A Comprehensive Summary of the Application of Machine Learning Techniques for CO₂-Enhanced Oil Recovery Projects

Round 1

Reviewer 1 Report

This literature review paper provides a thorough overview of applied machine learning studies in CO2-EOR projects, encompassing a significant portion of relevant publications. I find the paper to be comprehensive and well-constructed, with no major suggestions for improvement. I recommend accepting it in its current form. While there are additional more recent papers in the field that have not been mentioned, considering the time constraints and the already extensive coverage in the paper, I believe it is satisfactory. The inclusion of tables summarizing sample papers is especially valuable for readers.

In addition to that, there is only one typo that the authors can fix easily. In Section 4.5., paragraph 2: (Table 5Table6), remove “Table 6”.

Author Response

Please see the attachment, thank you!

Author Response File: Author Response.pdf

Reviewer 2 Report

In this review, the authors conducted a comprehensive summary on how machine learning and modeling (such as optimization methods) are applied in CO2-EOR research in the context of petroleum industry. I see this paper would serve as a handful for future research especially when sourcing the data for CO2-EOR related work. I would highly recommend the publication of this work after some minor comment being addressed:

1. In this work, the authors have the rating for each paper summarized in the table. Even it could be subjective, it would be helpful if the authors can explain how the score comes and what metrics are used/evaluted when such rating score is given for each individual work.
2. Can the authors briefly discuss if any/how uncertainty quantification are used in those different work and if feature importance analysis tool (such as SHAP) are used in any of those paper to help promote interpretability and transparency of developed model?

Author Response

Please see the attachment, thank you!

Author Response File: Author Response.pdf

Reviewer 3 Report

A research article should not simply be a summary of existing approaches. It is necessary to focus more on some contributions in the field in question. The authors' own opinion is visible only in a few bullet points at the end of the article. They recommend significantly enriching the article with new and beneficial observations of the authors, which resulted from the relatively extensive literature review. My other comments are as follows:

1) Avoid using abbreviations in the title and abstract.

2) A grouped list of references may be less reader-friendly. Providing a brief justification for each individual reference would enhance clarity.

3) Check the quality of the figure. For some figures, the quality does not seem to be very good (eg fig. 3).

4) Clarify the novelty of your work in comparison to existing literature. Provide a more comprehensive analysis of the strengths and limitations of prior publications and thoroughly assess the uniqueness and contribution of your research.

5) In the case of the rating in tables, it would be good to indicate the scale and what value means very good or very bad.

6) The rating given in the tables should be more described. What factors do the authors take into account, what is given significant weight, etc.?

7) The article describes a lot of commonly known information. It would be good to pay more attention to the actual contribution of this article.

Author Response

Please see the attachment, thank you!

Author Response File: Author Response.pdf

Reviewer 4 Report

This manuscript provides a comprehensive review of machine learning (ML) techniques used to analyze different parameters affecting the enhanced oil recovery (EOR) through CO2 injection. In general, this work demonstrates the advantages and shortcomings of applying ML in CO2-EOR area and could be served as a useful guidance for the future research in the oil and gas industry. Thus, I recommend the publication of this manuscript with some minor modifications.

1. In section 3, I think it would be more helpful if the authors could provide some comments on the advantages and disadvantages of several ML methods, especially those widely used for CO2-EOR area. As a result, it would be easier for readers to understand why certain methods are used more frequently in the literature.

2. In section 4, I think the authors should explain more about how they give rating score to each literature. It would be better if the authors provided some key points that distinguish a high and a low score paper.

3. In section 5, I was wondering if the authors could provide some suggestions and future directions regarding the limitations of ML so that the paper can be more valuable.

4. It seems that Figure 1 is not mentioned in the main text. Please check.

5. In line 155, it should be Figure 4 instead of Figure 5. Please fix the text.

Author Response

Please see the attachment, thank you!

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

The authors improved the article. However, I still do not find any major contribution in the text. This is a list of existing articles on the given topic. I did not find new conclusions in the article.

Author Response

Thank you for your comments. Please see the attachment.

Author Response File: Author Response.pdf