MnDOT Contract No.1003326 WO 2 NRRA LT1: Developing Best Practices for Rehabilitation of Concrete with Hot Mix Asphalt (HMA) Overlays related to Density and Reflective Cracking Task -4: Asphalt Density Evolution Over Time Task Report Prepared by: Katie E. Haslett, Eshan V. Dave, and Jo E. Sias Department of Civil and Environmental Engineering, University of New Hampshire May 2020 Published by: Minnesota Department of Transportation Research Services Section 395 John Ireland Boulevard, MS 330 St. Paul, Minnesota 55155-1899 This report represents the results of research conducted by the authors and does not necessarily represent the views or policies of the Local Road Research Board, the Minnesota Department of Transportation or the University of New Hampshire. This report does not contain a standard or specified technique. The authors, the Local Road Research Board, the Minnesota Department of Transportation and the University of New Hampshire do not endorse products or manufacturers. Any trade or manufacturersβ names that may appear herein do so solely because they are considered essential to this report.
43
Embed
NRRA LT1: Developing Best Practices for Rehabilitation of ......Total Reflective Cracking (RCTotal) rate for all test sections..... 24. Figure 13: Normalized Total Reflective Cracking
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
MnDOT Contract No.1003326 WO 2
NRRA LT1: Developing Best Practices for Rehabilitation of Concrete with Hot Mix Asphalt (HMA) Overlays related to
Density and Reflective Cracking
Task -4: Asphalt Density Evolution Over Time
Task Report
Prepared by:
Katie E. Haslett, Eshan V. Dave, and Jo E. Sias
Department of Civil and Environmental Engineering, University of New Hampshire
May 2020
Published by:
Minnesota Department of Transportation Research Services Section
395 John Ireland Boulevard, MS 330 St. Paul, Minnesota 55155-1899
This report represents the results of research conducted by the authors and does not necessarily represent the views or policies of the Local Road Research Board, the Minnesota Department of Transportation or the University of New Hampshire. This report does not contain a standard or specified technique.
The authors, the Local Road Research Board, the Minnesota Department of Transportation and the University of New Hampshire do not endorse products or manufacturers. Any trade or manufacturersβ names that may appear herein do so solely because they are considered essential to this report.
2
Table of Contents List of Abbreviations ...................................................................................................................... 5
1. Introduction and Background ................................................................................................. 6
2. Materials and Methodology .................................................................................................... 7
Field Performance Indices ........................................................................................................ 37
Field Cores ................................................................................................................................ 39
Density Profile System ............................................................................................................. 43
3
List of Tables
Table 1: Summary of mixtures design properties. ......................................................................... 7 Table 2: Summary of mixtures and corresponding field sections at MnROAD. ........................... 8 Table 3: Monthly traffic schedule on test sections. ..................................................................... 10 Table 4: Summary of traffic loading on field section lanes from September 2017 to November 2019............................................................................................................................................... 11 Table 5: Ranking of test cells using distress survey data (Nov 2019) based on percentage of reflective cracking. ........................................................................................................................ 19 Table 6: Ranking of test cells using distress survey data (Nov 2019) based on reported transverse cracking in linear feet. ................................................................................................. 21 Table 7: Average reflective cracking (AvgRC) performance ranking of cells before and after normalization for Nov 2019 survey data. ..................................................................................... 23 Table 8: Reflective cracking (RCTotal) performance ranking of cells before and after normalization. ............................................................................................................................... 25 Table 9: Number of transverse cracks reported with low severity. ............................................. 33 Table 10: Number of transverse cracks reported with medium severity. .................................... 33 Table 11: Number of transverse cracks reported with high severity. .......................................... 34 Table 12: Transverse cracking in linear feet reported as low severity. ........................................ 35 Table 13: Transverse cracking in linear feet reported as medium severity. ................................ 35 Table 14: Transverse cracking in linear feet reported as high severity. ...................................... 36 Table 15: AvgRC for Cells 984-995. ........................................................................................... 37 Table 16: Normalized AvgRC for Cells 984-995. ....................................................................... 37 Table 17: TotalRC for Cells 984-995. ......................................................................................... 38 Table 18: Normalized TotalRC for Cells 984-995....................................................................... 38 Table 19: Field core thickness measurements. ............................................................................. 39 Table 20: Field core bulk specific gravity (Gmb) and corresponding air void. ............................ 40
4
List of Figures
Figure 1: Asphalt concrete overlay design cross sections (Test Cells 984-995). .......................... 9 Figure 2: Example of distress crack map for Cell 994. ............................................................... 12 Figure 3: Density profile system (DPS) data collection. ............................................................. 13 Figure 4: Laboratory dielectric constant measurement set up. .................................................... 14 Figure 5: Example of core location selection. ............................................................................. 15 Figure 6: Percent reflective cracking at joints after 8, 14, 19 and 26 months in service by test cell. ................................................................................................................................................ 17 Figure 7: Percent reflective cracking with respect to time in service. ......................................... 18 Figure 8: Transverse cracking in linear feet after 8, 14, 19 and 26 months in service by test cell........................................................................................................................................................ 20 Figure 9: Transverse Cracking (Linear feet) with respect to time in service. ............................. 20 Figure 10: Average reflective cracking (AvgRC) rate for all test sections. ................................ 22 Figure 11: Average reflective cracking (AvgRC) rate normalized by overlay thickness for all test sections. .................................................................................................................................. 22 Figure 12: Total Reflective Cracking (RCTotal) rate for all test sections. .................................. 24 Figure 13: Normalized Total Reflective Cracking (RCTotal) rate for all test sections. .............. 24 Figure 14: Comparison of air void levels with time in wearing course of in-situ density sections........................................................................................................................................................ 26 Figure 15: Change in air void content from 2017 to 2019 by test section with number above bars showing the rate of change for in-situ air void content per year. ................................................. 26 Figure 16: Average field dielectric value with time for Cells 988-991. ...................................... 28 Figure 17: Comparison of field dielectric to (a) average lab dielectric, (b) average lab air void and (c) Gmb. ................................................................................................................................... 29 Figure 18: Average dielectric by test cell section where error bars represent the standard deviation in measured dielectric constant. .................................................................................... 41 Figure 19: Lab measured air void versus field dielectric for wearing courses in Cell 988-991. . 41 Figure 20: Lab measured air void versus field dielectric for SPWEC440E overlay mixture. ..... 42 Figure 21: Dielectric constant with time. ..................................................................................... 43
5
List of Abbreviations β’ AV: Air Void β’ AvgRC: Average Reflective Cracking Rate Performance Index β’ Des. Air Void: Design Air Void Level β’ Des. Gyrations: Design Gyrations β’ Des. Total AC: Design Total Asphalt Content β’ DPS: Density Profiling System β’ ESAL: Equivalent Single Axle Loads β’ Gmb: Bulk Specific Gravity β’ HMA: Hot Mix Asphalt β’ MDR: Materials Design Recommendation β’ Mix ID: Mixture Identification β’ MnDOT: Minnesota Department of Transportation β’ NMAS: Nominal Maximum Aggregate Size β’ PASSRC: Permeable Asphalt Stabilized Stress Relief Coarse β’ PCC: Portland Cement Concrete β’ QA: Quality Assurance β’ RAP = Reclaimed Asphalt Pavement β’ %RC: Percent Reflective Cracking β’ RCTotal: Total Reflective Cracking Performance Index
6
1. Introduction and Background Task-4 of the βDeveloping Best Practices for Rehabilitation of Concrete with Hot Mix Asphalt (HMA) Overlays related to Density and Reflective Cracking,β project focused on gathering and analyzing in-situ density as well as field performance data from MnROAD test sections from the time of construction to present. The objective of this report is to summarize and compare field performance results in terms of distress surveys, Density Profiling System (DPS) and field cores with an emphasis on monitoring density evolution with traffic and time. Density data acquired through field cores was plotted using Google Earth and provided to the project TAP as an electronic deliverable accompanying this report; this Google Earth layer allows for graphical representation of density within each test section. Relationships between density and performance for a select number of sections were explored to analyze the effectiveness of the different construction and mixture design techniques to achieve enhanced density in the field at the time of construction and evaluate how density changes over time for various mixtures. Analysis performed in Task-4 includes field performance data from all distress surveys with an emphasis on the most recent survey date (Nov 2019) and provides critical information for the development and calibration of performance curves as part of Task-5.
This Task-4 deliverable provides a brief summary of the corresponding MnROAD sections within this study (Cells 984-995), followed by the methodology used to collect the various types of field performance data and relevant post-processing of the data. Next, results and discussion on field performance is provided with a focus on the in-situ density sections (Cells 988-991). Lastly, a summary and conclusion of key findings as part of this task deliverable are presented.
7
2. Materials and Methodology 2.1 Pavement Cross Sections This study investigates 12 full-scale pavement test sections (Cells 984-995) located at the MnROAD test facility on the original alignment of I-94 westbound. All pavement test sections are asphalt concrete overlays on 241 mm thick Portland Cement Concrete (PCC) slabs over 127 mm MnDOT Class 5 aggregate base material (MnDOT Grading and Base Manual, 2015). A brief summary of the asphalt mixtures and corresponding field sections evaluated as part of the fulfillment of Task-4 are presented in Table 1 and Table 2, respectively. Four of the test sections (Cells 988-991) are dedicated to a compaction study evaluating the impact of in-situ density (as well as mix design approaches regarding design air void levels) on reflective cracking performance. These specific test sections have the same overall pavement structure but contain varying surface course materials designed at different air void levels. Figure 1 shows a schematic of the designed cross sections.
A variety of field data has been collected including traffic loading, distress surveys (specifically crack maps) and in-situ density from both density profiling system (DPS) and field cores. Four condition surveys are available to date, in-situ density measurements have been performed twice, and four field cores from each transition section were taken in July 2019 after approximately two-years in service. In this task report, Fall 2019 field performance data will be used as the most current survey date for comparison of test sections, however field performance from all surveys dates is included in the report. The following subsections discuss in more detail each collection method and any post-processing techniques used on the collected data. Field performance results are presented along with discussion of test section rankings, focusing on the in-situ density sections (Cells 988-991).
2.2 Field Performance Data Collection Traffic Loading
Field test sections have been subject to approximately 600,000 flexible equivalent single axle loads (ESALs) and 900,000 rigid ESALs from the time of construction (September 2017) to the end of the year in 2019. Table 3 provides a visual representation of the monthly trafficking schedule at MnROAD on the respective test cells (typically for a week duration of the month), while Table 4 breaks down the volume of traffic loading by lane (driving or passing) over the past 3 years. Car traffic is classified as having MnDOT C1-C3 vehicles, while truck traffic contains C4-C13. The sum of traffic loading from all vehicle classes on both driving and passing lanes is approximately 7.1 million vehicles. It should be noted that traffic numbers reported in Table 4 are approximate values as they are recorded before traffic enters the lanes on the original alignment of I-94 westbound, and do not account for cases where vehicles change lanes before arriving at a particular test cell.
Table 3: Monthly traffic schedule on test sections.
Month Year
2017 2018 2019
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sept
11
Oct
Nov
Dec
Table 4: Summary of traffic loading on field section lanes from September 2017 to November 2019.
TOTAL 2,546,157 3,346,296 747,800 453,894 3,293,957 3,800,190
Sum of traffic loading on both lanes (2017-2019): 7,094,147
Distress Surveys
Figure 2 provides an example a distress crack map recorded by MnDOT staff for test cell section 994. Crack distress maps provided by MnDOT researchers for each cell were converted into a percentage of joints cracked within each test cell (excluding the transition zones between the cells). This was accomplished by first taking an inventory of all crack maps and quantifying the total crack length at each joint location. The percent of reflective cracking (% RC) within a test cell was calculated by taking the sum of crack lengths (at underlying joints) within a test cell and normalizing by the total length of PCC joints within the test cell (Equation 1).
12
Figure 2: Example of distress crack map for Cell 994.
Distress data from the MnDOT automated pavement distress survey van was also used to compare the reported transverse cracking (linear feet) in each test section. The total low, medium and high severity designations were added together for both the driving and passing lanes. In other words, each severity level was weighted equally and the total number of linear feet of transverse cracking reported.
Density Profile System (DPS)
Dielectric constant data was collected using the MnDOTβs density profile system (DPS) for all 12 MnROAD test sections in the driving and passing lanes (Figure 3). The dielectric constant refers to a materialβs ability to transmit electromagnetic waves through its medium. The dielectric constant of an asphalt mixture is derived from the dielectric values its constituents (air, binder and aggregate). Air and binder both maintain properties of an electrical insulator and have lower dielectrics reflective of a limited ability to translate electromagnetic waves. Bitumen, independent of viscosity, normally has a dielectric ranging between 2.6 to 2.8, while the value for air is even lower. Aggregate, due to its mineral deposits which readily transmit electrons, tends to have a higher dielectric constant ranging from 4.5 to 6.5 (Saarenketo, 2009). It is expected that a mixture with a higher density would have fewer air voids and a higher dielectric value.
Figure 3: Density profile system (DPS) data collection.
14
Data analysis for Task-4 focused primarily on the in-situ density test sections (Cells 988-991). Correlations between field DPS data and laboratory measurements on the field cores were made. Comparisons were made between field dielectric constant, lab measured dielectric constant, bulk specific gravity (Gmb), and laboratory air void content measurements. Measurement of the lab dielectric constants was performed at UNH using the single antenna set up shown in Figure 4.
Figure 4: Laboratory dielectric constant measurement set up.
Field Cores
Field cores were acquired in the summer of 2019 from transition zones of test sections. A total of 48 cores were sampled, with four cores taken from each test cell: two in the wheel path and two near the joint location. Core locations along the wheel path and joint were selected based on preliminary DPS measurements taken within test sections (by performing a serve test) that indicated these locations were appropriate for sampling representative low and high density areas within each test section. Equations 2 and 3 were used to calculate the low and high target dielectric constant values respectively. DPS equipment was then used to locate areas within the transition zone of each test cell that corresponded to the target low and high dielectric values for coring.
Figure 5 provides an example of marking the low (along joint) and high (wheel path) core locations for test Cell 989 prior to extraction. All field cores were measured for in-place thickness prior to trimming for Gmb measurements.
Figure 5: Example of core location selection.
2.3 Reflective Cracking Performance Indices To quantitatively compare reflective cracking performances of field sections, it is important to use cracking performance measures that take into account the rate of cracking, onset of cracking early versus later in overlay life, variable pavement cross sections and service life durations. A study by Oshone et al. (2019) proposed several field cracking performance measures for making recommendations regarding asphalt mix and thickness designs for pavements in cold climates. These indices focused primarily on transverse cracking in cold climates, therefore these indices were modified slightly for the implementation in the current study to evaluate reflective cracking performance. As a result, cracking measures considered in this study include average reflective cracking rate (AvgRC, Equation 4) and the total reflective cracking performance index (RCTotal, Equation 5). AvgRC calculates the average reflective cracking rate a pavement
16
experiences per month of service. The RCTotal index uses a concept of cracking work (similar to fracture work), where the total area of the curve between percent cracking and pavement service life is calculated and normalized with respect to the time of the last survey. More details on these indices were provided in the Task 3 deliverable of this project.
Where: AvgRC = average reflective cracking rate, % reflective cracking per month RCTotal = total reflective cracking performance index, % reflective cracking per month
Performance indices were also evaluated using a normalized thickness approach to take into account pavement structure. This approach provides a more objective comparison of test sections comprised of varying overlay structures (number of lifts and thickness) and provides the ability to account for the contribution of each individual mixture or layer in the overall structural performance of the pavement.
17
Results and Discussion 3.1 Field Performance
Overall performance, in terms of the reflective cracking amounts (%RC) reported at joints after 8, 14, 19 and 26 months in service, for all study test cells is summarized in Figure 6 . Another way to visualize the field performance data is to plot %RC with respect to time in service. Figure 7 clearly shows the importance of monitoring field performance with time as test sections with good early performance may not necessarily have good long-term performance. Test sections dedicated to studying the impact of in-situ density (Cells 988-991) on reflective cracking are denoted using dashed lines in Figure 7. In general, these test sections have experienced less reflective cracking than other sections to date, with the exception of Cell 991 which is the traditional Superpave 4% design with 9.5mm NMAS. Cell 989 (1.75 inch, 12.5 mm (5% AV) and 2.25 inch HMA, 19 mm) is showing the best performance followed by Cell 990 (1.75 inch, 12.5 mm (3% AV) and 2.25 inch HMA, 19 mm) and Cell 988 (1.75 inch, 12.5 mm (4% AV) and 2.25 inch HMA, 19 mm).
Figure 6: Percent reflective cracking at joints after 8, 14, 19 and 26 months in service by test cell.
Figure 7: Percent reflective cracking with respect to time in service.
Table 5 summarizes the most current field distress survey data and ranking with respect to reflective cracking performance from 1 (best performer) to 12 (worst performer). It can be observed that Cell 989 has the lowest percentage of joint reflective cracking (37.5%), while there are 7 out of 12 test sections that have over 75% joint reflective cracking.
0
10
20
30
40
50
60
70
80
90
100
0 3 6 9 12 15 18 21 24 27 30
% R
C
Time in Service (Months)
984 985 986987 988 989990 991 992993 994 995
19
Table 5: Ranking of test cells using distress survey data (Nov 2019) based on percentage of reflective cracking.
Test Cell %RC Rank
984 96.7 10
985 98.6 11
986 92.5 8
987 44.0 3
988 48.8 4
989 37.5 1
990 43.3 2
991 82.1 6
992 63.9 5
993 83.8 7
994 98.8 12
995 95.7 9
Distress data was collected using MnDOT survey van at the four different survey dates. Figure 8 summarizes the reported transverse linear feet of cracking by test section, while Figure 9 shows the same information but with respect to time in service. Table 6 provides a summary of the ranking of test sections as of the most recent survey date. It should be emphasized that a key difference between Figure 6 (% RC) and Figure 8 (transverse cracking) is that %RC focuses on cracking reported only at joint locations while transverse cracking includes cracking reported within the entire test section cross sectional area. For both performance measures, Cell 989 with the Superpave 5 design mixture was the best performer. The worst performing test sections were Cell 994 and Cell 984 in terms of %RC and transverse cracking. Both test sections contain the same 9.5 mm NMAS 1.5-inch thick wearing course (SPWEA440E), however what varied between these two sections was that a pre-treatment was applied to PCC in Cell 994 at time of construction. The overlay in Cell 994 was preceded by a polyurethane compaction grouting and void filling process. The idea behind the pre-treatment was to reduce the number of voids under the PCC slab to lower the deflections at joints and cracks in the concrete pavement. By filling these voids prior to an overlay, the differential movement between slabs may be significantly reduced or eliminated which, in turn, may reduce or eliminate reflective cracking and allow for a thin, single-lift overlay treatment to be used. Based on the ranking of %RC and %RC with time (Figure 7), it does not appear that the pre-treatment was effective in reducing the percentage of cracks reflected at joints in Cell 994 as compared to Cell 984. Overall, neither test section did particularly well compared to the other 10 test sections in terms of %RC or reported transverse cracking.
20
Figure 8: Transverse cracking in linear feet after 8, 14, 19 and 26 months in service by test cell.
Figure 9: Transverse Cracking (Linear feet) with respect to time in service.
3.2 Reflective Cracking Performance Indices In addition to evaluating the percentage of reflective cracking in each test section, AvgRC and RCTotal field performance measures were computed. To interpret results of AvgRC, the lower the bar (value) the better the performance. If the height of bars remain constant that means that the rate of cracking is fairly stable, however if an increase in AvgRC is observed between survey dates (sharp spike in bar height), that indicates a higher cracking rate between the two consecutive survey dates. On the contrary, if a decrease in AvgRC is observed that indicates a lower cracking rate between consecutive survey dates. It should also be emphasized that it is important to make comparisons of test sections with time rather than at a single point. For example, test sections that performed well early by having a lower AvgRC may appear to be performing worse in the current survey date comparison due to more joints being βavailable to crackβ during the given analysis period.
Figure 10 shows the AvgRC results for all test cells at the four distress survey collection points since placement. Cells 984, 985 and 994 had comparable performance with approximately 3.8% average reflective cracking reported in Nov 2019. One difference between these test sections was the early field performance of Cells 984 and 985 was better than Cell 994 (where approximately 8.5% average reflective cracking rate was reported). Comparing only the in-situ density sections, Cell 991 reported the highest AvgRC while Cell 989 reported the lowest amount in Nov 2019.
22
Figure 10: Average reflective cracking (AvgRC) rate for all test sections.
Figure 11 also shows the average reflective cracking rate (AvgRC) but adjusted with respect to overlay thickness (by multiplying the cracking performance measure by overlay thickness in meters).
Figure 11: Average reflective cracking (AvgRC) rate normalized by overlay thickness for all test sections.
Taking a closer look at the performance ranking based on the most current distress survey date (Nov 2019) before and after normalization of each test section, Table 7 shows that the ranking of test sections varies. Prior to thickness normalization, Cell 989 was ranked as the best performer;
after accounting for overlay pavement structure Cell 995 was ranked as the best performer. This is in part due to substantial change in overlay structure between the two cells: 4 inches (Cell 989) and 0.75 inch (Cell 995). A summary of AvgRC and normalized AvgRC can be reviewed in Appendix A for all four distress survey dates.
Table 7: Average reflective cracking (AvgRC) performance ranking of cells before and after normalization for Nov 2019 survey data.
Figure 12 shows the RCTotal results at the four distress survey collection points in time since construction. From the most recent distress survey time the worst performing Cell is 994 while the best performing test Cell is 990. Figure 13 shows the normalized RCTotal results with the thickness adjustment. Unlike average reflective cracking rate, since RCTotal encompasses cracking performance over the life of the pavement, the performance measure from latest survey is most reliable for comparisons between various cells. This parameter gives credit to test sections that crack later and have early good performance.
24
Figure 12: Total Reflective Cracking (RCTotal) rate for all test sections.
Figure 13: Normalized Total Reflective Cracking (RCTotal) rate for all test sections.
Table 8 provides a closer look at the ranking of test sections based on RCTotal versus the normalized RCTotal. The best performing test section was Cell 990 in both cases, while the worst performing test section was Cell 994 for RCTotal and Cell 985 after thickness normalization. Again, further comparison of all test sections over the four different survey dates can be reviewed in Appendix A.
0.01.02.03.04.05.06.07.08.09.0
984 985 986 987 988 989 990 991 992 993 994 995
RC
Tot
al (%
/mon
th)
Test Cell Section
May-18 Nov-18 Apr-19 Nov-19
Density Sections
0.0
0.1
0.1
0.2
0.2
0.3
0.3
0.4
984 985 986 987 988 989 990 991 992 993 994 995
Nor
mal
ized
RC
Tot
al
(%-m
/mon
th)
Test Cell Section
May-18 Nov-18 Apr-19 Nov-19
Density Sections
25
Table 8: Reflective cracking (RCTotal) performance ranking of cells before and after normalization.
Lift thickness was verified for each test section by taking the average of the four cores extracted in each transition zone. In general, the in place thickness was higher than the designed thickness. However, it should be noted that these comparisons are made with transition zone cores and not with cores taken in the actual field test section themselves, therefore lift thickness within the actual section may vary. Appendix A can be reviewed for actual measured thickness of field cores prior to cutting for bulk specific gravity measurements.
MnDOT staff used the Corelok method (AASHTO T 331) to perform bulk specific gravity measurements on field cores that were sampled after approximately two years in service. A summary of individual Gmb values and corresponding percent air void for each field core can be found in Appendix A. Most of the air void percentages were within expected ranges compared to design air void levels by mix type. The only exception was the base course (SPWEC440E) used in Cell 987 having a slightly higher average air void content of 11.3% compared to other test sections that contained the same base course mixture with an average air void content of 6.0%.
Researchers then used quality assurance (QA) field cores collected at the time of construction (2017) in transition areas and the field cores collected in 2019 to investigate the change in air void percentage with traffic and time. Figure 14 shows the measured air void percentage for wearing course mixtures used in Cells 988-991 with time. Similar trends and rankings were observed when plotting the measured air void with respect to traffic instead of time.
26
Figure 14: Comparison of air void levels with time in wearing course of in-situ density sections.
Figure 15 shows the air void content broken down by test section, with the number above the bar representing the rate of change from 2017 to 2019. Cell 989, which contained the Superpave5 mixture, had the lowest rate of change (-0.20 %/yr). In comparison, Cell 988 had the highest rate of change (-1.71 %/yr), corresponding to a decrease in air void level of approximately 3.4% after only two years in service. To date, the Superpave5 (Cell 989) and regressed air void (Cell 990) mixtures have experienced less densification than the traditional 4% mixture designs (NMAS of 12.5 mm and 9.5mm ) used in Cells 988 and 991 respectively.
Figure 15: Change in air void content from 2017 to 2019 by test section with number above bars showing the rate of change for in-situ air void content per year.
0123456789
1011
2016 2017 2018 2019 2020
Air
Voi
d (%
)
Time (Years)
988 989990 991
0123456789
1011
988 989 990 991
Air
Voi
d (%
)
Test Section
2017 2019-1.71 %/yr
-0.20 %/yr
-0.53 %/yr
-0.88 %/yr
27
Density Profiling System (DPS)
Dielectric constant was measured using DPS following overlay construction in 2017 and again two years later in 2019. Typically, DPS is performed on newly constructed pavements and not on aged pavements. One reason for this is due to the higher potential for moisture presence during the service life of a pavement compared to immediately following lay down of a pavement and the impact on acquiring an accurate dielectric constant reading. Regardless of time in service, it is accepted that an increase in dielectric constant corresponds to an increase in density.
Density evolution of Cells 988-991 from 2017 to 2019 is shown in Figure 16 using the average dielectric constant collected along the length of each test section (in both driving and passing lanes). Appendix A may be referenced to view the density evolution of all test sections. While there are currently only two survey points of reference, it is anticipated that DPS measurements will be performed in 2020 allowing researchers to better track density evolution with time. From Figure 16 it can be seen that all in-situ sections experienced an increase in dielectric constant after two years of service, however the rate of increase (slope of the line) varies by overlay material type. All in-situ density sections have the same base course mixture (SPWEC440E) and total overlay thickness (4 inches) but vary by wearing course mixture type. The rate of increase in dielectric was slightly higher (slope of line) for Cell 989 (Superpave 5 wearing course) and Cell 990 (regressed air void wearing course) test sections. This would indicate higher potential densification of those mixtures compared to tradition mix designs and is at odds with results presented from field cores in the prior section. It is hypothesized that the difference in conclusion regarding density evolution with time for these mixtures is due to the fact that Figure 14 is showing data collected from transition zones (single point measurements), while Figure 15 is showing averaged data collected from within the actual field test section. The density evolution of these sections will continue to be monitored with DPS and it is anticipated that with further data collection researchers will be able to provide more insightful comparisons of these mixtures.
The percent difference in dielectric constant from 2017 to 2019 was calculated to verify if a significant difference existed between these test sections. The highest percent difference in dielectric was observed in Cell 990 with the regressed air void mixture (SPWEB430E) at 3.8% followed closely by Cell 989 with the Superpave5 design mixture at 3.7% difference. Meanwhile, Cell 988 (traditional 4% air void design, 12.5mm NMAS) and Cell 991 (traditional 4% air void, 9.5mm NMAS) changed by approximately 2.6% and 2.5% respectively. The average standard deviation in dielectric constant measurement was fairly consistent in 2017 and 2019 with values of 0.09 and 0.10 respectively. Therefore, only a minimal difference (approximately 1%) in dielectric constant measurement variability from 2017 to 2019 is observed among the in-situ density sections. Appendix A can be reviewed for further comparison of variability in dielectric constant measurements by test cell section.
28
Figure 16: Average field dielectric value with time for Cells 988-991.
In addition to looking at the evolution of density with time among in-situ density sections, researchers also explored the relationship between average dielectric measurements taken in 2019 at field core locations (measurement taken at a single point using time collect method for 10 seconds to produce and average dielectric constant value) with lab dielectric measurements taken on extracted field cores, calculated air voids percentage and Gmb (Figure 17). Lab dielectric measurements on field cores were taken using Time of Flight (TOF) method as outlined in Hoegh et al., 2019. However, it should be noted that lab dielectric measurements shown below are for wearing mixtures only from multiple lift field cores after trimming in preparation to perform Gmb (may have lower thickness and represent single lift measurement). After cutting cores, thicknesses on individual wearing lifts in Cell 988-991 ranged from 1.7 inches to 2.9 inches, thereby causing some cores (mainly the wearing course in Cell 991) to not meet the recommended minimum thickness of 2 inches. Despite this fact, good correlation between field and lab measured dielectric constant was observed and is shown in Figure 17a. A potential for some discrepancy between field and lab dielectric measurement can be due to field dielectric measurements being taken in place with test sections that contain multiple lifts (wear and base course).
As expected, an inverse relationship is observed between lab measured air voids on field cores and field dielectric constant taken at the location of core (Figure 17b). Meaning, as the dielectric constant increase the air void percentage decreases. Finally, a positive linear relationship is observed between field dielectric measurements and average Gmb values for wearing course mixtures obtained from field cores corresponding to each test section (Figure 17c). Further comparison was done between field dielectric measured at the same location that field cores were extracted and their respective lab measured air void percent by mixture type and can be reviewed in Appendix A.
4.70
4.75
4.80
4.85
4.90
4.95
5.00
5.05
2016 2017 2018 2019 2020
Ave
rage
Die
lect
ric C
onst
ant
Time (Years)
988 989 990 991
29
Figure 17: Comparison of field dielectric to (a) average lab dielectric, (b) average lab air void and (c) Gmb.
4.8
4.85
4.9
4.95
5
5.05
5.1
5.05 5.1 5.15 5.2 5.25
Ave
rage
Fie
ld D
iele
ctric
Average Lab Dielectric
4.8
4.85
4.9
4.95
5
5.05
5.1
0 2 4 6 8
Ave
rage
Fie
ld D
iele
ctric
Average Lab Air Void (%)
(b)
4.8
4.85
4.9
4.95
5
5.05
5.1
2.320 2.330 2.340 2.350 2.360
Ave
rage
Fie
ld D
iele
ctric
Gmb
(c)
(a)
Cell 988 (NMAS 12.5mm, AV 4%)
Cell 989 (NMAS 12.5mm, AV 5%)
Cell 990 (NMAS 12.5mm, AV 3%)
Cell 991 (NMAS 9.5mm, AV 4%)
30
3. Summary and Conclusions Task-4 summarized and compared field performance results in terms of distress surveys, Density Profiling System (DPS) and field cores with an emphasis on monitoring in-situ density sections (Cells 988-991). The following key conclusions are made based on the results and discussion presented in task deliverable.
β’ In general, in-situ density test sections have experienced less reflective cracking than other sections to date, with the exception of Cell 991. Cell 989 (1.75 inch, 12.5 mm (5% AV) and 2.25 inch HMA, 19 mm) is showing the best performance followed by Cell 990 (1.75 inch, 12.5 mm (3% AV) and 2.25 inch HMA, 19 mm) and Cell 988 (1.75 inch, 12.5 mm (4% AV) and 2.25 inch HMA, 19 mm).
β’ Field performance indices can be useful in evaluating the percentage of reflective cracking reported in each field section while taking into account variable pavement cross section structures and their service life durations. Comparing all test sections, the best performer in terms of AvgRC and RCTotal varies before and after thickness normalization. Comparing only the in-situ density sections (which all have the same overall thickness), Cell 991 reported the highest (worst) AvgRC while Cell 989 reported the lowest (best) amount in Nov 2019. In terms of RCTotal, Cell 990 was the best performer and Cell 988 was the worst.
β’ The use of DPS to measure dielectric constant allowed for the tracking of density evolution with time. All in-situ density sections experienced an increase in dielectric constant after two years of service, however the rate of increase (slope of the line) varies by overlay material type. Results from DPS collected within the test section show that Superpave 5 and regressed air void mixtures experienced more densification compared to traditional 4% air void design mixtures under the same traffic conditions. In contrast, results based on field cores sampled in 2017 during the time of construction and comparing measured air voids on field cores acquired in 2019 (from transition zones), show that the Superpave 5 and regressed air void mixtures experienced less densification compared to a traditional 4% air void design mixtures. This will be continuously monitored and conclusions revised as more DPS data is collected with time.
β’ Comparing trends for in-situ density sections based on DPS dielectric constant data and measured air voids on field cores with only two reference points show different conclusions. An increase in dielectric was observed by all in-situ density test sections with minimal difference (approximately 1%) in dielectric constant from 2017 to 2019 among the in-situ density sections. Meanwhile, the change in air void level based on field cores was lower in Cell 989 (Superpave 5 design mixture wearing course) and Cell 990 (regressed air void wearing course) compared to other test section with traditional mix designs after two years of service. Cell 989 exhibited the lowest annual rate of air-void reduction (-0.20 %/year) while Cell 988 had the highest rate of air-void reduction (-1.71 %/yr). Future DPS and field core data collection will help to distinguish the evolution of density with time and determine if there is more densification in one test section compared to another.
A continued effort will be made to monitor field performance of all test sections in the coming years. Similar field performance data collection and post processing techniques will be
31
used on upcoming collected data. These efforts will play a vital role in the development of relationships among field performance and predicted performance curves in subsequent tasks.
32
4. References Hoegh, K.; Roberts, R.; Dai, S.; Zegeye Teshale, E. Toward Core-Free Pavement Compaction Evaluation: An Innovative Method Relating Asphalt Permittivity to Density. Geosciences 2019, 9, 280. Oshone, Mirkat, Eshan V. Dave, and Jo E. Sias. "Asphalt mix fracture energy based reflective cracking performance criteria for overlay mix selection and design for pavements in cold climates." Construction and Building Materials 211 (2019): 1025-1033. Saarenketo, Timo. Measuring Electromagnetic Properties of Asphalt for Pavement Quality Control and Defect Mapping. Roadscanners. 2009.
33
Appendix A Transverse Cracking (Number of Cracks) Tables 9 through 11 provide a summary of the number of transverse cracks reported with low, medium and high severity at each of the four distress survey dates. Only Cells 991 and 989 reported having medium or high severity transverse cracking respectively.
Table 9: Number of transverse cracks reported with low severity.
Cell Number
Number of Transverse Cracks (Low) May-18 Nov-18 Apr-19 Nov-19
Transverse Cracking (Linear feet) Tables 12 through 14 provide breakdown of the reported transverse cracking in each test section in terms of linear feet with low, medium and high severity. Results are presented at all four distress survey dates.
Table 12: Transverse cracking in linear feet reported as low severity.
Cell Number
Transverse Cracking in Linear Feet (Low) May-18 Nov-18 Apr-19 Nov-19
Figure 18 breaks down the average dielectric by test cell showing the variability in measured dielectric constant using error bars.
Figure 18: Average dielectric by test cell section where error bars represent the standard deviation in measured dielectric constant.
Figure 19 summarizes the relationship between field dielectric and measured air voids on field cores by overlay mixture types used in wearing courses for Cells 988-991, while Figure 20 for the base course mixture used in all in-situ density sections. Only a mild correlation exists for the base course mixture (SPWEC440E) used in the lower pavement layer.
Figure 19: Lab measured air void versus field dielectric for wearing courses in Cell 988-991.
4.00
4.25
4.50
4.75
5.00
5.25
5.50
988 989 990 991
Ave
rage
Die
lect
ric
Test Cell
2017 2019
RΒ² = 0.0048
RΒ² = 0.0098
RΒ² = 0.3122
RΒ² = 0.158
2
3
4
5
6
7
8
9
4.7 4.8 4.9 5 5.1 5.2
Lab
Air
Voi
d
Field Dielectric
A440E
B450E
B430E
B440E
42
Figure 20: Lab measured air void versus field dielectric for SPWEC440E overlay mixture.
RΒ² = 0.5022
3.03.54.04.55.05.56.06.57.07.58.0
4.7 4.8 4.9 5 5.1 5.2
Lab
Air
Voi
d
Field Dielectric
43
Density Profile System Figure 21 shows a comparison of dielectric with time by test section. Note that DPS data was not collected for Cell 995 in 2017.