Prediction of Fracture Toughness of Intermediate Layer of Asphalt Pavements Using Artificial Neural Network

Round 1

Reviewer 1 Report

The article titled: "Prediction of Fracture Toughness of Intermediate Layer of Asphalt Pavements Using Artificial Neural Network" is in line with the current trend of research on the applicability of Neural Networks in engineering sciences. The following elements should be improved and changed:

- The structure of the abstract is incorrect it does not contain, among else, information about the applied methodology of the conducted research. This part of the article requires improvement.

- The aim of the research conducted should be clearly indicated in the text of the article.

- The structure of the article needs to be improved as it does not comply with the guidelines of the journal.

- The bibliography is not formatted according to the guidelines and publication in the MDPI journal.

- In the article there should be a part related to the discussion of the results of already published studies and in the similar subject with a reference to the studies conducted by the authors. This will enrich the publication and allow for a comparison of the methods used to exploit ANN and the results obtained using them.

Overall, this paper is an interesting contribution to the development of research on the use of ANNs in studies related to the construction and road construction sectors. The application of the above will allow to increase the scientific soundness of the article.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Summary:

The manuscript describes the development of an artificial neural network (ANN) model to predict the condition of the intermediate layer in asphalt pavements. The model was developed in function of the following properties: international roughness index, rut depth, surface distress and equivalent single axle load. Predictions using three potential models were compared with fracture toughness (load/displacement) data from 66 core samples taken randomly from an expressway. The authors report that the selected model showed small errors with actual results,  and that the distribution between predicted and measure was comparable.

Comments:

The literature review in the manuscript is concise, however, it is sufficient and the authors appropriately connect their work with previous studies on the subject. The subject of the manuscript is quite relevant for predicting maintenance in asphalt pavements

The research methodology followed in the study was thorough and reflects the most relevant factors in in assessing the condition of intermediate layer on asphalt pavements. The collection of measure data was adequate. The authors followed industry-accepted standards for performing the measured properties.

The utilization of an ANN to include the utilized variables seem adequate. The authors described the development of the model thoroughly with their assumptions and differences between the three proposed models. The research methodology was thorough and the statistical methods in their analysis was appropriate for the type of data they have. The results and analysis properly support their conclusions, however, the manuscript can benefit in expanding on the observed outliers and in their decision of eliminating them. It is not critical for the publication of the manuscript but it will benefit from it.

The objectives of the study and the research methodology are well described and performed. The results of the study are informative and relevant. I think the authors’ efforts should be recognized and the manuscript should be published as it is.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

 

 1、 In line 104, the authors stated that “…using the IRI, RD, SD, and ESAL as the independent variables.” As we know, the IRI, RD, SD are NOT independent with ESAL, in the meanwhile, the IRI, RD and SD are the functions of EASL, the structural parameters such as the thickness of asphalt layers, especially the mechanical properties of each pavement layers. The authors used limited date to obtain an ANN model to predict the fracture toughness (FT) . The models trained seems pure mathematical models, I am skeptical about the generality and applicability of the final model for the other asphalt pavements.

2、 Suggest the authors simplify and summarize the main conclusions of this study.

3、 Please explain the reason the model was complemented by setting the upper and lower limits for FT. please provide the predicted results of Nd144 model without setting upper and lower limits.

4、 What is the objectives and highlights of this study? Provide more reviewing report regarding the application of ANN in asphalt pavement

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Thank you for your revisions and corrections to the manuscript. 

There are still some minor mistakes in English grammar that need to be corrected.

Reviewer 3 Report

there is no need to show figure1, suggest delet them;
format the font style of all figures as the text, provid clear figures in appropriate format; add vertical axis lines in  figure 4 and 5, overall, improve the qulity of all figures and tables;