Review Reports - Development of an Adaptive Model for the Rate of Steel Corrosion in a Recirculating Water System

Round 1

Reviewer 1 Report

The manuscript entitled "Adaptive water quality modeling for corrosion rate prediction of refinery circulating water" is scientifically sound and robust, and can be published after addressing the following minor revisions.

In several locations, such as Page 2, Line 93, the citation brackets are placed after the period. Please correct all similar issues.
In Figure 1, the font size is small in some elements, and hard to read.
I recommend changing the title of Figure 1, and remove "A flowsheet..."
P4, L158: Font size issue.
P6, L194: Font size issue.
P8, L266: Font size issue.
In Conclusions, Line 299, in parentheses mention what the six modeling methods have been.
Use past tense throughout the Conclusions section, where you report your findings.
Please use the following article in your manuscript, where appropriate, to enhance the quality of your paper:

Comparison and Optimization of the Performance of Natural Based Non-Conventional Coagulants in a Water Treatment Plant. Journal of Water Supply: Research and Technology-AQUA, 2019, 69(1), 28-38. DOI: 10.2166/aqua.2019.075

Author Response

Journal: Processes (ISSN 2227-9717)

Manuscript ID: processes-1308628

Title: Adaptive water quality modeling for corrosion rate prediction of refinery circulating water

Authors: Xiaochuan Huang , Yan Gao , Ling Zhu , Ge He *

Dear Editor and Reviewers:

We are truly grateful to your critical comments and thoughtful suggestions. Based on your comments and suggestions, we have made careful modifications in the revised manuscript. Please note that in the revised manuscript, all the remarks/changes made according to your suggestions are marked in blue. We also try best to indicate where the changes in the manuscript have been made. Please refer to the point-by-point responses to your comments and suggestions, as listed in the following pages.
We appreciate your effort to help improve our work! Thanks a lot!

Comments and Suggestions for Authors

[Reply]:

Dear reviewer, thank you for your very positive comments of our work.

In several locations, such as Page 2, Line 93, the citation brackets are placed after the period. Please correct all similar issues.

[Reply]:

Thanks for your comments and suggestions. I have been modified all similar issues that the citation brackets are placed after the period.

In Figure 1, the font size is small in some elements, and hard to read.

[Reply]:

Thanks for your comments and suggestions. I have been modified Figure 1.

I recommend changing the title of Figure 1, and remove "A flowsheet..."

[Reply]:

Thanks for your comments and suggestions. I have been modified the title of Figure 1.

“Figure 1 Proposed AIGA-RF water quality modeling methodology”

P4, L158: Font size issue.
P6, L194: Font size issue.
P8, L266: Font size issue.

[Reply]:

Thanks for your comments and suggestions. I have been modified all the font size issue of this paper.

In Conclusions, Line 299, in parentheses mention what the six modeling methods have been.

[Reply]:

Thanks for your comments and suggestions. I have been modified in this paper.

“Comparison with other six modeling methods (AIGA-BP, AIGA-SVR, IGA-RF, GA-RF, PCA-RF, RF) showed that AIGA could better capture the system features under different modeling situations, thereby improving the accuracy of model prediction.”

Use past tense throughout the Conclusions section, where you report your findings.

[Reply]:

Thanks for your comments and suggestions. I have been modified in this paper.

“In the present study, the WQIs were calculated based on WQPs and used as input, and then a machine learning-based corrosion rate prediction modeling method was adopted, which employs AIGA feature selection strategy to automatically select the most suitable WQPs for modeling, and the RF method was used for water quality modeling. Taking the power plant of a petrochemical enterprise in Northwest China as an example, the corrosion rate was predicted by using two-year analysis data of 12 LIMS WQPs from 19 water fields. The results of the case analysis prove the applicability and effectiveness of the model. Comparison with other six modeling methods (AIGA-BP, AIGA-SVR, IGA-RF, GA-RF, PCA-RF, RF) showed that AIGA could better capture the system features under different modeling situations, thereby improved the accuracy of model prediction. Therefore, the establishment of a water quality model for corrosion rate prediction using the AIGA-RF method and with the calculated WQI values as input could be adopted by production enterprises to evaluate the circulating water quality and thus monitor and control water quality in real time.”

Please use the following article in your manuscript, where appropriate, to enhance the quality of your paper:

[Reply]:

Thanks for your comments and suggestions. I have been cited this paper in Line 54 of my manuscript.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors deal with very interesting and necessary in industrial practice issues of water quality control with the possibility of suitable model control with prediction for determining the rate of corrosion attack in heat exchangers and similar devices.

However, in the present article I lack a deeper meaning with regard to the narrowly described experiment, which only generally describes possible advantages, which, however, are not really substantiated in the text.

The title of the article itself should be a prediction of the rate of corrosion attack. The text does not describe the material on which corrosion acts, from which it is difficult to estimate which general model would be suitable for predicting the rate of corrosion and thus causing serious damage.

The introduction and theoretical part of the article are sufficiently described.

The corrosion rate should be described or expressed somewhere in the text based on the model prediction and the experiment used. It is not possible to assess from the findings what corrosion rate is suitable for a given material and a given aqueous environment. Although the model description is sufficiently explained, it brings only general and known laws. After all, different materials have different corrosion rates and this is influenced each time by different parameters and quantities.

I recommend the authors to add a specific and more analyzed experiment to the professional article, from which the findings follow.

Author Response

Journal: Processes (ISSN 2227-9717)

Manuscript ID: processes-1308628

Title: Adaptive water quality modeling for corrosion rate prediction of refinery circulating water

Authors: Xiaochuan Huang , Yan Gao , Ling Zhu , Ge He *

Dear Editor and Reviewers:

Comments and Suggestions for Authors

[Reply]:

Dear reviewer, thank you for your very positive comments of our work. In the revised manuscript, we have tried our best to make it more clear for readers to capture the new information of our work. Thanks!

[Reply]: Thanks for your comments and suggestions. We ignored this important statement when writing this paper. All corrosion rates in this paper are based on carbon steel, and the corrosion coupons of stainless steel and brass are not considered. Therefore, in addition to the modification of the title, the material is clearly stated where the corrosion rate is expressed in the text.

“Adaptive modeling towards the quality of circulating water for the steel corrosion rate prediction”

“Under the same factors, different materials have different corrosion rates. Therefore, in order to unify the benchmark, the corrosion coupon material we selected is carbon steel.”

The introduction and theoretical part of the article are sufficiently described.

[Reply]:

thank you for your very positive comments of our work.

[Reply]: Thanks for your comments and suggestions. We have already been added the text about corrosion materials. Thank you again for this suggestion.

“Under the same factors, different materials have different corrosion rates. Therefore, in order to unify the benchmark, the corrosion coupon material we selected is carbon steel.”

I recommend the authors to add a specific and more analyzed experiment to the professional article, from which the findings follow.

[Reply]: Thanks for your comments and suggestions.

Firstly, the data of this paper are from the actual industrial experimental analysis data. Secondly, the process of modeling and predicting the corrosion rate of carbon steel is as follows, after data preprocessing and WQI calculation, the time dimension attributes of WQI and corrosion rate were unified on a monthly basis by summing up the WQIs of the current month according to the analysis frequency, finally yielding 271 datasets for subsequent modeling. In the modeling, the dataset was randomly divided using a ratio of 8:2 to obtain a training data matrix with a size of 12×216 and a test data matrix with a size of 12×55. Finally, AIGA-RF modeling was performed. Finally, we get our final conclusion by comparing the accuracy of different models on the test set, that is, 55 actual industrial experimental analysis data. This is also the common mode of data science machine learning algorithm, that is, learning based on a large amount of data and characterizing the prediction performance of the test set to illustrate the effect of the model.

In order to clarify this point, we have revised the corresponding content in my manuscript.

“After data preprocessing and WQI calculation, the time dimension attributes of WQI and corrosion rate were unified on a monthly basis by summing up the WQIs of the current month according to the analysis frequency, finally yielding 271 datasets for subsequent modeling. In the modeling, the dataset was randomly divided using a ratio of 8:2 to obtain a training data matrix with a size of 12×216 and a test data matrix with a size of 12×55. Finally, AIGA-RF modeling was performed.”

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

I thank the authors for supplementing and specifying their contribution according to the attached review of the article. I do not have any further additions or modifications.

Author Response

Dear reviewer, we are very grateful to your positive comment and careful suggestions. On the issue of "English language and style", we once again asked professional English editing service company to polish it and provided "certificate of English language editing".

Thanks for you again！

Author Response File: Author Response.pdf