, 238–251. [Google Scholar] [CrossRef]
  42. Patel, S.J.; Panchal, A.K.; Kheraj, V. Extraction of solar cell parameters from a single current–voltage characteristic using teaching learning-based optimization algorithm. Appl. Energy 2014, 119, 384–393. [Google Scholar] [CrossRef]
  43. Askarzadeh, A.; Rezazadeh, A. Extraction of maximum power point in solar cells using bird mating optimizer-based parameters identification approach. Sol. Energy 2013, 90, 123–133. [Google Scholar] [CrossRef]
  44. Askarzadeh, A.; Rezazadeh, A. Artificial bee swarm optimization algorithm for parameters identification of solar cell models. Appl. Energy 2013, 102, 943–949. [Google Scholar] [CrossRef]
  45. Jiang, L.L.; Maskell, D.L.; Patra, J.C. Parameter estimation of solar cells and modules using an improved adaptive differential evolution algorithm. Appl. Energy 2013, 112, 185–193. [Google Scholar] [CrossRef]
  46. Ma, J.; Ting, T.O.; Man, K.L.; Zhang, N.; Guan, S.-U.; Wong, P.W.H. Parameter estimation of photovoltaic models via cuckoo search. J. Appl. Math. 2013, 2013, 362619. [Google Scholar] [CrossRef]
  47. Hachana, O.; Hemsas, K.E.; Tina, G.M.; Ventura, C. Comparison of different metaheuristic algorithms for parameter identification of photovoltaic cell/module. J. Renew. Sustain. Energy 2013, 5, 053112. [Google Scholar] [CrossRef]
  48. Askarzadeh, A.; Rezazadeh, A. Parameter identification for solar cell models using harmony search-based algorithms. Sol. Energy 2012, 86, 3241–3249. [Google Scholar] [CrossRef]
  49. AlHajri, M.F.; El-Naggar, K.M.; AlRashidi, M.R.; Al-Othman, A.K. Optimal extraction of solar cell parameters using pattern search. Renew. Energy 2012, 44, 238–245. [Google Scholar] [CrossRef]
  50. El-Naggar, K.M.; AlRashidi, M.R.; AlHajri, M.F.; Al-Othman, A.K. Simulated annealing algorithm for photovoltaic parameters identification. Sol. Energy 2012, 86, 266–274. [Google Scholar] [CrossRef]
  51. AlRashidi, M.R.; AlHajri, M.F.; El-Naggar, K.M.; Al-Othman, A.K. A new estimation approach for determining the I-V characteristics of solar cells. Sol. Energy 2011, 85, 1543–1550. [Google Scholar] [CrossRef]
  52. Ye, M.; Wang, X.; Xu, Y. Parameter extraction of solar cells using particle swarm optimization. J. Appl. Phys. 2009, 105, 1–8. [Google Scholar] [CrossRef]
  53. Shuhui, X.; Yong, W. Parameter estimation of photovoltaic modules using a hybrid flower pollination algorithm. Energy Convers. Manag. 2017, 144, 53–68. [Google Scholar]
  54. Kunjie, Y.; Xu, C.; Xin, W.; Zhenlei, W. Parameters identification of photovoltaic models using self-adaptive teaching-learning-based optimization. Energy Convers. Manag. 2017, 145, 233–246. [Google Scholar]
  55. Derick, M.; Rani, C.; Rajesh, M.; Farrag, M.E.; Wang, Y.; Busawon, K. An improved optimization technique for estimation of solar photovoltaic parameters. Sol. Energy 2017, 157, 116–124. [Google Scholar] [CrossRef]
  56. Gnetchejo, P.J.; Essiane, S.N.; Ele, P.; Wamkeue, R.; Wapet, D.M.; Ngoffe, S.P. Important notes on parameter estimation of solar photovoltaic cell. Energy Convers. Manag. 2019, 197, 111870. [Google Scholar] [CrossRef]
  57. Ishaque, K.; Salam, Z.; Taheri, H. Accurate MATLAB Simulink PV System Simulator Based on a Two-Diode Model. J. Power Electron. 2011, 11, 179–187. [Google Scholar] [CrossRef]
  58. Elbaset, A.A.; Ali, H.; Abd-El Sattar, M. Novel seven-parameter model for photovoltaic modules. Sol. Energy Mater. Sol. Cells 2014, 130, 442–455. [Google Scholar] [CrossRef]
  59. Rubežić, V.; Lazović, L.; Jovanović, A. Parameter identification of Jiles-Atherton model using the chaotic optimization method. Int. J. Comput. Math. Electr. Electron. Eng. 2018, 37, 2067–2080. [Google Scholar] [CrossRef]
  60. Ćalasan, M.; Mujičić, D.; Rubežić, V.; Radulović, M. Estimation of Equivalent Circuit Parameters of Single-Phase Transformer by Using Chaotic Optimization Approach. Energies 2019, 12, 1697. [Google Scholar] [CrossRef]
  61. Coelho, L.S. Tuning of PID controller for an automatic regulator voltage system using chaotic optimization approach. Chaos Solitons Fractals 2009, 39, 1504–1514. [Google Scholar] [CrossRef]
  62. Jovanović, A.; Lazović, L.; Rubežić, V. Adaptive Array Beamforming Using a Chaotic Beamforming Algorithm. Int. J. Antennas Propag. 2016, 2016, 8354204. [Google Scholar] [CrossRef]
  63. Jovanović, A.; Lazović, L.; Rubežić, V. Radiation pattern synthesis using a Chaotic beamforming algorithm. COMPEL Int. J. Comput. Math. Electr. Electron. Eng. 2016, 35, 1814–1829. [Google Scholar] [CrossRef]
  64. Yang, D.; Li, G.; Cheng, G. On the efficiency of chaos optimization algorithms for global optimization. Chaos Solitons Fractals 2007, 34, 1366–1375. [Google Scholar] [CrossRef]
  65. Valdivia-González, A.; Zaldívar, D.; Cuevas, E.; Pérez-Cisneros, M.; Fausto, F.; González, A. A Chaos-Embedded Gravitational Search Algorithm for the Identification of Electrical Parameters of Photovoltaic Cells. Energies 2017, 10, 1052. [Google Scholar] [CrossRef]
  66. Oliva, D.; Ewees, A.A.; Aziz, M.A.E.; Hassanien, A.E.; Peréz-Cisneros, M. A Chaotic Improved Artificial Bee Colony for Parameter Estimation of Photovoltaic Cells. Energies 2017, 10, 865. [Google Scholar] [CrossRef]
  67. Oliva, D.; Aziz, M.A.E.; Hassanien, A.E. Parameter estimation of photovoltaic cells using an improved chaotic whale optimization algorithm. Appl. Energy 2017, 200, 141–154. [Google Scholar] [CrossRef]
  68. Bertuglia, C.S.; Via, F. Nonlinearity, Chaos, and Complexity; Oxford University Press: Oxford, UK, 2005. [Google Scholar]
  69. Tress, W. Interpretation and evolution of open-circuit voltage, recombination, ideality factor and subgap defect states during reversible light-soaking and irreversible degradation of perovskite solar cells. Energy Environ. Sci. 2017. [Google Scholar] [CrossRef]
  70. Wetzelaer, G.A.H.; Kuik, M.; Lenes, M.; Blom, P.W.M. Origin of the dark-current ideality factor in polymer:fullerene bulk heterojunction solar cells. AIP Appl. Phys. Lett. 2011, 99, 153506. [Google Scholar] [CrossRef]
  71. Perovich, S.M.; Djukanovic, M.; Dlabac, T.; Nikolic, D.; Calasan, M. Concerning a novel mathematical approach to the solar cell junction ideality factor estimation. Appl. Math. Model. 2015, 39, 3248–3264. [Google Scholar] [CrossRef]
Figure 1. (a) SDM, and (b) DDM of a solar cell.
Figure 1. (a) SDM, and (b) DDM of a solar cell.
Energies 12 04209 g001
Figure 2. COA flowchart.
Figure 2. COA flowchart.
Energies 12 04209 g002
Figure 3. I-V and P-V characteristics of R.T.C. France solar cell.
Figure 3. I-V and P-V characteristics of R.T.C. France solar cell.
Energies 12 04209 g003
Figure 4. I-V characteristics of BPSolar MSX-60 module.
Figure 4. I-V characteristics of BPSolar MSX-60 module.
Energies 12 04209 g004
Figure 5. P-V characteristics of BPSolar MSX-60 module.
Figure 5. P-V characteristics of BPSolar MSX-60 module.
Energies 12 04209 g005
Figure 6. Experimental setup.
Figure 6. Experimental setup.
Energies 12 04209 g006
Figure 7. Current-voltage characteristics for two different insolation values and for temperature T = 42 °C.
Figure 7. Current-voltage characteristics for two different insolation values and for temperature T = 42 °C.
Energies 12 04209 g007
Figure 8. Power-voltage characteristics for two different insolation values and for temperature T = 42 °C.
Figure 8. Power-voltage characteristics for two different insolation values and for temperature T = 42 °C.
Energies 12 04209 g008
Figure 9. Current-voltage characteristics for two different temperatures and insolation values.
Figure 9. Current-voltage characteristics for two different temperatures and insolation values.
Energies 12 04209 g009
Figure 10. Power-voltage characteristics for two different temperatures and insolation values.
Figure 10. Power-voltage characteristics for two different temperatures and insolation values.
Energies 12 04209 g010
Table 1. Calculated SDM parameters for the R.T.C France solar cell.
Table 1. Calculated SDM parameters for the R.T.C France solar cell.
No.AlgorithmReferenceFirst Author, YearIpv (A)I0 (μA)nRs (Ω)Rp (Ω)RMSE
Proposed Method—COA0.76077450.32300181.48117740.036377553.739.860221 × 10−4
1.HISA *[15]Dhruv, 20190.7607078 0.31068459181.47726778 0.03654694 52.889794269.8911 × 10−4
2.HCLPSO *[16]Dalia, 20190.76079 0.31062 1.47710.036548 52.8851.12009 × 10−3
3.OBWOA *[17]Abd, 20180.76077 0.3232 1.5208 0.0363 53.68361.1417 × 10−3
4.MPSO *[18]Manel, 20180.760787 0.310683 1.475262 0.036546 52.889717.33007 × 10−3
5.ER-WCA[19]Kler D, 20170.7607760.3226991.4810800.03638153.691009.8609 × 10−4
6MSSO[20]Lin P, 20170.7607770.3235641.4812440.03637053.7424651.0599 × 10−3
7BPFPA *[21]Ram JP, 20170.76000.31061.47740.036657.71511.2536 × 10−3
8ICA [22]Fathy A, 20170.76030.146501.44210.038941.15771.1582 × 10−1
9GOTLBO[23]Chen X, 20160.7607800.3315521.4838200.03626554.1154269.8744 × 10−4
10CSO [24]Guo L, 20160.760780.32301.481180.0363853.71859.8612 × 10−4
11.NM-MPSO[25]Hamid N, 20160.760780.323061.481200.0363853.72229.8620 × 10−4
12.PCE [26]Zhang Y, 20160.7607760.3230211.4810740.03637753.7185251.0606 × 10−3
13.TONG [27]Tong NT, 20160.76100.36351.49350.0366062.5742.3859 × 10−3
14.MABC [28]Jamadi M, 20160.7607790.3213231.4813850.03638953.399992.7610 × 10−3
15.MVO [29]Ali EE, 20160.76160.320941.52520.036559.58841.2680 × 10−1
16.DET [30]Chellaswamy C, 20160.7510.3151.4870.03654.5322.4481 × 10−2
17.WCA [31]Jordehi AR, 20160.7609080.41355401.5043810.03536357.6694887.6069 × 10−3
18.TLBO 0.7608090.3122441.475780.03655152.84057.2723 × 10−3
19.GWO 0.7609960.24303881.4512190.03773245.1163097.2845 × 10−3
20.TVACPSO 0.7607880.31068271.4752580.03654752.8896447.3438 × 10−3
21.PPSO [32]Ma J, 20160.76080.32301.48120.036453.71859.9161 × 10−4
22.CARO [33]Yuan X, 20150.760790.317241.481680.0364453.08938.1969 × 10−3
23.LI [34]Lim LHI, 20150.76094380.34565721.487991690.0361423349.4822051.3462 × 10−3
24.MBA [35]El-Fergany A. 20150.76040.23481.48900.038844.611.1672 × 10−1
25.FPA *[36]Alam DF, 20150.760790.3106771.477070.036546652.87711.2121 × 10−3
26.LMSA [37]Dkhichi F, 20140.760780.318491.479760.0364353.326449.8649 × 10−4
27.DE [38]Niu Q, 20140.760680.355151.490800.0359856.55331.0035 × 10−3
28.BBO 0.760980.861001.587420.0321478.85552.3929 × 10−3
29.BBO-M 0.760780.318741.479840.0364253.362279.8656 × 10−4
30.STLBO [39]Niu Q, 20140.760780.323021.481140.0363853.71879.9763 × 10−4
31.TLBO 0.760740.323781.481360.0364154.40291.0016 × 10−3
32.ABC [40]Oliva D, 20140.76080.32511.48170.036453.64331.0967 × 10−3
33.HPEPD [8]Laudani A, 20140.76078840.31024821.47696410.0365530452.8590561.1487 × 10−3
34.MPCOA [41]Yuan X, 20140.760730.326551.481680.0363554.63282.3131 × 10−3
35.TLBO [42]Patel SJ, 20140.76080.32231.48370.036453.760279.6960 × 10−3
36.BMO [43]Askarzadeh A, 20130.760770.324791.481730.0363653.87169.8622 × 10−4
37.ABSO [44]0.760800.306231.475830.0365952.29039.9125 × 10−4
38.IADE [45]Jiang LL, 20130.76070.336131.48520.0362154.76439.9076 × 10−4
39.CS [46]Ma J, 20130.76080.3231.48120.036453.71859.9161 × 10−4
40.ABSO[47]Hachana O, 20130.760800.306231.479860.0365952.29031.4169 × 10−2
41.ABCDE0.760770.323021.479860.0363753.71854.8548 × 10−3
42.DE0.760770.323021.480590.0363753.71852.3423 × 10−3
43.MPSO0.760770.323021.470860.0363753.71853.9022 × 10−2
44.GGHS[48]Askarzadeh A, 20120.760920.326201.482170.0363153.06479.9089 × 10−4
45.HS0.760700.304951.475380.0366353.59469.9515 × 10−4
46.IGHS0.760770.343511.487400.0361353.28451.0335 × 10−3
47.PS[49]AlHajri MF, 20120.76170.99801.60000.031364.102561.4936 × 10−2
48.SA[50]El-Naggar KM, 20120.76200.47981.51720.034543.103451.8998 × 10−2
49.GA[51]AlRashidi MR, 20110.76190.80871.57510.029942.372881.9078 × 10−2
50.PSO [52]Ye M, 20090.7607980.3227211.483820.036394053.79659.6545 × 10−3
* for this method, a real RMSE are given [56].
Table 2. Calculated DDM parameters for the R.T.C France solar cell.
Table 2. Calculated DDM parameters for the R.T.C France solar cell.
No.AlgorithmRef.First Author, YearIpv(A)Io1(μA)Io2(μA)Rs(Ω)Rp(Ω)n1n2RMSE
Proposed Method—COA0.760781050.22597420.7493460.0367404355.48542361.4510167329.82484852 × 104
1.GOFPANM[53]X Shuhui, 20170.76078110.74934760.22597430.036740455.48544921.45101689.82485 × 10−4
2.SATLBO[54]Y Kunjie, 20170.760780.250930.5454180.0366355.1171.459821.999419.82941 × 10−4
3.MSSO[20]P Lin, 20170.7607480.2349250.6715930.03668855.7146621.4542551.9953051.059101 × 10−3
4.WDO[55]M Derick, 20170.76060.25310.04820.03743352.6608151.1621.384341.095213 × 10−3
5.CSO[24]L Guo, 20160.760780.227320.727850.03673755.38131.451511.997699.82532 × 10−4
6.GOTLBO[23] X Chen, 20160.7607520.8001950.2204620.03678356.07531.9999731.4489749.83152 × 10−4
7.PCE[26]Y Zhang, 20160.7607810.2260150.7493400.0367455.4831601.450.92329.8248 × 10−4
8.MABC[28]M Jamadi, 2016 0.76078210.241029920.63069220.0367121554.75500941.45685732.0000.5389.8276 × 10−4
9.FPA[36]DF Alam, 2015 0.7607950.3000880.1661590.036334252.34751.4747721.24239 × 10−3
10.BMO[43]A. Askarzadeh, 20130.760780.21110.876880.03682558.0811.445331.99.9979.82661 × 10−4
11.ABSO[44]A. Askarzadeh, 20130.730780.267130.381910.0365754.62191.465121.981529.8359 × 1004
12.IGHS[48]A. Askarzadeh, 20120.760790.973100.167910.0369056.83681.921261.428149.86572 × 10−4
Table 3. Time per iteration comparison.
Table 3. Time per iteration comparison.
AlgorithmMean Value of Requested Time (s)Maximal Value of Requested Time (s)Minimal Value of Requested Time (s)
COA0.0164160.0170230.015871
ER-WCA [19]0.0210630.0241450.019492
CS [46]0.0291790.0371770.027130
HS [48]0.0211030.0232640.020393
Table 4. Calculated DDM parameters for the BPSolar MSX-60 module.
Table 4. Calculated DDM parameters for the BPSolar MSX-60 module.
ParameterAnalytical Method [13]Numerical Method [13]Iteration Method [57]Newton Method [58]COA
Ipv (A)3.87523.80463.83.80843.8418
Io1 (A)3.6129 × 10−103.9901 × 10−104.704 × 10−104.8723 × 10−104.95821 × 10−8
Io2 (A)9.3773 × 10−64.033 × 10−64.704 × 10−106.1528 × 10−109.54961 × 10−9
Rs (Ω)0.30840.33970.350.36920.2495
Rp (Ω)280.6449280.2171176.4169.0471267.57
n110.9985911.00031.2569
n222.00141.21.99971.9345
RMSE0.03580.05170.12110.16360.0194
Table 5. Estimated value of experimentally tested solar module parameters.
Table 5. Estimated value of experimentally tested solar module parameters.
SDMDDM
Rs (Ω)0.2283Rs (Ω)0.2513
Rsh (Ω)439.55Rsh (Ω)782.9911
Io (A)10.56 × 10−8Io1 (A)6.8452 × 10−8
Ipv (A)0.2987n10.3342
n0.3441Ipv (A)0.2972
RMSE4.3418 × 10−4Io2 (A)6.0643 × 10−8
n21.9906
RMSE4.146 × 10−4
Back to TopTop