Machine Learning for Additive Manufacturing
Round 1
Reviewer 1 Report
This paper reports the application of machine learning in additive manufacturing. It is interesting and important. And the paper structure is written well. A few mistakes should be corrected as following：
- "2.2.2" in line 138 on page 4 should be 2.2.3.
- “Figure 1" in line 367 on page 7 should be removed.
- It will be better to review more papers about this topic to enrich your report. As far as I know, there are a lot of related papers published in recent years.
Author Response
The authors thank the Reviewer for their feedback and comments. Below, we respond to your points in turn and hope that they provide the additional information or reassurance that you seek.
This paper reports the application of machine learning in additive manufacturing. It is interesting and important. And the paper structure is written well.
The authors thank the Reviewer for acknowledging that the paper is interesting and well-structured.
A few mistakes should be corrected as following：
- "2.2.2" in line 138 on page 4 should be 2.2.3.
Thank you for pointing out this mistake. This has been corrected.
- “Figure 1" in line 367 on page 7 should be removed.
After checking, there was no reference to Figure 1 on line 367 or page 7.
- It will be better to review more papers about this topic to enrich your report. As far as I know, there are a lot of related papers published in recent years.
Thank you for the suggestion. Additional papers have been reviewed to broaden the scope of the paper. The section on monitoring has been expanded to discuss optical monitoring in binder jetting and material extrusion. In addition, the discussion around dimensional deviation has been significantly extended while the section on parameter optimisation has been broadened to include DED and binder jetting.
Reviewer 2 Report
The article presents information on machine learning for additive manufacturing. The article's authors described an exciting topic, and nowadays, more often included in the 3D printing process. Here are the main comments about the paper which need to be considered:
- The description of the first paragraph in the introduction should be expanded. Authors should also consider which areas printed models are most often used (e.g., medicine, architecture, automotive, or aviation industry) and refer to specific items in the literature. Additionally, the problems with repeatability were mentioned (please refer to more items in the literature);
- There are typing errors like Aadditive in subchapter 2.2. Please check all work for similar errors;
- The resolution of the figures included in the work should be improved;
- In subsection 2.2.1, the authors discuss optimization in the context of complex geometries. It would also be worth describing the optimization process in terms of print accuracy and, in particular, surface roughness. The following publications can help with this:
OMAIRI, Amzar; ISMAIL, Zool Hilmi. Towards Machine Learning for Error Compensation in Additive Manufacturing. Applied Sciences, 2021, 11.5: 2375.
KHANZADEH, Mojtaba, et al. Quantifying geometric accuracy with unsupervised machine learning: Using self-organizing map on fused filament fabrication additive manufacturing parts. Journal of Manufacturing Science and Engineering, 2018, 140.3 - Subsection 4.2.1. it is exciting but needs an extension. The authors present in Figures 2 and 3 the parameters that optimize the printing process concerning PBF and Material Extrusion technology. It would be worth describing other techniques mentioned in the article similarly, such as Binder jetting or Directed energy deposition. Examples of publications can help with this:
GÜNTHER, Daniel, et al. Condition monitoring for the Binder Jetting AM-process with machine learning approaches. In: 2020 IEEE Conference on Industrial Cyberphysical Systems (ICPS). IEEE, 2020. p. 417-420.
CHOI, Tae-Yang. Machine learning based predictive modeling of dimensional quality in direct energy deposition with SUS316L. 2020.
ZHANG, Ziyang; LIU, Zhichao; WU, Dazhong. Prediction of melt pool temperature in directed energy deposition using machine learning. Additive Manufacturing, 2021, 37: 101692.
NARAYANA, Pasupuleti Lakshmi, et al. Optimization of process parameters for direct energy deposited Ti-6Al-4V alloy using neural networks. The International Journal of Advanced Manufacturing Technology, 2021, 1-15.
Author Response
The authors thank the Reviewer for their feedback and comments. Below, we respond to your points in turn and hope that they provide the additional information or reassurance that you seek.
The article presents information on machine learning for additive manufacturing. The article's authors described an exciting topic, and nowadays, more often included in the 3D printing process. Here are the main comments about the paper which need to be considered:
The description of the first paragraph in the introduction should be expanded. Authors should also consider which areas printed models are most often used (e.g., medicine, architecture, automotive, or aviation industry) and refer to specific items in the literature.
Medical, automotive, and aerospace industrial application motivations have been included with appropriate references. An extended discussion on these is seen by the authors to be outside of the scope of this paper in the context of this Encyclopedia entry.
Additionally, the problems with repeatability were mentioned (please refer to more items in the literature);
A suitable reference for repeatability concerns has now been provided but the authors view in depth discussion of this to be extraneous to the core discussion.
There are typing errors like Aadditive in subchapter 2.2. Please check all work for similar errors;
We have now thoroughly checked the manuscript and believe that no further typographical errors remain. This specific section has been reformatted in response to another reviewer’s comments.
The resolution of the figures included in the work should be improved;
Figures have been remade at a higher resolution and with improved formatting for clarity.
In subsection 2.2.1, the authors discuss optimization in the context of complex geometries. It would also be worth describing the optimization process in terms of print accuracy and, in particular, surface roughness. The following publications can help with this:
Thank you for this point. We agree that discussion of print accuracy discussion is important and have added more in-depth discussion of error reduction into Section 4.3. Parameter optimisation for surface roughness reduction has also now been discussed in Section 4.2.1.
OMAIRI, Amzar; ISMAIL, Zool Hilmi. Towards Machine Learning for Error Compensation in Additive Manufacturing. Applied Sciences, 2021, 11.5: 2375.
Thank you for this reference. It has been used to give context to the newly-added section on dimensional deviation.
KHANZADEH, Mojtaba, et al. Quantifying geometric accuracy with unsupervised machine learning: Using self-organizing map on fused filament fabrication additive manufacturing parts. Journal of Manufacturing Science and Engineering, 2018, 140.3
This source has now been discussed in Section 4.3.1.
Subsection 4.2.1. it is exciting but needs an extension. The authors present in Figures 2 and 3 the parameters that optimize the printing process concerning PBF and Material Extrusion technology. It would be worth describing other techniques mentioned in the article similarly, such as Binder jetting or Directed energy deposition. Examples of publications can help with this:
The authors thank the Reviewer for the suggestions of additional references. We have reviewed these suggestions and where we agree that they are in context to the manuscript we submitted, we have included the citation. Where sections have been extended, further references to the literature have been made.
GÜNTHER, Daniel, et al. Condition monitoring for the Binder Jetting AM-process with machine learning approaches. In: 2020 IEEE Conference on Industrial Cyberphysical Systems (ICPS). IEEE, 2020. p. 417-420.
The authors have included reference to this paper, but it has limited usefulness due to not comprehensively describing the model employed and not analysing the performance of the employed algorithm. As such, it has only been used as evidence that machine learning assisted monitoring techniques may be applied to the binder jetting process.
CHOI, Tae-Yang. Machine learning based predictive modeling of dimensional quality in direct energy deposition with SUS316L. 2020.
The authors have now discussed this thesis alongside other works on dimensional deviation and accuracy in DED.
ZHANG, Ziyang; LIU, Zhichao; WU, Dazhong. Prediction of melt pool temperature in directed energy deposition using machine learning. Additive Manufacturing, 2021, 37: 101692.
This paper looks at melt pool prediction rather than optimisation but has been added to the relevant discussion in Section 4.2.2.
NARAYANA, Pasupuleti Lakshmi, et al. Optimization of process parameters for direct energy deposited Ti-6Al-4V alloy using neural networks. The International Journal of Advanced Manufacturing Technology, 2021, 1-15.
Paper has been included in the now-added paragraph (Section 4.2.1) on parameter optimisation for DED processes.
Reviewer 3 Report
The manuscript “Machine Learning for Additive Manufacturing” has a very general tittle which might cover a wide range of topic.
Reviewer suggest authors to check some mistakes as follow:
• Line 22- 23: This is not the definition of AM.
• Line 55- 56: Do you have reference for this example?
• Line 72-87: you cite only one source, so you can put the citation in only one place.
But why do you need a section just to repeat from one reference?
• Line 347: AM design, process and production was developed without ML. Considering change the writing.
and then in line 368, authors claim "most ML applications for AM are not robust or trusted enough to be adopted in industry. Can you resolve the conflict here?
Minor error:
line 88: typo “Aadditive”
Author Response
The authors thank the Reviewer for their feedback and comments. Below, we respond to your points in turn and hope that they provide the additional information or reassurance that you seek.
The manuscript “Machine Learning for Additive Manufacturing” has a very general tittle which might cover a wide range of topic.
We thank the author for this comment, however; the publisher allows a maximum of five (5) words in the article title. We therefore do not have the opportunity to amend the title given.
Reviewer suggest authors to check some mistakes as follow:
Line 22- 23: This is not the definition of AM.
The definition for AM has been amended to bring it more in line with that provided in ISO/ASTM 52900:2015(E).
Line 55- 56: Do you have reference for this example?
Thank you for this comment. The authors have reworded this section to make it more concise and provided a reference as requested.
Line 72-87: you cite only one source, so you can put the citation in only one place. But why do you need a section just to repeat from one reference?
Thank you for the close attention paid to our formatting. Section 2 has now been reformatted in-line with this suggestion. Specifically, the definitions have been moved to the section introduction and the referencing has been made more concise.
Line 347: AM design, process and production was developed without ML. Considering change the writing.
Thank you for this comment. Upon re-reading this section, the authors agree that it was misleading/ambiguous. This has been re-written and fixed.
and then in line 368, authors claim "most ML applications for AM are not robust or trusted enough to be adopted in industry. Can you resolve the conflict here?
Thank you for pointing out this contradiction. This has been amended to differentiate between successful implementation in research and less adoption in industry.
Minor error: line 88: typo “Aadditive”
The typographical error has been amended and the rest of the paper has now been thoroughly proof-read.
Round 2
Reviewer 1 Report
The manuscript can be accepted as the comments have been addressed completely.
Reviewer 2 Report
Dear Authors,
Thank you for submitting your comprehensive responses to my comments. I accept the manuscript in its present form.
Kind regards,
Reviewer
Reviewer 3 Report
The revised manuscript is well written and could be recommended for publication.
