University of California Pavement Research Center

Moisture-related shrinkage is regarded as one of the phenomena that has the largest impacts on the performance of jointed plainconcrete pavements. Still, most mechanistic-empirical design methods oversimplify or ignore predictions of moisture-relatedshrinkage and its effects on concrete pavements. This study evaluates how moisture-related shrinkage accumulates in concretepavements and the structural response of the concrete pavement slabs to the shrinkage action. The experimental data comefrom six thin concrete overlay of asphalt pavements that were instrumented with sensors to measure the structural andhygrothermal response of the slabs due to temperature and moisture-related actions. After an analysis of the predictions madeby current shrinkage models, a new shrinkage model was developed. This new model, which is based on the incremental-recursive application of the B4 shrinkage model, provided an excellent prediction of the moisture-related shrinkage measured inthe field. In addition, the structural response of the concrete pavement slabs under the moisture-related shrinkage action wasanalyzed using the finite element method (FEM). The FEM analysis based on the standard practice for concrete pavementmechanistic-empirical modeling resulted in unrealistically high tensile stresses. However, much smaller stress values were foundwhen the time-dependent (viscoelastic) behavior of concrete and asphalt was considered.

Currently, no performance-related test exists for fatigue cracking for use in routine asphalt mix design to approve job mix formula (JMF) and quality control and quality assurance (QC/QA) in California. The existing four-point bending (4PB) test was developed to evaluate the fatigue performance of asphalt pavement, but it is not necessarily appropriate for use in routine JMF and it is too slow for QC/QA. The objective of this study is to evaluate potential surrogate fatigue performance-related testing methods and identify a test that is simple and easy to perform but also able to provide guidance for asphalt mix design on routine projects and QC/QA on all projects. This report compares potential performance-related tests with 4PB tests for simplicity, repeatability (variability), and their relationship to stiffness and fatigue life. Tests evaluated in this study included the semicircular bend (SCB) test, indirect tensile asphalt cracking test (IDEAL-CT), and fatigue testing on fine aggregate matrix (FAM) mixes with linear amplitude sweep (LAS) analysis. These tests were conducted on a variety of asphalt mixtures. Fracture parameters obtained from SCB and IDEAL-CT tests and fatigue parameters from FAM mixes were included as potential fatigue cracking indicators. Linear regression analysis was used to correlate these indicators with the stiffness and fatigue life from 4PB tests. The comparison analysis demonstrates that SCB and IDEAL-CT tests are providing the same fracture information. Fracture parameters from SCB and IDEAL-CT tests are well correlated with the initial flexural stiffness from 4PB tests, and the initial flexural stiffness has a moderate inverse nonlinear correlation with the fatigue life from the controlled-strain 4PB tests. As the IDEAL-CT test is faster and requires less specimen preparation, the recommendation is that attention focus on this test. All the fracture tests indicate that the Strength parameter has low variability and good correlation with 4PB flexural stiffness and a moderate correlation with flexural fatigue, and it is proposed as a surrogate indicator for flexural stiffness and an indication of fatigue life. The FAM test showed promise regarding matching 4PB fatigue life as well as stiffness. The relationships identified in this study between flexural stiffness and flexural fatigue life and between flexural stiffness and the Strength parameter from the IDEAL-CT tests were used to develop a preliminary approach to using the Strength parameter to place upper and lower boundary limits on the stiffness of mixes for use in routine mix design and QC/QA. Examples of this approach are presented in the appendices based on flexural fatigue testing and on setting of those limits without 4PB tests. The sensitivity of performance for both approaches is demonstrated by mechanistic-empirical simulation using CalME. Recommendations are made to further develop the IDEAL-CT Strength parameter for routine mix design and QC/QA, with limits set following the approach developed in this study. Review of pavement management system field cracking data and indirect tensile strength from the AASHTO T 283 test in Caltrans databases is recommended to help this development. Further development of FAM mixes LAS testing is also recommended.

Use of recycled materials, such as reclaimed asphalt pavement (RAP) and reclaimed asphalt shingles (RAS), is gaining widespread interest. Currently the California Department of Transportation standard specifications do not allow the use of RAS and limit the use of RAP contents to a maximum of 25% by dry weight of aggregate, though there is a non-standard special provision that allows use of up to 3% RAS by mass of aggregate and 40% RAP content by binder replacement. Nevertheless, mixes with high RAP and RAS are being produced across California for local agencies and commercial use. This study investigated the performance of four plant-produced high RAP or RAS mixes collected from different regions in California. The mixes were not designed and produced following Caltrans specifications. However, they provide insight into the effects of silo storage time on blending of virgin and RAP binder, the performancerelated properties of these mixes, and the measurement of properties by accepted performance-related tests. The mixes were collected before silo storage and after hours in the silo. Testing of the mixes included the following tests: four-point flexural beam stiffness and fatigue, Hamburg Wheel-Track (HWT), confined and unconfined repeated load triaxial (RLT), semicircular bending (SCB), and indirect tensile asphalt cracking test (IDEAL-CT). Additional testing was also conducted on the fine aggregate matrix (FAM) mixes to characterize fatigue and stiffness. This report presents preliminary findings from this study, including results related to the effects of silo storage time on stiffness, cracking resistance, and rutting resistance using performance-related tests as well as an initial comparison of the results from alternative cracking test types for these mixes. The results showed that silo storage time can increase stiffness on the order of 50% to 60%, with corresponding negative effects on fracture resistance and controlled-strain flexural fatigue life. The fatigue performance of the mixes reduced with increased silo storage, particularly at high strain levels, as measured by the flexural beam test. Use of a high rejuvenator dose could also potentially lead to rutting problems and poor fatigue performance. The FAM mix testing showed promising results in terms of characterizing fatigue. However, in its current form, it is not yet practical for use as a quality control/quality assurance test. The recommendation is that the effect of aging and blending of high RAP or RAS mixes be further investigated to understand the full impact of silo storage on these types of mixes. Performance-related specifications should consider the variation in mix properties due to silo storage.

This research report summarizes a literature review update, construction of a test track to assess various aspects of gap-graded rubberized asphalt concrete (RHMA-G) mixes with and without the addition of reclaimed asphalt pavement (RAP) as aggregate replacement, a first-level analysis of the results from five Heavy Vehicle Simulator (HVS) tests, and a first-level analysis of laboratory test results on the four mixes.

Four different RHMA-G mixes were placed on seven cells on the test track at the UCPRC. Mixes differed by nominal maximum aggregate size (NMAS, 1/2 and 3/4 in.) and the addition of 10% RAP by weight of the aggregate as an aggregate replacement. Single and double lifts of each mix were placed. Apart from the addition of RAP, the mix designs all met current Caltrans specifications. Although Caltrans currently does not permit more than one lift of RHMA-G on projects, the placement of each lift of each mix on the test track met current Caltrans specifications for RHMA-G layers.

The five HVS tests discussed in this report covered the control section (0.2 ft. [60 mm], 1/2 in. NMAS with no RAP), a section with two lifts (0.4 ft. [120 mm]) of the 1/2 in. mix with no RAP, a section with one lift of 1/2 in. mix with RAP, a section with one lift of 3/4 in. mix with no RAP, and a section with two lifts (0.5 ft. [150 mm]) of 3/4 in. mix with RAP. The untested sections included a section with two lifts of 1/2 in. mix with RAP, and a section with two lifts of 3/4 in. mix with no RAP. Findings include the following:

• Performance of all four mixes was satisfactory in terms of the level of trafficking required to reach a terminal average maximum rut of 0.5 in. (12.5 mm).• Differences in nominal maximum aggregate size and/or the addition of RAP as a coarse aggregate replacement did not appear to have any significant influence on the HVS and laboratory test results. The time that the mix was stored in the silo and the interval between construction and start of HVS testing (i.e., degree of aging of the RHMA-G) appeared to have a larger influence on results.• The backcalculated stiffnesses of the RHMA-G layer on each section before and after HVS testing indicate that trafficking generally caused some damage on the sections, as expected. An exception to this observation was noted on the first test, which was attributed in part to stiffening of the mix through diffusion of small amounts of RAP binder, which possibly countered the effect of damage by trafficking, and potentially in part to the method used for backcalculation.• A hydraulic oil spill on one of the sections had a notable negative effect on rutting performance.• No cracks or other distresses were observed on any of the sections after trafficking.

Recommendations, if justified, for changes to limits for nominal maximum aggregate size in relation to RHMA-G lift thickness, RHMA-G lift thickness, whether more than one RHMA-G lift can be considered in pavement designs, and the use of RAP as aggregate replacement in RHMA-G mixes will be made in a separate report that documents, second-level analysis, and mechanistic simulations.

The work discussed in this interim report is part of a larger study, funded by the California Department of Transportation. The study objective focuses on developing and recommending testing procedures and criteria for performance-based specifications of asphalt rubber binders used in gap-graded and open-graded mixes using current Superpave performance grade (PG) equipment. Work in this phase covered the testing of 19 plant-produced binders and the base binders used to produce them. Plant-produced gap-graded rubberized hot mix asphalt mixes from five of the projects were also tested. The following important observations from the binder rheology tests were made:

--Although the low-temperature performance grades appeared to be reasonable, the high-temperature grades appeared to be unrealistically high, while the intermediate-temperature grades appeared to be potentially lower than anticipated, when comparedto the base binders. Fourteen of the binders tested with concentric cylinder geometry and 13 tested with parallel plate geometry had PGs higher than the maximum grade of 82°C listed in the AASHTO M 320 standard. Grades higher than 82 are considered to be unrealistically high and probably not a true indication of likely high-temperature performance (i.e., rut resistance under heavy loads on hot days).

--A comparison of the concentric cylinder and parallel plate (3 mm gap) geometries indicated that the results between the two geometries are different and are likely to be higher than the precision and bias of the individual procedures. Precision and bias statements for these procedures had not been developed at the time of preparing this report. These results indicate that the two geometries cannot be used interchangeably.

--No consistent trends in results were observed between any of the parameters tested.

--Observations from previous testing and during this phase of the study indicated that incompletely digested rubber particles—which have different sensitivities to temperature, aging, and applied stress and strain than the base asphalt binder—appeared to have a dominant influence on results and caused variability between results, regardless of the testing geometry used. Considering these incompletely digested particles as part of a homogenous binder may therefore not be appropriate when determining performance grades. Work is continuing in Phase 3 of this study to adjust testing procedures to account for the influence that these incompletely digested particles have on results.

The proposed modifications to short- and long-term aging procedures (i.e., rolling thin-film oven and pressure aging vessel) and to the bending beam rheometer specimen preparation procedures developed in Phase 2 are considered to be more aligned with the original intent of the tests and will likely reduce the variability between replicate specimens during testing.

Preliminary test results indicate that Fourier transformed infrared spectroscopy is a potentially valid method for quantifying rubber content in rubber-modified binders.

The California Department of Transportation (Caltrans) has been using full-depth recycling (FDR) as a pavement rehabilitation strategy since 2001. Early projects were recycled with foamed asphalt and cement. Cement-only treatments were permitted from 2015 to improve the properties of more marginal materials. However, shrinkage cracking associated with the hydration and curing of the cement-treated layers remains a concern, especially with regard to crack reflection through asphalt concrete surfacings and the related problems caused by water ingress.Crack mitigation has been studied for decades, and a range of measures related to improved mix designs and construction practices have been implemented by road agencies. One of the most promising measures, used in conjunction with appropriate mix designs, is that of microcracking the cement-treated layer between 48 and 72 hours after construction. In theory, this action creates a fine network of cracks in the layer that limit or prevent the wider and more severe block cracks typical of cement-treated layers. Limited research to assess microcracking as a crack mitigation measure has been completed on a number of projects in Texas, Utah, and New Hampshire. Recommendations from these studies were first implemented by the Texas Department of Transportation and then later by other state departments of transportation. However, longer-term monitoring on a range of projects in Texas, California, and other states had revealed that microcracking has not always been successful in preventing cracking. Some projects showed reflected transverse and block cracks in a relatively short time period, attributable to a number of factors including but not limited to cement content, cement spreading, the method of curing, and the interval between base construction and placement of surfacing.Discussions with researchers in Texas indicated that additional research was necessary to better understand the microcracking mechanism, and to identify the key factors influencing performance, including but not limited to aggregate properties, cement content, the time period before microcracking starts, layer moisture contents, roller weights and vibration settings, the number of roller passes, the field test methods and criteria used to assess the degree of microcracking, and the effects of early opening to traffic. A multiphase project was therefore initiated at the University of California Pavement Research Center (UCPRC) to investigate these outstanding issues. The second phase of this study is discussed in this report. This phase covered the design, construction, monitoring, and associated laboratory testing of a 37-cell test road to evaluate shrinkage crack mitigation procedures. The study found that microcracking is an effective mitigation measure, provided that design strengths do not exceed 600 psi (4.1 MPa) and that microcracking is done between 48 and 56 hours after compacting the layer.

The work discussed in this interim report is part of a larger study, funded by the California Department of Transportation, with the objective of developing and recommending testing procedures and criteria for performance-based specifications of asphalt rubber binders used in gap-graded and open-graded mixes using current Superpave performance grade (PG) equipment. Work covered the testing of five plant-produced binders, the base binders used to produce them, and the gap-graded rubberized hot mix asphalt mixes produced with them. The following important observations from the binder rheology tests were made:

Although the low-temperature performance grades appeared to be reasonable, the high-temperature grades appeared to be unrealistically high, while the intermediate-temperature grades appeared to be potentially lower than anticipated when compared to the base binders. A comparison of the concentric cylinder and parallel plate (3 mm gap) geometries indicated that the results between the two geometries are different and are likely to be higher than the precision and bias of the individual procedures. Precision and bias statements for these procedures had not been developed at the time of preparing this report. Consistent trends in results were observed between high-temperature PG/true grade, Delta TC, and non-recoverable creep compliance at 3.2 kPa. Observations from previous testing and during this phase of the study indicated that incompletely digested rubber particles—which have different sensitivities to temperature, aging, and applied stress and strain than the base asphalt binder—appeared to have a dominant influence on results and caused variability between results, regardless of the testing geometry used. Considering these incompletely digested particles as part of a homogenous binder may therefore not be appropriate when determining performance grades. Work is continuing in Phase 3 of this study to adjust testing procedures to account for the influence that these incompletely digested particles have on results.

The proposed modifications to short- and long-term aging procedures (i.e., rolling thin film oven and pressure aging vessel) and to the bending beam rheometerspecimen preparation procedures developed in Phase 2 are considered to be more aligned with the original intent of the tests and will likely reduce the variability between replicate specimens during testing.

Preliminary test results indicate that Fourier transformed infrared spectroscopy is a potentially valid method for quantifying rubber content in rubber-modified binders.

This report summarizes the work completed to develop the concrete overlay on asphalt (COA) tables of the new Caltrans Highway Design Manual (HDM) Rigid Pavement Design Catalog. The tables consider the different pavement structures that are candidates for rehabilitation with COA with short transverse joint spacing on the Caltrans road network. The tables were developed using Pavement ME (version 2.5.5) with the nationally calibrated COA cracking model. Pavement ME inputs were determined by considering the state’s climate, traffic, materials, and construction practices. The design tables reflect the recommendations from previous Caltrans research about COA, including slab size, shoulder type, and load transfer efficiency. The Pavement ME inputs for developing the tables include a design life of 20 years, 10% target cracking, and 95% design reliability. The tables will be included in the printed version of the new HDM Rigid Pavement Design Catalog.

This report summarizes the work conducted to develop the jointed plain concrete pavement (JPCP) tables of the new Caltrans HighwayDesign Manual (HDM) Rigid Pavement Design Catalog. The tables consider the different pavement structures that are expected toperform properly on the Caltrans road network. The tables were developed using Pavement ME (version 2.5.5) with the nationallycalibrated transverse cracking model. Pavement ME inputs were determined by considering the state’s climate, traffic, materials, andconstruction practices. A design life of 40 years, 10% target transverse cracking, and 95% design reliability were chosen for developmentof the tables. Transverse joint faulting and the International Roughness Index (IRI) were also determined for the sections in the JPCPtables using Pavement ME (version 2.5.5) nationally calibrated models and compared to Caltrans faulting and IRI limits of 0.15 in. and170 in./mi., respectively. The tables will be included in the printed version of the new HDM Rigid Pavement Design Catalog.

Weigh-in-motion (WIM) devices measure and record highway vehicle axle loads. The data they collect include axle loads and spacing, vehicle classification and gross weights, and travel speed. These WIM data are used for pavement design, management, and performance studies. In 2016, there were 123 WIM devices operating on sites throughout the California state-owned highway network, one of the densest and best maintained in the United States. The University of California Pavement Research Center (UCPRC) has studied California’s WIM data with the California Department of Transportation (Caltrans) since 2007 to advance mechanistic-empirical (ME) pavement design procedures. Previously, the UCPRC completed the first analysis of the WIM data collected from 1998 to 2003 to discover similarities in axle load distributions at the WIM sites and then grouped them using cluster analysis to generate default traffic inputs for pavement design software. For this research, the UCPRC processed WIM data collected from 80 California WIM-measuring sites from 2004 to 2015, identified the axle load distribution for each site, and updated the earlier grouping using hierarchical cluster analysis. A decision tree was developed to classify WIM data into five WIM axle load spectra, which have been implemented in the Caltrans pavement management system database, PaveM, to generate truck traffic inputs in the CalME and Mechanistic-Empirical Pavement Design Guide (MEPDG) pavement design software tools used by Caltrans for asphalt and concrete pavement design, respectively.