Decomposition of a Cooling Plant for Energy Efficiency Optimization Using OptTopo
Round 1
Reviewer 1 Report
This work deals with the optimization methods of complex industrial processes, so important in modern business. The authors derive their approach from chosen parameters, the energy performance indicators ( EnPI) and describe an optimization algorithm OptTopo, that could give a valuable help.
In the second paragraph the authors summarize same of the most used method to evaluate optimal set-points for complex systems and after they describe their own approach and the developed algorithm. The OpTopo algorithm is compared to three well established and quite known optimizers, COBYLA and SDPEN, which are local optimizers, and the global optimizer COUENNE.
The experimental results show the good performance of the present approach on both local and global optimization case, that make this algorithm of general applicability.
The conclusion is that OpTopo algorithm can successfully compete in term of efficiency with two of the most known software for local optimization problem solvers and even with a well established global optimizer: these are unique characteristics.
A wider description of the mathematical method (paragraph 3) could be appreciated by non expert readers, increasing the readability of the work and deepening with more details the characteristics of the developed algorithm (that from the claim of the authors sounds to be quite unique).
I think the present work can be published in the present form: it is clear and well written.
I have no corrections to suggest.
Author Response
Thank you very much for your positive feedback. Please find attached the revision with highlighted changes and a clear version.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This paper uses the optimization algorithm OptTopo to optimize a complex system. There are some comments:
1. The introduction needs to be improved.
2. The algorithm principle and system description should be improved.
3. In page 9, line 361 "can bei"should be "be".
4. One table or figure of the final results of optimized parameters for the whole system should be provided.
Author Response
Thank you very much for your feedback. Please find attached the revision with highlighted changes. I hope, the changes suit your expectations.
1) Introduction and abstract are revised
2) I added a schema in order to illustrate the system. This should also help to understand the traversal.
3) Check.
4) I added a table with to intervals of cooling power requested as well as a table with the solutions achieved by OptTopo.
Author Response File: Author Response.pdf
Reviewer 3 Report
My comments are as follows.
1. Abstract is too vague. It does not provide any methodology or quantified results. It also does not refer to the specific topic of the paper.
2. Avoid use of first person. Re-write in third person passive past tense like other journal papers.
3. Line 18 – avoid links in text. Make it a reference.
4. The introduction does not properly motivate the problem. As an outsider, I have no idea why this is important or why I should bother reading the rest of the paper.
5. The literature review would be better served by eliminating the subtitles that aren’t labeled.
6. The literature review is good, but it could be improved. Present an evolution of optimization approaches, from the metaphorical beginning. With each subsequent paper, cite what was missing or deficient and then how later works (or your work) furthered the research. As it currently stands, the cited papers read like a collection of random facts. There is no direction or cohesion.
7. Line 361 – be
8. Put the tables after they are referred to in the text.
9. A significant weakness of this paper is the lack of tangibility. Everything, from the problem setup to the solution, is abstract and extremely hard to visualize. Provide a schematic of the theoretical layout of the cooling systems in the optimization problem – provide the sizes for each and list the relevant variables from the tables in their appropriate locations. This will help the reader gain more traction with this paper.
10. Explain COP in Figure 3.
11. Use percentages when comparing different optimizers. Put these in the conclusions as well.
12. The prognosis does not seem to be clear. The authors conclude that their approach works well, but don’t say by how much. In the plots, it looks like some optimizers produce better results and their cost is way lower. Why should the reader adopt the proposed method?
Author Response
Thank you very much for your feedback. Please find attached the revision with highlighted changes and a clear version. I hope, the following changes suit your expectations.
1) The abstracted is revisited and extended now.
2) The author did not find any specification by MDPI but the MDPI style guide is consistently written in first person plural: https://res.mdpi.com/data/mdpi-author-layout-style-guide.pdf . In this review paper http://proc.iscap.info/2015/pdf/3487.pdf , the use of the first person is recommended. Nevertheless, I tried to adapt the formulations according to your remark.
3) Check
4) Introduction is extended
5) We changed paragraphs to subsections.
6) The state oft the art is further explained.
7) Corrected
8) Thanks for the advice regarding the table positions. I included the latex package "flafter" which ensures that all floating objects are after their first reference.
9) Now, a schema of the considered system helps to understand the approach better.
10) I added an explanation.
11) Percentage values also in results section and conclusion section.
12) I tried to improve the argumentation all over the paper. Main advantages are a) reduced number of dimension due to decomposition b) no new runs necessary afer an initial computation for all operation points c) compatible with different model types.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
This paper has been improved a lot. The reviewer has no comments.
Reviewer 3 Report
Thank you to the authors for replying to my comments and for making the changes to their manuscript. The paper can now be accepted for publication.
