Acoustic Transmission Characteristics and Model Prediction of Upper and Lower Completion Pipe Strings for Test Production of Natural Gas Hydrate

Round 1

Reviewer 1 Report (New Reviewer)

Comments and Suggestions for Authors

Table 1 provides a helpful comparative overview between previous studies and the current work on acoustic transmission in pipe strings. However, the table would be strengthened by explicitly including a column or final row that clearly highlights the novel contributions and methodological innovations proposed in this paper. While the Current studies column outlines the tools and context, it does not sufficiently isolate what is new or original in this particular work beyond the integration of existing methods (COMSOL, LightGBM, NSGA-II). A summary of the unique approach, insights, or results proposed by the authors such as the optimisation framework, parameter sensitivity analysis, or specific design recommendations—would help distinguish this study from prior literature and better support the novelty claims.

Additionally, I recommend that the authors include explicit citations in Table 1 to support the claims made under the Previous studies column. At present, the table provides generalised summaries of past research without referencing specific sources, which makes it difficult for readers to verify the comparisons or consult the relevant works. Including references—either as inline citations in the table or footnotes—would significantly improve the scientific rigour and traceability of the comparative analysis. It would also allow the authors to more clearly position their contribution in relation to existing literature.
----------------------------------------------------------------------------------------
The definitions and formulas presented in Section 2.2.2 for evaluating acoustic transmission characteristics particularly the expressions for sound pressure amplitude ratio, tubing displacement ratio, and transmission loss—are standard in the field of acoustic and structural analysis. As such, I strongly recommend that the authors provide appropriate references to support these equations. Citing relevant literature will clarify that these formulations are not novel contributions of the paper but established metrics, and will enhance the scientific rigour and transparency of the methodology.
----------------------------------------------------------------------------------------
In Section 2.4.1, the notation (n = 30) is included in the title, but its meaning is not clearly explained in the accompanying text. I recommend that the authors explicitly clarify what n = 30 refers to. Does it indicate the number of simulated frequencies, test cases, mesh nodes, or experimental repetitions? Providing this clarification is essential to ensure transparency in the experimental design and to allow readers to correctly interpret the scope and reliability of the results presented in this section.
---------------------------------------------------------------------------------------
Figure 12 presents a basic flowchart of the LightGBM algorithm, which appears to depict the standard training process of the method. Given that LightGBM is a well-established algorithm introduced by Ke et al. (2017), the authors should clearly state that this flowchart is a summarised representation of the known methodology and not an original contribution. Additionally, I recommend citing the foundational reference for LightGBM in the figure caption or the main text to ensure proper attribution and avoid any misunderstanding about the novelty of the algorithm itself.
---------------------------------------------------------------------------------------
The accompanying description of Table 4 is helpful in presenting the R² scores and LightGBM parameter settings. However, I recommend several clarifications and improvements to strengthen the scientific validity of this section. Firstly, the term model coefficients R² is somewhat misleading—R² is a model evaluation metric, not a model coefficient. The authors should revise this phrasing for clarity. Secondly, while R² values of 0.8879 and 0.8508 are promising, the analysis should include at least one complementary error metric such as RMSE or MAE to assess predictive performance in real units. Thirdly, the text refers to interpretation of coefficients and extent of the effects of independent variables, but LightGBM is an ensemble of decision trees that does not provide interpretable coefficients in the conventional sense. The authors might consider discussing feature importance rankings instead. Lastly, it would be beneficial to explain how the optimal parameters were selected (search method and validation approach), and to define the variables used as input features—including their physical meanings and units—to make the results more reproducible and interpretable.
---------------------------------------------------------------------------------------

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report (New Reviewer)

Comments and Suggestions for Authors

This paper presents a detailed study integrating numerical simulation, machine learning (LightGBM), and multi-objective optimization (NSGA-II) to evaluate and enhance acoustic signal transmission in downhole pipe strings. The topic is technically relevant and potentially valuable for specialized applications in hydrate reservoir monitoring. The work is promising; however, I suggest the following revisions be addressed before the paper can be considered for publication:

• While the integration of COMSOL simulation, LightGBM prediction, and NSGA-II optimization is technically interesting, the manuscript would benefit from a clearer statement of its novel contribution beyond prior work. Please elaborate in the Introduction or Discussion how your study significantly extends or differs from existing approaches to downhole acoustic modeling.
• The simplification of the fluid domain as a homogeneous gas-water mixture and the adoption of a 2D axisymmetric model are understandable for feasibility. However, it would help readers if you could discuss the potential implications or limitations of these assumptions, especially given the complexity of multiphase flow and downhole conditions.
• The application to hydrate production is clear, but the practical implications could be better emphasized. The results and conclusions are not entirely convincing or well-supported. There is no experimental or field validation of the simulation findings – all results are from modeling, so their real-world validity is unproven. Consider including a brief section or paragraph discussing how your findings might be implemented in actual field operations or how they compare in value to existing monitoring technologies.
• Since the results rely heavily on simulation and prediction, it would strengthen the work to briefly comment on any available field data or prior studies that could qualitatively validate your findings. If none exist, a sentence acknowledging this limitation would suffice.
• The manuscript provides a comprehensive list of prior works, but the critical comparison with recent studies could be made more explicit. Consider referencing and discussing specific differences with recent works on acoustic telemetry using machine learning or on multi-layer completion modeling to better position your contribution.
• Line 131,” is illusrrated in Fig. 1” change to illustrated
• Fig 7 和 8 better centered.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report (New Reviewer)

Comments and Suggestions for Authors

Review

With the advancement of hard-to-recover and offshore resource extraction, including gas hydrates, the importance of reliable downhole-to-surface data transmission increases. This article presents a comprehensive study focused on numerical modeling and optimization of acoustic transmission in multilayer tubular strings for natural gas hydrate production monitoring. The work integrates finite element analysis (COMSOL), machine learning (LightGBM for predicting acoustic characteristics), and multi-objective optimization (NSGA-II optimizing for: transmission range; attenuation; amplitude ratio). The article presents results analyzing the influence of frequency, coupling geometry, and tubular string length on acoustic signal attenuation and transmission efficiency. The pipes are modeled as multilayer systems with fluid media, making this research highly relevant.

Remarks:

1. While the article analyzes acoustic transmission in a multilayer pipe system, it does not consider the impact of pipe inhomogeneities and defects, such as corrosion, threaded coupling wear, pump-induced vibration, or cement ring presence, which could significantly affect signal transmission. The authors should evaluate incorporating such factors via parametric perturbations or supplementing simulations with experimental studies on prototypes featuring defects.
2. Transmission characteristics prediction employs a LightGBM model, yet the analysis lacks feature importance results influencing the outcome. The authors should add a feature importance diagram to identify the most critical parameters of the pipe system (e.g., frequency, coupling length, number of cascades) and enhance model interpretability for engineering practice.
3. The work proposes a simulation model for studying acoustic transmission in tubular strings and presents computational experiment results. However, it lacks information on the experimental design plan, such as the methodology exemplified in: Oleg Bezyukov, Dmitry Pervukhin, Dmitry Tukeev. Experimental Study Results Processing Method for the Marine Diesel Engines Vibration Activity Caused by the Cylinder-Piston Group Operations. Inventions 2023, 8(3), 71;  https://doi.org/10.3390/inventions8030071.
4. The article uses a simplified gas-water mixture model based on macroscopic parameters, neglecting microstructural effects like gas bubble dynamics, turbulence, and medium heterogeneity. This may lead to inaccuracies in attenuation and sound speed calculations. Consequently, the authors should state in the conclusions that future work should employ multiphase CFD models and experimental calibration of acoustic parameters for specific gas/water ratios.
5. The influence of environmental parameters—such as temperature, pressure, and fluid composition changes (gas/water)—is not accounted for in the numerical model, despite their potential impact on acoustic properties. The authors should assess extending the model to incorporate multicomponent media and temperature-dependent characteristics of materials and fluids, particularly when simulating real deepwater drilling conditions.
6. The authors should clarify the rationale for selecting the acoustic frequency range boundaries (20–2000 Hz). This range covers only part of the spectrum used in modern acoustic telemetry systems (e.g., ultrasonic methods extend up to 20 kHz). This limitation may stem from the dominance of low frequencies in long pipelines due to lower attenuation and COMSOL constraints for high-frequency modeling (requiring finer meshing and greater computational resources). Therefore, hybrid methods would be advisable to extend the operational acoustic frequency range.

 

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

Please add structure of paper at end of introduction section.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1.  Error! Reference source not found., please elaborate this
2.  please put all ref. for equations and tables.
3. please put latest ref from 2020 onwords.
4. in introduction part a summary of table can be made to compare works done.
5. at the end of introduction part organization of paper must be presented.
6. what are model coefficients in table 3.
7. in 3.2.2 what is Pareto optimal solution
8. conclusions seem ok

Author Response

Please check the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The paper proposed for review discusses the potential of numerical modeling in tubing columns. The authors outline the purpose of utilizing such devices and identify the challenges related to acoustic emission investigations in tubing columns. The research study includes the construction of numerical models and the application of optimization algorithms to enhance predictive performance.

While the proposed topic appears interesting and novel, the current state of the paper does not warrant consideration for publication due to significant deficiencies in its scientific format, content preparation, and overall language quality. The paper is excessively long, with substantial sections that could be condensed as they are not directly relevant to the primary topic. The language is poor, and the authors have not provided a significant amount of editing work, as evidenced by numerous formatting errors in the references, including multiple instances of "Error! Reference not found" present in the text. Additionally, the text does not comply with the layout requirements of the journal.

Due to these considerable concerns and general remarks regarding the quality of the text, I have decided not to provide a comprehensive review or detailed comments, as numerous and broad corrections are needed before reconsideration. Below are my general remarks:

1. The content of the manuscript should be condensed to focus on the proposed topic while avoiding extensive descriptions of the device and a lengthy introduction that is not directly connected to the subject. Consider using appendices to relocate less relevant sections, thereby maintaining a strict and focused main body of the paper.

2. The language throughout the paper needs to be revised, as it is currently difficult to understand the sentences presented by the authors.

3. The description of the proposed FEM model is fragmented and does not enable direct reproduction, which is required for scientific papers of this nature. Please revise the model description to include details such as the exact options used in COMSOL to simulate the entire environment. This is particularly essential since the FEM model is the main focus of the paper.

4. Many plots presented in the paper hinder separate result analysis due to an abundance of data curves and their overlaps. The authors should provide a data presentation method that facilitates analysis, which is currently not possible. For references, see Figures 9 and 12. The plots need to be revised to ensure correct axis labeling and an appropriate font size that allows for content recognition; as they stand, most are too small to be analyzed effectively.

5. The 22 pages of the paper are summarized in half a page of conclusions, reinforcing my first point that significant portions of the paper are irrelevant. Please modify the content and adjust the conclusions to reflect the actual findings, highlighting the most important results of the research.

Comments on the Quality of English Language

text need to be almost entirely rewritten

see the main review text

Author Response

Please check the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

This study presents a relevant topic; however, it requires major revisions in different areas, including the abstract, methodology, literature review, data processing, and model validation. Therefore, a major revision is required to address the following comments:

1. The abstract should concisely present the main findings, including quantitative results and conclusions. However, the current abstract focuses only on the methodology. Please revise it significantly to reflect the key outcomes of the study.
2. Lines 27-28: What are the input parameters used for predicting the acoustic transmission characteristics curve of the column? Please specify.
3. Lines 29-32: Given that multi-objective optimization generates a set of optimal Pareto front solutions, how is a single optimal solution selected? Please clarify the decision-making process.
4. Why was NSGA-II chosen despite the availability of more advanced variants of this algorithm? Please justify the selection with supporting references.
5. Several references are incorrectly cited, displaying "Error! Reference source not found." Please correct these citations.
6. The numerical model must be properly calibrated and validated against experimental results. The validity and reliability of the results depend on rigorous model validation. Please provide details on how this was conducted.
7. The literature review lacks a thorough discussion on the application of machine learning in combination with multi-objective optimization (MOO) algorithms. While studies on ML applications exist, the specific focus on ML and MOO is limited. Please expand the literature review by incorporating relevant studies in this area.
8. The methodology has critical flaws that require substantial revisions. Please address these concerns in detail.
9. The explanation of the NSGA-II algorithm in Section 3.2.1 needs substantial revision to provide a more comprehensive overview.
10. The study lacks a clear explanation of the decision-making process in multi-objective optimization scenarios. Please refer to Section 2.4 and Section 3.
11. The study selects only LightGBM for modeling. However, prior research emphasizes the importance of evaluating multiple machine learning algorithms, from simple to complex models, to identify the most accurate and reliable approach. Please justify why only LightGBM was chosen and compare its performance with other algorithms.
12. The discussion on the boosting algorithm in Section 3 needs corrections. LightGBM is based on boosting; however, the description in the paper does not adequately explain this. Please revise accordingly.
13. Proper data preprocessing is a crucial step in machine learning model development, as the reliability of the model depends on data quality. However, this study does not provide sufficient details on data processing. Please include a thorough discussion on data preprocessing steps and provide figures illustrating the dataset.
14. The study does not detail the hyperparameter optimization process, which is a critical step in developing machine learning models. Systematic optimization using techniques like grid search, random search, or Optuna should be included.
15. The results clearly demonstrate the inadequacy of the developed model and its flaw in its methodology.
16. Has the study conducted any reliability analysis for design applications? If not, please discuss its relevance and potential implications.
17. Since the study does not deploy the developed model as a graphical user interface (GUI) tool, how can other researchers or engineers use it? The authors should at least reference studies that deploy ML models as practical tools and acknowledge this limitation.
18. The conclusion needs significant improvement. It should briefly summarize the study’s significance, novelty, objectives, methodology, and key findings, including quantitative results.
19. The study should explicitly discuss its limitations and provide recommendations for future research directions.
20. The database and developed machine learning algorithms should be made publicly available in a GitHub repository for transparency and to facilitate further research.

Author Response

Please check the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors submitted a reviewed version of their manuscript. Unfortunately, despite their claims in the response to the review, they failed to address most of my comments. Therefore, I support the decision to reject the manuscript due to a lack of scientific content, poor organization, and insufficient rigor in its preparation. Below are some specific remarks:
1. The authors stated that they revised the entire paper, which is not true. Typographical errors remain, such as "error reference not found" in line 124 and line 146, among others. Additionally, the language still suffers from poor formatting.
2. The "novelty proof" provided in Table 1 is very brief and definitely insufficient.
3. The revised model description does not allow for the recreation of the study and is inadequate.
4. Figures 8-10 are still impossible to analyze due to overlapping curves and poor preparation.

Author Response

Please see the attachment.

Author Response File: Author Response.docx