Evaluation of Bonded Concrete Overlays on Asphalt Pavements (2022)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

129   BCOAs have been successfully used to rehabilitate rutted and shoved but otherwise struc- turally sound asphalt pavements. BCOAs are relatively easy to construct compared with other overlay methods, require only simple preoverlay preparation of the existing asphalt pavement, expedite construction through placement of thinner concrete layers directly on existing pave- ment, and provide a cost-competitive rehabilitation alternative. Several BCOA design methods have been developed, allowing the designer to determine layer thickness and slab size on the basis of site-specific traffic and climate conditions, as well as on concrete material properties. This research initiative summarized BCOA practices through a literature review and SHA survey, conducted automated distress and GPR surveys on 20 in-service BCOA projects, con- ducted detailed field surveys on 19 of the 20 in-service BCOA projects, and compared predicted results from the BCOA-ME, Colorado DOT, Illinois DOT, and PaveME design methods. Conclusions Past studies determined several key criteria for ensuring BCOA performance. The condition of the asphalt layer to remain in place is to be a minimum of 3 in. and structurally sound (e.g., potholes and severe fatigue and transverse-cracked areas repaired). Early projects focused on rapid construction techniques and early opening to traffic, resulting in high cement contents and earlier than expected failures. More recent projects have incorporated conventional con- crete mixtures with or without synthetic macrofibers. The incorporation of synthetic macro- fibers slightly increases the flexural capacity of the concrete and minimizes slab migration. Durable concrete mixture properties include the use of well-graded aggregate (maximum aggregate size of 0.75 to 1.0 in.), Type I or Type II cement (w/cm ratio of 0.40 to 0.42), and synthetic macrofibers (3 to 4 lb/yd3). BCOAs can be placed and finished using conventional con- crete paving equipment and specifications. Current slab size recommendations are greater than 4- × 4-ft, and the longitudinal joints are not to align with the wheelpaths. Finally, joint sealing is expected to improve performance, but only if joint seals are maintained. Predominant BCOA maintenance includes joint resealing, crack sealing, and removing and replacing distressed slabs. A survey of SHAs identified the extent of BCOA use, project selection criteria, methods for evaluation of existing conditions and BCOA design methods, and construction and maintenance practices. Of agencies that responded to the survey (27 SHAs), nearly half have constructed one to 15 projects, while two agencies have constructed more than 15 BCOA projects. The majority of projects have been on rural state and U.S. routes with low to moderate traffic volumes. SHAs based project evaluation primarily on coring, pavement distress surveys, and deflection testing and backcalculation analyses. The most common combinations of BCOA thicknesses C H A P T E R 6 Conclusions and Further Research

130 Evaluation of Bonded Concrete Overlays on Asphalt Pavements and slab sizes were 6-in. thick 6- × 6-ft slabs, 4-in. thick 4- × 4-ft slabs, and 5-in. thick 6- × 6-ft slabs. Approximately 40% of the responding agencies included synthetic macrofibers in the concrete mixture. Agencies indicated that the predominant preoverlay and surface preparation activities are patching, asphalt pavement milling, and sweeping. Noted maintenance includes crack sealing, spall repair, diamond grinding, and full slab replacements. From the SHA survey, 19 projects were selected for site evaluations. The 19 projects range in age from 7 to 26 years and have been subjected to a broad range of truck loadings (rural low volume roads to interstate). BCOA design layer thickness ranged from 4 to 6 in., with slab sizes ranging from 4- × 4-ft to 12- × 12-ft. Automated pavement condition surveys were conducted to determine IRI, faulting, and cracking on all 20 projects (approximately 175 mi). In general, most of the segments evaluated were in good condition. Approximately 90% of the tested segments had IRI less than 170 in./mi, with nearly 50% of segments having IRI less than 95  in./mi. Most segments had maximum faulting of 0.10 in. or less (average 0.04 in.), and less than 5% of slabs showed corner breaks, longitudinal cracks, and transverse cracks. An analysis of automated pavement condition data indicated a statistically significant differ- ence in intersection performance compared with nonintersection locations. Intersections were found to be more faulted and rougher (17% and 36% higher, respectively) than noninter section locations. However, total cracking at intersection locations was not found to be significantly dif- ferent than at nonintersection locations (3% versus 1%, respectively). The projects with 6- × 6-ft slab sizes are performing better in relation to IRI, faulting, and cracking than the projects with 4- × 4-ft and 12- × 12-ft slab sizes. All but two projects with 4- × 4-ft slab sizes have IRI greater than 170 in./mi. Projects with 12- × 12-ft slab sizes have cracking in more than 5% of slabs. GPR surveys were conducted on all 20 projects, covering the same locations as the auto- mated distress surveys. BCOA and asphalt layer thicknesses were determined on each 0.10-mi good, fair, and poor segment selected for the detailed site evaluations. GPR-determined BCOA layer thicknesses agreed well with extracted core thicknesses (with an average difference of only 2.5%). GPR-determined asphalt layer thicknesses were significantly higher than extracted core thicknesses, with a nearly 40% difference. A statistical analysis of GPR and core results indi- cated no statistical significance for 84% of BCOA layer measurements or 57% of asphalt layer measurements. Detailed site evaluations were conducted on the 0.10-mi good, fair, and poor segments from 19 projects. Site evaluations included distress surveys in accordance with the Distress Identifica- tion Manual for the Long-Term Pavement Performance Program (Miller and Bellinger 2014), fault- meter testing, ultrasonic tomography testing, FWD testing, coring, DCP testing, and sampling of unbound layers. Results of the detailed distress surveys identified three projects having significantly more distressed slabs than all other projects: IL SR-53 (4- × 4-ft slabs, underdesigned for the antici- pated traffic), and CO US-6 and MN TH-30 (12- × 12-ft slabs that have been in service the longest of all projects evaluated). Removing these three projects, the primary distress types for the 4- × 4-ft slabs were longitudinal cracking, transverse cracking, and corner breaks (4% of slabs for each distress type). Distress types for 6- × 6-ft slabs included longitudinal cracking and transverse cracking (less than 4% of slabs for both distress types). The 12- × 12-ft slabs were predominantly distressed with longitudinal cracking (23% of slabs). Conducting automated and visual distress surveys on the same roadway segments provided an opportunity to evaluate how well the two methods compare in condition assessment (i.e., fault- ing, longitudinal and transverse cracking, corner breaks). On the basis of the analysis, a statistical

Conclusions and Further Research 131   difference exists between automated distress and visual distress collection methods for the proj- ects evaluated. Faulting measurements were 93% different, corner breaks were 33% different, transverse cracking was 20% different, and total cracking was 26% different. No statistical dif- ferences were found between the two methods in relation to longitudinal cracking. Although the two methods were found to be statistically different, a linear correlation does exist. Using the results of the visual distress surveys, several statistical analyses were conducted to assess whether various factors (i.e., in-service age, layer thickness, ESALs, synthetic macrofibers, and joint sealing) affect BCOA performance. Several conclusions were made on the basis of a multifactor statistical analysis of service age, layer thickness, ESALs, synthetic macrofibers, and joint seal: • Faulting is influenced by slab size, BCOA layer thickness, and the interactions between slab size and BCOA layer thickness. Increasing the asphalt layer thickness reduces the potential for faulting. Synthetic macrofibers and joint sealing showed no effect on faulting. • Corner breaks are influenced by slab size, in-service age, synthetic macrofibers, ESALs, and the interactions of slab size–BCOA thickness and asphalt layer thickness–BCOA thickness. However, the number of corner breaks on in-service pavements was minimal. • Longitudinal, transverse, and total cracking and transverse joint spalling are not influenced by any of the factors evaluated. • Longitudinal joint spalling is influenced by in-service age, synthetic macrofibers, and the interaction of BCOA–asphalt layer thickness. As in-service age increases, increased spalling is observed, but the use of synthetic macrofibers reduces the amount of spalling. • Transverse joint spalling is not influenced by any of the factors evaluated. • Total spalling is influenced by in-service age. The statistical analysis was further modified by normalizing all projects to 20-year ESALs. The results of the 20-year ESAL normalization showed • Faulting, cracking, and joint spalling are influenced by slab size and layer thickness; and • Joint spalling is influenced by slab size. Finally, a statistical analysis was conducted specifically to analyze the effects of joint sealing and synthetic macrofibers. From this analysis, joint sealing and synthetic macrofibers had no effect on performance. Faultmeter results ranged from –0.07 to 0.30 in. and averaged 0.04 in. The 6- × 6-ft slabs had the least amount of faulting, with all measurements less than 0.20 in. and average fault depth of 0.03 in. The faultmeter results for the 4- × 4-ft slabs were all less than 0.20 in., with an average fault depth of 0.04 in. The 12- × 12-ft slabs had the highest measured faultmeter results (up to 0.30 in.) with an average measured fault depth of 0.06 in. Ultrasonic tomography tests were conducted to evaluate layer thickness and assess the degree of bond between the concrete and asphalt layers. In total, 15 tests were conducted on each 0.10-mi good, fair, and poor segment, summarized, and compared with extracted core layer thicknesses. The comparison analysis indicated that ultrasonic tomography can effectively estimate BCOA layer thickness. The standard ultrasonic tomography algorithm cannot reliably calculate the asphalt layer thickness. FWD tests were conducted to evaluate LTE and to estimate effective thickness as a proxy for assessing bond condition. Most segments (74%) had LTE values between 80% and 90% and only 9% of segments had LTE values less than 80%. Results of FWD testing were used to estimate effective thickness. If the effective thickness exceeds the BCOA thickness, it suggests at least a partial bond is present. The effective thicknesses were greater than the measured BCOA thick- nesses for 78% of the core locations, indicating at least a partial bond for most core locations.

132 Evaluation of Bonded Concrete Overlays on Asphalt Pavements However, estimated bond condition and actual core condition were confirmed on only 41% of the cores, as the coring operation may have caused debonding of some cores. Cores extracted from each good and poor segment were used to determine layer thickness and bond condition. DCP testing and sampling of unbound materials were conducted. All extracted cores and unbound materials were shipped to a laboratory for further tests that included AASHTO soil classification, the Atterberg limits, particle size distribution, concrete compressive and split tensile strength, concrete CTE, concrete–asphalt bond (conducted only on intact cores), asphalt dynamic modulus, and Hamburg wheel tracking. The majority of aggregate base and subgrade materials were classified as granular materials (AASHTO soil classification A-1 to A-3), with nearly half of the unbound materials classified as plastic. For unbound materials classified as plastic, PIs ranged from 1% to 17% (average of 7%). Tests for concrete compressive strength and split tensile strength were modified to incorporate the thinner concrete samples. Compressive strength for all cores ranged from 3,133 to 14,504 psi (average of 6,310 psi), and split tensile strength ranged from 276 to 812 psi (average of 447 psi). Tests for concrete CTE were also modified to accommodate the thinner concrete sample, with results ranging from 4.6 to 6.3 × 10-6 in./in./°F (averaging 5.6 × 10-6 in./in./°F). Of the nearly 150 extracted cores, only 20 (from nine projects) were intact for shear-strength testing. Bond shear strength ranged from 38 to 178 psi and averaged 111 psi (standard deviation of 42 psi). Results of the laboratory testing were included in the evaluation of the BCOA design methods. In addition, the statistical analysis indicated that liquid limit and PI are significant factors for corner breaks and longitudinal cracking, and compressive and split tensile strength are signifi- cant factors for longitudinal and transverse cracking. The BCOA design methods evaluated were BCOA-ME, Colorado DOT, Illinois DOT, and PaveME. • BCOA-ME incorporates ESALs, asphalt layer thickness, k-value, dynamic asphalt modulus as a function of site-specific temperatures, preoverlay asphalt pavement distress informa- tion, concrete MOR and modulus, and FRC in the determination of BCOA layer thickness to meet the criteria of percentage of cracked slabs. • The Colorado DOT method predicts the transverse cracking potential for various BCOA thicknesses on the basis of ESALs, concrete MOR and E, concrete slab temperature gradient, asphalt layer thickness, k-value, asphalt and concrete fatigue, and slab size combinations; however, it does not include the ability to evaluate FRC. • The Illinois DOT method incorporates ESALs, concrete slab temperature gradient, concrete MOR and E, asphalt layer thickness, k-value, and FRC in the BCOA thickness determination. • PaveME uses axle load spectra, site-specific climate, existing pavement layer thickness and properties, the BCOA-ME longitudinal cracking model, and the existing bonded concrete over asphalt transverse cracking model to evaluate user-specified BCOA layer thickness and slab dimensions. All design methods determine BCOA layer thickness, except for PaveME, which predicts longitudinal cracking. In addition, the slab sizes and BCOA layer thicknesses evaluated vary by design method. BCOA-ME evaluates joint spacing ≥ 2- × 2-ft and ≤ 15- × 12-ft with BCOA thicknesses ranging from 2 to 8 in. However, BCOA-ME recommends slab sizes of 4- × 4-ft and 6- × 6-ft with a BCOA layer thickness of 3.0 to 5.5 in. and 12- × 12-ft with a 5.5- to 6.5-in. thick BCOA layer. Colorado DOT joint spacing ranges from 6- × 6-ft (recommended) to 12- × 12-ft with a min imum BCOA thickness of 3 in. and recommended BCOA layer thicknesses of 5- to 8-in. Joint spacing for Illinois DOT includes 4- × 4-ft or 6- × 6-ft with a BCOA layer thicknesses from 3- to 6-in. PaveME

Conclusions and Further Research 133   evaluates joint spacing ranging from 5- × 5-ft to 8- × 8-ft and BCOA layer thicknesses from 4- to 8-in. Results of the sensitivity analysis indicated that design methods are sensitive to changes in asphalt layer thickness (except BCOA-ME), truck loading, subgrade condition (except Illinois DOT method), and concrete strength. Each of the four BCOA design methods was conducted using data obtained from the detailed site investigations for each good and poor segment. Recommended BCOA layer thicknesses were determined from BCOA-ME, Colorado DOT, and Illinois DOT methods and predicted longitudinal cracking from PaveME. In general, BCOA-ME recommendations for 4- × 4-ft slabs are close to as-constructed BCOA layer thicknesses, and recommended thicknesses for all other slab sizes tended to be less than as-constructed BCOA layer thicknesses. The Colorado DOT and Illinois DOT design methods, in general, recommended thicker BCOA layers than as constructed. For PaveME, a stronger correlation existed between predicted and observed longitudinal cracking. On the basis of the results of this project, BCOAs are a viable option for the rehabilitation of applicable asphalt pavements. Many of the projects evaluated as part of this study are in fair or better condition after more than 15 years of service. Further Research Research in support of future use and application of BCOAs might include the following. • PaveME could be adjusted to bring together a unifying theoretical model with the significant inputs and appropriate distress failure mechanisms seen in the field (e.g., corner breaks, trans- verse cracks, faulting, and IRI). • More extensive and controlled research is needed to assess the influence of joint sealing on BCOA performance. • More extensive and controlled research is needed to assess the influence of synthetic macro fibers on a potential reduction in BCOA layer thickness and its impact on long-term performance. • More extensive and controlled research is needed on the initiation and progression of fault- ing in BCOA pavements. • Improved characterization of the bond interface between the concrete overlay and the asphalt needs to be developed and included in the design models. • The ultrasonic tomography algorithm could be adjusted to more accurately predict the thick- ness of the underlying asphalt layer. • Projects with higher than typical cracking could be assessed through forensic investigation. • Asphalt joint sealants could be enhanced to address concerns with long-term elasticity from panel shrinkage (i.e., joint openings can be substantial). • Comprehensive pavement management data might be collected to support improvement of performance prediction curves and subsequently improve design procedures.