Research on the Migration and Settlement Laws of Backflow Proppants after Fracturing Tight Sandstone

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear authors, I have read this text carefully, but unfortunately I must admit that it is poorly written.

There is a lot of chaos, mixed theses and some abbreviated information without providing sources. Finally, in the results we have some information about how the neural network was programmed, and then graphs showing how great everything works and... that's all, there is no data showing various aspects of fracking and substance flow that we could compare with the prediction based on the neural network. Then the article ends with conclusions that are the authors' wish rather than the result of their work because they are not included in the text. I'm very sorry, but in my opinion this text is not suitable for printing, it requires a fundamental change, also in the definition of where the goals should be described, where the theory should be described, etc. Moreover, there are no real results that could be discussed, the chart is arranged in such a way that it follows that the model works in over 90% is too laconic information to judge the validity of the authors' theses on this basis. There are no case studies and no real conclusions drawn from them.

Below I am sending other comments, going from the beginning to the end of the text.

You can't write like that in an international magazine: "The current research status on the transport and settlement laws of proppants both domestically and internationally shows a sustained and diverse trend. In terms of research on settlement laws, scholars at home and abroad have conducted a large amount of ex-perimental and theoretical derivation work.” There are more such expressions in this text.

When you hover your mouse over the images, Chinese alternate description characters appear. This needs to change.

The given formulas do not always have references to the literature or are therefore the result of the authors' calculations, but this must be indicated in the text or it has been forgotten.

In the explanations to formula (1), the acceleration due to gravity is marked with a capital letter G, and in the formula with a lowercase letter.

The presented literature review is laconic, it does not indicate the need for further research, the authors do not show any scope for further research in the data quoted in this paragraph, but emphasize that they believe that this issue will be further developed in the future. This is not a scientific approach.

The article contains not all of 9 pages (not including literature) and the theory lecture takes up 3 pages, which constitutes 30% of the text. Apart from some brief information, it is not clear why the authors cite these relationships. However, the authors did not write anything about the conditions for training the neural networks, what the structure of the algorithm is, etc. On page 7 (out of 10) we learn about the purpose of the manuscript, this should have been included in the introduction.

Below is information about neural networks, this should be transferred to the methodology.

Fig. 2. It is unnecessary, it contains text that is described earlier anyway.

Fig. 3 is constructed in such a way that the graph corresponds to the function y=x, which is to prove that the model fits, and in addition, all 4 parameters are presented in the form of 4 separate graphs. If so, you can instead determine the model's % fit or construct graphs to show the model's shortcomings at a level that can be discussed. This is poorly selected graphics.

And after introducing the way the neural network works, illustrating some sample data and showing the model fit, the authors should now basically show how this network works in various non-standard cases. Meanwhile, at this stage the article ends, the "results" chapter has the following sentence: "Through continuous correction of multi-layer net-works, the mapping logic between multiple engineering parameters and transport laws is fitted, and its effectiveness is verified through actual operational data .”, which is supposed to mean a priori, because the results end with this sentence!

The theses contained in the conclusions are rather declarations of the authors, but they are not supported by actual results, which are not included in the text.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The article examines the migration and settlement laws of backflow proppants after fracturing tight sandstone. It proposes a fitting method based on a multi-task learning network (MTLN) to address interference from various physical parameters during the transport and settlement process of proppants. The paper introduces the characteristics of tight sandstone reservoirs, analyzes the sedimentation rate model of proppants, and the equilibrium height of proppants under multiple interference factors. Using actual production data, the proposed model demonstrates the ability to fit the input-output relationship effectively, supporting the study of proppant transport and settlement laws.

To provide a solid foundation for the calibration methodology in your study, you can refer to the work of Lupo et al., "Calibration of DEM for Cohesive Particles in the SLS Powder Spreading Process. 10.3390/pr9101715" This paper details a comprehensive DEM calibration procedure that includes the evaluation of interfacial adhesive surface energy and the coefficient of rolling friction, which are critical for simulating the behavior of cohesive particles accurately. By adopting similar approaches and referencing their findings, you can align your methodology with established practices and provide a robust comparison for your results.

Consideration of Field-Specific Parameters:

Suggestion: Include additional operational parameters specific to the field, such as injection pressure and the chemical composition of the fracturing fluid, to enhance model accuracy.

Reference: "During on-site construction, the construction displacement, time, sand ratio, and total operation time are often determined based on the technical personnel's reference to the fracturing pump injection program table" (line 177).

Wider Scale Validation:

Suggestion: Extend the model validation across a broader range of reservoirs with varying geological characteristics to ensure robustness.

Reference: "This article trained and tested the model using actual production data" (line 153).

Optimization of Model Parameters:

Suggestion: Use advanced optimization techniques to calibrate the MTLN model parameters, thereby improving the fitting accuracy.

Reference: "The hidden layer can be determined by c = √a + b + qL" (line 175).

Consideration of Thermal Effects:

Suggestion: Integrate the effect of temperature on the fracturing fluid's viscosity and the proppants' settling velocity.

Reference: "The viscosity of the fluid in the fracturing fluid can also affect the settling speed of the proppant" (line 292).

Improvement of Data Visualization:

Suggestion: Implement advanced visualization tools to represent the relationships between operational parameters and model outcomes, facilitating data interpretation by engineers.

Reference: "Figure 3 shows the training fitting effect of the proposed MTLN algorithm" (line 343).

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Title:

It is not coherent with what is done in the article, it must be specific in terms of what is being done, objectives, method, etc.

Abstract:

lines 5, 14, 18 indicate assumptions.

Keywords:

They are not coherent with the variables studied, they must be specific and address what has been investigated.

Introduction and literature review:

It is suggested to write the text following a structure, since the ideas are shown in a disorderly way.

Lack of citations and references, these should be shown in the text and help to support the ideas that are presented.

Materials and methods

A formula is presented, is this the methodological basis?, how is it inferred or deduced?, where is the methodology of reservoir analysis, sedimentation rate model analysis, multitask learning network, model training and testing described?

The text must be descriptive as to what was developed but not inferential or hypothetical: I quote "... In accordance with the derivation of formulas (2) and (3), we believe that the principals..."

The methodology used is not described, it is mentioned that the research consisted of several parts and each of them corresponds to a different field, it is necessary to mention each of them and how it is addressed and articulated in terms of a general methodological structure.

Results

It is suggested to show formulas outside the text

Conclusions:

They are not coherent with the methodology, nor the variables studied, the aforementioned information is not addressed and others that do not correspond to the initial mentioned are inferred.

Bibliography:

It is suggested to follow some style of reference corresponding to the text where the idea is mentioned or supported.

See previous comments. 

Comments on the Quality of English Language

See comments made to the authors

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have made the minimum requested improvements.