An ANN Model for Predicting the Compressive Strength of Concrete

Round 1

Reviewer 1 Report

            The article discusses about the predication of compressive strength using ANN model from literatures. ANN model used in this study is back propagation network with one hidden layer in prediction of compressive strength. Authors used data’s to be trained from literatures for ANN. But authors didn’t discuss about how many mixes they used? How they arrived their compressive strength? What are all the factors from mix influences the compressive strength? Why author used 3, 7 and 12 hidden layers for training? Author want to explain table 1 properly, how they obtained data. Any reason behind it. There is no enough introduction supporting the current study. There is no clear scope and objective of research. Article lack in lot of references from literatures throughout the article. Author want to discuss about their experimental project briefly with proper citation. Author want to rewrite the conclusion part.

Comments for author File: Comments.docx

Author Response

Thank you very much for the comments. Please see our responses in the uploaded file.

Author Response File: Author Response.docx

Reviewer 2 Report

The approach in your research is very interesting and the result is promising. I hope that in the future you will continue this research and adapt your model to predict also other mechanical and/or physical properties of concrete.

Additional comments: My rating may have been a little generous. The introduction to the article is rather brief. The reflection by the authors, or discussion in relation to other references, is limited. Presumably this is because the authors relied on results and research from another source - ref 1 & 2. Unfortunately, this is a reference from a Chinese university and the information is mainly written in Chinese. This research is actually a continuation and extension of earlier research by another author. Due to the large amount of data, the authors have omitted this information. This makes it difficult to assess the article, but the authors have referred to their source material and how they have used this data, although it is not always very clear how they used or processed the data. A more extensive explanation of the methodology and use of the data would be welcome - if possible within the allowed amount of pages for the article. The research and conclusions of these authors are certainly interesting because a lot of calculation methods for structural applications in concrete (e.g. finite element-methods) starts from models contained in recommendations or standards. Mechanical properties of concrete are used in the development of new calculation models for structural analysis. Unfortunately, there is often insufficient data available on concrete mixes and their mechanical properties, so the use of a good prediction model such as the one presented in this article can offer a solution. Hence my positive assessment.

Author Response

Thank you very much for the comments. Please see our responses in the uploaded file.

Author Response File: Author Response.docx

Reviewer 3 Report

The authors are requested to improve the manuscript based on the following comments:

Comment 1

The citations in the abstract should be removed

Comment 2

The introduction should be more elaborate

 Comment 3

The purpose of the work should be more highlighted in the Introduction

Comment 4

line 68, Data and their ranges should be presented in the table

Comment 5

Was the scope of the data selected using the experiment planning method? All data should be shown in the appendix

Comment 6

Did the authors optimize the ANN topology? What software was used at work?

How many samples are assigned to the testing, training and validation set?

Comment 7

Tables 4 should be corrected or deleted

Comment 8

The conclusion should explain why the ANN model is better?

Comment 9

Literature should be more developed

For example include literature:

https://doi.org/10.1007/s00521-016-2801-y

Author Response

Thank you very much for the comments. Please see our responses in the uploaded file.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The article discusses about the predication of compressive strength using ANN model from literatures. ANN model used in this study is back propagation network with one hidden layer in prediction of compressive strength. Authors used data’s to be trained from literatures for ANN. Authors are requested to still write the introduction part briefly. Discussing about back propagation network with one hidden layer. Authors updated the revised article with recent literatures which makes article in good shape. Hence, I accept the article for the publication.

Reviewer 3 Report