Infrastructure Lifecycle Corrosion Management Using AI Analytics and Digital Twins

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. General assessment: This article addresses a relevant and timely topic on corrosion management and optimisation of infrastructure lifetime using artificial intelligence (AI) and digital twins. The paper is professionally sound and has an extensive literature base. The authors discuss in detail AI-based corrosion management methods, predictive modelling and decision support systems. However, the article needs further clarification in some areas.
2. Scientific quality and originality: The article illustrates the importance of using AI and digital twins for corrosion management. The novelty value is the detailed analysis of AI-based predictive modelling and the role of digital twins. The authors identify the challenges that the industry faces and demonstrate the benefits of AI-based solutions well. However, further experimental or case study-based results could strengthen the argument.
3. Structure and clarity: The structure of the article is logical and easy to follow. Each section is well structured, and key concepts are clearly explained. Figures and tables support the textual explanations well, but in some places, more precise lists of figures or examples would help improve clarity.
4. Methodological strengths and weaknesses: The authors describe the different data management strategies and analytical tools of AI. However, the description of the methodology would require more technical details, especially regarding the exact parameterisation of the AI models and the selection criteria for the training datasets. An analysis of the validation methods and the reliability of the prediction models would also be useful.
5. Sources and references: The article contains an extensive list of references based on relevant literature. However, some references to more recent publications could help to modernise the context. The inclusion of recent research in the field of AI and digital twins would further strengthen the credibility of the paper.
6. Style and language: The language is academic and well-structured, but in places, a little too technical for non-specialist readers. The explanation of each AI algorithm could be made more understandable with simpler examples or visual illustrations.
7. Suggestions for improvement:
   - Present further case studies or experimental results.
   - Details of the exact parameterisation and validation methods for AI models.
   - Integration of the most recent literature.
   - Add more detailed explanations to the figures and tables.
   - Making some sections more accessible to a wider readership.
8. Final assessment: The article is a valuable contribution to the understanding of the interface between corrosion management and AI. After some minor modifications, it is recommended for publication.

 

Author Response

Reviewer #1 Comments & Responses

Reviewer #1- Comment 1: The authors discuss in detail AI-based corrosion management methods, predictive modelling and decision support systems. However, the article needs further clarification in some areas.

Response: The authors’ responses to specific comments follow here.

Reviewer #1- Comment 2: The authors identify the challenges that the industry faces and demonstrate the benefits of AI-based solutions well. However, further experimental or case study-based results could strengthen the argument.

Response: We acknowledge further experimentation, in addition to that presented in the existing paper, as a fruitful topic for future research in addition to the examples provided in the paper and modification to the paper in the conclusions and recommendations section of the paper Section 4.3.1, Page 15 Lines 547-549 and Section 5.1, Page 16, Lines 620-625

Reviewer #1- Comment 3: Figures and tables support the textual explanations well, but in some places, more precise lists of figures or examples would help improve clarity.

Response: We acknowledge this supporting information in the paper and for future research as provided in Section 5.1, Page 16, Lines 620-625.

Reviewer #1- Comment 4: The authors describe the different data management strategies and analytical tools of AI. However, the description of the methodology would require more technical details, especially regarding the exact parameterisation of the AI models and the selection criteria for the training datasets. An analysis of the validation methods and the reliability of the prediction models would also be useful.

Response: We acknowledge the reviewer's comment and have added further discussion on this topic to include…”AI approaches have divided into those with integrated non-transparent models and those with documented machine learning models.  In both cases, the user must practice due diligence in validating the outputs the same as any engineering model.”  This added discussion is presented and discussed in the comparison of the analytics models discussed in Section 4.3.1, Page 15, Lines 547-549 of the paper.

Reviewer #1- Comment 5: The article contains an extensive list of references based on relevant literature. However, some references to more recent publications could help to modernise the context. The inclusion of recent research in the field of AI and digital twins would further strengthen the credibility of the paper.

Response: We acknowledge the reviewer's comment on recent references and would like to clarify that our analysis includes recent research. Some quick analytics on references in the paper indicate almost 40% are current year (2025), 59% are 3 years or less and 76% are less than 5 years.  The analytics discussion includes a significant number of recent references for the new and rapidly developing AI field applicable to infrastructure management. Two additional recent references are added with discussion in Section 4.3, Page 15, Lines 578-582.  The older references are associated with the legacy life-cycle management approaches, that are still relevant within the context of the overall theme of the paper.

Reviewer #1- Comment 6: The language is academic and well-structured, but in places, a little too technical for non-specialist readers. The explanation of each AI algorithm could be made more understandable with simpler examples or visual illustrations.

Response: We have referenced the AI algorithms for additional details of the specific AI applications discussed in the paper in Section 4.1, Page 11, Lines 442 and 458.  The applicable AI parameters vary greatly from one source and application to another depending on the specific application are presented in modifications of Section 5.2, Page 16, Lines 620-625.

Reviewer #1- Comment 7: Suggestions for improvement:
- Present further case studies or experimental results.
- Details of the exact parameterisation and validation methods for AI models.
- Integration of the most recent literature.
- Add more detailed explanations to the figures and tables.
- Making some sections more accessible to a wider readership.

Response: We acknowledge the reviewer's comments as a summary of prior comments and responses provided individually.

Reviewer 2 Report

Comments and Suggestions for Authors

General Comments

The paper intends to investigate current limitations in the research on corrosion management using AI analytics and Digital Twins, in a way to bridge the gap between the guidance contained in the codes of practice and current research in the literature. In this way, the paper intends to align with the scope of the Special Issue “Applied Infrastructure Corrosion Science for Construction Practice Advancement”.

However, I am afraid that the content of the paper does not seem to conform with the scientific structure of “research review” articles published by MDPI journals or other prestigious scientific journals for that matter.  The literature review on current research is very limited, current limitations on corrosion management are not properly addressed and research questions are not stated. Furthermore, the methods used to find out quality research to answer to the questions are not stated and suggestions for further development are not drawn.

I would suggest that the study should be rearranged in a way to be submitted as a “Project Report” although this might be of the responsibility of the Editors/Guest Editors.

Author Response

Reviewer #2 Comments & Responses

Reviewer #2- Comment 1: The paper intends to investigate current limitations in the research on corrosion management using AI analytics and Digital Twins, in away to bridge the gap between the guidance contained in the codes of practice and current research in the literature. In this way, the paper intends to align with the scope of the Special Issue “Applied Infrastructure Corrosion Science for Construction Practice Advancement”.

However, I am afraid that the content of the paper does not seem to conform with the scientific structure of “research review” articles published by MDPI journals or other prestigious scientific journals for that matter. The literature review on current research is very limited, current limitations on corrosion management are not properly addressed and research questions are not stated. Furthermore, the methods used to find out quality research to answer to the questions are not stated and suggestions for further development are not drawn.

I would suggest that the study should be rearranged in a way to be submitted as a “Project Report” although this might be of the responsibility of the Editors/Guest Editors.

Response: We acknowledge the reviewer's comment and would like to clarify that our paper is written as a focused response to the special issue theme of research benefiting infrastructure life cycle management and from an industry perspective with examples in analytics that bridge AI to Digital Twins (DT) in the context of infrastructure lifecycle management. While there is literature on both AI and DTs primarily in manufacturing, the integration into infrastructure lifecycle management is relatively new and literature sparse. The authors agree the paper type should be a decision by the editors. This stated objective is included in the Conclusions and Recommendations Section 5.0, Page 16, Lines 596-601.

Reviewer 3 Report

Comments and Suggestions for Authors

The paper discusses various types of corrosion databases, including flat files, three-dimensional databases, cloud-based platforms, and digital twin systems. Each format capability to store, analyze, and apply corrosion-related data is evaluated. Digital twins are highlighted as a emerging solution capable of representing real-time operational conditions and enabling predictive maintenance through data simulation and structural analysis. The review also outlines the role of strategic sampling, lifecycle analysis frameworks, and risk-informed maintenance scheduling. The authors stress that while AI and digital twins present opportunities for cost-effective, proactive corrosion management, challenges still remain, particularly when it comes to data availability and integration complexity.

This article needs a major revision before it can be reconsidered for publication. Below are some comments to the authors:

1. How does this review differentiate itself from prior reviews on corrosion monitoring or digital twin integration? I think a more clear articulation of the paper unique contribution would help position its relevance and novelty within the broader literature.
2. The authors need provide a more technical explanation or conceptual framework of how AI analytics are integrated into digital twin systems. Diagrams, model architectures, or system workflows would improve clarity and practical value.
3. To strengthen the real-world applicability of the review, I would like to see it includes case studies (for instance USCG cutters, sewer systems) into more comprehensive analyses with discussions on implementation challenges, benefits, and lessons learned.
4. How are uncertainty and data sparsity addressed when applying AI and Bayesian inference models to corrosion data?
5. The “recommendations” section should be expanded to include more targeted guidance with a roadmap for future research and implementation. For example, it can include suggestions such as open dataset creation, benchmarking models, or pilot programs, which I believe would improve the article impact.

Author Response

Reviewer #3 Comments & Responses

Reviewer #3- Comment 1: How does this review differentiate itself from prior reviews on corrosion monitoring or digital twin integration? I think a more clear articulation of the paper unique contribution would help position its relevance and novelty within the broader literature.

Response: The current paper differentiates itself from other AI and Digital Twins (DT) DT reviews by highlighting the application of analytics to corrosion management in infrastructure.  The majority of AI and DT efforts in the literature (Refs 24-27) are focused on inspections for corrosion as indicated in the paper.  Recent literature added (Refs 30, 31) is sparce on this topic.

Reviewer #3- Comment 2: The authors need provide a more technical explanation or conceptual framework of how AI analytics are integrated into digital twin systems. Diagrams, model architectures, or system workflows would improve clarity and practical value.

Response: We acknowledge the reviewer's concern and have added additional text in the paper Section 5.2, Page 16, Lines 620-625 as further recommendations in this new field of research and that the specifics of execution varies for each application.

Reviewer #3- Comment 3: To strengthen the real-world applicability of the review, I would like to see it includes case studies (for instance USCG cutters, sewer systems) into more comprehensive analyses with discussions on implementation challenges, benefits, and lessons learned.

Response: We have included two examples of analytics are provided in the paper, one human manually developed, and the same application of AI.  Differences in execution are presented in modified Section 4.1, Page 12, Lines 453-458.    

Reviewer #3- Comment 4: How are uncertainty and data sparsity addressed when applying AI and Bayesian inference models to corrosion data?

Response: Bayesian networks and deep learning models, address uncertainty and data sparsity in AI by leveraging probabilistic modeling and regularization techniques. This allows them to not only make predictions but also quantify the uncertainty associated with those predictions, making them suitable for scenarios with limited or noisy data.  The authors have added clarification to the paper in Section 4.2, Page 14, Lines 511-514.

Reviewer #3- Comment 5: The “recommendations” section should be expanded to include more targeted guidance with a roadmap for future research and implementation. For example, it can include suggestions such as open dataset creation, benchmarking models, or pilot programs, which I believe would improve the article impact.

Response: We acknowledge the reviewer's concern and would like to offer these topics as examples for further research and have included this in the paper as modifications in Section 5.2, Page 16, Lines 620-625

 

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors

I understand that it could possibly be other ways to assess the relevance of the paper. However, the revised version of the paper does not address the underlying comments I made in the first round of review. Thus, the suggestion (reject) I made in the first round of review still holds.

 

Author Response

1. Clarification of Paper Type:

 

The manuscript was submitted as an original research article (i.e., an “Article”), but it appears to have been processed and reviewed under the category of a Review paper. This misclassification is central to the misunderstanding surrounding the manuscript’s content and structure.

 

While the paper includes references to relevant literature to contextualize the proposed approach, it does not aim to provide a comprehensive review of the state of the art. Instead, it offers a novel methodology that integrates AI analytics and digital twins into corrosion lifecycle management, supported by technical examples and discussions of implementation strategies.

 

We therefore respectfully request that the manuscript be reclassified as an “Article” rather than a “Review,” which would more accurately reflect its intended contribution and structure.

 

 

Reviewer 3 Report

Comments and Suggestions for Authors

This paper can be published after revision. 

Author Response

Author Response and Request for Reconsideration

Manuscript ID: cmd-3539098

Title: Infrastructure Lifecycle Corrosion Management using AI Analytics and Digital Twins

Authors: Bilal Ayyub, Karl Stambaugh

Submitted to: Corrosion and Materials Degradation

 

Dear Editor,

 

We are grateful for the opportunity to respond to the reviewers' and editorial comments on our manuscript, and we sincerely appreciate the time and effort that have been devoted to evaluating our submission.

 

We respectfully request reconsideration of the editorial decision, which currently recommends a "Reject and encourage resubmission" outcome. Our request is based on the following points of clarification and justification:

 

1. Clarification of Paper Type:

 

The manuscript was submitted as an original research article (i.e., an “Article”), but it appears to have been processed and reviewed under the category of a Review paper. This misclassification is central to the misunderstanding surrounding the manuscript’s content and structure.

 

While the paper includes references to relevant literature to contextualize the proposed approach, it does not aim to provide a comprehensive review of the state of the art. Instead, it offers a novel methodology that integrates AI analytics and digital twins into corrosion lifecycle management, supported by technical examples and discussions of implementation strategies.

 

We therefore respectfully request that the manuscript be reclassified as an “Article” rather than a “Review,” which would more accurately reflect its intended contribution and structure.

 

2. Summary of Reviewer Feedback:

 

We are encouraged by the feedback received from Reviewer 2, who awarded the manuscript a 5-star rating and recommended it for publication without any changes. Reviewer 2's comments indicate a clear understanding of the scope and contribution of the manuscript as a technical article.

 

Reviewer 3, on the other hand, appears to have been confused by the manuscript’s classification. Their comments suggest an expectation aligned with that of a review paper. Nonetheless, we addressed the substantive concerns raised by Reviewer 3 in our prior response, and we believe that their key issues have been resolved given the correct categorization.

 

3. Response to Editorial Comments:

 

The editorial summary appears to mirror the expectations set forth for a review article, including extensive literature surveys, numerous figures and tables summarizing prior works, and an exhaustive critical discussion of trends in AI and digital twins across multiple infrastructure sectors. While these are certainly valuable elements in a review, they fall outside the scope and intent of our paper.

 

To meet those expectations would necessitate a complete rewrite, a substantial expansion of the paper’s length, and a diversion from the core methodology and findings that form the paper’s original contribution. Such a transformation would constitute an entirely different paper.

 

4. Manuscript Status and Resubmission:

 

Given the clarification above, and having already addressed the relevant technical feedback from Reviewer 3, we are submitting the unchanged manuscript once again under the appropriate classification of Article. We believe the manuscript is now appropriately positioned for re-evaluation based on the correct article type.

 

In conclusion, we respectfully ask the editorial team to reconsider the decision in light of the clarification of the manuscript's type and the positive feedback already received. We remain available to make minor editorial or formatting adjustments should they be required under the revised classification.

 

Thank you once again for your time, attention, and the opportunity to contribute to Corrosion and Materials Degradation.

 

Warm regards, Bilal Ayyub and Karl Stambaugh

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

General Comments

I am aware that this research is submitted to the Special Issue “Applied Infrastructure Corrosion Science for Construction Practice Advancement. As I stated in the first round of review, the paper intends to bridge the gap between the guidance contained in the codes of practice and current research in the literature. So, the last version of the paper fits well within the scope of the SI.

However, I am of the opinion that there are some minor points that need to be addressed to improve the quality and relevance of the research.

1. The main objectives of the paper need to be clarified in the “Introduction” section, focusing on how the paper addresses current limitations in the academic research literature;
2. Related to the above, the authors should elaborate, in section 5, on how the stated objectives have been achieved;

• I would also suggest that a paragraph stating the structure of the paper be added at the end of the “Introduction” section.

Detailed Comments

In the “Introduction”, lines, first paragraph, lines 25-37, it is stated that “…The global cost of corrosion on infrastructure [1] is estimated to be around $2.5 trillion annually which represents approximately 3.4% of the world's Gross Domestic Product (GDP)…”  The authors should clarify the year this cost representation refers to;

In the “Introduction”, lines, first paragraph, lines 25-37, it is stated that “…The United States total annual estimated direct cost of corrosion [2] is a staggering $276 billion—approximately 3.1% of the nation’s GDP..”. The authors should clarify the year this cost representation refers to.

Author Response

Dear Editor,

 

We sincerely appreciate the opportunity to submit a revised version of our manuscript for consideration in the journal. Your dedication, along with the reviewers' thoughtful feedback, has been instrumental in improving our work. We are grateful for the constructive suggestions that have led to significant enhancements in our paper. We have carefully addressed and incorporated all the recommendations provided by the reviewers. To assist with your review, we have used track changes within the manuscript. Additionally, we have prepared a detailed point-by-point response to the reviewers’ comments and concerns, which can be found below in blue text. All sections and page numbers correspond to those in the revised manuscript file containing tracked revisions. We are confident that these revisions have strengthened our manuscript, and we eagerly await your and the reviewers’ evaluation.

 

Reviewer #3 Comments & Responses

Reviewer #3- Comment 1: The main objectives of the paper need to be clarified in the “Introduction” section, focusing on how the paper addresses current limitations in the academic research literature.

Response: We acknowledge the reviewer's concern and have added additional text to clarify the objective of the paper (Page 2, Lines 49-58) as follows:

“The objective of this paper is to present new technology for bridging the gap between applicable standards and reality for infrastructure corrosion by monitoring and using the latest technology available for data analysis and decisions for forecasting inspection and maintenance needs. This paper begins with an overview of engineering approaches used to manage and analyze corrosion data along with risk-based approaches used to support decisions for mitigating and managing risk from corrosion failure in infrastructure.  The paper then presents new technology involving DTs, data analytics and Artificial Intelligence (AI) used to support these engineering activities along with an example application of AI in data analytics.  The synergistic benefits of AI and AI analytics are discussed as part of the new research perspectives. Finally, recommendations are provided for further research on this topic. “

 

Reviewer #3- Comment 2: Related to the above, the authors should elaborate, in section 5, on how the stated objectives have been achieved;

 

Response: We acknowledge the reviewer's concern and have added additional text in the paper Section 5.0, Page 16, Lines 603-617 as follows:

“This paper is a focused response to the special issue theme of research benefiting infrastructure life cycle management and presented from an industry perspective. Examples are presented using analytics that bridge AI to Digital Twins in the context of infrastructure lifecycle management. While there is literature on both AI and Digital Twins primarily in manufacturing, the integration of analytics into infrastructure lifecycle management is relatively new and literature sparse.

This focused paper begins with an overview of engineering approaches used to manage and analyze corrosion data along with risk-based approaches used to support decisions.  This paper then presents new technology involving the synergistic combination of Digital Twins, data analytics and Artificial Intelligence (AI) used together to support these engineering activities, risk mitigation and decision support. An example AI application in data analytics is presented for corrosion in ship structure is new research. The benefits of AI and AI analytics are discussed in context of necessary resources to implement them. Recommendations are provided here for further research for continued development of the emerging technologies to mitigate and manage the risks from corrosion in infrastructure.”

 

Reviewer #3- Comment 3: In the “Introduction”, lines, first paragraph, lines 25-37, it is stated that “…The global cost of corrosion on infrastructure [1] is estimated to be around $2.5 trillion annually which represents approximately 3.4% of the world's Gross Domestic Product (GDP)…”  The authors should clarify the year this cost representation refers to;

Response: We have included the date of data referenced by NACE international [1] as 2017 and have modified the text in Section 1, Page 1, Line 36.   The data referenced is the most recent available on the NACE International website indicating the relevance of needed research in this area.

 

 

Reviewer #3- Comment 4: In the “Introduction”, lines, first paragraph, lines 25-37, it is stated that “…The United States total annual estimated direct cost of corrosion [2] is a staggering $276 billion—approximately 3.1% of the nation’s GDP..”. The authors should clarify the year this cost representation refers to.

Response: We have included the date of data referenced by NACE [2] as 2002 and have modified the text in Section 1, Page 1, Line 40    The data referenced is the most recent available on the NACE website indicating the relevance of needed research in this area.

 

Author Response File: Author Response.docx