Research on Performance Prediction of Elbow Inline Pump Based on MSCSO-BP Neural Network

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. The script uses 70% of the 50 data sets as samples to train the network model, with 10% of the sample data used for testing and verification separately. However, it does not explain the use of the remaining 10% of the data. If 10% of the data is used for training, the number of training samples in Figure 13(a) should only be 5, but why is the number of training samples in Figure 13(a) much greater than 5? Similarly, why are the numbers of test samples in Figures 13(b) and 13(c) much greater than 5?
2. The interpretation of Figure 12 mentions that the determination coefficient (R-squared) of the MSCSO-BP prediction model, whether for the training set, test set, or verification set, exhibits a correlation coefficient R value smaller than that of SCSO and closer to 1. This explanation is incorrect and does not match the content of the image.
3. The article selects DS-C, DS-D, DS-G, and LA of the elbow-shaped bend as input parameters for the prediction model. Please elaborate on the significance of the experimental design. Why can't parameters related to the impeller and other parts of the pump be selected as input parameters?
4. The script uses the efficiency at the pump design point as the prediction target. Why is there no mention of a comparison of efficiency before and after in the result analysis?
5. The explanation of the optimized pressure and velocity contour plots is not convincing. It only briefly analyzes the changes in pressure and velocity gradients, without providing a comprehensive evaluation and verification of the pump's performance after optimization. For example, in Figure 15, the pressure gradient improves, but the maximum velocity in the throat area increases. Could this affect the pump efficiency? Section 5.2 needs more detailed verification and theoretical analysis.
6. The paper uses CFX for numerical simulation of the pump, but there is no detailed explanation of the governing equations for fluid flow, which is necessary in CFD.
7. SST turbulence mode is chosen for simulation, however, the relevant equations for this model should be provided.

Author Response

"Please see the attachment."

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

I've read the manuscript (water-3552745) in detail and my comments are as follows: 

General Comments: 

In the manuscript, authors conduct a research on the optimization of the design and efficiency of vertical inline pumps, which are widely used in marine engineering. The performance of elbow inline pumps are predicted using an improved Sand Cat Swarm Optimization algorithm combined with a Backpropagation neural network. Numerical simulations are conducted on a detailed 3D model of the pump to analyze pump performance. Pump performance tests are obtained experimentally on a closed-loop test bed to validate numerical simulation outcomes. Results show that the optimized elbow inlet design improved velocity and pressure distribution, diminished hydraulic losses, reduced velocity gradients, and enhanced pump efficiency. I believe that the research contributes to the design and optimization of high-efficiency inline pumps, with potential applications in marine engineering and fluid transport systems. The research design is appropriate, but the flow in the paper may be improved. The introduction section provides sufficient background information from the literature. In my humble opinion, the quality of paper may be increased further if the points below are considered/corrected.     

Specific Comments: 

- Line 121-122: "...so unstructured grids were used to perform hexahedral grid generation on the flow field, with local encryption applied to relevant parts of the tongue." Please check the sentence because Figure 2 shows tetrahedral elements in the flow domain.

- Section 2.3. Please show governing equations and report all numerical solution details, such as the discretization methods of the PDEs in the governing equations.  

- Line 169-170: Please report the method to compute 0.4% error between the experimental test and numerical simulation (CFD) performance.

- Line 275-276: "...Using Matlab code was written to train the network model using 70% of the 50 sets of data as samples." Check the meaning of the sentence.

- Line 331-336: The paragraph is too long to understand, so its message is lost. I think this paragraph should be divided in 2-3 sentences. 

- Line 362-366: The paragraph is too long to understand, so its message is lost. I think this paragraph should be divided in at least 2 sentences.

- Line 366-369: Please check the meaning of the sentences and simplify the expression.

- Line 387-391: I think possible reasons may be discussed in 2-3 sentences for future studies.

- Is it really necessary to compare original design and modified model in terms of pressure distribution because the manuscript is already very long. 

- Please include a paragraph to elaborate on the novel contributions this research makes to the existing body of the literature, as well as the implications of its findings for future research endeavors. Additionally, a discussion of the strengths and weaknesses of the proposed optimization method should be provided, along with a comparative analysis of the results in relation to existing studies.

- Please include a paragraph to discuss the capabilities of the MSCSO-BP model regarding its current advantages and limitations, as well as its implications for future research.

Thanks

Author Response

"Please see the attachment."

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

I've read the manuscript REVISED (water-3552745) in detail and my comments are as follows:

Authors made substantial changes on the first version of the manuscript submitted. The manuscript design is much better and the flow in the paper is improved. The manuscript revised is now much more consistent with the current literature.

Thanks