A PSO-Based Recurrent Closed-Loop Optimization Method for Multiple Controller Single-Output Thermal Engineering Systems

Round 1

Reviewer 1 Report

The paper describes an innovative method of optimizing control parameters in an operating plant and, based on the result section, Figs 11 and 12, can be extended to include inter-connected devices.  The paper could be improved by providing more specific information about the test case illustrated.  The data is obviously from an operating plant, but the method used to extract the data into the model is very superficial, more details could be provided.  Specifically the lines 251-253 should be expanded and referred to the shown data as to what is left and what is right.  The authors should also describe the random number generator that they are using.  Lines 158-159 also have some confusing superscript or out of alignment variables.   The results shown in Fig. 8 are quite dramatic, but are only superficially commented on.  The proposed method provides much better control. These comments are intended to improve the paper, the basic approach is scientifically sound.

One other area that should be included is the improvement in the performance that would result in implementing this approach.  This is the goal of the method, but the authors do not mention it.  It could be that extending the improved control of the super heated steam process is beyond the scope of the paper, but a mention of the potential improvement would provide another contribution from the paper. 

Author Response

Dear Reviewer:

Thank you very much for your valuable comments. Please see the attachment for our responses.We are looking forward to hearing from you soon.

Best regards！

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear authors,

The research idea pursued in this work is interesting and build upon previous work in the literature addressing similar optimal control problems.

The research layout is well-constructed with concrete examples and the contributions are clearly pointed out. However, may I suggest elaborating on the research contributions more explicitly using a separate subsection - the challenges addressed, the extent to which this work compares with peers and why you chose to use a particular metaheuristic PSO in the first place.

Please also address the choice of assumed values in the cost function - the choices seemed fairly ad-hoc.

There are a few typos within the article eg. in line 321 RMSE is mis-spelt. 

The general trend within the SI community is to let processes reach convergence before plotting the fitness values - you might want to look at figs 11 and 12 and run them for several more iterations to effect in convergence.

Also, the experiments are incomplete without any information on the time complexity involved - can you compare using reported CPU times for the tests?

Given that many readers do not know about the PSO algorithm in general please also provide a separate section dedicated to PSO and explain how it works. Include a figure showing the personal and global best attractors pulling particles in hyperspace while the inertia maintaining exploration. Please cite, in this regard the following recent survey of PSO:  

Sengupta, S.; Basak, S.; Peters, R.A., II. Particle Swarm Optimization: A Survey of Historical and Recent Developments with Hybridization Perspectives. Mach. Learn. Knowl. Extr. 2019, 1, 157-191.

How much do your results vary? PSO is a random search technique and will terminate with different parameter values for every run given convergence is not reached/problem space is compelx, etc. Can you report mean and standard deviation of tests?

One last thing: please elaborate on how the implications of your findings - i.e a way to meta-control the optimization of system parameters, benefit peer research - i.e. include a subsection on future work/extensions etc. Here you may also outline the limitations of the baseline methods to highlight what advantages heuristic evolution of parameters using PSO/similar random search algorithms may provide researchers.

Please make these changes and upload the revised manuscript. 

Author Response

Dear reviewer,

Thank you very much for your comments. Please see the attachment for our responses. We are looking forward to hearing from you soon.

Best regards! 

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript describes the use of a particle swarm optimization-based procedure to control thermal engineering systems. The subject is of interest to the journal and scientific community. The mathematical model is scientifically sound. However, some issues must be addressed. The major concern is that the manuscript does not present any comparison with other control systems/algorithms performances. The comparison with other algorithms would strengthen the work and its scientific relevance. In addition, the authors should present evolving indicators of the performance (e.g., number of fitness function evaluations, etc.). The manuscript lacks a deeper discussion, due to lack of comparison with other methods, and meaningful conclusions. Other minor issues are:   1) Abbreviations should be avoided in the title. Specially the one between parenthesis as PSO is already used in the title before. 2) In the introduction avoid lump sum citations (e.g., in line 41, “[3-7]”). Please critically review each citation individually. 3) Please add nomenclature and abbreviations table. 4) Several blank spaces should be added between values and units, values and words, bullet items and words. e.g., line 108, 307, 313, etc. 5) In text formulae does not math fully with the equation formulae. Please correct. e.g., lines 119-120. 6) Please clarify if (1) is relative to Eq. (1) or bullet point (1) from Principle 2. If it refers to Eq. (1), please add Eq. before it. 7) Please describe Figs. 3 and 4 to the reader in the first paragraph on section 3. 8) Figs 5 to 8 have y-axis labels depicting °C with the wrong character. 9) Results and Discussion sections should be merged into a single one, as the Discussion section does not convey additional information per se. 10) References should be avoided in the Conclusion sections.

Author Response

Dear reviewer,

Thank you very much for your comments. Please see the attachment for our responses. We are looking forward to hearing from you soon.

Best regards!

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The revised manuscript reads much better and I am glad to see that important issues such as variability of results and comparison with existing approaches have been put forward in detail. Given that the paper's organisational issues are taken care of and the paper is spell and grammar checked - I find it to be a solid contribution with detailed analysis for the sake of reproducibility.

I am happy to recommend this paper for publication. 

 

Author Response

Dear reviewer,

Thank you very much for your valuable comments. All of them are very useful for improving our paper. We have checked through the entire manuscript and corrected a little misuses on capitalized letters and a typo. Please see the attachment.

Best regards,

Lei Pan

--

Professor

School of Energy & Environment, Southeast University

Sipailou 2# , Nanjing, China

Author Response File: Author Response.docx

Reviewer 3 Report

The authors have addressed all comments. The manuscript has been improved significantly.

Some last comments:

1) References to figures are different in several parts of the text. E.g., in line 311 and line 342. Check all text.

2) According to IS units the value must been separated from the unit. E.g., in line 311, 330MW should be 330 MW; line 515, 80% should be 80 %, line 442 1.36°C should be 1.36 °C, etc.

Author Response

Dear Reviewer:

Thank you for your comments concerning our manuscript. The modified places are highlighted in the revised paper. Following please find the responses to each points.

Point 1. References to figures are different in several parts of the text. E.g., in line 311 and line 342. Check all text.

Response 1: Thank you for the comment. We have changed the brief texts on figure reference in line 342 into Figure 6, Figure 7 and Figure 8. And we have also checked the entire manuscript to make sure that there is no more such difference.  

Point 2. According to IS units the value must been separated from the unit. E.g., in line 311, 330MW should be 330 MW; line 515, 80% should be 80 %, line 442 1.36°C should be 1.36 °C, etc.

Response 2: Thank you for the comment. We have modified the three places mentioned above. And we have also checked through the entire manuscript and did not find this type of mistake anymore.

The authors sincerely appreciate all the valuable comments from the reviewer, which are very useful for modifying and perfecting our paper.

Best regards!

Lei Pan

--

Professor

School of Energy & Environment, Southeast university

Sipailou 2#, Nanjing, China

 

Author Response File: Author Response.pdf