Performance Assessment of Heuristic Genetic Algorithm (HGA) for Electrochemical Impedance Spectroscopy Parameter Estimation
, Gladis G. Suarez-Velázquez, Eddie N. Armendáriz-Mireles, Carlos A. Calles-Arriaga and E. Rocha-Rangel *Round 1
Reviewer 1 Report
I liked the topic of the manuscript and the processing of the issue using the described tools for HGA.
Such methods are used in energy systems, but not only in them.
Nevertheless, I have a few comments:
Formal errors:
-In the text of the manuscript, "Rs, Rp, and Cp" are written without a subscript, please edit as it is in mathematical relations.
-Mathematical relations (1) and (2) are a different color, if it does not have a special meaning, it would be good to keep the same color for all mathematical equations.
-In the mathematical relation for "C" there is some error in the denominator, it is an operator and something seems to be missing after it.
-Figure 2- the algorithm does not have a marking of the decision block for fulfillment (True) and non-fulfillment (False) of the condition
-in line 272 is probably an incorrectly marked dimension of the value "1000 ?"
Other:
When reading the manuscript, the uninterested reader will not know why the values for the research, which are listed in Table 1, were chosen. Is it possible to choose other values and what impact would this have on the investigation of the mentioned problems?
The algorithm in Figure 2 is described briefly, but it refers to the notation of equation (4), from which only the statement about the composition of the function is accepted. A specific function for the model or a reference to it is not defined.
Apparently, the solution to the problem is based on information from the literature, but when reading the manuscript, the reader will learn only the results of calculations and measurements.
Since the Matlab toolbox was used for the solution, I would welcome a more detailed description of it in the text or in chapter 2.
I have no serious comments on the results and evaluation.
Author Response
Reviewer 1
Comment 1:
-In the text of the manuscript, "Rs, Rp, and Cp" are written without a subscript, please edit as it is in mathematical relations.
Response to comment 1:
We appreciate this comment by the reviewer. The document was modified according to the reviewer's commend, and also italic style was used considering that Rs, Rp and Cp are mathematical variables.
Comment 2:
Mathematical relations (1) and (2) are a different color, if it does not have a special meaning, it would be good to keep the same color for all mathematical equations.
Response to comment 2:
Thankful for the comments. The color does not have a special meaning was a format error. The document was modified by putting the same color.
Comment 3:
In the mathematical relation for "C" there is some error in the denominator, it is an operator and something seems to be missing after it.
Response to comment 3:
We appreciate this comment by the reviewer. It is correct we miss the imaginary operator j. Equation 3 was corrected.
Comment 4:
Figure 2- the algorithm does not have a marking of the decision block for fulfillment (True) and non-fulfillment (False) of the condition
Response to comment 4:
Figure 2 was modified by adding the Yes/No decision.
Comment 5:
in line 272 is probably an incorrectly marked dimension of the value "1000 ?"
Response to comment 5:
The document was corrected, and the unit is ohms.
Comment 6:
When reading the manuscript, the uninterested reader will not know why the values for the research, which are listed in Table 1, were chosen. Is it possible to choose other values and what impact would this have on the investigation of the mentioned problems?
Response to comment 6:
Thankful for the comments. This is an interesting point because this work was developed to show undergraduates and practitioners the potential of genetic algorithms. The proposed broad values were chosen to study the power of Heuristic algorithm, and because it is a simple circuit no important effect on time iteration was observed.
Comment 7:
The algorithm in Figure 2 is described briefly, but it refers to the notation of equation (4), from which only the statement about the composition of the function is accepted. A specific function for the model or a reference to it is not defined.
Response to comment 7:
The multi-objective function was adapted considering equation 3, and two different equations are analyzed for angular frequency (real and imaginary).
Comment 8:
Apparently, the solution to the problem is based on information from the literature, but when reading the manuscript, the reader will learn only the results of calculations and measurements.
Response to comment 8:
The problem's solution was achieved considering the values of Zreal and Zim of the experimental data. This information was used to feed the multi-objective function; the function is presented in equation 5.
Comment 9:
Since the Matlab toolbox was used for the solution, I would welcome a more detailed description of it in the text or in chapter 2.
Response to comment 9:
The manuscript was modified by adding information about the Matlab optimization toolbox.
Reviewer 2 Report
The authors use GA to identify parameters for a simple Randles circuit, which is verified using experiments. None of the work is novel. The reviewer suggests that more literature work is needed to identify a problem to solve.
Author Response
Reviewer 2
Comment 1:
- The authors use GA to identify parameters for a simple Randles circuit, which is verified using experiments. None of the work is novel. The reviewer suggests that more literature work is needed to identify a problem to solve
Response to comment 1:
We agree with the reviewer comment. Although some papers have been published, mainly on the battery applications, on the topic of impedance analysis by genetic algorithm, it should be mentioned that most of them use intended program software to solve the optimization problem. So, practitioners or engineers need to have in deep level of knowledge in programming structures. The use of Matlab Toolbox open avenue for scientific and practitioners to solve multi-objective functions by several methods in an easy way. In order to improve the manuscript some references has been added were genetic algorithm was adopted to solve the specified problem.
Reviewer 3 Report
The paper is interesting but i have some comments
What is novelty of the paper?
The conclusion should be more direct points to tge major findings
The English language should be edited
A future work section should be added to the paper
After all the paper can be accepted
Author Response
Reviewer 3
Comment 1:
What is novelty of the paper?
Response to comment 1:
This comment has been addressed considering the comment of Reviewer #2. Response: Although some papers have been published, mainly on battery applications, on the topic of impedance analysis by genetic algorithm, it should be mentioned that most of them use intended program software to solve the optimization problem. So, practitioners or engineers need to have in deep level of knowledge in programming structures. The use of Matlab Toolbox open avenues for scientific and practitioners to solve multi-objective functions by using several methods easily. In order to improve the manuscript, some references have been added where a genetic algorithm was adopted to solve the specified problem.
Comment 2:
The conclusion should be more direct points to tge major findings
Response to comment 2:
The manuscript has been modified according to this comment.
Comment 3:
The English language should be edited
Response to comment 3:
The manuscript has been modified, and misspelling and grammar errors have been corrected.
Comment 4:
A future work section should be added to the paper
Response to comment 4:
The previous manuscript contains a paragraph for future work in the conclusion section. To clarify, we create a new section to put the future work
Round 2
Reviewer 2 Report
Unfortunately, the authors haven't addressed the reviewer's concerns in the first submission. The authors should explore complicated circuits or an existing electrochemical system to validate their suggested HGA approach. The simple RC circuit (which could be validated with a simple multimeter) is a suitable test case to start research, but more work is needed for this to be accepted as a research paper.
Author Response
Reviewer #2:
Comment_1:
- Unfortunately, the authors haven't addressed the reviewer's concerns in the first submission. The authors should explore complicated circuits or an existing electrochemical system to validate their suggested HGA approach. The simple RC circuit (which could be validated with a simple multimeter) is a suitable test case to start research, but more work is needed for this to be accepted as a research paper.
Response to comment 1:
We appreciate the reviewer's suggestion and apologize for the comment's misinterpretation in the first round. The document was modified according to the reviewer's comment by incorporating a case for battery parameter estimation (nine unknown parameters). Two Randles circuits were connected with a resistor and a constant phase element (to simulate the behavior of non-ideal capacitors in the electrochemical process, ZCPE=Q-1(jw)-a. Furthermore, some relevant references were added to improve the understanding of the Randles circuit and its implementation in battery technology.
Round 3
Reviewer 2 Report
The reviewer appreciates the efforts by the researchers to improve the manuscript. The discussion on Li-ion and the applicability of the method for a larger parameter set. Can the authors comment on using this method over other simple and well-known optimization techniques such as LHS, SA, PSO, and LM? It would be worthwhile to quantitatively show that GA is better than other optimization techniques. Also, many of these are available in the Matlab toolbox the authors are using, so this should be a simple exercise and would improve the manuscript a lot.
Author Response
Reviewer #2:
Comment_1:
The reviewer appreciates the efforts by the researchers to improve the manuscript. The discussion on Li-ion and the applicability of the method for a larger parameter set. Can the authors comment on using this method over other simple and well-known optimization techniques such as LHS, SA, PSO, and LM? It would be worthwhile to quantitatively show that GA is better than other optimization techniques. Also, many of these are available in the Matlab toolbox the authors are using, so this should be a simple exercise and would improve the manuscript a lot..
Response to comment 1:
Thanks for your comment. The manuscript was modified according to the reviewer's comment by comparing the heuristic genetic algorithm with another multiobjective optimization function available in the Matlab optimization toolbox. The result shows that HGA has a better fitness function, demonstrating this algorithm's versatility to solve multiobjective problems with several variables. On the other hand, we believe that by adding these suggestions, practitioners and researchers will have a broader picture of the benefits of using the Matlab Optimization Toolbox, which may positively impact future investigations.
Round 4
Reviewer 2 Report
The authors have improved the article reasonably.
