Application of Physics-Informed Neural Networks to River Silting Simulation

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1.The topic of this paper is the application of neural network based on physical information in river sedimentation simulation. However, this paper only uses PINN method and numerical simulation method to estimate the velocity, pressure and density indexes in water environment for comparative analysis. The research content is only in the scope of hydrodynamic research and cannot directly reflect the prediction content of river siltation.The topic are overstated.

2.The difference between the calculation results of velocity and pressure may also be due to the calculation difference caused by the one-dimensional compressible Euler model used in PINN and the two-dimensional Euler equation model used in numerical simulation. Please focus on the difference between model selection and calculation results.

 

3.There is too much basic conventional knowledge in the article, and there are a large number of research backgrounds directly related to the research in this paper. At the same time, the implicit solution and explicit solution in the ANSYS fluid calculation process do not need too much explanation.

4.The structure of the article is somewhat chaotic. The PIINN model calculation in the article is mixed with ANSYS calculation, and the advantages of PINN are not highlighted. It is recommended to reduce the ANSYS solution process and only list the results, highlighting the comparison and advantages of the two model.

5.The main innovation of this paper should be PINN modeling and application, and PINN such as loss function should be highlighted. At the same time, only one calculation model is used in different time, which is difficult to highlight the prediction universality of PINN modeling. It is suggested to add simulation cases.

Comments on the Quality of English Language

The quality of English needs improving to highlight your research objectives.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

In this work, the authors use PINN's to predict the solution of the 1D compressible Euler equations, with application to river pollution due to sediment silting. The main topic of the work is very interesting for the journal readers, the text is adequately written and but the structure can be improved. The introduction section and the literature review are adequate, although I miss some important explanations. The methodology is suitable for the goals of the study but the conclusions are not fully supported by the numerical results. Some important shortcomings can be found in this manuscript and must be addressed compulsorily before this work can be considered for publication.

My major concert is regarding the model application. In the conclusion section, the author claim: "Thanks to the PINN method, numerical results obtained on time can be used for subsequent numerical modeling of silting of river channels and channels." This statement is not supported for the obtained results in any way, mainly because the mathematical model cannot be directly applied to the physical problem. Furthermore, the benchmark test reported is suitable to demonstrate the PINN performance but it has not relation with the physical problem introduced the in title, river silting. The authors must adapt the PINN to the physical problem and a new test related to sediment silting in rivers should be included. From my point of view, this correction is mandatory for the article can be consider for publication.

Other important concerns and typos are listed in the attached pdf file.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

1、Please continue to explain the principle and existing research basis of identifying the occurrence of siltation by density parameters and implicit identification by velocity flow. And explain the innovation of this paper

2、Please continue to explain the predictive universality of PINN modeling. In this paper, only one case is used to analyze whether it is universal, and whether the training set of PINN can accurately predict the occurrence of siltation in other cases. It is suggested to increase the simulation cases in this paper, or further explain the representativeness of the existing cases.

3、The introduction of this article is too lengthy, it is recommended to streamline the content of the introduction again.

4、The table format needs to be revised.

The reviewing process will move forward only when the above issues will have been well addressed.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have adddressed all my previous comments. In my opinion, the manuscript can be published in its current version.

Author Response

The authors are grateful to the Reviewer for the thorough reading of the work and the comments made. His review has significantly improved this article.