Modeling and Simulation of Photovoltaic Modules Using Bio-Inspired Algorithms
, Rafael Krummenauer and Cid Marcos Gonçalves Andrade *
Round 1
Reviewer 1 Report
The authors present a work on modelling and simulation of PV modules using bio-inspired algorithms. Although the paper is interesting, I am sorry to say that my recommendation is to reject it but to encourage the authors to submit a new version.
General Comments
The results are based on a single module and are therefore too narrow to validate the conclusions. It would be desirable to replicate the work with modules of other crystalline technologies.
It would be interesting to compare the curves obtained from the simulations with the parameters calculated with the IEC 60891 standard, at least for the best algorithm.
Specific Comments:
Introduction:
- The importance of PV is highlighted in lines 52-60. However, the most recent data possible should be presented.
- A brief outline of the structure of the article is useful at the end of the introduction.
Materials and methods:
- Even if the objective of this work is to extract the parameters using only the data provided in the PV module datasheet, a brief overview of photovoltaic technologies and the most suitable models for each would be interesting.
- Likewise, a description of the technology and structure of the module being worked on would be useful.
- The exposition of the bio-inspired algorithms is clear and sufficiently detailed.
Results
- With regard to the conditions of the simulations (line 390 onwards), although they may be convenient as a quality control of the method, they do not reflect the usual working conditions of a PV module. In particular, it is strange that NOMT (formerly known as NOCT) conditions are not used.
Conclusions
- The last sentence (lines 538-) should be refined. It is not clear to me whether this refers to the extraction of parameters or to the results of simulations that could be done with them.
There are some (minor) spelling mistakes, typos and expressions that should be corrected.
There are some (minor) spelling mistakes, typos and expressions that should be corrected.
Author Response
Reviewer #1
“General Comments: The results are based on a single module and are therefore too narrow to validate the conclusions. It would be desirable to replicate the work with modules of other crystalline technologies. It would be interesting to compare the curves obtained from the simulations with the parameters calculated with the IEC 60891 standard, at least for the best algorithm.”
Answer:
Dear Reviewer:
Thank you for reviewing our manuscript and for your valuable comments and suggestions to improve the manuscript quality. Based on your good suggestions, the manuscript has been carefully corrected. We are looking forward to receiving your positive responses.
With best wishes,
Cid Marcos G Andrade.
About the General Comments:
Answer:
A similar study was carried out for the CANADIAN CS-3W-440-P photovoltaic module composed of polycrystalline cells since the CANADIAN CS-3W-450-MS module presented in Section 3 uses monocrystalline technology. However, the authors chose to allocate this study to the supplementary materials section to make the text less repetitive. If this extra study of the polycrystalline modules would be added to the results, this Section would have many similar Figures and Tables that could confuse the reader.
The aforementioned supplementary study is presented below in the same manner as the results were presented in Section 3 of the manuscript:
“Complementary research of the module Canadian CS3W-440Wp P Poly-crystalline (CANADIAN, 2020).
|
Table S1. Parameters extracted by each method from the CANADIAN CS-3W-440-P module. |
|||||||
|
|
|||||||
|
PSO 1D5P |
7.3759*10-8 |
1.4539*10-2 |
1.4695 |
3478.6876 |
11.405 |
|
|
|
PSO 2D7P |
1.8627*10-7 |
4.0683*10-2 |
1.5449 |
4658.635 |
11.501 |
1.141*10-8 |
1.9779 |
|
GA 1D5P |
1.00*10-6 |
7.656*10-3 |
1.704 |
4999.98 |
11.539 |
|
|
|
GA 2D7P |
3.6455*10-7 |
1.689*10-2 |
1. 690 |
4999.99 |
11. 5638 |
1.00.10-6 |
1.7622 |
|
DE 1D5P |
4.6268*10-8 |
7.499*10-2 |
1. 43352 |
3076.919 |
11. 5493 |
|
|
|
DE 2D7P |
3.9558*10-7 |
2.2470*10-9 |
1. 8170 |
4199.99 |
11.6051 |
7.788*10-8 |
1. 48605 |
(a)
(b)Figure S1. Superposition of I x V curves generated by PSO, GA and DE for (a) 1D5P and (b) 2D7P under conditions G = 1000 W / m² and Tc = 25° C.
Table S2. Points of interest for the different methods under conditions G = 1000 W / m² and Tc = 25 ° C.
|
Method |
|||||
|
Data provided |
440.000 W |
10.920 A |
40.300 V |
11.400 A |
48.700 V |
|
by Manufacture |
|||||
|
PSO 1D5P |
440.8775 W |
10.675 A |
41.3 V |
11.405 A |
48.611 V |
|
e% |
0.20% |
2.24% |
2.48% |
<0.10% |
0.18% |
|
GA 1D5P |
435.3234 W |
10.770 A |
40.420 V |
11.539 A |
48.604 V |
|
e% |
1.06% |
1.37% |
0.30% |
1.22% |
0.20% |
|
DE 1D5P |
441.929 W |
10.803 A |
40.908 V |
11.540 A |
48.650 V |
|
e% |
0.44% |
1.07% |
1.51% |
1.23% |
0.10% |
|
PSO 2D7P |
438.247 W |
10.713 A |
40.908 V |
11.500 A |
48.488 V |
|
e% |
0.40% |
1.90% |
1.51% |
0.88% |
0.44% |
|
GA 2D7P |
433.879 W |
10.734 A |
40.421 V |
11.57 A |
48.486 V |
|
e% |
1.39% |
1.70% |
0.30% |
1.49% |
0.44% |
|
DE 2D7P |
442.829 W |
10.825 A |
40.908 V |
11.480 A |
48.525 V |
|
e% |
0.64% |
0.87% |
1.51% |
0.70% |
0.36% |
(a)
(b)Figure S2. Superposition of I x V curves generated by PSO, GA and DE for (a) 1D5P and (b) 2D7P under conditions G = 700 W / m² and Tc = 40° C.
Table S3. Points of interest for the different methods under conditions G = 700 W / m² and Tc = 40 ° C.
|
Method |
|||||
|
Data provided |
291.811 W |
7.604 A |
38.376 V |
8.040 A |
45.785 V |
|
by Manufacture |
|||||
|
PSO 1D5P |
287.0525 W |
7.480 A |
38.376 V |
8.0443 A |
45.688 V |
|
e% |
1.63% |
1.63% |
<0.10% |
<0.10% |
0.21% |
|
GA 1D5P |
280.105 W |
7.550 A |
37.100 V |
8.139 A |
45.588 V |
|
e% |
4.18% |
0.72% |
3.44% |
1.22% |
0.43% |
|
DE 1D5P |
288.204 W |
7.510 A |
38.376 V |
8.145 A |
45.725 V |
|
e% |
1.24% |
1.24% |
<0.1% |
1.31% |
0.13% |
|
PSO 2D7P |
283.273 W |
7.636 A |
37.097 V |
8.101 A |
45.508 V |
|
e% |
2.93% |
0.42% |
3.33% |
0.76% |
0.61% |
|
GA 2D7P |
278.747 W |
7.514 A |
37.097 V |
8.15 A |
45.255 V |
|
e% |
4.48% |
1.18% |
3.33% |
1.37% |
1.16% |
|
DE 2D7P |
285.972 W |
7.453 A |
38.370 V |
8.145 A |
45.548 V |
|
e% |
2.00% |
1.99% |
<0.10% |
1.31% |
0.52% |
(a)
(b) Figure S3. Superposition of I x V curves generated by PSO, GA and DE for (a) 1D5P and (b) 2D7P under conditions G = 700 W / m² and Tc = 40° C.
Table S4. Points of interest for the different methods under conditions G = 400 W / m² and Tc = 60 ° C.
|
Method |
|||||
|
Data provided |
151.763 W |
4.352 A |
34.872 V |
4.640 A |
41.600 V |
|
by Manufacture |
|||||
|
PSO 1D5P |
145.155 W |
4.306 A |
33.710 V |
4.643 A |
41.380 V |
|
e% |
4.35% |
1.06% |
3.33% |
<0.10% |
0.53% |
|
GA 1D5P |
139.143 W |
4.275 A |
32.548 V |
4.697 A |
40.740 V |
|
e% |
8.32% |
1.77% |
6.67% |
1.23% |
2.07% |
|
DE 1D5P |
146.301 W |
4.340 A |
33.710 V |
4.699 A |
41.420 V |
|
e% |
1.63% |
1.63% |
<0.10% |
<0.10% |
<0.10% |
|
PSO 2D7P |
142.214 W |
4.220 A |
33.700 V |
4.700 A |
41.150 V |
|
e% |
6.29% |
3.03% |
3.36% |
1.29% |
1.08% |
|
GA 2D7P |
138.004 W |
4.240 A |
32.548 V |
4.700 A |
40.685 V |
|
e% |
9.07% |
2.57% |
6.67% |
1.29% |
2.20% |
|
DE 2D7P |
143.166 W |
4.247 A |
33.710 V |
4.697 A |
41.160 V |
|
e% |
5.66% |
2.41% |
3.33% |
1.23% |
1.06% |
References
CANADIAN SOLAR, 2020. Canadian Solar HiKu CS3W 440Wp Poly-Crystalline Datasheet; 2
Accessed July 31 2023 at https://www.canadiansolar.com/test-au/wp-content/uploads/sites/2/2020/04/Canadian_Solar-Datasheet-HiKu_CS3W-P_v5.59_AU-2.pdf
It is important to note that the results were remarkably similar for both cases. The 1D5P model had greater accuracy for the polycrystalline module than the 2D7P model. In the same way as the previous case, the more the conditions deviated from the STC conditions, the greater the error. Another similar result is that the PSO 1D5P and DE 1D5P combinations obtained the highest accuracy. Regarding the relative error values, note that the results were practically identical. However, the schemes had a slightly higher accuracy for the extra study (polycrystalline).
It is important to mention that a comment about this investigation was allocated at the end of Section 3. It is highlighted in blue.
About the Specific Comments:
“1) Introduction: The importance of PV is highlighted in lines 52-60. However, the most recent data possible should be presented.”
Answer:
The authors agree with the comment. Therefore, we have restructured this paragraph and we added a recent reference. Changes are highlighted in blue.
“2) Introduction: A brief outline of the structure of the article is useful at the end of the introduction.”
Answer:
This suggestion was accepted and added in the revised MS. This brief outline is at the end of the introduction and highlighted in blue.
“3) Materials and Methods: Even if the objective of this work is to extract the parameters using only the data provided in the PV module datasheet, a brief overview of photovoltaic technologies and the most suitable models for each would be interesting. Likewise, a description of the technology and structure of the module being worked on would be useful.”
Answer:
The authors agree with the comment. We improved Section 2 by adding a brief overview of photovoltaic technologies, modeling process and the most suitable models. These changes are highlighted in blue.
“4) Materials and Methods: The exposition of the bio-inspired algorithms is clear and sufficiently detailed”
Answer:
Thanks for the comment.
“5) Results: With regard to the conditions of the simulations (line 390 onwards), although they may be convenient as a quality control of the method, they do not reflect the usual working conditions of a PV module. In particular, it is strange that NOMT (formerly known as NOCT) conditions are not used.”
Answer:
The authors agree with the comment. We added one more temperature and irradiation case to the results, the NOMT case (formerly known as NOCT). We have restructured the results, adding Figure 10 and Table 7 for comparative purposes in usual cases of temperature and irradiation in the cell.
Now our results cover a wide range of samples. That varies from situations close to the reference (STC), situations that are more common to be found (NOMT and Medium case), and situations that differ significantly from the reference in temperature and irradiation. Therefore, it is the worst case for simulation. Furthermore, to finish a case with a higher output power than the reference. In this way, the effectiveness of the methods in a wide range of operations is verified.
“6) Conclusion: The last sentence (lines 538-) should be refined. It is not clear to me whether this refers to the extraction of parameters or to the results of simulations that could be done with them.”
Answer:
The authors agree with the comment. So we improve final of conclusion to make the interpretation easier. These changes are highlighted in blue.
“7)There are some (minor) spelling mistakes, typos and expressions that should be corrected.”
Thanks for the suggestion. A revision of the English Writing in the entire text was carried out.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The performance of PV modules is simulated and assessed in this paper under different environmental conditions of irradiation and temperature in the models of one diode and five parameters (1D5P) and two diodes and seven parameters (2D7P), using three bio-inspired algorithms: Particle Swarm Optimization, Genetic Algorithm and Differential Evolution.
The subject is worth investigating and the paper may be considered for publication after a revision addressing the following issues.
The main concern is related to the choice of simulations situations described in lines 392-393. The choice of low solar irradiance and high cell temperature is not usual as well as high solar irradiance and low cell temperature. Thus, the relevance of the results is limited. To increase the relevance of the results, the simulations must be performed and validated on usual situations too.
The literature survey related to the installed PV capacity is outdated (Haegel et al., 2019) and must be updated. Similarly should be done for the references (Cengiz and Mamiş, 2015; Rehman et al., 2007).
Line 57: “consumption of photovoltaic panels” should be reconsidered;
The use of lumped references as in (Da Rocha Queiroz et al., 2020; Provensi et al., 2022; Tifidat et al., 2023; Yahya-Khotbehsara and Shahhoseini, 2018), (Chin et al., 2015; Laudani et al., 2014; Yahya-Khotbehsara and Shahhoseini, 2018), (Liang et al., 2020; Reis, 2018; Tifidat et al., 2023; Yousri et al., 2019) should be avoided to provide sufficient representation of the main contribution of each referenced paper.
Line 357: the citation Storn, 1997 must be correctly done
Line 362: the measurement unit of the solar irradiance must be corrected
The quality of figures 6 to 9 must be improved.
The use of more than one or two decimals in Tables 3 to 6 is not justified.
The English must be carefully checked for typos, grammar, spelling and syntax. (e.g. „Among the equivalent electric circuit models is the single diode model, which is one of the most used equivalent circuits in the literature to represent a photovoltaic cell.”).
Author Response
Reviewer #2
“The performance of PV modules is simulated and assessed in this paper under different environmental conditions of irradiation and temperature in the models of one diode and five parameters (1D5P) and two diodes and seven parameters (2D7P), using three bio-inspired algorithms: Particle Swarm Optimization, Genetic Algorithm and Differential Evolution.
The subject is worth investigating and the paper may be considered for publication after a revision addressing the following issues.
Answer:
Dear Reviewer:
Thank you for reviewing our manuscript and for your valuable comments and suggestions to improve the manuscript quality. Based on your suggestions, the manuscript has been carefully corrected. We are looking forward to receiving your positive responses.
With best wishes,
Cid Marcos G. Andrade
“1) The main concern is related to the choice of simulations situations described in lines 392-393. The choice of low solar irradiance and high cell temperature is not usual as well as high solar irradiance and low cell temperature. Thus, the relevance of the results is limited. To increase the relevance of the results, the simulations must be performed and validated on usual situations too.”
Answer:
The authors agree with the comment. We added one more temperature and irradiation case to the results, the NOMT case (formerly known as NOCT). We have restructured the results, adding Figure 10 and Table 7 for comparative purposes in usual cases of temperature and irradiation in the cell.
Now our results cover a wide range of samples. That varies from situations close to the reference (STC), situations that are more common to be found (NOMT and Medium case), and situations that differ significantly from the reference in temperature and irradiation. Therefore, it is the worst case for simulation. Furthermore, to finish a case with a higher output power than the reference. In this way, the effectiveness of the methods in a wide range of operations is verified.
“2) The literature survey related to the installed PV capacity is outdated (Haegel et al., 2019) and must be updated. Similarly, should be done for the references (Cengiz and Mamiş, 2015; Rehman et al., 2007).”
Answer:
The authors agree with the comment. Therefore, we have restructured the paragraph with Haegel reference and we added a recent reference. Changes are highlighted in blue. In addition, we updated the Cengiz and Rehman reference. These last changes are highlighted in red.
“3) Line 57: “consumption of photovoltaic panels” should be reconsidered;”
Answer:
This suggestion was accepted and corrected in the revised MS. We have restructured this paragraph by adding recent references and changing this term. Changes are highlighted in blue.
“4) The use of lumped references as in (Da Rocha Queiroz et al., 2020; Provensi et al., 2022; Tifidat et al., 2023; Yahya-Khotbehsara and Shahhoseini, 2018), (Chin et al., 2015; Laudani et al., 2014; Yahya-Khotbehsara and Shahhoseini, 2018), (Liang et al., 2020; Reis, 2018; Tifidat et al., 2023; Yousri et al., 2019) should be avoided to provide sufficient representation of the main contribution of each referenced paper.”
Answer:
The authors agree with the comment. The use of lumped cited references was avoided, with the exception of the last one, as it proves the argument that several studies that apply heuristic algorithms for PV simulation use a wide variety of current-voltage (I x V) data external to the datasheet use.
“5) Line 357: the citation Storn, 1997 must be correctly done.”
Answer:
This suggestion was accepted and corrected in the revised MS. We highlight in red, this correction.
“6) Line 362: the measurement unit of the solar irradiance must be corrected.”
Answer:
This suggestion
n was accepted and corrected in the revised MS. We highlight in red, this change (kWh/m² à W/m²).
“7) The quality of Figures 6 to 9 must be improved.”
Answer:
Thanks for the suggestion. We improved the quality of Figures: 6, 7, 8, 9 to 300 DPI.
“8) The use of more than one or two decimals in Tables 3 to 6 is not justified.”
Answer:
Thanks for the suggestion. To maintain uniformity, we have standardized current, voltage, and power points to 3 decimal places and relative errors to 2 decimal places in all quoted tables. The choice of two decimal places for e% is more usual and differentiates what information is and what is an error.
Once again, many thanks for the reviewers, we are sincerely grateful for the significance given by the reviewers to our work, which will certainly encourage us to continue to develop and improve researches like this.
“9)The English must be carefully checked for typos, grammar, spelling and syntax. (e.g. „Among the equivalent electric circuit models is the single diode model, which is one of the most used equivalent circuits in the literature to represent a photovoltaic cell.”).”
Thanks for the suggestion. A revision of the English Writing in the entire text was carried out.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
I hope you will continue to develop this line of research because I find it very interesting.
Author Response
Reviewer #1
“I hope you will continue to develop this line of research because I find it very interesting.”
Answer:
Dear Reviewer:
Thank you for reviewing our manuscript and for your valuable comments and suggestions to improve the manuscript quality. The group of authors intends to continue with this line of research.
With best wishes,
Cid Marcos G Andrade.
Author Response File: Author Response.docx
Reviewer 2 Report
The authors addressed the queries.
The following issues must be addressed.
Line 34: “a relative error greater than 8%” must be corrected since the maximum error was 7.9%
Line 459: DE acronym must be consistently used in Table 3 (instead of DEA).
Line 569: the value of maximum error of PSO 1D5P must be consistent with Table 7.
Author Response
Reviewer #2
“The authors addressed the queries. The following issues must be addressed. “
Answer:
Dear Reviewer:
Thank you for reviewing our manuscript and for your valuable comments and suggestions to improve the manuscript quality. Based on your suggestions, the manuscript has been carefully corrected. We are looking forward to receiving your positive responses.
“1) Line 34: “a relative error greater than 8%” must be corrected since the maximum error was 7.9%
Answer:
The GA 1D5P data needed to be clarified with the GA 2D7P data. The maximum error value of 9.59% (Table 5, GA 2D7P scheme) was put instead of using "was greater than 8%" to be more direct for future readers.
“2) Line 459: DE acronym must be consistently used in Table 3 (instead of DEA).
Answer:
This suggestion was accepted and corrected in the revised MS.
“3) Line 569: the value of maximum error of PSO 1D5P must be consistent with Table 7.”
Answer:
This suggestion was accepted and corrected in the revised MS.
Best regards
Cid Marcos G. Andrade
Author Response File: Author Response.docx
