Finite Element-Based Machine Learning Model for Predicting the Mechanical Properties of Composite Hydrogels

Round 1

Reviewer 1 Report

Manuscript Number: applsci-1991851

Finite element based Machine Learning Model for Predicting the Mechanical Properties of Composite Hydrogels

Comments on Manuscript: applsci-1991851

The authors of the present work combined a machine learning approach with finite element method to predict the mechanical properties of bioglass-collagen composite hydrogels. The paper fits well the aims of the journal. The paper is well written and structured. However, some concerns, comments, and suggestions should be addressed before publication.

 1.      There are numerous researches about the application of the Machine Learning in the literature that the authors can use for reviewing in the Intro or use them for the relations. Some of them are suggested in the following:

·         Predicting elemental stiffness matrix of FG nanoplates using Gaussian Process Regression based surrogate model in framework of layerwise model. Engineering Analysis with Boundary Elements, (2022) 143, 779-795.  https://doi.org/10.1016/j.enganabound.2022.08.001

·         Machine learning models for predicting the compressive strength of concrete containing nano silica. Computers and concrete, (2022), 30(1), 33-42 https://doi.org/10.12989/cac.2022.30.1.033

·          Machine-learning based design of active composite structures for 4D printing. Smart Materials and Structures, (2019), 28(6), 065005. https://doi.org/10.1088/1361-665X/ab1439

2.      The contributions, applications, and especially, novelties of the paper should be clearly outlined in the last paragraph of the Intro section to justify the motivation for this study. It should be clearly highlighted how the proposed model will fill the gap in the existing literature. It is not very clear to a general reader.

3.      The abstract should be improved and include clear statement about objectives, methodology and findings. 

4.      The authors must explain the deficiencies or shortcomings of other studies to make a bridge to introducing the novelty of their work. The advantage, benefits, and basic assumptions of the present solving procedure must be declared.

5.      In many places, the writing is an acceptable standard, but in other please there are some typographical errors, misspellings, and punctuation, and sentences that are not formulated properly, which is beyond what can be corrected in typesetting and significantly undermine the work. The whole paper should be rechecked, carefully.

6.      Please, always make sure that every symbol is described before or immediately after its first appearance in the text, and that all the notations for all the quantities are introduced and explained in the main text.

7.      To verify the accuracy of the numerical solution, the authors should added the comparison tables with referential data for the validation.

8.       The combination of the finite element analysis with ML is very confusing for general readers. It should be enhanced further by adding more detail.

Author Response

Please see attached

Author Response File: Author Response.pdf

Reviewer 2 Report

This manuscript presents the prediction of mechanical properties of bioglass/collagen composites using the convolutional neural network method. This manuscript is well-written and expected to be of great interest to the scientific community. However, a minor revision is required before it can be accepted for publication.

1. Avoid using first-person pronouns in a research article. Many first-person pronouns have been used throughout the manuscript. The research article should be expressed in the third person.

2. On page 3; line 121, why did the authors apply the axial strain of 0.20 to measure the Young's modulus? How did the authors know the materials at the axial strain of 0.20 still fall within the elastic region?

3. On page 5, the captions of Figure 2, Figure 4 and Figure 5 are overlong. Please separate the description from the caption of Figure 2.

4. On page 6; lines 203 – 210, how about the prediction errors of Young's modulus and Poisson's ratio of the materials with other volume fractions? Are they in the acceptable range?

5. On page 8; lines 266 – 269, have the authors predicted thermal conductivity, thermal expansion coefficient, fatigue life, toughness and stress-strain curves using the same method? How did the authors come to this conclusion if the prediction for other properties had not been performed?

6. The conclusion of this work is missing. It should be added at the end of the manuscript to summarize the significance of the study and the major findings obtained in this work.

Comments for author File: Comments.pdf

Author Response

Attached

Author Response File: Author Response.pdf

Reviewer 3 Report

This paper performed to Finite element based Machine Learning Model for Predicting the Mechanical Properties of Composite Hydrogels. The article is, in general, well written but there are some issues that authors should consider to revise in order to improve its quality. Some comments were done in this way:

 

·         Abstract should be expanded sentences related to the results. The results of the study should be given as numerical percentages.

·         The paper should be also supported by a literature search including relevant and recent papers. Only the most relevant and up-to-date articles on the study should be given. The recent articles related to FEM and machine learning may be cited.

·         Let's fix grammatical errors throughout the article.

·         Give the finite element analysis parameters and machine learning parameters in a table.

·         The article should be edited completely according to the journal writing guide.

·         Throughout the article, the words table and figure should start with capital letters (Table, Figure).

·         Fractions should be given with dots throughout the article, including figures and tables.

·         Explain section 2.2 in more detail.

·         The value of R2 in Figure 4.b is 0.83. This value is too low. Modeling can be revised for higher correlation value.

·         Expand the discussion section using current literature.

·         Conclusions should be written in more detail adding numeric data.

 

 

Author Response

Attachment

Author Response File: Author Response.pdf