On the Use of Indirect Measurements in Virtual Sensors for Renewable Energies: A Review

Round 1

Reviewer 1 Report

The work is an impressive attempt to bring together the state of the art of indirect sensing in the renewable energy sector. Here my suggestion for improvement:

- When mentioning prognosis (e.g., lines 45-60), I would also mention remaining useful life (RUL) estimation. Reference can be found e.g. for rotating machinery in the research groups of profs. Jerome Antony  (INSA Lyon) and Konstantinos Gryllias (KU Leuven).

- Line 273: when mentioning "lack of generalizability" I would add that the issue is on extrapolating results outside the range in which the models have been trained (as you correctly mention later in line 298).

- I suggest to explicitly explain what it is meant by "empirical data" (lines 286 and 343), as to me it seems that they are just measured data. In line 343, are the Authors hinting at model update?

- In section 2.2 and 2.3, I recommend defining better the difference between physics based methods and hybrid methods based on state space. I believe that some confusion can arise. Specifically, I would ask the Authors to (1) avoid using "correction" in line 323, as this term is often associated to Kalman filters (prediction and correction phases), that are now listed among the hybrid methods (2) expand the sentences in lines 339-341 and 342-343

- Add the section numbers to Figure 1.

- Lines 421-422: please add a reference for the NAR.

- When referring to [111], I suggest to also check a more recent publication: Bosmans, J., Kirchner, M., Croes, J., Desmet, W. with Bosmans, J. (2023). Validation of a wind turbine gearbox strain simulation model in service to virtual sensing. Forschung Im Ingenieurwesen-Engineering Research, 87, 107-117. doi: 10.1007/s10010-023-00635-0 

The paper reads very easily.

Minor comments:

- Line 1-2: underscores --> underscore (should be plural)

- Use of "precision" instead of "accuracy" in many parts of the paper: is this an author choice indicating that the results are not converging to the real value, but an offset is present? Please clarify this choice and make a statement in the paper.

Author Response

Dear Reviewer,

Thank you very much for your thorough review. We have made the following revisions to the manuscript, highlighted in red, based on your feedback:

• Regarding the mention of RUL: This has been addressed (please check the revised lines 45-48).
• Concerning "the lack of generalizability": This has been addressed (please check the updated line 276).
• We initially used the term "empirical" to avoid repetition; however, we have now replaced "empirical" with "measured" throughout the document to ensure clarity. 
• We have not specifically discussed model updates; rather, we have maintained a more general approach. We define a hybrid method as a physics-based model that incorporates a statistical model to account for real system fluctuations.
• To clarify any ambiguities, we propose now that a physics-based model should only use data for validation purposes or to quantify model inputs. Please review lines 329-334.
• For Figure 1, we have added section numbers to differentiate between data-based, physics-based, and hybrid methods.
• The NAR reference has been updated as requested.
• Regarding the publication by Bosmans et al., it presents an interesting physics-based method. However, it does not employ an indirect measurement approach, or at least does not explain how torque is estimated from strain. The focus of this study is solely on static strain, which I personally think that it doesn not fully characterize the real behavior of the gearbox. If you consider this publication highly relevant, we can find ways to include it into our paper.
• Correction from "underscores" to "underscore" has been made.
• Lastly, the term "precision" was used in place of "accuracy" due to a language transfer error. We have corrected this by using "accuracy" throughout the document.

Thank you again for your valuable input.

Reviewer 2 Report

The proposed work for review is very interesting. The authors have collected in one place, in a systematized form, an awful lot of information on the subject of the work. The theme of the work is very current and modern. The work has the following very strong points:
• A very thorough review of the existing issues on the topic under consideration was made;
• The authors have presented in detail what are the advantages and disadvantages of the considered indirect measurements, used methods and where they are applicable;
• In the "Discussion" section, the authors have made a very thorough comment and analysis of the current state of the subject, based on their overview, and what the possible future development of the virtual sensors and the indirect measurements are.
Notes:
1. The abstract should be edited and include briefly what was done in the presented work;
2. Section 1 "Introduction" should be edited. It should include a brief explanation of the state of the problem and how the presented work solves the problem or adds to existing knowledge. I propose, of course the authors may not agree with this proposal, that the present abstract (lines 1 - 15) together with lines 170 to 203, be the new section 1 - Introduction, of course making the appropriate editing of the text to is meaningful;
3. The current section 1 "Introduction", without lines from 170 to 203, be renamed to section 2, for example "State of the problem" or another, according to the authors, more appropriate title;
4. In Section 2, new 3, "Indirect Measurements", it is good that the authors give some simple examples of Indirect Measurements. In this way, this type of measurements will be understood more easily by other readers of the work;
5. In sub-sections, now 2.1; 2.2 and 2.4, new 3.1; 3.2 and 3.3, several simple examples of application of the considered methods should be given to make them understandable to a wider range of readers.

Author Response

Dear Reviewer,

Thank you very much for your thorough review. We have made the following revisions to the manuscript, highlighted in blue, based on your feedback:

• We have revised the abstract as requested.

• Regarding the introduction: We have previously presented the state of the problem in the earlier version of the document (please check the updated lines 172-198). To facilitate direct access to this part, we have added a new subsection titled "Problem Statement". Additionally, we have previously described how our work contributes to existing knowledge in the earlier version of the document (please review lines 193-198).

• The content previously spanning lines 170-203, now revised to lines 172-198, contains conclusions derived from the entire development presented in lines 1-172. Consequently, it cannot serve as an introduction. We believe that incorporating it within the "Problem Statement" subsection aligns with your suggestions and fits the manuscript’s structure more aptly.

• In Section 2 (Indirect Measurements), we have previously presented simple examples of indirect measurement techniques used in industry in the earlier version of the document. These can be reviewed in lines 246-262.

• Regarding the inclusion of simple examples in Subsections 2.1, 2.2, and 2.3, this has been addressed (please review new lines 318-322, 335-341, and 362-366).

Thank you again for your valuable input.