Identification of Hydrodynamic Coefficients of the SUBOFF Submarine Using the Bayesian Ridge Regression Model
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsI recommend the publication of this article in its present form.
Why??? Exactly. I’ve been asking myself that question over and over again. This paper has some flaws, which I will explain in detail, but overall, it is a very well-written article. I think the authors used some proofreading service, software, or A.I. The text is so clear and easy to read that I really enjoyed reading it.
I assume other reviewers will have more complaints about this article, but here are a few of mine:
1) Between lines 168 and 209, references to FLOEFD are needed.
2) This article will have very limited interest for the readers. How many researchers and engineers are working on submarine development? Unfortunately, there are many more than we think, but it is still a negligible number compared to the scientists working on renewable energy solutions, climate change, disease prevention, and medicine development. The SUBOFF model developed by DARPA looks very similar to the Los Angeles-class fast-attack submarine. If the researchers used the geometry of a deep-sea submersible like the recently lost "Titan,” one could argue that their research would contribute to the exploration of ocean depths, the discovery of new species, etc. Instead, the only practical application of this research is the development of a torpedo guidance system, in which a Bayesian ridge regression model can be used to defeat countermeasures launched by a targeted submarine, and to determine its actual location.
The main question addressed by the research is the determination of submarine hydrodynamic coefficients using the Bayesian ridge regression model and data obtained by CFD simulations.
This topic is not original since Xue et al. [20] used the Bayesian method to identify the four-degree of freedom hydrodynamic coefficients of surface ships through simulated motion data, and this research would represent an extension of this idea into the 3D space of ocean depths, but this research is still relevant because it facilitates the determination of hydrodynamic coefficients of submarines using CFD simulation without the need for experimental testing. The authors addressed this specific gap in the field by using the Bayesian Ridge Regression Model on data obtained by numerous CFD simulations.
Compared with other published material, this approach enables the prediction of submarine trajectory without CFD simulation, making it suitable for guidance systems.
As stated in the conclusion, in their future work, the authors should verify this methodology using experimental data.
Conclusions are consistent with the evidence and arguments presented in the paper and address the main question addressed by this research; however, a more detailed discussion should be given regarding the results of the Bayesian ridge regression method and CFD simulation and the maximum error of 43%.
References appropriately describe the state of the art in the field of submarine hydrodynamic coefficient determination.
The tables in this paper are well placed and very effectively summarize key data, but in Table 2, the order of grid level should be reversed (a fine grid must have the most cells).
Figures are clear with an appropriate font size.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for Authors- It is necessary to show more abour the process to estimate the coefficients using Bayesian Ridge Regression method. There are only the genenral principles of BRR in the manuscript.
- Long time ago, I performed the identification of hydrodynamic coefficients using extended Kalman Filter method which was done by MIT Abkowitz lab and Estimation-Before-Modeling technique. During the process, so called simultaneous drift phenomenon which is inevitable for ship maneuvering motion made me trouble. How did you avoid this problem ? I think this is the intrinsic problem which is not dependent on the identification tool. Of course, it could be avoided if we use not structural model like neural network,...
- I wrote some comments using a red pen on the manuscript.
Comments for author File: Comments.pdf
Good.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe authors presented an interesting paper on the Identification of Hydrodynamic Coefficients of SUBOFF Using the Bayesian Ridge Regression Model.
The paper is generally well prepared, has good scientific soundness and can be accepted after addressing the following points:
The main quantitative findings are to be mentioned in the abstract.
The introduction is relatively short and may be extended by adding recently published papers.
What is the used CFD software?
The solved governing equations and turbulence equations are to be mentioned.
The boundary conditions are to be expressed mathematically.
The used turbulence model is to be justified.
What are the dimensions of the computational domain? To be justified.
A validation/verification of the numerical model is to be performed by comparing with previously published results.
What is the convergence criterion?
What is the time step?
Information about the performances of the used computer and computational time are to be provided.
The paper is to be checked for misprints and grammatical errors.
Comments on the Quality of English Language
The paper is to be checked for misprints and grammatical errors.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
