Evaluation of Machine Learning Models for Enhancing Sustainability in Additive Manufacturing
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsMachine Learning Models are the useful tools for manufacturing, including additive manufacturing. This article lets me feel the theoretical methodology is occupied in stead of the experiments. I am wondering that are there any experiments to verify the points from the article?
Comments on the Quality of English LanguageThe language is acceptable. But it needs further improvement.
Author Response
Comment 1:
Machine Learning Models are the useful tools for manufacturing, including additive manufacturing. This article lets me feel the theoretical methodology is occupied in stead of the experiments. I am wondering that are there any experiments to verify the points from the article?
Response:
We thank the reviewer for the insightful comment. We would like to clarify that our work is grounded in experimentally obtained data. The machine learning models (ML) are applied after conducting physical experiments, not as a substitute for experimental validation.
The primary objective of this work, as specified in the updated Section 1 (lines 90–99), is to highlight and model the underlying relationships between Additive Manufacturing (AM) process parameters and key sustainability metrics, such as energy consumption, printing time, part weight, and scrap weight, by using ML. These relationships are often complex and not easily discernible through experimental analysis alone, making ML a valuable tool for uncovering and quantifying them.
By first performing experiments and then applying ML, we ensure that the models are developed and tested based on real-world data, thereby enhancing both the credibility and practical relevance of our findings.
Reviewer 2 Report
Comments and Suggestions for AuthorsResearch Objective:
This study applies machine learning to optimize sustainability in additive manufacturing by predicting energy use, material waste, and production time based on key process parameters. Among tested models, Random Forest showed the best performance. The research offers an efficient method to link process settings with environmental and economic outcomes.
Originality and Innovation:
This research introduces a novel integration of advanced machine learning models with additive manufacturing process optimization for sustainability. It uniquely applies the L-BFGS-B algorithm for efficient hyperparameter tuning, demonstrating superior performance over conventional methods. The study also offers a scalable framework directly linking process parameters to sustainability outcomes—bridging a key gap in current literature.
It is worth emphasizing that the manuscript combines two emerging digital technologies — Machine Learning and Additive Manufacturing — which aligns well with the scope of the journal Technologies.
Pros:
+ The research problem is clearly and appropriately defined (Section 1. Introduction)
+ The literature review is comprehensive and well-aligned with the research problem and subject (Section 2. Literature Review)
+ The research methodology is clearly structured and easy to follow (Section 3. Methodology)
+ The analysis of research results is thorough and detailed
+ The reference list is extensive, citing 36 publications from the last 5 years (68%) and 46 from the past decade (90%)
Cons:
- The research objective is either missing or placed inappropriately. If the passage “While specific models like Regression [30], Classification [31], and Neural Networks (NN) [26, 27] have been employed, the overarching goal remains the same: optimizing AM processes to reduce waste and energy consumption” refers to the objective of this study, it is defined in the wrong section (lines 155–157)
- The manuscript lacks clear and specific conclusions derived from the conducted research
Should be clarified:
• It is recommended to explicitly define the research objective and place it in Section 1. Introduction, preferably at the beginning or end of the section. Another suitable location could be Section 3. Methodology, as the research objective is an integral part of the scientific research methodology.
• Since Section 2. Literature Review is intended as a review of existing literature, it is unclear why no citations are included
• Adding concrete and in-depth conclusions (e.g., in Section 5. Conclusions) would significantly enhance the value of the manuscript
Author Response
Comment 1: It is recommended to explicitly define the research objective and place it in Section 1. Introduction, preferably at the beginning or end of the section. Another suitable location could be Section 3. Methodology, as the research objective is an integral part of the scientific research methodology.
Response:
We appreciate the reviewer’s suggestion to explicitly define the research objective and indicate a suitable placement within the manuscript. In response, we have clearly articulated the research objective in the lines 90–99 at the end of Section 1: Introduction. This placement is chosen to provide a natural conclusion to the introductory section, contextualizing the study within the broader field and leading into the subsequent methodology. We believe this positioning enhances the clarity and flow of the manuscript while aligning with academic conventions.
Comment 2: Since Section 2. Literature Review is intended as a review of existing literature, it is unclear why no citations are included.
Response:
We appreciate the reviewer’s observation. In response, we have expanded Section 2 to include relevant citations and a broader review of the literature. While several studies have examined the relationship between AM parameters and mechanical properties, energy consumption, or part mass using ML, there remains a gap in directly linking AM parameters to comprehensive sustainability metrics through ML. To address this, we have incorporated additional studies, some employing ML and others utilizing G-code approaches, into the literature review (lines 125–147), thereby enhancing the depth and relevance of the section.
Comment 3: Adding concrete and in-depth conclusions (e.g., in Section 5. Conclusions) would significantly enhance the value of the manuscript.
Response:
We thank the reviewer for the valuable suggestion to enhance the conclusion with more concrete and in-depth insights. In response, we have revised Section 5: Conclusion (lines 664–682) to clearly summarize the key findings, highlight the role of dominant AM parameters, and emphasize the practical implications of the ML models used. The revised conclusion now reflects the broader significance of the study by outlining its contribution to data-driven, sustainable manufacturing practices. We believe this strengthens the manuscript’s overall impact and better aligns with the journal’s objectives.
Reviewer 3 Report
Comments and Suggestions for AuthorsThe paper shows an optimisation of the additive manufacturing process using AI tools. Through the manuscript, I found the following issues.
- Abstract: Please shorten the abstract to the required length. Also, the first part (lines 10-15) is completely unnecessary. Provide what is new in the suggested approach and provide the most important outcomes of your work.
- For what kind of AM technique are these analyses valid? Please point out the specific technique from the ISO 529000 classification.
- Figure 1 - The text in the figure is almost invisible!
- Figure 2 - provide a properly prepared 2D drawing - not a screen from software!
- Figure 4 - Why does this chart have linear courses? Are you sure that you can connect each point to the line ???
- There is too little data about the printing process - provide more specific information.
- I cannot see the validation of the results.
Author Response
Comment 1: Abstract: Please shorten the abstract to the required length. Also, the first part (lines 10-15) is completely unnecessary. Provide what is new in the suggested approach and provide the most important outcomes of your work.
Response:
We appreciate the reviewer’s feedback regarding the abstract. In response, we have carefully revised the abstract (lines10-27) to emphasize the novelty of our approach and highlight the key outcomes of the study. The revised abstract now directly outlines the experimental foundation of the work, the application of machine learning (ML) to model sustainability metrics, and the performance of the evaluated models. The initial lines have been refined to ensure they provide relevant background in alignment with the journal’s structured abstract guidelines.
We would also like to clarify that the original abstract was 171 words, within the journal’s 200-word limit. The updated version is now 182 words and remains compliant with the required length while better reflecting the unique contributions and findings of the research.
We hope the revised abstract meets the reviewer’s expectations.
Comment 2: For what kind of AM technique are these analyses valid? Please point out the specific technique from the ISO 529000 classification.
Response:
We appreciate the reviewer’s insightful comment. The analyses in this study are valid for the Fused Filament Fabrication (FFF) technique, which falls under the material extrusion category according to the ISO/ASTM 52900 classification. This clarification has now been included in the lines 407-408 of the revised manuscript.
Comment 3: Figure 1 - The text in the figure is almost invisible!
Response:
We appreciate the reviewer’s observation regarding the visibility of Figure 1. The figure has been updated, in the lines 422-423, accordingly with larger and clearer text annotations, ensuring that all dimension labels and details are now easily readable. This change improves the overall clarity of the figure and supports the explanation provided in the corresponding section of the manuscript.
Comment 4: Figure 2 - provide a properly prepared 2D drawing - not a screen from software!
Response:
We appreciate the reviewer’s suggestion regarding Figure 2. The previously used software screenshot has been replaced with a clean, properly formatted 2D technical drawing of the ASTM D638 Type I tensile specimen, in the lines 423-424. This updated figure adheres to professional presentation standards and enhances the clarity and technical accuracy of the manuscript.
Comment 5: Figure 4 - Why does this chart have linear courses? Are you sure that you can connect each point to the line ???
Response:
We thank the reviewer for raising this important point. To clarify, the lines in Figure 4 are not intended to represent linear or continuous relationships between the variables. They are included solely to guide the reader’s eye and help follow the distribution of each property across discrete sample indices.
In response, we have revised the figure description in Lines 428-429 to clearly indicate the use of the Fused Filament Fabrication (FFF) process. Additionally, we have included a corresponding statement within the main text in lines 417-418 and the figure caption to ensure consistency and clarity.
Comment 6: There is too little data about the printing process - provide more specific information.
Response:
We appreciate the reviewer’s comment. In response, we have revised the lines 403-412 to provide more detailed information about the 3D printing process used in the study. The updated text includes the specific 3D printer model, its build volume, minimum achievable layer thickness, extrusion temperature, printing speed, and nozzle configuration. We also clarified that a standard PLA filament was used and noted the classification of the process under ISO/ASTM 52900. These additions aim to enhance the reproducibility and clarity of the experimental procedure.
Comment 7: I cannot see the validation of the results.
Response:
We thank the reviewer for raising this important point regarding model validation. To clarify, our research is grounded in experimentally obtained data, derived from a carefully designed and executed series of additive manufacturing (AM) trials. These experiments generated a robust dataset comprising both AM process parameters and corresponding sustainability metrics.
The primary objective of this work, as specified in the updated Section 1 (lines 90–99), is to highlight and model the underlying relationships between AM process parameters and key sustainability metrics, such as energy consumption, printing time, part weight, and scrap weight, using ML. These relationships are often complex and not easily discernible through experimental analysis alone, making ML a valuable tool for uncovering and quantifying them.
Accordingly, the ML models are applied post hoc to this experimentally generated dataset. Rather than replacing experimentation, ML in this work serves to interpret and explain the data more comprehensively, revealing nuanced interactions that are otherwise difficult to capture through traditional methods.
While the models are not validated using new, independent experimental data beyond the original dataset, we conducted a rigorous internal validation process. This included evaluating model performance using standard metrics such as the coefficient of determination (R²) and Mean Squared Error (MSE), as well as assessing model stability across multiple randomized training and testing splits. These measures provided confidence in the reliability and generalizability of the ML-based insights.
Round 2
Reviewer 3 Report
Comments and Suggestions for AuthorsThe paper can be published