Using NDT Data to Assess the Effect of Pavement Thickness Variability on Ride Quality

Round 1

Reviewer 1 Report

The main problem for this paper is that more field tests and numerical modelling should be done to more convincingly support their hypothesis. 1. The main question is to correlate the GPR-based pavement thickness with roughness-based road quality index. 2. This topic is hot in current field, however, this paper cannot address a specific gap. Because the idea was done by others. 3. As far as I see, it only proposed an equation or used an exiting method to process the data, like statistical analysis. However, this was not stressed in the introduction. As far as I see, this paper is lack of novelty. 4. the authors are encouraged to add more situations and field validation for their proposed statistical processing on raw data. 5. The evidence to support the conclusion is superficial. 6. The references are appropriate, however, it lacks of so so so so so many relevant studies. Because if the authors show these studies, their few novelties will be more obvious.

In the literature, the correlation between GPR and IRI was studied.

It is better to stress on the novelty and importance of this study.

 

It is not clear what is length in Figure 8, what is the Axis X?

 

From Figure 8, no clear correlation was observed.

 

It is not clear what is the sample size of hac.

 

If totally 100 m was measured, with 10 m as interval, then only 10 is not enough.

English language is sufficient to understand.

Author Response

The main problem for this paper is that more field tests and numerical modelling should be done to more convincingly support their hypothesis.

1. The main question is to correlate the GPR-based pavement thickness with roughness-based road quality index.

 

2. This topic is hot in current field, however, this paper cannot address a specific gap. Because the idea was done by others.

Reply: Thank you for the comment. Please see lines 161-180.

 

3. As far as I see, it only proposed an equation or used an exiting method to process the data, like statistical analysis. However, this was not stressed in the introduction. As far as I see, this paper is lack of novelty.

 

Reply: Thank you for the comment. Please see lines 290-318.

 

4. the authors are encouraged to add more situations and field validation for their proposed statistical processing on raw data.

 

Reply: Thank you for the comment. Please see lines 290-318.

 

5. The evidence to support the conclusion is superficial.

Reply: Thank you for the comment. Please see lines 162-180, 185-204, 469-470 and 482-487.

 

6. The references are appropriate, however, it lacks of so so so so so many relevant studies. Because if the authors show these studies, their few novelties will be more obvious.

In the literature, the correlation between GPR and IRI was studied.

It is better to stress on the novelty and importance of this study.

 

Reply: Thank you for the comment. Please see lines 108-131 and 290-315.

 

7. It is not clear what is length in Figure 8, what is the Axis X?

 

Reply: Thank you for the comment. Please see lines 185-186, 371-372 and Figure 6.

 

8. From Figure 8, no clear correlation was observed.

Reply: Thank you for the comment. Please see lines 373-376 and 378-379.

 

9. It is not clear what is the sample size of hac.

 

Reply: Thank you for the comment. Please see lines 366-369 and 361.

 

10. If totally 100 m was measured, with 10 m as interval, then only 10 is not enough.

 

Reply: Thank you for the comment. Please see lines 185-186 and 371-372.

 

Reviewer 2 Report

Pavement condition evaluation and management are very important to the transportation system. The paper proposed a non destructive testing approach to assess pavement condition by combining the data from two independent NDT systems. And the authors further investigated whether the expected variations in asphalt layer thickness may have an impact on pavement roughness as expressed in International Roughness Index (IRI) values. The real data has been collected and the methodology has been validated convincingly. In this way, I think the paper is valuable and the quality is good to be published in this journal. To further improve the paper, I have the following comments:

-In the introduction section, please highlight the contributions of this paper. As of now, only the problem or limitations of the current literature was discussed but failed to discuss the contributions of this work. Another limitation is that the literature is only about the ground penetrating radar sensors. However, with the fast development of intelligent transportation systems, multiple sensors including LiDAR and cameras are available on the ground/aerial autonomous vehicles that have the potential to be used to assess the pavement conditions. Even if with the GPS/IMU data, the localization of the road can be recorded as well. That being said, I highly suggest the authors to discuss the potential possibilities from these sensors and include the following works: an automated driving systems data acquisition and analytics platform; yolov5-tassel: detecting tassels in rgb uav imagery with improved yolov5 based on transfer learning; autonomous vehicle kinematics and dynamics synthesis for sideslip angle estimation based on consensus kalman filter; improved vehicle localization using on-board sensors and vehicle lateral velocity; pothole mapping and patching quantity estimates using LiDAR-based mobile mapping systems.

- The section 2.4 data processing can be elaborated more. The current form did not give much technical points of the methodology or algorithm.

-From figure 2-figure 5, the equipment needs more information. At least the caption should include more details of the hardware that was used to collect data.

-The quality of the figure having the results can be improved.

Author Response

Pavement condition evaluation and management are very important to the transportation system. The paper proposed a non destructive testing approach to assess pavement condition by combining the data from two independent NDT systems. And the authors further investigated whether the expected variations in asphalt layer thickness may have an impact on pavement roughness as expressed in International Roughness Index (IRI) values. The real data has been collected and the methodology has been validated convincingly. In this way, I think the paper is valuable and the quality is good to be published in this journal. To further improve the paper, I have the following comments:

-In the introduction section, please highlight the contributions of this paper. As of now, only the problem or limitations of the current literature was discussed but failed to discuss the contributions of this work.

Reply: Thank you for the comment. Please see lines 161-180.

Another limitation is that the literature is only about the ground penetrating radar sensors. However, with the fast development of intelligent transportation systems, multiple sensors including LiDAR and cameras are available on the ground/aerial autonomous vehicles that have the potential to be used to assess the pavement conditions. Even if with the GPS/IMU data, the localization of the road can be recorded as well. That being said, I highly suggest the authors to discuss the potential possibilities from these sensors and include the following works: an automated driving systems data acquisition and analytics platform; yolov5-tassel: detecting tassels in rgb uav imagery with improved yolov5 based on transfer learning; autonomous vehicle kinematics and dynamics synthesis for sideslip angle estimation based on consensus kalman filter; improved vehicle localization using on-board sensors and vehicle lateral velocity; pothole mapping and patching quantity estimates using LiDAR-based mobile mapping systems.

Reply: Thank you for the comment. Please see lines 46-60.

- The section 2.4 data processing can be elaborated more. The current form did not give much technical points of the methodology or algorithm.

Reply: Thank you for the comment. Please see lines 290-357.

-From figure 2-figure 5, the equipment needs more information. At least the caption should include more details of the hardware that was used to collect data.

Reply: Thank you for the comment. Please see lines 213, 237, 255.

-The quality of the figure having the results can be improved.

Reply: Thank you for the comment. Please see figure 6.

 

Reviewer 3 Report

The manuscript is worth publishing once the raised issues are addressed (see attached file).

 

Comments for author File: Comments.pdf

Author Response

An interesting effort in which the authors have attempted to relate asphalt pavement thickness to pavement roughness. As the authors are aware, pavement roughness is affected by so many parameters in a pavement structure, way beyond asphalt pavement thickness. Thus, the way the manuscript is presented appears to be more of a "case study," where such relationship seems to be “casual,” (i.e., specific to these limited data) rather than having a long-lasting value in pavement evaluation and the potential use of blended NDTs.

Thus, the manuscript is still worth publishing once the following issues are addressed.

The authors need to make sure that the following important study limitations are further emphasized, and /or explained/defended within the manuscript:

(i) recognize upfront, in the “Introduction” section, the study limitations and discuss on the influence of many parameters affecting pavement roughness, yet not considered in this study (initial smoothness, construction and materials quality and uniformity of surface, base/subbase, influence of age/traffic and environmental parameters over time, transition from one subgrade type to the next – if applicable, and many other. While the authors do refer in some cases only briefly to these, there is not a detailed discussion on how such additional parameters are important, and how may or may not be relevant to this study.

Reply: Thank you for the comment. Please see lines 108-121 and 162-173

 

(ii) identify the need for further validating the results and findings with a more systematic approach considering a wider range of roadways in terms of the above parameters.

Reply: Thank you for the comment. Please see lines 482-487

 

It is critical in recognizing and further emphasizing such aspects within the manuscript so that readers can attest of: (i) the extended knowledge the authors do have in this area, and, (ii) the value and reasons of publishing this manuscript even though the study has it’s limitations.

In many instances within the manuscript there are either specific comments, and/or often generalized conclusions that the authors have to relate to the study limitations. Also, many clarifications are needed. Among the various areas of the manuscript, following are some cases mentioned herein as examples:

Abstract

“In particular, the objective of this study is to determine whether the expected variations in asphalt layer thickness may have an impact on pavement roughness as expressed in International Roughness Index (IRI) values.”

What are the “expected variations in asphalt layer thickness” the authors refer to? Need to be specific. Are these related to construction uniformity? potential in-service densification of layer thicknesses due to traffic? age/deterioration of roadway producing changes and settlement of pavement structure and thus potential rutting produced from subgrade, subbase, base and surface layer? else?

Reply: Thank you for the comment. Please see lines 17-18 and 162-173.

 

1. Introduction

“The deterioration rate of asphalt pavements depends on several factors, including not only material properties and pavement design, but also external factors such as the environment and traffic loading.”

How these relate to, or were considered in this study?

Reply: Thank you for the comment. Please see lines 185-204

 

The authors need to further expand the literature review on (i) GPR for flexible pavements, and (ii) pavement roughness, and report the challenges encountered in each case. A lot has been done over the years in both areas of pavement evaluation.

Reply: Thank you for the comment. Please see lines 61-85 and 108-131.

 

“The IRI represents the vibrations caused by a typical vehicle.”

The authors need to explain, is it “vibrations” or vertical accelerations due to uneven roadway surface (i.e., vertical elevations representing deviations from a true horizontal profile). What type of roughness devices are they referring to with such a statement?

Reply: Thank you for the comment. Please see lines 123-131

 

“It is determined based on the geometric characteristics of the road surface and can be measured using modern laser profilometers.”

Not only! Need to also add references, and expand on the alternative methods beyond laser profilometers (inertial profilometers, profilographs, etc).

Reply: Thank you for the comment. Please see lines 123-131.

 

2. Materials and Methods

This section needs to be expanded significantly to provide relevant and meaningful assessment in regards to the study limitations and constraints, as well as value of the analysis and findings.

Reply: Thank you for the comment. Please see lines 184-204

What were the design and as-built constructed pavement layer thicknesses?

Reply: Thank you for the comment. Please see Table 1 and Table 2.

 

What was the thickness variability in each case?

Reply: Thank you for the comment. Please see Table 2.

 

Figure 1 does not provide much of meaningful input into what pavement sections were considered in the study. Pavement layer material properties should be also mentioned.

Reply: Thank you for the comment. Please see lines 184-204

 

Reporting differences in AC layer thickness of up to 9cm implies that there is a variety of alternative pavement designs involved. Thus, the data need to be carefully analyzed in relation to each case (thin, medium, thick pavement sections). Looking at ranges of hAC might not necessarily do so, as mentioned later.

Reply: Thank you for the comment. Please see lines 184-204 and 373-375

 

2.2 GPR Survey

“Principally, the GPR system measures the time intervals it takes for the electromagnetic waves to penetrate the pavement materials and return to the receiver.”

The authors need to better clarify that is the “two way travel” of emitted and reflected EM wave, rather than simply referring to ”time intervals.”

Reply: Thank you for the comment. Please see lines 221

 

Figure 2 does not provide any meaningful info, should be excluded.

Reply: Thank you for the comment. It is excluded.

 

Figure 3, would be more meaningful to show an actual measurement from the study with specifics on depths, signal amplitudes, rather than the typical general sketch shown in many references.

Reply: Thank you for the comment. Please see Figure 2

 

The authors need to provide more info on what the “special software” is? It is important to identify/discuss the principles of signal processing, it’s capabilities and limitations.

Reply: Thank you for the comment. Please see lines 241-242

2.3 RSP Survey

In terms of the profile measurements need to provide further details: where they collected from the two wheel-paths or else? Where such measurements averaged?

Reply: Thank you for the comment. Please see lines 282-283

 

What was the EDM system used?

Reply: Thank you for the comment. Please see lines 283-286

 

“The sprung mass (ms), which is the mass of the part of the vehicle that burdens the suspension and includes a percent of the weight of the suspension. The unsprung mass (mu), which is the mass corresponding to the weight that does not burden the suspension system, but is supported by the wheel or the tire and follows its displacements.”

The authors need to include units for such parameters.

Reply: Thank you for the comment. Please see lines 266-273

 

“The aforementioned parameters equal to specific values in the process of calculating the IRI index. These values constitute the “golden car.” The principal idea for the IRI calculation is to model the movement of the quarter car on the pavement surface and the calculation of the total displacement of the suspension. IRI values equal to zero corresponds to a totally smooth pavement, while IRI values around ten indicates a practically impassable road surface.”

What does it mean the “golden car”? Need to explain.

Reply: Thank you for the comment. Please see line 274-277

 

What are the IRI units? What is an IRI =10 means?

Reply: Thank you for the comment. Please see line 259, 280

 

Who determines that an IRI=10 is an “impassable” road surface? What are the reference standards used?

Each country may have different standards of acceptable and unacceptable categories of road roughness.

The authors need to expand on this statement with specifics.

Reply: Thank you for the comment. Please see line 281 and the related reference

 

2.4 Data Processing

“Identification of the stratigraphy.”

What does it mean "identification of the stratigraphy" and how was this done? Need to provide specifics.

Reply: Thank you for the comment. Please see lines 290-315

 

“In order to accomplish the maximum possible accuracy in estimating the thicknesses, limited core sampling is needed. Essentially, the real pavement data (layer thickness) measured by coring, parallels the corresponding data recorded by the GPR system. This way, it is possible to calculate the parameters that are needed for converting time to thickness. In GPR data processing phase, signals from raw pavement data, end reflection and metal reflection were elaborated. During the processing phase data filtering was also done to improve the data quality.”

It is not clear from this statement if the authors did site inspections and/or coring. This would have revealed the "ground truth" conditions in both layer thickness and "material properties" to compare with the GPR response, in addition to the standard calibration with metal plate.

Reply: Thank you for the comment. Please line 344-346

 

“In Figure 7 raw data as well as data after processing are presented. The asphalt layer thickness was estimated at an interval of 10m and was found to lie between 13 and 31 cm along the various segments.”

This is a significant range. At what length of roadway was this observed? What were the reasons? What type of roadways were monitored, when they were build and under what construction guidelines? What was the age of the roadways, traffic and specific materials?

This raises concerns whether the monitored pavements had relatively controlled and uniform quality during construction, affecting thus significantly GPR response shadowing to some degree repeatability, and/or aggregating below surface variability and/or deterioration of materials, density and moisture effects, other.

This goes back to the insufficient description pertinent to the Materials and Methods section mentionedabove.

Reply: Thank you for the comment. Please line 185-204 and Table 1 & 2

 

“During the processing phase data filtering was also done to improve the data quality.”

Need to expand of what data filtering entails and how was it achieved?

Reply: Thank you for the comment. Please line 335-339

 

3. Results

“Figure 8 presents an overall of the data used for the analysis. Differences in hAC within the segments vary from 0 to 9 cm, while IRIav values range between 0.16 and 2.65.”

Are these all the data used in the analysis? If not, how all the data look like?

Are these from one stretch of roadway, or how many? Authors indicated that considered testing on "various flexible pavement cross sections of two different highways in operation."

Further details are needed.

Reply: Thank you for the comment. Please line 185-204

 

“The high value of the standard deviation (>0.10) indicates that there is a large variation in IRIav values and therefore the pavement sections are non-homogenous in terms of smoothness. Thus, the riding quality cannot be represented by the mean value.”

It is evident from the range of thicknesses reported that quality control was an issue for these pavement sections. The same is reflected on the IRI variations which very well reflect contractor's ability to produce smooth roads (quality of paver, roller compaction practice, personnel training, asphalt temperature during compaction, and several other parameters) which may or may not influence pavement thickness. Clearly a stronger base/subbase layer, but also a smoother base, will provide eventually as smoother surface layer.

The authors need to address this dilemma right-on and defend their attempt to relate pavement smoothness to asphalt pavement layer thickness, other than just looking on how to present and analyze the data.

A more controlled and extensive study with strict QA/QC construction guidelines and alternative pavement sections and as-built smoothness levels will eventually provide a good assessment on whether there is such a relationship between pavement smoothness and pavement thickness. The authors do refer to one study that has shown that pavement thickness is the prevailing factor, but not the only one! Again, in pavement roughness, equipment quality and operators’ expertise are critical in achieving a smooth pavement independently of pavement design thicknesses. This has been shown by many studies over time. In fact, pavement performance specifications include both pavement thickness and smoothness, among other, for this reason. The authors need to debate this argument.

Reply: Thank you for the comment. Please line 108-121, 185-204, 401-403.

 

Figure 10 and Table 2.

Again, it is not clear if these results are based on a single/two roadways (and so coincidental) or a more systematic study where alternative pavement design thicknesses are considered and with the same or alternative paving smoothness quality during construction.

Thus, the conclusions of this study are of limited value (perhaps coincidental to this exploration until further explored systemically). Thus, findings need to be further validated to be able to generalize such conclusions. The authors need to address extensively this limitation from the start as well as in the conclusions and their findings.

Reply: Thank you for the comment. Please line 185-204

 

4. Discussion

“In this study, the possible relationship between asphalt layer thickness variation and pavement roughness was investigated. GPR and profilometer measurements were performed in different sections of two different highways.”

Still not clear what were the different sections? traffic loads, differences in pavement design and materials, age, deterioration rates?

Reply: Thank you for the comment. Please line 185-204

 

The analytical results

“The analysis presented in this paper show that the pavement performance in terms of pavement roughness is related to the variation of asphalt layer thickness.”

As mentioned above such conclusion is based on too few data and limited experimental parameters to judge and generalize conclusions. Further validation is needed in a more controlled conditions considering all the parameters affecting construction quality in terms of (i) thickness uniformity (randomness/variability vs significant meaningful levels of design thickness), quality of smoothness levels by a contractor (reflecting both equipment and operator(s) abilities); many other factors such as deterioration rates, traffic, environment, etc. Otherwise such conclusions remain "casual" (i.e., specific to this experiment and limited data) and thus cannot be generalized or consider to provide a meaningful contribution to the state of knowledge in pavement evaluation. Thus, further validation is critical for making such claim that thickness significantly affects smoothness.

Keep in mind that a good contractor can produce a very smooth pavement for thin lifts or thick pavement layers. Smoothness is the result of many factors, not just thickness, and thus authors need to clearly indicate that such conclusions are specific to this study.

Reply: Thank you for the comment. Please line 185-204 and 482-487

 

“Consistent with the above, it is argued that GPR capabilities can be extended beyond pavement thickness determination and structural evaluation to an initial assessment of expected pavement ride quality.”

Risky statement. GPR, perhaps can be used to assess pavement roughness if "surface elevation" is considered. Recent studies have looked into that and should be referenced.

 Reply: Thank you for the comment. Please line 135-158, 459

 

5. Conclusions

“Variation in asphalt layer thickness was shown to affect pavement roughness. It is assumed that the layer thicknesses accumulate around the design value during the construction phase, especially during paving and compaction of the asphalt mix.”

There is some disconnect of this statement “thicknesses accumulate around the design value during the construction phase,” versus the significant range of thickness variability reported, (i.e., differences in asphalt layer thickness up to 9cm). The authors examined the data for three different levels of hAC, but these may not necessary reflect groups of “thin”, “medium” and “thick” asphalt pavement layers (i.e., you may have thick pavement sections with minimal thickness variability - when a good contractor is on the job, and vice versa). This implies that more info or data are needed to assess range/variability of build thickness quality in each case (thin, medium, thick sections).

Reply: Thank you for the comment. Please Table 2

 

Reviewer 4 Report

This study fuses two NDT methods to investigate the correlation between GPR-measured thickness and IRI of pavement. The research topic is novel with potential engineering applications. The conclusions are consistent with the materials presented in the manuscript. I suggest considering the effect of vibration during high-speed NDT survey in the future study. For example, the thickness measured from GPR may be affected by vibration noise in the calculated dielectric constant profile. The authors may check the following reference:   https://www.sciencedirect.com/science/article/pii/S0950061822030902   https://ieeexplore.ieee.org/document/8302505   Overall, it is well-structured and well-written. I suggest minor revision before publication.

 

Well-written.

Author Response

This study fuses two NDT methods to investigate the correlation between GPR-measured thickness and IRI of pavement. The research topic is novel with potential engineering applications. The conclusions are consistent with the materials presented in the manuscript. I suggest considering the effect of vibration during high-speed NDT survey in the future study. For example, the thickness measured from GPR may be affected by vibration noise in the calculated dielectric constant profile. The authors may check the following reference:   https://www.sciencedirect.com/science/article/pii/S0950061822030902   https://ieeexplore.ieee.org/document/8302505   Overall, it is well-structured and well-written. I suggest minor revision before publication.

 

Reply: Thank you for the comment. Please see lines 482-487

Round 2

Reviewer 1 Report

The paper is improved, however, some questions are still not answered. Anyway, it can be accepted, but I have to say the novelty of this paper is very limited.

Reviewer 3 Report

Accept in current form.