Next Article in Journal
A Comprehensive Review on the Performance of Structural Lightweight Aggregate Concrete for Sustainable Construction
Previous Article in Journal
A Historical Review of Impressed Current Cathodic Protection of Steel in Concrete
 
 
Article
Peer-Review Record

Development of an Interface Shear Strength Tester and a Model Predicting the Optimal Application Rate of Tack Coat

Constr. Mater. 2021, 1(1), 22-38; https://doi.org/10.3390/constrmater1010002
by Dowan Kim and Sungho Mun *
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Constr. Mater. 2021, 1(1), 22-38; https://doi.org/10.3390/constrmater1010002
Submission received: 28 February 2021 / Revised: 25 March 2021 / Accepted: 30 March 2021 / Published: 2 April 2021

Round 1

Reviewer 1 Report

The provided manuscript concerns the development of new device for testing the bonding strength between the layers of bituminous pavements as well as a development of new mathematical models for predicting the strength of this bond. The paper is well written, however, there are some minor language errors present. Also, in a few cases authors need to be more specific in their statements. The specific remarks are given in the attached PDF. The work is of high significance to the discipline and the subject is up to date. The employed methods are appropriate, and the research is scientifically sound. I would like to give some additional comments regarding the produced predictive models:

  • It should be recognized by the authors, that in modelling most phenomena encountered in engineering practices simple mathematical models are usually sufficient and also desirable; utilization of high-order polynomials often leads to overfitting and such case is perfectly seen in figure 10, particularly 10a and 10b (the 4th and 6th order polynomials) but also in 10c and 10d; in order to use the models as a predictive tool, they need to be produced with a certain number of degrees of freedom (i.e. not producing a 3rd order polynomial based on only 4 data points, etc.), some inaccuracies in model fit and predictive errors are unavoidable due to measurement errors, changes in the material properties, human errors etc.; consider this as a mild remark and a note for future investigations and model validation;
  • Information about the number of samples tested should be provided, given that the intention of the research was to enable predicting the bond strength;
  • In p. 6.3 authors state that the models were evaluated; was the model evaluation (as in figure 14 with measured vs. predicted values) conducted on the same data as the model was trained or was it evaluated using a separate dataset reserved only for model testing? This is not clearly stated; it is obvious that a high-order, complex model will near-perfectly fit the data to which it was fitted in the first place, however, this will not be a prediction; in this case it is especially important to distinguish the evaluation of model fit and evaluation of its potential for predicting future results; if there was no distinguishing between training and test data sets, then the predictive potential of the models was not evaluated at all, hence: “prediction” in fig. 14 should be changed to “fitted value”, in fig 13. “predictive curve” should be changed to “fitted curve”, changes should be made elsewhere in the text;
  • The calculated and further used parameters of the models 5a, 5b and 6 should be disclosed similarly to the parameters used in the regression models (table 4) so that the models can be evaluated by other researchers.

 

Comments for author File: Comments.pdf

Author Response

List of Changes and Rebuttal

 Reviewer #1

The provided manuscript concerns the development of new device for testing the bonding strength between the layers of bituminous pavements as well as a development of new mathematical models for predicting the strength of this bond. The paper is well written, however, there are some minor language errors present. Also, in a few cases authors need to be more specific in their statements. The specific remarks are given in the attached PDF. The work is of high significance to the discipline and the subject is up to date. The employed methods are appropriate, and the research is scientifically sound. I would like to give some additional comments regarding the produced predictive models:

 

Responses to the reviewer’s comments:

Thanks to the reviewer’ remarks, we change the errors in the revised paper, based on the specific comments given in the attached PDF.

 

Reviewer’s comments:

  • It should be recognized by the authors, that in modelling most phenomena encountered in engineering practices simple mathematical models are usually sufficient and also desirable; utilization of high-order polynomials often leads to overfitting and such case is perfectly seen in figure 10, particularly 10a and 10b (the 4th and 6th order polynomials) but also in 10c and 10d; in order to use the models as a predictive tool, they need to be produced with a certain number of degrees of freedom (i.e. not producing a 3rd order polynomial based on only 4 data points, etc.), some inaccuracies in model fit and predictive errors are unavoidable due to measurement errors, changes in the material properties, human errors etc.; consider this as a mild remark and a note for future investigations and model validation.

 

Responses to the reviewer’s comments:

Thanks for reviewer’s remark, we have a chance to improve our conclusion remarks included as follows:

 

Some inaccuracies in model fit and predictive errors may be unavoidable due to measurement errors, changes in the material properties, and human errors; thus, future investigations and model validation should be required in order to distinguish the model fit and predictive errors.

 

Reviewer’s comments:

  • Information about the number of samples tested should be provided, given that the intention of the research was to enable predicting the bond strength.

 

Responses to the reviewer’s comments:

The number of tested samples are provided in page 8 of the revised paper as follows:

 

The three samples of each RS(C)-4, BD-Coat, QRS-4, and AP-3 were prepared; application rates of 0.24, 0.3, 0.36, 0.45, 0.48, 0.6, and 0.8 L/m2 were used for evaluating the RS(C) and BD-Coat; furthermore, application rates of 0.3, 0.45, 0.6, and 0.8 L/m2 for QRS-4 and AP-3.

 

Reviewer’s comments:

  • In p. 6.3 authors state that the models were evaluated; was the model evaluation (as in figure 14 with measured vs. predicted values) conducted on the same data as the model was trained or was it evaluated using a separate dataset reserved only for model testing? This is not clearly stated; it is obvious that a high-order, complex model will near-perfectly fit the data to which it was fitted in the first place, however, this will not be a prediction; in this case it is especially important to distinguish the evaluation of model fit and evaluation of its potential for predicting future results; if there was no distinguishing between training and test data sets, then the predictive potential of the models was not evaluated at all, hence: “prediction” in fig. 14 should be changed to “fitted value”, in fig 13. “predictive curve” should be changed to “fitted curve”, changes should be made elsewhere in the text.

 

Responses to the reviewer’s comments:

Thanks to the reviewer’s remark, we can correct the Figures 13 and 14. Please check the revised paper consisting of the corrected terminology of fitted curve and fitted value. Furthermore, all commented figures are corrected in the revised paper.

 

Reviewer’s comments:

  • The calculated and further used parameters of the models 5a, 5b and 6 should be disclosed similarly to the parameters used in the regression models (table 4) so that the models can be evaluated by other researchers.

 

Responses to the reviewer’s comments:

Thanks to the reviewer’s remark, in page 16, we can provide the values of Eq. (6) used in this study as follows:

 

The determined alpha and beta parameters in Eq. (6) were resulted in 0.711 and 0.16 (RSC-4), 0.903 and 0.148 (BD-Coat), 0.697 and 0.194 (QRS-4), and 0.720 and 0.280 (AP-3) in this study.

Author Response File: Author Response.doc

Reviewer 2 Report

The adhesion of the pavement layers is important because it largely ensures the integrity of the structural layers of the road surface. Too high a stiffness or disobedience of the adhesive bond between pavement layers, or vice versa, too much displacement can adversely affect the performance of the road pavement. The publication could mention, for example, with reference to DOI: 10.22616/ERDev2019.18.N400  (see. Fig. 5., and table No.5.), a layer shear test according to German specifications ALP A-StB, which shows the adhesion as shear strength in kN between BBTM and down layer AC. It is important to note that the adhesion of interlayers is affected by tensile stresses resulting from the cyclic loading of heavy transport loads. Actual tensile stresses read by the sensor are given, for example, DOI: 10.1155/2021/8850368 (see table No.10.). Due to the cyclical nature of transport, it would be good to note the possibility of further developing this test or model in order to simulate real conditions as much as possible.

Author Response

Reviewer #2

Reviewer’s comments:

The adhesion of the pavement layers is important because it largely ensures the integrity of the structural layers of the road surface. Too high a stiffness or disobedience of the adhesive bond between pavement layers, or vice versa, too much displacement can adversely affect the performance of the road pavement. The publication could mention, for example, with reference to DOI: 10.22616/ERDev2019.18.N400  (see. Fig. 5., and table No.5.), a layer shear test according to German specifications ALP A-StB, which shows the adhesion as shear strength in kN between BBTM and down layer AC. It is important to note that the adhesion of interlayers is affected by tensile stresses resulting from the cyclic loading of heavy transport loads. Actual tensile stresses read by the sensor are given, for example, DOI: 10.1155/2021/8850368 (see table No.10.). Due to the cyclical nature of transport, it would be good to note the possibility of further developing this test or model in order to simulate real conditions as much as possible.

 

Responses to the reviewer’s comments:

Thanks to the reviewer’s remark, we can provide the statements and references suggested by the reviewer in the revised paper. Please check the page 7 in the revised paper.

 

In other researches, a layer shear test according to German specifications ALP A-StB, which shows the adhesion as shear strength in kN between BBTM and down layer of asphalt concrete, notes that the adhesion of interlayers is affected by tensile stresses resulting from the cyclic loading of heavy transport loads [23-24]. Due to the cyclical nature of transport, it would be good to note the possibility of further developing this test or model in order to simulate real conditions as much as possible. However, this research focuses on the non-cyclic loading shear strength.

 

  1. Riekstins, A.; Haritonovs, V.; Abolins, V.; Straupe, V.; Tihonovs, J. Life cycle cost analysis of BBTM and traditional asphalt concretes in Latvia. Engrg. for Rural Dev.2019, 22.-24.05., 1065-1072.

 

 

  1. Braunfelds, J.; Senkans, U.; Skels, P.; Janeliukstis, R.; Salgals, T.; Redka, D.; Lyashuk, I.; Porins, J.; Spolitis, S.; Haritonovs, V.; Bobrovs, V. FBG-based sensing for structural health monitoring of road infrastructure. J. Sens. 2021, Article ID 8850368.

Author Response File: Author Response.doc

Back to TopTop