Machine Learning Algorithm to Predict CO2 Using a Cement Manufacturing Historic Production Variables Dataset: A Case Study at Union Bridge Plant, Heidelberg Materials, Maryland
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis study uses six machine learning methods to model the different stages of the cement calcination process with the aim of determining the direct link between the carbon dioxide produced during the production of raw materials and process factors. Overall, the paper is interesting and does have novelty. The authors are also advised to consider the following suggestions to further improve the paper quality.
1. It is recommended that authors emphasize the novelty or contribution of this paper in the introduction section.
2. At the end of the "Introduction" section, authors are advised to describe the layout of the following sections.
3. Image quality could be further improved, e.g. sharpness.
4. Normalized RMSE is recommended because RMSE has units and normalized RMSE has no units.
5. There are some typo errors. For example, this paper has Figure 82.
6. How do you think about the points’ distribution, such as in Figure 13? The points are not evenly distributed.
7. It is worth noting that for points with very small values, small differences are still large relative errors. In the predictions, as in Figure 13, I see that many of the predictions have large relative errors, which suggests that the model is not well trained.
Comments on the Quality of English LanguageThe writing needs to be further polished.
Author Response
Please see the attached document as a response to the comments
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe submitted manuscript is about understanding the impact of manufacturing raw materials (cement clinkers) that influence CO2generation in the calcination process by using machine learning tools. I kindly ask authors to prepare a revised version, considering the following comments:
1) It would be interesting to discuss briefly about dissolution process for cement rich in belite clinker as is beneficial solution due to its lower limestone demand, associated CO2 emissions, energy demand in nano and KMC atomistic upscaling approach. The following literatures can help: https://doi.org/10.3390/ma15186388, https://doi.org/10.3390/ma15196716.
2) In table 1 and other sections of the manuscript, the numbers associated with compositions should be subscript.
3) "After Figure 7, Figure 9 is included, while Figure 8 is missing. Furthermore, Figure 82 (is there a total of 82 figures in this manuscript?) is found on page 16, and Figures 10 and 11 are repeated. After Figure 14, Figure 18 is presented, but Figures 15, 16, and 17 are missing. Additionally, Figure 3 appears twice. These errors in figure numbering and presentation are quite unusual and should be addressed for the accurate referencing of figures."
4) The vertical and horizontal labels, titles and legends must be more obvious for easier visibility (please increase fonts or bold them). Such as Figures 3, 11, 82, 10, ….!
5) In Table 2, C3S formation is as result of reaction 3CaO with SiO2!
6) Can you elaborate on the specific machine learning techniques employed in the study and their respective advantages or limitations in modeling the calcination process?
7) What is the significance of the CO2 molecular composition (dry basis) in the study, and how does it relate to the broader goal of reducing carbon emissions in cement production?
8) Could you describe the sensitivity analysis methodology in more detail and provide insights into the selection of input variables from the historical manufacturing health data points?
9) What are the implications of the Root Mean Squared Error (RMSE) in the context of evaluating the accuracy of regression models, and how did the various models perform in this study?
10) How were the independent variables in cement manufacturing identified, and what criteria were used to assess their impact on the dependent variables?
Author Response
Please see the attached with response to the comments
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe identification of manufacturing process parameters/factors is an important topic to guide the reduction of CO2 generation. This work uses massive historical manufacturing health data to train several machine-learning (ML) techniques for this aim. The paper is easy to follow and falls within the scope of this journal. There are some points to improve before a possible recommendation for publication.
(1) The literature review needs to stress the logic and novelty, as the two paragraphs present too much information (mixed), and do not clearly indicate what this paper has done while other studies has not. So the authors need to write four to five paragraphs that include more comparison with other studies, and cite more references to highlight known parameters/factors, traditional methods and ML techniques.
(2) The workflow of the used ML techniques should be provided in terms of preparation of datasets, identification of input and output parameters, random split, and performance indicators, e.g.10.1016/j.jrmge.2022.10.014.
(3) The authors are suggested indicate how they choose ML techniques and add a table of performance metrics for different ML models, considering RMSE, R, etc.
(4) The Conclusions also blend too much information. The first paragraph is too long. The authors need to present the most relevant, important, individual points clearly. Otherwise, potential readers may not get to or agree with “the adoption of digitization, AI, and machine learning will play a key role in the cement industry over time”.
Comments on the Quality of English LanguageRequire minor editing of English language.
Author Response
Please see the attached responses to all the comments
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsPresent form of manuscript can be accepted for publication
Reviewer 3 Report
Comments and Suggestions for AuthorsThe authors have improved the manuscript. It can be accepted for publication in the current form.
Comments on the Quality of English LanguageMinor editing is required in the proofing.
