Classifying Degraded Three-Dimensionally Printed Polylactic Acid Specimens Using Artificial Neural Networks based on Fourier Transform Infrared Spectroscopy
Round 1
Reviewer 1 Report
This manuscript reports research on the classification of thermal degradation of printed PLA using FTIR and ANN method. There are several places need to be considered or clarified before publishing.
- In the introduction, the author mentioned that FTIR is one of the best methods for understanding the degradation of PLA. The author should list more references and give a short discussion about different methods and what the advantage is when using FTIP method.
- Section 2.1 Does the filling pattern affect degradation? If yes, did author try other patterns?
- Section 2.1 How to choose the measurement locations on the PLA specimens?
- Section 2.2, author mentioned that there are two control factors, one is time and the other is temperature. However, the time for all sample is set to 24h. So here should be only one control factor.
Author Response
Dear Reviewer,
I appreciate your valuable comments. I modified my manuscript under your guidelines.
Sincerely,
Sung-Uk Zhang.
Reviewer Comments | Responses |
- In the introduction, the author mentioned that FTIR is one of the best methods for understanding the degradation of PLA. The author should list more references and give a short discussion about different methods and what the advantage is when using FTIP method. | FTIR spectroscopy is a non-destructive and quick technique to access molecular-level change of materials so it is easy to generate big data. In order to observe changes in the molecular level, there are several other technologies such as Raman spectroscopy, mass spectrometry, nuclear magnetic resonance spectroscopy (NMR), and X-ray photoelectron spectroscopy[11]. |
- Section 2.1 Does the filling pattern affect degradation? If yes, did author try other patterns? | I could say ‘yes’ in aspect of the mechanical property. I do not think so for FTIR spectroscopy. |
- Section 2.1 How to choose the measurement locations on the PLA specimens? | We randomly selected them under the specific portion. I substituted three measurement locations to three random measurement locations. |
- Section 2.2, author mentioned that there are two control factors, one is time and the other is temperature. However, the time for all sample is set to 24h. So here should be only one control factor. | To solve the misunderstanding, I added one sentence after that. “This study focuses on one control factor of the temperature.” |
Author Response File: Author Response.pdf
Reviewer 2 Report
Taking into account all the several features, the accuracy, scientific quality, scientific content and the interpretation of the results are interesting.
- The approach is interesting and the topic is appropriate for the journal.
- The work has a very clear structure and all the sections are well written in a way that is easy to read and understand. In addition, the structure of the paper is very good.
- The paper deals with classifying degraded three-dimensionally printed polylactic acid specimens using Fourier Transform Infrared Spectroscopy and Artificial Neural Networks, reporting interesting results. In the introduction of the work, the authors state “Additive manufacturing or three-dimensional (3D) printing is widely used in several industrial areas such as automotive, aerospace, mechanical, medicine, biological systems, and food supply chains [1]. Based on 3D CAD data, additive manufacturing can be used to fabricate complex geometries economically with a wide variety of materials. Polymers are the most commonly used materials in the 3D printing industry owing to their diversity and ease of adaptation to different 3D printing processes [2]…”. Accordingly, I also suggest to BRIEFLY introduce some progresses in the field of Additive manufacturing/three-dimensional (3D) printing for different applications such as micro/nanocomposite scaffolds for tissue engineering (i.e., “Three-dimensional printed bone scaffolds: The role of nano/micro-hydroxyapatite particles on the adhesion and differentiation of human mesenchymal stem cells”. Proceedings of the Institution of Mechanical Engineers, Part H: Journal of Engineering in Medicine 2017, 231(6), 555-564 …) as well as the possibility to use recycled polymers to fabricate 3D printed structures (i.e., “A comparison between mechanical properties of specimens 3D printed with virgin and recycled PLA” Procedia CIRP, Volume 79, 2019, Pages 143-146 …).
- The introduction should be improved.
- The list of references should be improved.
- It seems that the paper does not contain repetitions.
- The quality of figures must be improved.
- The title is adequate and appropriate for the content of the article.
- The abstract contains information of the article.
- Figures and captions are essential and clearly reported.
Author Response
Dear Reviewer,
I appreciate your valuable comments. I modified my manuscript under your guidelines.
Sincerely,
Sung-Uk Zhang.
Reviewer Comments | Responses |
The paper deals with classifying degraded three-dimensionally printed polylactic acid specimens using Fourier Transform Infrared Spectroscopy and Artificial Neural Networks, reporting interesting results. In the introduction of the work, the authors state “Additive manufacturing or three-dimensional (3D) printing is widely used in several industrial areas such as automotive, aerospace, mechanical, medicine, biological systems, and food supply chains [1]. Based on 3D CAD data, additive manufacturing can be used to fabricate complex geometries economically with a wide variety of materials. Polymers are the most commonly used materials in the 3D printing industry owing to their diversity and ease of adaptation to different 3D printing processes [2]…”. Accordingly, I also suggest to BRIEFLY introduce some progresses in the field of Additive manufacturing/three-dimensional (3D) printing for different applications such as micro/nanocomposite scaffolds for tissue engineering (i.e., “Three-dimensional printed bone scaffolds: The role of nano/micro-hydroxyapatite particles on the adhesion and differentiation of human mesenchymal stem cells”. Proceedings of the Institution of Mechanical Engineers, Part H: Journal of Engineering in Medicine 2017, 231(6), 555-564 …) as well as the possibility to use recycled polymers to fabricate 3D printed structures (i.e., “A comparison between mechanical properties of specimens 3D printed with virgin and recycled PLA” Procedia CIRP, Volume 79, 2019, Pages 143-146 …). | Copinet et al. [5] tried to estimate the biodegradation of a co-extruded starch/poly (lactic acid) polymeric material. Lanzotti et al. [6] studied the mechanical properties of virgin and recycled PLA. Currently, fused filament fabrication (FFF) using PLA is commonly used by the public [7]. |
The list of references should be improved. |
Author Response File: Author Response.pdf
Reviewer 3 Report
The article entitled Classifying degraded three-dimensionally printed polylactic acid specimens using Fourier Transform Infrared Spectroscopy and Artificial Neural Networks is an interesting example of ANN application in the field of FTIR spectra analysis, however, it has some drawbacks that need to be eliminated prior to publication.
1. First of all the Author seems to mix the term classifier with the name of data used to the classification – FTIR is not a classifier, the only thing is using data from FTIR spectroscopy as input for ANN. The Author classifies data using ANN, not FTIR. That mistake is consistently repeated in several places:
· Lines 3 – 5 (article title): the title needs to be clarified, as current one suggests classifying with FTIR and ANN, for example, the form: Classifying degraded three-dimensionally printed polylactic acid specimens using Artificial Neural Networks on the basis of/based on Fourier Transform Infrared Spectroscopy can be used.
· Lines 17-18: the sentence needs to be rewritten
· Line 55: the sentence needs to be corrected
· Line 57: the sentence needs to be corrected
· Line 148: the sentence needs to be corrected
2. Generally, data used for ANN teaching especially when some signal is analysed, need to be normalized, which means that the Author should conduct pre-analysis of FTIR spectra covering baseline correction and normalising the spectra using the greatest intensity or the area under the main peak. Not doing it may result in a less reliable and less stable output. Of course, normalisation is not mandatory when we have got binary data, however in that case normalisation should be done. When the ranges of values for each input are the same, tested model is more stable and modelling results are more reliable. The Author should treat that comment as a clue for future work.
3. Section 2.4: What was the way in which the particular network architecture was chosen – has the range of ANNs was tested, if yes, which algorithm has been applied for testing and on which basis the best one was chosen. Has the one which was used proved to be the best for that task? Was the only one architecture of MLP (presented in figure 4) used to classify all the data? Which learning algorithm was applied for the chosen ANN (standard backpropagation or other)? The exact number of neurons in all the layers should be given either in the text or in the description of figure 4 (inputs number depending on the division of the training set, the number of neurons in the hidden layers and in the output one).
4. Line 90: Artificial Neural Networks is the name for the big family of networks, while Multilayer Perceptron is only one of the members of that family, next to Radial Basis Function Networks, Kohonen networks etc. The sentence needs to be rewritten.
5. Lines 98-99: Could you explain why as an activation function rectifier linear unit was used, not the other one? Has any other been tested? Please, develop the shortcut used.
6. In the description of figure 6 please add the symbols of degradation levels (D01-D08).
7. Line 128: I do not understand why did you use only four outputs as you need to distinguish among eight degradation levels. Have networks been not able to conduct such a classification? Why only D01, D02, D07 and D08 were chosen? Were the FTIR spectra to similar in other groups? I do not see any probable explanation of that problem, which seems to be crucial for designing reliable and independent classifier.
8. The technique used for finding the most proper wavenumber ranges for the classification is quite tricky. Scanning the spectra seems to be the correct attitude, especially if you do not know characteristic peaks. However, have you tried to use as an input only those ranges that cover the main peaks for the neat PLA, for example, 750-760 cm-1, 865 to 875 cm-1 or others?
Author Response
Dear Reviewer,
I appreciate your valuable comments. I modified my manuscript under your guidelines.
Sincerely,
Sung-Uk Zhang.
Reviewer comments | Responses |
Lines 3 – 5 (article title): the title needs to be clarified, as current one suggests classifying with FTIR and ANN, for example, the form: Classifying degraded three-dimensionally printed polylactic acid specimens using Artificial Neural Networks on the basis of/based on Fourier Transform Infrared Spectroscopy can be used. | “Classifying degraded three-dimensionally printed polylactic acid specimens using Artificial Neural Networks based on Fourier Transform Infrared Spectroscopy” |
Lines 17-18: the sentence needs to be rewritten | “A classification methodology using artificial neural networks (ANNs) based on Fourier transform infrared (FTIR) and was developed.” |
Line 55: the sentence needs to be corrected | “Zhang [14] reported that the thermal degradation of 3D-printed ABS and PLA could be classified using ANNs based on FTIR.” |
Line 57: the sentence needs to be corrected | “The present study focuses on the thermal degradation of 3D-printed PLA and classifies it by using ANNs based on FTIR.” |
Line 148: the sentence needs to be corrected | “By using experimental data, this study demonstrated herein that thermal degradation of 3D-printed PLA could be classified using ANNs based on FTIR.” |
Generally, data used for ANN teaching especially when some signal is analysed, need to be normalized, which means that the Author should conduct pre-analysis of FTIR spectra covering baseline correction and normalising the spectra using the greatest intensity or the area under the main peak. Not doing it may result in a less reliable and less stable output. Of course, normalisation is not mandatory when we have got binary data, however in that case normalisation should be done. When the ranges of values for each input are the same, tested model is more stable and modelling results are more reliable. The Author should treat that comment as a clue for future work. | “Currently, this methodology has used FTIR datasets without any preprocessing such as normalization. This study assumed that ANNs could classify raw FTIR spectra. However, the preprocessing of the datasets may give better performance to ANNs so that it could be further work.” |
Section 2.4: What was the way in which the particular network architecture was chosen – has the range of ANNs was tested, if yes, which algorithm has been applied for testing and on which basis the best one was chosen. Has the one which was used proved to be the best for that task? Was the only one architecture of MLP (presented in figure 4) used to classify all the data? Which learning algorithm was applied for the chosen ANN (standard backpropagation or other)? The exact number of neurons in all the layers should be given either in the text or in the description of figure 4 (inputs number depending on the division of the training set, the number of neurons in the hidden layers and in the output one). | I need to reduce the number of hyperparameters to solve this study. So I need to fix several hyperparameters guided by J.Heaton’s book “ Introduction to Neural Networks for Java”. According to the rules to generate the ANN model, the number of neurons varied with the data size. There were five cases as shown in Figure 7. |
Line 90: Artificial Neural Networks is the name for the big family of networks, while Multilayer Perceptron is only one of the members of that family, next to Radial Basis Function Networks, Kohonen networks etc. The sentence needs to be rewritten. | “Multi-layer perception (MLP) as artificial neural networks (ANNs) was used to classify the degree of thermal degradation of the PLA specimens used herein.” |
Lines 98-99: Could you explain why as an activation function rectifier linear unit was used, not the other one? Has any other been tested? Please, develop the shortcut used. | “ In this study, an activation function for the nodes was ReLU[15] which is currently known as the most successful and widely-used activation function[16]. “ |
In the description of figure 6 please add the symbols of degradation levels (D01-D08). | I added the symbols for figure 6. |
Line 128: I do not understand why did you use only four outputs as you need to distinguish among eight degradation levels. Have networks been not able to conduct such a classification? Why only D01, D02, D07 and D08 were chosen? Were the FTIR spectra to similar in other groups? I do not see any probable explanation of that problem, which seems to be crucial for designing reliable and independent classifier. | The maximum number of classifications was four under this study. The ANN models could only classify D01, D02, D07, and D08, when the number of divisions in FTIR spectra ranged between four and sixteen. More datasets and preprocessing of FTIR spectroscopy may increase the number of classifications. It will be further work. |
The technique used for finding the most proper wavenumber ranges for the classification is quite tricky. Scanning the spectra seems to be the correct attitude, especially if you do not know characteristic peaks. However, have you tried to use as an input only those ranges that cover the main peaks for the neat PLA, for example, 750-760 cm-1, 865 to 875 cm-1 or others? | This classification method was initially designed and suggested for novices who are not familiar with FTIR spectroscopy. If we want to use the small range (750-760 cm-1 or 865 to 875 cm-1 ), it appears that we need to apply any data processing such as normalization. |
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The author has addressed all of my concerns, and I recommend it for publication.