An AI-Extended Prediction of Erosion-Corrosion Degradation of API 5L X65 Steel
Round 1
Reviewer 1 Report
This manuscript presents an approach to erosion-corrosion predictions utilising Artificial Neural Networks (ANNs), adeptly addressing the prevailing practical challenges in the field. The method and results are lucidly articulated, demonstrating contributions to the subject. However, there are specific areas where targeted enhancements could further elevate the overall quality and impact of the manuscript:
1. On Page 3, Lines 69-72, the manuscript introduces the challenges inherent in new AI-based techniques. A more nuanced explanation of how ANNs address these problems would strengthen this section. Specifically, consider moving the content from Lines 77-80 to follow Line 72 to create a more cohesive narrative.
2. In equations (3) and (4), it would enhance the reader's understanding to include more specific definitions and units for the variables involved. Providing this context will ensure clarity and precision in the mathematical formulation.
3. In Page 13, Lines 356-358, the process of combining experimental and synthetic data is mentioned but requires more detailed elaboration. An explicit explanation of the methodology and rationale for this integration will help readers to appreciate the approach's complexity and novelty.
4. The commentary on Figure 13 may benefit from more depth and specificity. Elucidating the insights derived from the figure with detailed analysis and interpretation would enrich the overall understanding of the relevant findings. This would contribute to a more comprehensive and engaging presentation of the results.
Author Response
Please see the attachment
Author Response File: 
Author Response.docx
Reviewer 2 Report
Comments to the Author
Title: AN AI-EXTENDED PREDICTION OF EROSION-CORROSION DEGRADATION OF API 5L X65 STEEL
Comments: Because of the apparently unpredictable nature of the wear caused by E-C, it is considered mandatory to develop a prediction tool capable of estimating the wear rate of materials exposed to slurry flow. In this paper, a novel ANN method is proposed, structured to a small training dataset and trained with synthetic data to produce E-C neural network, which can make better predictions. Therefore. I recommend the manuscript be published after minor revision.
1. Why these 6 parameters (flow velocity, particle content, temperature, pH, content
of dissolved oxygen, and content of copper ions) were chosen for the ANN method over other parameters in the erosion and corrosion data, and what is the basis?
2. The dataset is based on the results of others, where is the innovation of this article?
3. The corrosion or erosion time in the experiment is 75 minutes, why is this time uniformly chosen, does it have any impact on the final result?
4. How to verify the validity of the experimental data obtained by the interpolation of the polynomial model method?
5. The very latest references about the wear-resistant alloys, mechanisms and computations (e.g., Composites Part B 263 (2023) 110833; Friction 10(11): 1913-1926 (2022) may be of interest to the authors and of relevance to this work for citations.
6. What are the implications of this study for industrial applications?
Moderate editing of English language required
Author Response
Please see the attachment
Author Response File: Author Response.docx
Reviewer 3 Report
In this paper, an Artificial Neural Network (ANN) method is proposed to predict the wear rates of erosion-corrosion (E-C). The superiority of E-C neural network to predict the wear rates are demonstrated by considering individual parameters and their interactions.
Some minor revisions, motivated by the comments below, are necessary to improve the quality of the paper and to meet the journal criteria.
1. “Abstract”
The abstract lacks concluding remarks, and it is suggested that they should be added.
2. “Figure 3 and Figure 13”
It is recommended that Figures 3(A), 3(B), 13(A) and 13(B) be labelled.
3. “Page 5, Line 140”
“…, whereas the high levels of dissolved oxygen correspond to fully oxygenated and oxygen-free electrolyte, …”
It is recommended to check whether the formulation here is correct.
4. “page 6, line 179, 180”
“The optimal architecture was determined using experimental data, as schematized in Figure 6.”
Against the figure, please check if the description here is incorrect.
5. “Page 10, Line 283”
“The residual standard error of this model is 1.328mg·cm-2h-1, the multiple R-squared value is 0.8438, and the adjusted R-squared value, accounting for the number of predictors and degrees of freedom, was calculated to be 0.8276.”
It is proposed to check whether “1.328 mg·cm-2h-1” is the content of “3.2 MFA model of stand-alone erosion wear”.
How the multiple R-squared value and the adjusted R-squared value were derived, it is recommended that the procedure be given in the text.
6. “Page 11, Line 297,298”
“The variability of the corrosion rate at the central levels as assessed by the standard deviation to be approximately 1.292 mg·cm-2h-1, …”
This is inconsistent with the description of “…, stand-alone erosion and stand-alone corrosion to be 0.078, 0.173 and 1.328 mg·cm-2h-1, respectively.” on page 4, please check it.
7. “Page 16, Line 470”
“In particular, the synergy of high ph(11) and absence of Cu2+(0 ppm) predicted by MFA is overestimated, …”
There is no increase in ph(11) in the text and it does not match the description in Figure 13, please check it.
Minor editing of English language required
Author Response
Please see the attachment
Author Response File: Author Response.docx
