Abstract
Performance engineered concrete mixtures (PEM) include optimized mixture designs which, paired with advanced quality assurance methods, provide substantially improved durability, economy, and sustainability. One agency challenge to implementation of PEM has been engagement of contractors and other industry stakeholders in field deployment. In this case, an agency, North Carolina Department of Transportation (NCDOT), US, found a willing contractor partner to support an initial concrete paving PEM implementation project. The collaborative effort supported both agency PEM implementation initiatives and provided benefits to the contractor. PEM tests included the Box Test, Super Air Meter (SAM), and surface resistivity, as “shadow tests.” The contractor found the Box Test useful for mixture development and evaluation of modifications during the project. The SAM and resistivity meter were readily implemented into standard quality control (QC) testing without additional QC staff. SAM tests were performed over the course of the project, but the data was variable, and additional training is needed to support improved results. Surface resistivity test results indicated the contractor improved concrete consistency over the two project phases, and the concrete should provide good durability performance. The contractor intends to implement PEM tests for QC on future projects to help improve pavement quality. The agency found surface resistivity tests for acceptance easy to perform. For mixtures containing fly ash, research is needed to identify a 28-day surface resistivity target that generally correlates to a 56-day or 90-day resistivity that predicts good durability performance. This experience reinforced that collaboration between agency, industry, and academia will support advancement of PEM initiatives.
Keywords
The long service life expectations of infrastructure components are often difficult to meet by using typical tests for specification and acceptance used by agencies, which center around three criteria: slump, air content, and compressive strength. These three criteria are only loosely related to deterioration phenomena and therefore do not always ensure satisfactory field performance. Consistent with the focus of Moving Ahead for Progress in the 21st Century Act (MAP-21) legislation on performance, there is a desire by the Federal Highway Administration (FHWA), public agencies, and industry to move toward performance engineered construction materials ( 1 , 2 ). Performance engineered concrete mixtures (PEM) include optimized mixture designs (materials selection, gradation, cement content, etc.) which, paired with advanced quality assurance methods, provide substantially improved durability, economy, and sustainability. American Association of State Highway and Transportation Officials (AASHTO) PP 84 “Performance Engineered Concrete Pavement Mixtures” provides agencies “with tools to prepare a specification for concrete pavement mixtures that moves closer to measuring and basing acceptance on parameters that are truly critical to the long-term performance of the system” ( 3 ).
Although many agencies have made strides toward implementing PEM, one challenge has been securing the engagement of contractors and other industry stakeholders ( 4 , 5 ). North Carolina Department of Transportation (NCDOT), US, found a contractor partner willing to support an initial PEM implementation project, and the resulting collaboration both supported advancement of agency initiatives and provided benefits for the contractor in mixture development and quality control (QC).
Before the implementation project, researchers developed shadow specifications and performance targets for PEM tests of interest to NCDOT. The researchers supported the pilot project with data analysis documentation of the effort, including lessons learned that will be translated to future PEM pilot studies. The resulting synergy between agency, contractor, and academia supported a successful first implementation of PEM in North Carolina (NC).
Movement Toward PEM
NCDOT’s initial steps to move toward PEM were made through two research studies, one completed before this field study and one ongoing ( 6 , 7 ). Together, these studies included testing to establish performance-related criteria for several emerging PEM test methods and, based on these results, shadow specifications for use in pilot projects were established. Research focused on three PEM tests (Super Air Meter [SAM], surface resistivity, and shrinkage), although only two of these tests, SAM and surface resistivity, were used in this pilot project. The PEM initiative also includes improved tests for workability, one of which was used in this demonstration project. The Box Test, AASHTO TP 137, is intended for use by the contractor to evaluate the workability of a concrete paving mixture under vibration energy ( 8 , 9 ).
The SAM test, AASHTO PP 118, provides a measurement of the size and spacing (dispersion) of the air void system within concrete by measuring a change in response of concrete to a series of sequential pressures ( 10 , 11 ). The total volume (%) of air contained in fresh concrete is also measured. The SAM number has been shown by the device developers and other agencies and researchers to correlate with the air void spacing factor computed using ASTM C457, and with rapid freeze thaw durability testing per ASTM C666 ( 11 – 14 ). Other researchers did not find a correlation between the SAM number and spacing factor, or the SAM number and freeze-thaw durability, but found that a threshold SAM number could be identified that is, in most cases, comparable with an acceptable ASTM C 457 spacing factor ( 15 ).
Surface resistivity testing, AASHTO T 358, provides an indication of concrete’s permeability and its ability to resist the penetration of chloride ions and other deleterious agents ( 16 ). This test has been correlated to the AASHTO T 277 rapid chloride permeability test ( 17 , 18 ).
In 2018, NCDOT applied for funds to support PEM implementation as part of FHWA’s “Demonstration Project for Implementation of Performance Engineered Mixtures/AASHTO PP 84.” Funding to support three categories of implementation were secured ( 4 ):
Category A: Incorporating two or more AASHTO PP 84-17 tests in the mixture design/approval process ( 19 ). Shadow testing was acceptable.
Category B: Incorporating one or more AASHTO PP 84-17 tests in the acceptance process ( 19 ). Shadow testing was acceptable.
Category D: Requiring the use of control charts, as called for in AASHTO PP 84-17 ( 19 ).
Contractor personnel approached the research team early in 2018, indicating interest in using PEM tests to improve their QC and willingness to support NCDOT’s PEM initiatives. Based on this interest, the contractor was asked to partner in supporting the FHWA Implementation Funds PEM demonstration project. The contractor suggested a design-build urban interstate project, a stretch of I-85 widening north of Charlotte, NC. Although the contract had been let and awarded before the decision to use this project as a PEM demonstration, parties collectively agreed on scopes of work to support the implementation categories listed above.
Project Background
The project included widening of 5.3 mi of I-85 in Rowan County, NC. The existing four-lane interstate (two lanes in each direction) was widened to provide four additional lanes (two lanes in each direction) to support an eight-lane interstate. The project included approximately 500,000 square yards of concrete pavement construction at a total project cost of $140 million ( 20 ).
The mainline I-85 pavement was the focus of this PEM demonstration project. The existing 10 in. thick continuously reinforced concrete pavement on aggregate base course was constructed in 1978. The 2015 average daily traffic (ADT) for this segment of I-85 was 97,100, with a design year (2040) ADT of 179,500. Truck traffic was estimated in 2015 to be 19% [14% truck tractor semi-trailer (TTST) and 5% duals]. The new mainline pavement is 12 in. thick doweled jointed concrete, paved on a non-woven geotextile interlayer, and a 1½ in. asphalt surface course interlayer (SF9.5A) placed on stabilized subgrade. The travel lanes are each 12 ft wide with a 22 ft median (11 ft concrete inside shoulders separated by a T-wall) separating the four lanes in each direction.
Phase 1 construction shifted traffic to the outside shoulders and existing traffic lanes in both directions to construct two new travel lanes, an inside shoulder, and a new median barrier. Phase 2 included shifting traffic to the two newly constructed inside lanes in each direction, so the outer two lanes, outside shoulder, and roadside could be completed. Phase 1 mainline paving began in April 2018 and was completed in November 2018. Phase 2 mainline paving began during April 2019 and concluded in October 2019. PEM tests were primarily used on the mainline paving mixtures, although concrete used for some ramps and selected shoulder locations was also tested using the PEM tests. Additional background information, data, and analysis is provided in two other publications ( 21 , 22 ).
Concrete Mixtures
NCDOT Standard Specifications require that mixture designs for concrete pavements contain at least 526 lb per cubic yard (yd3) of cement, have a maximum water-to-cementitious materials (w/cm) ratio of 0.559, an air content ranging from 4.5% to 7.5%, and a maximum slump of 1.5 in. ( 23 ). Minimum strengths at 28 days are 650 pounds per square inch (psi) (flexural) and 4,500 psi (compressive). Fly ash can be substituted for up to 30% of cement at a replacement rate of 1:1 by weight.
Mixture designs for the project were developed by the contractor, and batched and tested at the contractor’s QC laboratory to develop the data required for NCDOT approval. Category A of the FHWA PEM Implementation Funds Program required incorporating two or more AASHTO PP 84-17 “shadow” tests in the mix design/approval process, with shadow testing being acceptable ( 19 ). All mixtures were accepted using NCDOT’s current process, using the information submitted a standard form. In addition to tests currently required by NCDOT for approval, PEM test data was collected using the Box Test, SAM, and surface resistivity.
Three mixture designs were used for the mainline paving for the project (Table 1). Water reducing, retarding, and air entraining admixtures were used as needed. The intention of obtaining multiple approved mixture designs was to ensure the required compressive strength could be met in a variety of construction and weather conditions. Paving began using the 472SLNS mixture design, but, as the production process became reliable and favorable weather prevailed, the primary mixture used for almost all of the mainline pavement became mixture ID 460SLNS, which had a lower cementitious materials content. The vast majority of PEM test data was collected for mixture 460SLNS. Mixture 496HPNS was used as the “hand-placed” mixture (3 in. slump concrete consolidated and finished by roller screed and hand vibrators), and had an increased paste and water content to achieve the desired workability for non-paver placement. To ensure strength at the higher water content the cementitious materials content of this mixture was increased, although the w/cm ratio was held roughly constant between the slipformed and hand-placed mixtures.
Mixture Proportions for Mainline Paving Mixtures
PEM Test Results Obtained During Mixture Development
The Box Test was performed intermittently during mixture development, assisting the contractor in assessing the changes in workability and edge slump resulting from changes in materials and proportions, until the final mixture designs were identified and submitted for approval. Figure 1 illustrates the progression of surface void reduction during mixture development. During construction, the contractor found the Box Test very useful in optimizing mixtures, particularly when admixture dosages were changed to accommodate changes in temperature, placement conditions, or material variability (such as moisture content, percent fines, and fly ash carbon content).
Example Box Test results obtained during mixture development phase, showing progression of trial mixtures toward suitable workability.
Since the contractor agreed to participate in the PEM demonstration project after the project was awarded, much of the preliminary work, including batch plant setup, concrete mixture development, and verification of batch plant operations, had already been performed by the time the contractor obtained the SAM and surface resistivity meter testing devices. However, once the contractor obtained the test equipment, QC personnel were able to perform a limited amount of SAM and resistivity tests before construction began.
SAM was used on one batch of one mixture during the plant verification and test batching stages. The mixture had a fresh air content of 6.4% with a SAM number of 0.23. Resistivity measurements made at ages of 1, 6, 7, and 14 days were 1.8, 3.2, 3.3, and 3.8 kΩ•cm, respectively. Specimens for QC resistivity testing were cylinders used for compressive strength testing, cured in the contractor’s QC laboratory in lime water curing tanks. Per AASHTO T 358, the averages of all surface resistivity readings were multiplied by 1.1 to account for this type of curing method.
Production and Construction
An on-site portable, central-mixed, double-drum batch plant with a 12 yd3 capacity was used, producing 10 yd3 batch loads. The plant could produce 280 to 320 yd3 of concrete per hour at full production. However, production was often limited because of the reduced availability of haul trucks during the exceptionally busy 2018/2019 construction seasons. Therefore, a typical production rate was approximately 200 yd3 per hour. Typical paver rates were approximately 200 ft per hour (fph), with a maximum of 325 fph on a peak day. Adjustments within allowable tolerances, such as admixture dosages, were performed periodically because of variations in haul time or weather.
PEM QC Test Results
A key component of FHWA’s PEM initiative is improving contractor QC ( 2 ). QC testing was performed at the field laboratory (SAM test and hardened concrete tests) and in front of the paver (slump, temperature, and air content via pressure meter). NCDOT’s QC requirements are outlined in specification section 1000-3 (E) Contractor’s Responsibility for Process Control ( 21 ). In addition to agency QC requirements, Category D of the FHWA PEM Implementation Funds required the contractor to use control charts. Although not currently required by NCDOT, this contractor typically prepares a spreadsheet-based database/control chart for air content, slump, unit weight, concrete temperature, and compressive strength, with one measurement recorded per lot. NCDOT specifications define a lot as 1,333.3 square yards of pavement, or fraction thereof, placed within 28 days ( 23 ). As part of this PEM demonstration project, SAM test results and resistivity test results were added to the chart.
SAM
SAM tests were performed intermittently (target of once per day) on fresh concrete sampled at the batch plant. This project marked the first use of the SAM by this contractor. At first, multiple users took measurements with the device, although roughly the last 70% of SAM measurements were made by a single user. Early in Phase 1, several SAM test results with very high values were recorded, indicative of the learning curve for technicians before becoming proficient with the SAM. Based on the more extensive experience of the research team personnel with typical NCDOT paving mixtures with 4.5% to 7.5% air contents, SAM numbers greater than 0.60 often indicate an error (such as a leak) during the test procedure. Therefore, before analysis, SAM numbers greater than 0.60 were removed from the dataset. All data is provided in the project report ( 19 ). A total of 78 SAM test results for mixture 460SLNS and 6 SAM test results for mixture 496 HPNS (n = 84 for the full dataset) were analyzed, with minimum, maximum, and average test results for each mixture and for the combined dataset shown in Table 2.
Air Content and Air Void System Tests
Note: SAM = Super Air Meter; Min. = minimum; Avg. = average; Max. = maximum; psi = pounds per square inch.
The SAM air content measured at the plant tended to consistently run slightly lower than the ASTM C231 air content measured at the paver using the Type B meter (Table 2 and Figure 2). The reason for this is not readily evident, and many studies, including those in previous research studies with North Carolina concrete, have shown strong agreement between the SAM and Type B meter ( 6 , 9 , 15 ). The air void system of concrete can change over time as a result of many materials- and construction-related factors, and the air content of concrete can increase with handling ( 24 ). SAM measurements made on concrete before it was loaded into the paver may have provided different (lower) SAM numbers, more consistent with the higher total air contents measured using the Type B pressure meter. It is recommended that, in future studies, SAM tests be performed at the same location as the Type B meter to provide a more direct comparison.
Air content using ASTM C231Type B pressure meter at paver versus Super Air Meter (SAM).
Previous research using typical NCDOT paving mixtures indicated a SAM number of approximately 0.30 corresponds to mixtures exhibiting satisfactory freeze-thaw performance in the ASTM C666 Procedure A test ( 25 – 27 ). Figure 3 shows the relationship between the SAM number and air content measured by the SAM device and Type B meter. For this project, little correlation was observed between the total air content and SAM number, and this variability has not been observed previously. However, it can be noted that the average SAM number of the slipformed paving mixture 460SLNS met the shadow specification target SAM number of 0.30 (Table 2).
Air content measured with Super Air Meter (SAM) and Type B pressure meter versus SAM number.
It is theorized that measurement of the air void system using the SAM at the paver (instead of at the plant) may have provided both total air contents closer to that of the ASTM C231 test and, potentially, lower SAM numbers. Additional field study is recommended, with SAM tests performed closer to the paver alongside the ASTM C231 testing. The dispersion of the data that can be observed in Figure 3 is likely attributable to the inexperience of the contractor with the device. High SAM numbers observed from high air content mixtures may be because of leaks occurring during the test, or other issues. Similar variability has been reported in other field studies, but agencies have reported improved results after additional training from the developer ( 12 , 28 , 29 ).
The “train the trainer” approach used by other agencies, where developers of the SAM visit the jobsite or agency facility, may be useful for future PEM shadow projects ( 29 ). The daily leak check procedure suggested by the SAM developer should be implemented, and the vibration procedure for consolidation of the concrete for the SAM could be used, since it could more consistently consolidate the concrete in the device. This vibration procedure is described in AASHTO TP 118 as an alternative to the rodding procedure, and may provide improved consolidation of the sample in the measuring bowl. Additionally, the reliability factor could be used to check the accuracy of future SAM tests ( 12 ). The algorithm used to develop the reliability factor uses machine-based learning to help detect issues such as leaks at the rim of the device or leaks in the upper chamber. Although the algorithm is still in development, some users of the SAM have found the currently available version (programmed into a spreadsheet) useful in training operators, reducing the number of incorrectly run tests, and identifying problematic measurements ( 12 ).
Surface Resistivity
Resistivity measurements were performed on cylinders cast for compressive strength testing throughout both Phase 1 and 2 paving, typically at ages of 3, 28, 56, and 90 days. Additional tests were made at different ages (e.g., 4, 8, 29 days) based on the contractor’s need to test cylinders for compressive strength on those days. Concrete was sampled at the batch plant, initially cured in the field laboratory, then stripped and lime-water cured. Per AASHTO T 358, the averages of all readings were multiplied by 1.1 to account for lime-water curing.
For both Phase 1 and 2, resistivity measurements were made at 3 and 28 days. During Phase 1, resistivity measurements were taken at 90 days, while during Phase 2 paving resistivity measurements were taken at 56 days. The reason the later-age tests were switched from day 90 (in Phase 1) to day 56 (in Phase 2) was driven by ongoing research to identify 56-day resistivity targets that appear suitable to assess the performance of higher (up to 30%) fly ash mixtures. It was desired to compare 56-day field resistivity measurements from this project with proposed 56-day resistivity targets being developed in the laboratory.
A total of 1360 resistivity measurements were made on the primary paving mixture 460SLNS, with 677 made during Phase 1 and 683 made during Phase 2. A total of 125 resistivity measurements were made on the hand-placed mixture 496HPNS, with 67 made during Phase 1 and 58 made during Phase 2. Measurements for both mixtures at each testing age are shown in Figure 4.
Surface resistivity versus age.
Mixtures contained fly ash, and as expected, experienced a significant increase in resistivity between 28 and 90 days. The proposed resistivity target for concrete pavement mixtures of 11 kΩ•cm, identified by the research team during previous studies is also shown on Figure 4 ( 6 , 30 ). This target was identified through laboratory testing by relating resistivity test results to rapid chloride permeability test results and agency experiences with adequate field performance of in-service structures. At this time, the target resistivity is anticipated to be reached by mixtures containing fly ash by 56 or 90 days, with ongoing research to support refinement of the proposed specification ( 7 ).
As can be observed, 85 of 106 (80.1%) of the 56-day test results of the lots of the 460SLNS concrete mixture tested at 56 days met the proposed resistivity target of 11 kΩ•cm, with 11 of the 21 lots having 56-day test results very close to the target (>10.5 kΩ•cm). All lots of 460SLNS concrete tested at 90 days met the proposed target. Two of the four lots of hand-placed 496HPNS concrete tested at 56 days did not quite meet the proposed resistivity target of 11 kΩ•cm, although all four lots were > 10 kΩ•cm at 56 days, and all eight lots of 496HPNS concrete readily met the target at 90 days.
The 90-day resistivity of many lots significantly exceeded the proposed resistivity targets for NC bridge concrete identified in previous research (15 to 16 kΩ•cm, depending on chloride exposure), indicating the concrete comprising this demonstration project should exhibit good durability ( 6 , 30 ). It is noted that there is significant spread in the resistivity readings, particularly at 90 days. The cause for this dispersion of the 90-day readings was not readily evident, but could be attributed to changes in the composition of the cement, fly ash, or both, during the 2-year project, temperature changes at the time of batching during the two construction seasons, or other reasons. Additional study of the variability of later age resistivity readings may be of interest to improve the specification, QC procedures, or both.
The 3-day resistivity of mixture 460SLNS produced during Phase 1 and 2 paving plotted over time in control charts are shown in Figure 5. The 28-day resistivity measurements of mixture 460SLNS produced during Phase 1 and 2 paving plotted over time in control charts are shown in Figure 6. Although control limits were not used by the contractor in real time, they were computed to support data analysis. Computation of typical control chart action (± 2σ) and suspension (± 3σ) limits for each phase shows the improvement in control that was achieved over the course of the project. Although the 3-day and 28-day surface resistivity achieved during Phase 1 paving was fairly consistent, it was highly consistent during Phase 2 paving, as evidenced by the narrowing of typical upper and lower action/suspension limits toward the central line in Phase 2. The standard deviation for 3-day resistivity for Phase 1 paving (0.75 kΩ•cm) was reduced by almost half to 0.40 kΩ•cm for the Phase 2 paving. Similarly, for 28-day tests, the standard deviation was significantly reduced from Phase 1 (1.09 kΩ•cm) to Phase 2 (0.79 kΩ•cm). Although some reduction in variability could be attributed to the increasing proficiency of the technicians, variability in the materials, or other causes, it may also indicate a reduction in the variability (increase in consistency) of the concrete produced.
Control chart showing 3-day resistivity for mixture 460SLNS during Phase 1 and Phase 2 paving.
Control chart showing 28-day resistivity for mixture 460SLNS during Phase 1 and Phase 2 paving.
Measurements of 56-day resistivity were primarily made during Phase 2 paving. The proposed resistivity target for NC pavement concrete is shown on these control charts for illustrative purposes, and it is noted that specification language on the age at which the target should be met is still being evaluated in an ongoing research study ( 7 ). These measurements plotted over time in a control chart are shown in Figure 7. As mentioned previously, it can be observed in Figure 7 that most concrete (79.6%) had met the proposed resistivity target by 56 days. Measurements of 90-day resistivity were primarily made during Phase 1 paving, and are plotted over time in a control chart and shown in Figure 8. It can again be seen in Figure 8 that all concrete met the proposed resistivity target by 90 days.
Control chart showing 56-day resistivity for mixture 460SLNS during Phase 2.
Control chart showing 90-day resistivity for mixture 460SLNS during Phase 1.
When comparing with the earlier age resistivity measurements, a larger spread in the data at 56 days (σ = 1.76 kΩ•cm) and 90 days (σ = 4.13 kΩ•cm) of age is evident. Should enhanced control chart techniques be developed and used by the contractor, a central line would be shown on the chart, with upper and lower control limits computed in real-time to prompt consideration of process changes.
PEM Acceptance Test Results
To obtain the funds associated with Category B of FHWA’s Demonstration Project for Implementation of Performance Engineered Mixtures program, one or more AASHTO PP 84-17 tests was to be used by the agency in the acceptance process (agency testing to determine the degree of compliance with requirements). Shadow testing was an acceptable approach. In this demonstration project, surface resistivity acceptance tests were performed as shadow testing by NCDOT at their laboratory on some cylinders used for compressive strength testing during Phase 1, typically at 14 and 90 days of age. Curing was performed in a moist room, and therefore no correction factor was applied to the average results, per AASHTO T 358. Resistivity tests were not performed by NCDOT during Phase 2. Comparing NCDOT’s acceptance tests of the 460SLNS mainline paving mixture at 28 days (n = 105) to the contractor QC Phase 1 28-day resistivity data (n = 262) revealed strong agreement, with NCDOT’s acceptance tests showing an average 28-day resistivity of 8.27 kΩ•cm, slightly higher than the contractor QC average of 7.29 kΩ•cm. NCDOT’s acceptance tests had a standard deviation (1.07 kΩ•cm) almost identical to that of the contractor’s QC tests (1.09 kΩ•cm) over the course of Phase 1, showing similar variability between the two sample populations.
Conclusions
The PEM demonstration project was a success. Contractor and agency personnel gained valuable experience with three PEM tests and plan to continue using them in the future. The Box Test was found by the contractor to be highly useful in mixture development and for mixture modifications during the project. From the agency perspective, NCDOT agrees the Box Test provides beneficial information, and could potentially add it as a requirement for mixture design submittals.
The SAM test was successfully performed, but data was variable, and additional training of agency and contractor personnel is needed. NCDOT is performing additional shadow testing to become comfortable with the SAM. A training session with the SAM developers was held before use on an ongoing follow-up PEM implementation project, and additional data is being used to refine the performance target and potential specifications.
The contractor found surface resistivity testing straightforward to perform and readily integrated it into their QC practices. The primary mixtures used for the pavement met the suggested resistivity target of 11 kΩ•cm by 90 days (often by 56 days), indicating that the pavement constructed should provide satisfactory durability performance.
From the contractor’s perspective, the training by the research team gave them unique exposure to new testing equipment and methods. Contractor personnel gained an improved understanding of the impact of concrete quality on pavement durability and longevity. The contractor easily integrated the SAM and resistivity meter into the standard testing procedures, found the tests easy to perform, and did not need to provide additional (scarce) QC staff to support the extra testing. The project schedule did not allow the contractor to extensively use PEM tests during the preliminary mixture design phase. However, the contractor intends to implement PEM tests on future projects.
NCDOT found implementation of resistivity testing for acceptance straightforward and noted that the agency can equip laboratories with the instrument for a low cost. One concern is the use of 56- or 90-day tests to capture durability performance of fly ash mixtures at later ages, since the additional samples may pose a storage issue. There would also need to be a plan to address concerns about low surface resistivity test results at 56 days, 90 days, or both. Ongoing research is being performed to identify a 28-day resistivity target that generally correlates to a 56-day or 90-day resistivity indicating satisfactory performance, which would likely address agency concerns.
Footnotes
Author Contributions
The authors confirm contribution to the paper as follows: study conception and design: T. Cavalline, B. Hunter, B. Tempest, F. White, C. Ange; data collection: F. White, P. Simpson; analysis and interpretation of results: T. Cavalline, F. White, B. Tempest, B. Hunter; draft manuscript preparation: T. Cavalline, B. Tempest, B. Hunter, F. White, C. Ange. All authors reviewed the results and approved the final version of the manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Research presented in this paper is part of the research project RP 2019-41 sponsored by NCDOT.
The contents of this paper reflect the views of the authors and not necessarily the views of NCDOT. The authors are responsible for the accuracy of the data presented here. Additionally, this paper does not constitute a standard, specification, or regulation and does not necessarily reflect the official policies of NCDOT.