Proposed AASHTO Practice and Tests for Process Control and Product Acceptance of Asphalt-Treated Cold Recycled Pavements (2021)

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.

7   CHAPTER 2 Research Approach Chapter 2 provides an overview of the work completed agency specifications review was completed with the goal of during this study. The objectives of the study were addressed identifying current traffic opening and surfacing requirements through three phases. In Phase I, current and emerging test and other material quality tests. The survey of academic, methods for quality assessment and process control of cold industry, and agency stakeholders was conducted to identify recycled materials where emulsified asphalt or foamed asphalt any additional tests not discovered during the literature review serves as the stabilizing/recycling agent were identified. The (such as those that might come from unpublished studies or three primary sources of information used were a literature from ongoing research) and to help the research team identify review, a review of agency specifications, and an online stake- the curing time(s) from which the stakeholders would seek to holder survey. Phase II, a laboratory-based experiment, was use the results of any proposed tests. conducted with test slabs of recycled materials fabricated in the laboratory using materials sampled from recycling proj- 2.1.1  Specification Review ects in the United States and Canada. Phase III, a field-based experiment, was conducted where the most representative The standard specifications and special provisions for tests identified in Phase II were used to assess the early-age asphalt-based CIR, CCPR, or FDR from U.S. states and properties of recycled materials on nine field projects in Canadian provinces were collected and reviewed. Of the the United States. 50 U.S. states, 41 states, in addition to FHWA’s Federal Lands Highway Division, had at least one of the relevant specifica- 2.1 Phase I—Current and Emerging tions. Of the 10 Canadian provinces/territories, three included Quality Tests asphalt-based CIR, CCPR, or FDR in their standard speci- fications. In addition, the specifications from three munici- From the literature review, agency specifications review, palities in the United States were collected and reviewed. and stakeholder survey, the research team identified relevant In total, 83 specifications were reviewed. Of these, material properties that were key to addressing the study approximately 54% (45 specifications) governed CIR, 17% objectives and candidate tests that could assess the relevant (14 specifications) governed CCPR, and 29% (24 specifica- material properties of interest. Through these steps, the fol- tions) governed FDR. Figure 2.1 is a map of the United States lowing key properties were identified: product uniformity, and Canada that identifies the locations of all state/province/ moisture, compaction, thickness, curing, strength/stiffness, municipal CIR, CCPR, and FDR specifications that were and raveling resistance. Assessing one or more of these proper- reviewed. ties could then be used to determine if a recycled material was ready to be opened to traffic or ready to be surfaced. The work 2.1.2  Stakeholder Survey focused on identifying those tests that could assess these prop- erties, were relatively inexpensive, were easy to operate in the An online stakeholder survey was conducted from October field, and were able to capture the anticipated material prop- through December 2017. The objectives of this survey were erty trends associated with changes in stabilizing/recycling to identify: agent types, the presence of active fillers, and field curing. The review of agency specifications for asphalt-stabilized • Tests that were being assessed but had not yet been incor- FDR, CIR, and CCPR and the stakeholder survey were con- porated in current agency standard specifications or ducted to identify and summarize current practices. The special provisions, 

8 Alberta (AB) is the only municipality in the “Municipality Specification Only” category, and California is the only state in the “State/Province and Municipality Specification” category. Color figure can be viewed in the online version of this report. Figure 2.1.  Location of all state, province, and municipality CIR, CCPR, and FDR specifications reviewed in this study. • Procedures used by practitioners for process control that AASHTO Committee on Materials and Pavements, selected are not standardized or published, TRB committees in the pavements and asphalt materials • Potential hurdles that exist with implementation of current sections, and in presentations at regional and national pave- methods of product acceptance and specifications, ment recycling conferences. A total of 84 survey responses • Rankings of the most important test method characteris- were received. tics by practitioners (e.g., time taken to perform the test, equipment required), 2.2  Phase II—Laboratory Testing • Potential field projects that could be used in Phase III of the study for evaluating the testing procedures developed Using the information collected during the literature in Phase II, and review and stakeholder survey, the research team developed • Recommended improvements to existing tests from a laboratory experiment conducted in Phase II of this study. practitioners. The experiment was conducted on test slabs of recycled materials fabricated in the laboratory. The slabs were fabri- The stakeholder survey questions are presented in cated from loose recycled materials sampled during construc- Appendix A. The online survey link was distributed to the tion of projects from across the United States and Canada, 

9   as shown in Figure 2.2. For this work, loose materials were 2.2.1 Objectives defined as the processed RAP and any other material required to produce an FDR, CIR, or CCPR layer but not including The objectives of the laboratory testing included assessing any stabilizing/recycling agent or active filler. Industry and tests that could be conducted easily, quickly, and inexpen- agency partners assisted the research team with identifying sively in a field setting and could quantify material property relevant projects, obtaining the mix designs, and shipping differences resulting from changes in curing time, type of approximately 500 lb to 600 lb of loose materials from each stabilizing/recycling agent, and presence of active filler. project. Along with the loose materials, the research team Preferably, tests would provide an immediate result. received the designated asphalt emulsion or binder to repli­ cate the mix design in the laboratory. The research team 2.2.2  Experimental Design followed the developed mix design, mixed the provided materials, and compacted slab specimens in the laboratory. To accomplish the objectives of the laboratory experi- These slabs were tested with the tests identified in Phase I to ment, the research team chose to conduct the laboratory determine the tests’ ability to discern differences in material experiment using a partial factorial design. A partial factorial behavior related to curing time, type of stabilizing/recycling design was selected because of the high number of potential agent, and presence of active filler. combinations given the possible factors (including recycling With Active Filler No Active Filler CIR CIR CCPR CCPR FDR FDR Figure 2.2.  Project locations for Phase II materials sampling. 

10 processes; stabilizing/recycling agent types; presence of active agency provided the research team with 10 to 12 5-gal buckets fillers; RAP, asphalt binder, and aggregate sources; and curing of the loose material (composed of RAP and sometimes time), multiple levels of each factor, and number of tests. unbound materials from underlying layers) and 4 to 6 gal The experiment was designed using factors and levels that of emulsified asphalt for those projects where emulsified were expected to yield the greatest range of results for the asphalt was used. For those projects where foamed asphalt testing conducted. For example, for a test that measures was used, a performance grade (PG) 64-22 binder already stiffness of a recycled material, rather than each possible com- available in the laboratory was included. To help develop bination of recycling process, stabilizing/recycling agents, a more complete matrix of material types, 24 and 48 5-gal chemical additives/active fillers, geographic location, and RAP buckets of loose material were sampled from two ongoing type being tested, an experimental design could be developed research studies in California and Virginia, respectively. These that studied the combinations expected to produce the least extra materials were used to produce additional mixtures and the greatest stiffness. From this, the effectiveness of each using both emulsified and foamed asphalt (each with and test (or more tests) could be studied in a way that was more without an active filler) from the same source. efficient while still technically valid. Table 2.1 describes the projects from which the loose Multiple sets of test slabs were fabricated to conduct all the materials were obtained. The stabilizing/recycling agent tests. This was necessary given (1) the limited size of the test dosage ranged from 1.2% to 4.5%, with 2.5% being the specimens, (2) the desire to conduct testing on undisturbed most common agent content. In addition, the active filler sections of the test specimens, and (3) the need to provide content ranged from 0% to 1.5%, with 0% and 1% being the replication. Each test specimen was fabricated in accordance most common. with the mix design, and the quantity of each ingredient was recorded. Subsequent test specimens from the same project 2.2.4  Slab Specimen Fabrication were fabricated using the same ingredient quantities. Compacted slab specimens (500 mm × 400 mm × 110 mm) were manufactured from field-produced materials. The 2.2.3 Source Projects and Materials slabs were prepared with the loose materials sampled from Sampling each field project upon which the various laboratory tests With the help of industry and agency partners, the research were conducted. Following the mix design from (or devel- team was provided with a mix design (which included an oped for) each project, the loose materials were mixed with optimum density, a stabilizing/recycling agent content, stabilizing/recycling agents and active fillers (where used) and an active filler content [if used]) and materials from in a Wirtgen WLM30 laboratory-scale twin-shaft pug mill. 14 recycling projects located across the United States and This equipment has the capacity to mix a batch of approx- Canada. From each project, the supporting contractor or imately 30 kg (66 lb). For those mixtures using emulsified Table 2.1.  Phase II source projects summary. Stabilizing/ Agent Active Target Mix Recycling Active Content, Filler Density, ID Agent Filler Process State Project Description % Content, % lb/ft3 1 IN SR 101 2.5 1.0 128.0 CCPR 2 Cement VA I-64 Segment II 2.5 1.0 128.0 3 FDR TX I-10 4.5 1.1 135.0 4 CA UCPRC Test Track 2.5 1.0 128.0 Emulsified 5 asphalt NY Courtland 3.0 0.0 134.0 CCPR 6 VA I-64 Segment II 2.5 0.0 128.0 No 7 cement CIR ON Huron County, Road 87 1.2 0.0 121.5 8 IN Shelby County, SR 252 2.5 0.0 118.0 FDR 9 CA UCPRC Test Track 2.5 0.0 128.0 10 CCPR VA I-64 Segment II 2.5 1.0 128.0 11 CA Hayward, Soto Road 2.0 1.0 124.8 CIR 12 Cement MA Southwick 2.5 1.0 129.5 13 TX FM 1245, Groesbeck 2.4 1.5 124.8 Foamed FDR 14 asphalt CA UCPRC Test Track 2.5 1.0 128.0 15 CCPR VA I-64 Segment II 2.5 0.0 128.0 16 No MI Jackson County, Rosehill Road 2.2 0.0 130.0 cement CIR 17 WI Douglas County, STH 35 2.0 0.0 121.5 18 FDR CA UCPRC Test Track 2.5 0.0 128.0 

11   asphalt, the emulsion (kept at a temperature of 40°C) was 40°C until a constant mass was reached, and then the mois- added directly during the mixing process. For those mixtures ture content was calculated. using foamed asphalt, a Wirtgen WLB10S laboratory foam- Each test slab had dimensions of 500 mm in length × ing unit produced the foamed asphalt at a temperature of 400 mm in width and a thickness of approximately 110 mm. 163°C. Once the emulsified or foamed asphalt, mixing water, To fabricate each slab, two batches of mixed materials were and active filler (if used) were combined, the mixed materials required. The batches were produced as follows: were transferred to a slab compactor to fabricate the slab specimens. 1. Calculate the mass of parent material required for desired The following example calculations demonstrate how the slab size and mix design density. quantities for mixing a batch of material were determined. 2. Using two mostly full 5-gal buckets of loose material, Two examples are shown, one for foamed asphalt materials add one-half of each bucket to an 18-gal tub, and mix and the other for emulsified asphalt materials. For each by hand. Transfer the contents of one 5-gal bucket into example, 1.0% cement was included, the RAP had an the other and empty the mixed contents of the 18-gal “in-the-bucket” moisture content of 2.4%, and the optimum tub into the empty 5-gal bucket. Add the contents of the moisture content was 4.0%. For the foamed asphalt example, second 5-gal bucket to the 18-gal tub and mix by hand. a binder content of 2.0% was assumed. For the emulsified Empty the 18-gal tub into the second 5-gal bucket. asphalt example, a binder content of 2.5% (2⁄3 residual 3. Add and mix the contents of each 5-gal bucket in the pug binder, 1⁄3 water) was assumed. The actual mix design from mill for 1 minute. After mixing, empty the contents into each project was used when the test specimens were made. a 50-gal tub. Foamed asphalt materials: 4. Based on the desired slab density, calculate the amount of mixed material required. Place approximately equal 1. Weigh RAP (assume a batch weight of 25 kg) and account portions of mixed material from the 50-gal tub into two for existing moisture of 2.4% = 25,000 g + 2.4% = 25,600 g. 5-gal buckets. 2. Add 1.0% cement based on mass of dry RAP = 25,000 × 5. Add enough loose material to the 5-gal buckets to 1.0% = 250 g × 1.0% = 253 g. account for the measured moisture content (determined 3. Add water to reach an optimum moisture content of as described previously). 4.0% = 25,600 × (4.0% − 2.4%) + 253 × 4.0% = 420 g. 6. Add the contents of one 5-gal bucket into the pug mill. 4. Add 2.0% foamed asphalt based on mass of dry RAP = 7. Add water if needed (calculated as described previously) 25,000 × 2.0% = 500 g. and mix for 1 minute. 8. Add cement if needed (calculated as described previously) Emulsified asphalt materials: and mix for 1 minute. 9. If emulsion is used, add emulsion directly to the pug 1. Weigh RAP (assume a batch weight of 25 kg) and account mill and mix for 1 minute. If foamed asphalt is used, for existing moisture of 2.4% = 25,000 g + 2.4% = 25,600 g. spray the foam into the pug mill and mix for 1 minute. 2. Add 1.0% cement based on mass of dry RAP = 25,000 × 10. Transfer contents to an empty 50-gal tub. 1.0% = 250 g × 1.0% = 253 g. 11. Repeat steps 6 through 10 for the second 5-gal bucket. 3. Calculate 2.5% emulsion mass based on mass of dry RAP = 25,000 × 2.5% = 625 g. The slab specimens were prepared using an IPC Global/ 4. Calculate water to reach an optimum moisture content of Controls Group Advanced Asphalt Slab Roller Compactor. 4.0% and subtract the water proportion (assumed as 1⁄3) The slab compactor used a roller head segment (having a of emulsion = 25,600 × (4.0% − 2.4%) + 253 × 4.0% = radius of 535 mm) and applied the compaction load to 420 g − 625 × (1⁄3) = 166 g. the material by the specimen mold carriage moving back 5. Add 166 g of water. and forth under the roller head with the load applied in a 6. Add 625 g of emulsion. pendulum-like action. The slab compactor was used since it could operate in a displacement control function so that For foamed asphalt materials, the research team used the the desired thickness of the test specimen could be set, and same PG 64-22 binder so that the foaming temperature and thus the approximate bulk density was controlled by adjust- amount of foaming water would be the same. For emulsi- ing the mass of material added. fied asphalt materials, the research team used the emulsion After mixing in the pug mill, the mixed materials were supplied by the agency/contractor. The day prior to mixing, transferred to the slab mold. The mixed material was added approximately 1,000 g of loose material was taken from a by hand, filling the corners first and then the lower edges to sealed 5-gal bucket and placed in a forced-draft oven set at reduce the chances of having lower density in these areas. 

12 The mixed material was rodded using a concrete molding rod until the material surface was below the maximum that could be accommodated by the slab compactor. For certain mixtures having higher densities, the height of the loose material exceeded the maximum that could be accommo- dated by the slab compactor (approximately 155 mm prior to compaction), and thus all slabs (regardless of initial height) were rodded. Following rodding, a sheet of heavy- duty aluminum foil was used to cover the rodded material to prevent it from sticking to the compactor roller head segment. No other lubricants or bond breakers were used. The slab compactor was set to compact using a displacement rate of 1 mm per pass until the machine had compacted the slab to the desired height of 110 mm. Prior to the testing, the research team needed to deter- mine the best way to produce the slabs and then handle them without causing damage. It was originally planned to demold the slabs after compaction but prior to testing. After the first few slabs were fabricated, it was observed that the slabs tended to crack during handling when removed from the mold, as shown in Figure 2.3. To counter this, a metal base Figure 2.4.  LWD testing of slab plate was placed in the slab mold before the mixed material confined using metal plates and was added. The idea was that the metal base plate would confined with tie-down strap and support the slab while the slab was removed from the mold. wood blocks. The plate did assist with reducing handling damage, but it was soon discovered that any testing away from the center of the slab (especially at early ages) caused the slab to crumble eter (LWD) showed a large difference in stiffness properties because of a lack of confinement. depending on the amount of pressure applied by the tie-down The research team next tried confining the slab by remov- strap. For these reasons, all testing was conducted within the ing the slab from the mold and using metal plates along the fabrication mold. edges of the slab. The metal plates were held together with a tie-down strap and wood blocks to hold the plates tight to the slab, as shown in Figure 2.4. It was discovered that the 2.2.5  Slab Specimen Testing confinement pressure was variable and difficult to replicate. In addition, exploratory testing with a lightweight deflectom- A series of existing and newly developed tests capable of assessing the early-age condition of recycled materials were identified in Phase I. These tests, listed in Table 2.2, were grouped into the following material property catego- ries: density, stiffness, penetration resistance, deformation Table 2.2.  List of properties and tests for Phase II testing. Property Suggested Test or Device Density Mass of dry material divided by slab volume Stiffness Soil stiffness gauge Lightweight deflectometer Penetration resistance Dynamic cone penetrometer Deformation resistance Marshall hammer Shear resistance Long-pin shear test* Raveling resistance Short-pin raveling test* Figure 2.3.  CCPR slab with arrow showing crack that Moisture Electromagnetic moisture probe formed during the demolding process. *Conceptual test proposed by the research team. 

13   resistance, shear resistance, raveling resistance, and mois- ture. The density was assessed by dividing the mass of dry material by the slab volume. The stiffness of the recycled materials was assessed by using a commercially available LWD and soil stiffness gauge (SSG). The penetration and deformation resistance were assessed using a commercially available dynamic cone penetrometer (DCP) and a Marshall hammer (MH) assembly having a 4-in.-diameter foot, respec- tively. Shear and raveling resistance were assessed using custom-fabricated fixtures developed during the project. Moisture was assessed using an electromagnetic moisture probe. Tests were conducted on single or replicate slabs; the number of replicates varied depending on the test conducted. All tests were conducted within the slab specimen com- paction mold. This was necessary since recycled materials tend to be susceptible to damage during handling, especially at early ages and when unconfined. It was recognized that there were likely some unaccounted edge effects that could influence the magnitude of the test results. However, the purpose of the Phase II laboratory testing was to assess the Figure 2.5.  LWD test conducted at the center of response of the various test methods with respect to changes a test slab. in material properties in the laboratory. It was expected that trends in the measured responses in the laboratory and during field testing would be similar. plete the test at early ages without plastically deforming the material was investigated, and it was found that drop heights 2.2.5.1  Stiffness Tests ranging from 4 in. to 12 in. produced similar test results Stiffness testing was completed using commercially avail- (using a 10-kg mass). All testing with the LWD was completed able LWD and SSG devices. Both devices were used to calcu- using a drop height of 12 in. This height/mass combination late the stiffness of the test slabs at 2 and 72 hours after slab applied a force of approximately 800 lbf and a pressure of fabrication. The time that the slabs were fabricated was used approximately 35 psi. to denote the start of the curing time. For LWD testing, a The SSG uses an electromechanical vibration to impart known load pulse was applied to induce a deflection on a test a small dynamic load as low-frequency sound waves on the slab surface, as shown in Figure 2.5. The vertical movement surface of a test slab. The resulting surface deflection as a of the surface was measured directly under the LWD with a function of frequency is measured. The test surface vibration 6-in.-diameter load plate and a fixed center deflection sensor. is applied between 100 Hz and 196 Hz at 4-Hz increments, The deflection measurements were then used to determine producing 25 steady-state frequencies. The magnitude of the applied force is about 9 N, and the induced deflections are the surface deflection modulus (stiffness) of the test slabs less than about 0.00005 in. The stiffness of the test slab is using Equation 1: determined for each of the 25 frequencies, and the average value from these measurements is reported as the stiffness. f ∗ s 0 ∗ a ∗ (1 − v 2 ) E0 = (1) Three replicates of the SSG test were performed at the d center of a test slab by rotating the device approximately 120° between tests, and the average stiffness value of the three where E0 is a surface deflection modulus; f is a factor for replicates was reported. stress distribution, taken as 2 for the measurements in this study; s0 is a stress under the LWD plate; ν is Poisson’s ratio 2.2.5.2  Moisture Content Tests (assumed to be 0.35); a is the radius of the plate; and d is the center deflection under the LWD. Moisture content testing was conducted using a recently The test variability was reduced when a total of 10 drops developed commercial electromagnetic moisture device. The were used for each test, with the average of the last three device was manufactured to be used in conjunction with a drops being reported. The drop height required to com- low-level NDG. Since the device could not be driven into the 

14 inserted into the test slab and is the same size as the probe rod of an NDG. 2.2.5.3  Penetration Resistance Tests The penetration resistance of the recycled material was measured using an MH assembly and a commercially avail- able DCP (conforming to the requirements of ASTM D6951). DCP testing was conducted by placing the DCP on top of the recycled slab and then dropping the 8-kg mass 575 mm and recording the penetration after each drop. Testing was conducted at 2 and 72 hours after compaction using the same slabs used during stiffness testing, and again at 1 and 24 hours after compaction for the same test slabs used during raveling testing (discussed in the following sections). DCP testing was conducted by placing the tip of a fixed cone on the recycled slab. The penetrated depth was recorded with each blow, starting at a penetration of zero. An MH assembly with a 4-in.-diameter foot and a 17.6-lb sliding weight falling 22.6 in. was used, as shown in Figure 2.7. The test procedure included dropping the weight Figure 2.6.  Moisture testing. on a location 20 times with recording of the penetration depth every five drops. The penetration depth was measured test slab itself, the probe end was inserted into a hole created using a digital caliper with an external depth blade. The by driving a metal rod (like that used for nuclear density test- penetration depth was measured at three locations along a ing in the direct transmission mode) into the test slab or by line (at approximately a 1-in. spacing) across the full diameter using the hole that remained after dynamic cone penetrom- of the penetrated area, as shown in Figure 2.7. MH testing eter testing. Figure 2.6 shows the device and data collection was conducted at 2 and 72 hours after compaction on the unit. Not shown in Figure 2.6 is the probe sensor, which is same test slabs used during stiffness testing. Figure 2.7.  Marshall hammer testing of a recycled slab (left) and penetrated area measurement (right). 

15   2.2.5.4  Shear and Raveling Tests Shear resistance was assessed using a developed fixture that could be driven using the upper assembly of a stan- dard DCP. The developed prototype shear fixture, shown in Figure 2.8, consisted of a steel base plate approximately 5 in. square with four outer pins (each 13⁄32-in. in diameter, extending 3.0 in. from the base plate) located along four points of a circle approximately 3.5 in. from a 1⁄2-in.-diameter, center pin that extended 3.0 in. from the base plate. The test performed with this fixture was termed a “long-pin shear test.” The term “long-pin” is used to differentiate this fixture from a similar-looking fixture with shorter pins used to assess raveling resistance. To conduct the test, the shear test fixture was driven into the test slab until the plate seated on the surface, using the upper assembly of a DCP that fits over the center shaft on top of the fixture base plate, as shown in Figure 2.9. The center shaft had a diameter of 1.0 in. with a hexagonal head milled into it so that a 3⁄4-in. socket could be attached to the center to accommodate a handheld torque wrench. After the fixture Figure 2.9.  Upper assembly from a was driven in and the number of blows until the base plate DCP used to drive the long-pin shear touched the slab surface was recorded, the operator used test fixture. a torque wrench to apply a rotational force, as shown in Figure 2.10. The maximum torque reading was recorded. The length of pins was intentionally chosen to match the this fixture is termed a “short-pin raveling test.” The length approximate minimum thickness of a recycled layer (approxi- of pins was chosen to be similar to the likely maximum par- mately 3 in.). Pins were included in the fixture design, rather ticle size of most recycled materials (approximately 1 in.). than solid vanes, to reduce damage to the recycled layer caused The test procedure was essentially the same as the long-pin by testing. shear test in that the upper assembly of a DCP was used to The raveling resistance of the recycled materials was drive the raveling test fixture into the test slab, as shown in assessed using a modified version of the shear test fixture. Figure 2.12. To maintain a constant normal force that kept The raveling fixture is similar to the shear test fixture, but the the short pins from riding up onto the slab surface, two 10-lb outer pins used for the raveling test extend 1.0 in. from the plates were added on top of the raveling test fixture to apply base plate, as shown in Figure 2.11. The test performed with a normal force. The operator used a torque wrench to apply Figure 2.10.  Measuring torque with long-pin shear Figure 2.8.  Prototype long-pin shear test fixture. test fixture. 

16 Figure 2.11.  Prototype short-pin raveling test fixture. a rotational force, as shown in Figure 2.13, and the maximum Figure 2.13.  Applying torque to measure raveling torque reading was recorded. resistance. Since the raveling fixture pins were of different lengths, two separate blow counts were recorded for the short-pin raveling test, as described in the following. The number of number of blows to these various positions was recorded blows required to drive the fixture to the tip of the shorter in case one measurement proved to be a more significant outer pins was counted and denoted N1. The cumulative predictor of performance than another. number of blows required to drive the entire fixture to the level of the base plate was recorded and denoted N2. The 2.2.5.5  Other Tests Considered Several other tests were considered for the laboratory experiment but ultimately were not chosen by the research team. These tests included penetration resistance tests using a rapid compaction control device, a stiffness test using a Clegg hammer, stiffness assessment using an ultrasonic pulse velocity and portable seismic pavement analyzer, a raveling/abrasion test using the Wet Track Abrasion Test and cohesion testing (ASTM D3910), and a cohesion test with a field-portable pneumatic cohesion tester (based on the ASTM D3910 cohesion test). Ultimately, these tests were not selected because of issues such as limited device availability, incompatibility with early-age properties of the recycled material, and difficulties with demonstrating expected per- formance trends with changes in material properties. These tests, which were not part of the laboratory experiment, are not discussed further in this report. 2.2.5.6  Test Arrangement Three sets of test slabs from each project were fabricated to accommodate all the tests. For each project, single slabs Figure 2.12.  Applying blows using DCP or replicates were fabricated depending on the amount of upper assembly with short-pin raveling loose material available. The first set of slabs was fabricated test fixture. to facilitate moisture, stiffness, and penetration resistance 

17   testing. The second set was used for stiffness (by LWD only) and shear resistance testing. The third set of slabs was used for stiffness (by LWD only), DCP, and raveling testing. The three sets of tests were conducted to maximize the amount of information that could be collected while reducing the potential for one test to influence another. Moisture, stiffness, and penetration resistance testing was 6-in.-diameter LWD conducted on the first set of slabs. The tests were arranged on the slab so that one slab could be used to support multiple tests at two curing times (2 and 72 hours after slab fabrication), Shear Fixture as shown in Figure 2.14. At the 2-hour test, LWD and then the SSG tests were conducted at the center of the slab. Next, the MH was used for penetration resistance testing at two corner locations on one side of the slab (e.g., upper left and lower left, as shown in Figure 2.14). The tests were con- ducted such that the MH foot was approximately 2 in. from Figure 2.15.  Test locations for laboratory-fabricated any edge of the slab. Following this, the DCP was used at slabs for shear tests (400 mm ë 500 mm), shown approximately the midpoint between the two MH tests to scale. and approximately 4 in. from the edge of the slab. Moisture content measurements with the moisture gauge were taken in the hole left after the DCP test. At the 72-hour test, the were prepared from each source project. Tests were conducted LWD and SSG tests were again conducted at the center of the on each slab at 1, 3, 6, and 24 hours after compaction. The slab. The MH, DCP, and moisture tests were then conducted LWD test was conducted first, followed by the shear test at at the end of the slab opposite to the end used in the 2-hour the same location. At each curing time, a different corner test. Moisture contents at the 2-hour test were compared to the of the slab was tested. Figure 2.15 is a schematic of the test moisture content during mixing, and the moisture content at locations. the 72-hour test was compared to the moisture content of a Stiffness, DCP, and raveling tests were conducted on the sample taken from the slab after all tests were completed and third set of slabs in a similar way to those performed on the then dried in an oven. shear test slabs. One or two replicate slabs were prepared Stiffness and shear resistance tests were conducted on the from each source project. Tests were conducted on each slab second set of slabs. These tests were arranged differently at 1, 3, 6, and 24 hours after compaction. The LWD test was since the shear test is destructive. Two or three replicate slabs conducted first, followed by the raveling test at the same location. At each curing time, a single measurement using the LWD and raveling fixture were conducted in a different corner of the slab. The DCP test was conducted adjacent to 4-in.-diameter MH the longest dimension of the slab at 1 and 24 hours only. Figure 2.16 is a schematic of the test locations. 2.3  Phase III—Field Testing Testing to assess the short-term properties of recycled materials on nine field projects in the United States was DCP and conducted at the locations shown in Figure 2.17. The field MH Test Line Moisture Gauge projects included CIR, CCPR, and FDR using either emulsi- fied or foamed asphalt as the stabilizing/recycling agent with and without cement as an active filler. The projects were 6-in.-diameter LWD completed by multiple contractors using different source materials and were located in different climatic regions. 8-in.-outer-diameter SSG 2.3.1 Objectives Figure 2.14.  Test locations for laboratory-fabricated slabs for stiffness and penetration resistance tests Phase III field testing was conducted to assess the most (400 mm ë 500 mm), shown to scale. applicable tests from Phase II and to determine the appropriate 

18 contractor representatives. All tests were performed within test blocks that were approximately 4 ft (in the direction of traffic) by 2 ft (perpendicular to the direction of traffic), as shown in Figure 2.18. Within each test block, one field density test and three replicates of stiffness (LWD and SSG), DCP, shear (torque and number of blows), and raveling (torque and number of blows) were conducted. Based on the results 6-in.-diameter LWD of laboratory testing, MH testing was not conducted at the field projects. For several of the early projects, moisture tests Raveling Fixture were also conducted in the same hole following the DCP tests. Where possible, replicate test blocks were used to gain DCP knowledge of test variability at multiple locations. An NDG was not available for every testing block, so selected test blocks were located near previously performed field density tests. All tests within each test block were completed within approximately 30 minutes. Figure 2.16.  Test locations for laboratory-fabricated slabs for raveling tests (400 mm ë 500 mm), shown to scale. 2.3.4  Testing Details The research team conducted at least one test block at limits of the tests for identifying time to opening or surfacing. selected curing times ranging from 0.5 hour to 48 hours The research team tested multiple locations within the same after the contractor completed the compaction process at the project when changes in material properties were observed. selected test location within the project. Table 2.4 shows the In addition to the test assessment, a field-based preliminary curing times and number of replicate test blocks for each interlaboratory study (ILS) was performed at one project field project. location to obtain an indication of the precision of the test Table 2.5 shows the distribution of testing by process and methods. It was impractical to ship (and receive undamaged) material type. From Table 2.5, the total number of sections test slabs produced in the laboratory to other research team exceeds the number of sites visited since multiple sections were members while maintaining curing conditions. So the ILS was tested at some projects. A total of eight unique process and conducted in the field, as has been done with previous studies material combinations were assessed from testing 14 sections on the DCP and for fresh properties of Portland cement con- and 51 test blocks. crete. Lessons learned, draft methods of test, and precision statements were developed from this Phase III testing. 2.4 Repeatability and Reproducibility of Field Raveling and Shear Tests 2.3.2  Field Project Summary A preliminary evaluation of repeatability and repro- Table 2.3 shows a summary of the Phase III field projects. ducibility of the raveling and shear tests developed in this All testing was done during the 2019 construction season. research was conducted by means of a ruggedness evaluation The additive and filler contents shown in Table 2.3 were and an ILS in accordance with ASTM C802. In accordance obtained from the material mix designs that were completed with ASTM E1169 and ASTM C1067, the ruggedness eval­ by the contractor or agency. Either a contractor or an agency uation was needed prior to the ILS since the raveling and representative used an NDG to measure the field density shear tests were newly developed. Once completed, a preci- within or near the test location; the measured field densities sion statement was prepared for each test. are shown in Table 2.3. A ruggedness evaluation is a controlled experiment where factors or test conditions are varied to evaluate their effect on the test response. Examples of factors that were initially con- 2.3.3  Test Block Layout sidered include test temperature, lift thickness, load/torque At each field project, the research team met with agency applied, and test equipment apparatus physical characteristic and contractor representatives to discuss the goals of testing dimensions or rates. For each factor, variations or levels were and the specific locations available that day. All testing was determined at the expected extreme values for each level, conducted by members of the research team, but consider- and the respective test was conducted. If the impact of level able logistical support was provided by the local agency and variation due to operating conditions and tolerances proved 

6 3 2 With Active Filler No Active Filler CIR CIR CCPR CCPR FDR FDR FDR, CIR, and CCPR (ILS) Numbers in shapes indicate number of projects/processes tested if more than one. Figure 2.17.  Project locations for Phase III. Table 2.3.  Phase III project summary. Agent Active Field Nearest Content, Active Filler Density, State Project Town Process Agent % Filler Content, % lb/ft3 Route 30 Andes Emulsion 2.5 129.1 CIR NY Route 28 Meredith Emulsion 3.0 None 0.0 141.0 Route 23A Prattsville Foam 2.8 131.9 Foam 2.3 Cement 1.0 133.8 CCPR Emulsion 3.5 None 0.0 131.3 Foam 2.3 Cement 1.0 130.6 MN 70th Street Albertville CIR Emulsion 2.8 None 0.0 129.0 FDR-HD Emulsion 3.0 Cement 1.0 133.0 FDR-LD Emulsion 3.0 Cement 1.0 117.3 SC SC 123 Clemson FDR Foam 2.3 Cement 1.0 119.4 CIR-GS Emulsion 2.5 None 0.0 122.5 IN SR 1 Ft. Wayne CIR-PS Emulsion 2.5 None 0.0 122.5 CA SR 178 Ridgecrest CIR Emulsion 3.4 Cement 0.5 121.6 NM U.S. 491 Tohatchi FDR Foam 2.0 Cement 1.0 134.0 CA SR 22 Woodland CIR Foam 2.2 Cement 1.0 127.9 Notes: FDR-HD = full-depth reclamation, high density; FDR-LD = full-depth reclamation, low density; CIR-GS = cold in-place recycling, good support; CIR-PS = cold in-place recycling, poor support. 

20 Table 2.5.  Phase III total number of sections and test blocks by process. 4 Feet Process Agent and Filler No. of Sections No. of Test Blocks F-C 1 3 LWD/SSG F-N – – Shear 1 Shear 2 Raveling 1 CCPR 1 E-C – – E-N 1 3 2 Feet F-C 2 6 LWD/SSG LWD/SSG Shear 3 NDG DCP 1 F-N 1 2 2 3 CIR E-C 1 3 E-N 4 18 Raveling F-C 2 10 Raveling 2 DCP 2 DCP 3 F-N – – 3 FDR E-C 2 6 E-N – – Figure 2.18.  Phase III test block layout. Total number of sections 14 Total number of test blocks 51 Total number of unique process/material combinations 8 Notes: F-C = foam plus cement; F-N = foam, no cement; E-C = emulsion Table 2.4.  Phase III project testing details. plus cement; E-N = emulsion, no cement. Curing No. of Replicate State Project Process Time Test Blocks 1 hour 2 to be too great, the test method needed to be refined further Route 30 CIR E-N 24 hours 2 or the tolerance reduced prior to performing an ILS. 1 hour 1 NY Route 28 CIR E-N Following the ruggedness evaluation, an ILS was con- 48 hours 1 1 hour 1 ducted to generate precision statements for the newly devel- Route 23A CIR F-N 18 hours 1 oped test methods in accordance with ASTM C802. For this CCPR F-C 1 hour 3 CCPR E-N 1 hour 3 study, multiple laboratories were represented by three CIR F-C 1 hour 3 institutions in the research team. Multiple materials were MN 70th Street CIR E-N 1 hour 3 assessed at the field-testing site at the MnROAD test track FDR E-C HD 3 hours 3 FDR E-C LD 3 hours 3 in August 2019 (as shown in Figure 2.17). 1 hour 2 The ILS is termed “preliminary” because it was not possible SC SC 123 FDR F-C 4 hours 1 to fulfill all of the requirements of ASTM C802, Standard 1 hour 1 3 hours 1 Practice for Conducting an Interlaboratory Test Program to IN SR 1 CIR E-N GS Determine the Precision of Test Methods for Construction 6 hours 1 24 hours 1 Materials, which include a valid written test method, a rugged- 1 hour 2 3 hours 1 ness test prior to the ILS, and a minimum of 10 participating IN SR 1 CIR E-N PS laboratories. Proposed test methods were written for each test, 6 hours 1 24 hours 1 ruggedness tests were conducted for each test, and at least 1 hours 1 CA SR 178 CIR E-C 1.5 hours 1 six materials were used per test method. However, only three 3 hours 1 laboratories participated because the tests were new and 0.5 hours 2 commercially available equipment was not available. Also, 2 hours 1 NM U.S. 491 FDR F-C 3 hours 1 the ILS was conducted in the field rather than in a laboratory. 4 hours 1 This was done because of the difficulty associated with pre- 24 hours 2 paring, shipping, and testing undamaged slabs. 2 hours 1 CA SR 22 CIR F-C 6 hours 1 24 hours 1 2.4.1  Ruggedness Evaluation Notes: CIR E-N = cold in-place recycling, emulsion, no cement; CIR F-N = cold in-place recycling, foam, no cement; CCPR F-C = cold central-plant A ruggedness evaluation was performed using a single recycling, foam plus cement; CCPR E-N = cold central-plant recycling, emulsion, no cement; FDR E-C = full-depth reclamation, emulsion plus material for both the shear and raveling test methods in cement; FDR E-C HD = full-depth reclamation, emulsion plus cement, accordance with ASTM C1067-2, Standard Practice for Con- high density; FDR E-C LD = full-depth reclamation, emulsion plus cement, low density; CIR E-N GS = cold in-place recycling, emulsion, ducting a Ruggedness Evaluation or Screening Program for no cement, good support; CIR E-N PS = cold in-place recycling, emulsion, Test Methods for Construction Materials. The laboratory no cement, poor support. testing conducted as part of Phase II and engineering judg- ment by the research team were used to identify potentially influential factors. Optimally, a ruggedness evaluation would 

21   Table 2.6.  Raveling test factors and levels for ruggedness evaluation. Factor Level Outer Pin Pin Tip Applied Angular Rate, Outer Pin Length, in. Angle, ° °/sec Tip Dullness Diameter, in. Level 1 (+1) 0.85 70 90 Sharp 1/2 Level 2 (−1) 0.65 50 60 Dull (0.1 in.) 13/32 Table 2.7.  Shear test factors and levels for ruggedness evaluation. Factor Level Length, in. Tip Angle, ° Angular Rate, °/sec Tip Dullness Outer Pin Diameter, in. Level 1 (+1) 3.1 85 90 Sharp 1/2 Level 2 (−1) 2.9 65 60 Dull (0.1 in.) 13/32 assess each influential factor both independently and inter- Tables 2.6 and 2.7 show the factors and levels for the ravel- dependently using full factorial designs. However, this was ing and shear tests, respectively. The factor and level com- not possible given the time and material resources required. binations were randomly assigned to different slabs as part Thus, a partial factorial experiment was designed in accor- of an experimental design. The factors investigated included dance with ASTM C1067. This standard provides clear direc- length of the outer pins, pin tip angle, angular rate applied tion for the design of a ruggedness evaluation for construction to the torque wrench, tip dullness, and outer pin diameter. materials using the Plackett-Burman design. The tip dullness was adjusted by first performing the tests with a sharp tip and then grinding off approximately 0.1 in. of the tip to a more flattened tip. Tables 2.6 and 2.7 indicate 2.4.1.1  Ruggedness Evaluation Factors and Levels two levels, denoted with plus (+) and minus (−) signs. A plus Eight specimens were prepared and tested using a single (+) sign for a given factor indicates that the measurement material for each evaluated test. Per a Plackett-Burman design, was made with that factor at the high level, and a minus up to seven factors could be considered, with each factor (−) sign indicates the factor was at a low level. The factors having two levels. In the ruggedness evaluation experimental and values for each level were determined based on the design, a partial factorial was developed by varying the com- results of concurrent laboratory testing in Phase II, limited binations of factor upper and lower levels among the eight field testing, and the engineering judgment of the research specimens. team. Tables 2.8 and 2.9 show the associated experimental Table 2.8.  Raveling test experimental design for ruggedness evaluation. Factor Specimen Length, in. Tip Angle, ° Angular Rate, °/sec Tip Dullness Outer Pin Diameter, in. Specimen 1 +1 +1 +1 −1 −1 Specimen 2 −1 +1 +1 +1 +1 Specimen 3 −1 −1 +1 +1 −1 Specimen 4 +1 −1 −1 +1 +1 Specimen 5 −1 +1 −1 −1 +1 Specimen 6 +1 −1 +1 −1 +1 Specimen 7 +1 +1 −1 +1 −1 Specimen 8 −1 −1 −1 −1 −1 Table 2.9.  Shear test experimental design for ruggedness evaluation. Factor Specimen Length, in. Tip Angle, ° Angular Rate, °/sec Tip Dullness Outer Pin Diameter, in. Specimen 1 +1 +1 +1 −1 −1 Specimen 2 −1 +1 +1 +1 +1 Specimen 3 −1 −1 +1 +1 −1 Specimen 4 +1 −1 −1 +1 +1 Specimen 5 −1 +1 −1 −1 +1 Specimen 6 +1 −1 +1 −1 +1 Specimen 7 +1 +1 −1 +1 −1 Specimen 8 −1 −1 −1 −1 −1 

22 designs for the raveling and shear test methods, respec- 2.4.2  Interlaboratory Study tively. The ruggedness study was conducted in accordance The research team conducted an ILS to develop pre­ with ASTM E1169, Standard Practice for Conducting Rugged- liminary precision statements for the shear and raveling ness Tests. tests developed in this study. The term “preliminary” is used since only three laboratories participated in the ILS. The 2.4.1.2  Slab Preparation for Ruggedness Test ILS was conducted in accordance with ASTM C802 and ASTM C670, Standard Practice for Preparing Precision and Slabs prepared using RAP from Lockwood, Nevada, and Bias Statements for Test Methods for Construction Materials. PASS-R emulsified asphalt were used to conduct the rugged- During the ILS, DCP tests were conducted in addition to the ness evaluation. Oven-dried RAP was mixed with 3% water shear and raveling tests, allowing precision statements to be and 4% emulsion (2.4% residual bitumen) until uniformly prepared for this test. coated. The optimum moisture content and emulsion content ASTM C802 outlines the following general requirements were selected by comparing strength test results for speci- for an ILS: mens prepared in a gyratory compactor. For the shear tests, a Vibroplate compactor was used to compact the mixture to 1. A valid and well-written test method; a target density of 130 lb/ft3 in a mold having dimensions 2. Established tolerances for various conditions in each test of 24 in. × 59 in. × 3.5 in. Test slabs for raveling tests were method (e.g., from a ruggedness study); fabricated in the same manner, but the molds had dimen- 3. Clearly defined and an available apparatus for perform- sions of 24 in. × 30 in. × 3.5 in. For the shear test, the length ing the test; of the mold was increased from 30 in. to 59 in. so that all 4. Personnel in participating laboratories with adequate the experiments could be performed on the same slab with experience; minimum disturbance from previous tests. The molds were 5. Preliminary knowledge of how changes in materials and conditions affect the test results; anchored to the concrete floor, as shown in Figure 2.19. All 6. Procedures and facilities for obtaining, preparing, and the shear and raveling tests were conducted after 4 hours of distributing test specimens; curing at ambient conditions. The slabs were cast outside 7. Randomized selection of test specimens for distribution and exposed directly to the sun. to laboratories; 8. Application of the test method on materials with a range of properties representative of the characteristics for which the method will be used; 9. Adequate number of participating laboratories, with at least 10 recommended; and 10. At least three materials or materials with three different average values of the measured test characteristic. For this study, the research team was unable to satisfy requirements 6, 7, and 9. Requirements 6 and 7 could not be satisfied because the research team could not distribute samples to participating laboratories without damaging the samples because of the nature of the material and the effects of transportation and aging. However, within ASTM C802, a provision exists that operators can convene at one location if material cannot be distributed. The research team incor- porated this provision by performing all testing near the MnROAD test track as part of another ongoing study on a portion of 70th Street in Albertville/Oswego, Minnesota. At this location, multiple recycling processes and stabilizing/ recycling agents were used on the roadway to better under- stand their performance as a pavement rehabilitation tech- Figure 2.19.  Compacted slab for ruggedness nique. An example of the preconstruction condition is shown evaluation. in Figure 2.20. 

23   Figure 2.21.  Members of research team conducting ILS. 2.4.2.1  Experimental Design The ILS was conducted on six unique pavement test sec- tions. The test sections included CIR, CCPR, and FDR using emulsified or foamed asphalt, and some contained cement Figure 2.20.  Preconstruction view of 70th Street, Albertville/Otsego, as an active filler. Table 2.10 illustrates the planned test sec- Minnesota. tions; each test section was 500 ft long. Note that the larger MnROAD experiment included mill and fill and thinlay sections that were not tested by the research team. The ini- The research team for this study was able to conduct the tial plan was to conduct testing on both the eastbound and ILS during this unique opportunity. Requirement 9 was left westbound lanes for cells 1 through 6. Because of weather unsatisfied since the number of member institutions on the restrictions and construction challenges, this was not feasible. research team was less than the required number and the Table 2.11 shows the recycling work that was completed and developed tests were not yet commercially available. A photo­ the five sections that were used during the ILS. The beginning graph of the research team members conducting the ILS is of one section (FDR E-C) was found to have lower density shown in Figure 2.21. than the remainder of the section using the nuclear density Table 2.10.  Proposed ILS test sections. Test Section Direction 1 2 3 4 5 6 7 8 Mill & fill Westbound FDR F-C FDR E-C CIR F-C CIR E-N CCPR E-N CCPR F-C Thinlay thinlay Mill & fill Eastbound FDR F-C FDR E-C CIR F-C CIR E-N CCPR E-N CCPR F-C Thinlay thinlay Notes: CIR E-N = cold in-place recycling, emulsion, no cement; CIR F-C = cold in-place recycling, foam plus cement. CCPR F-C = cold central-plant recycling, foam plus cement; CCPR E-N = cold central-plant recycling, emulsion, no cement; FDR F-C = full-depth reclamation, foam plus cement; FDR E-C = full-depth reclamation, emulsion plus cement. Table 2.11.  Actual ILS test sections. Test Section Direction 1 2 3 4 5 6 7 8 Westbound – – – – – – CCPR F-C – Eastbound FDR E-C – CIR F-C CIR E-N CCPR E-N – CCPR F-C – Notes: CIR E-N = cold in-place recycling, emulsion, no cement; CIR F-C = cold in-place recycling, foam plus cement; CCPR F-C = cold central-plant recycling, foam plus cement; CCPR E-N = cold central-plant recycling, emulsion, no cement; FDR E-C = full-depth reclamation, emulsion plus cement. 

24 gauge. The research team completed testing in this area and a way similar to that shown in Figure 2.18. Three replicate considered it a sixth material type. LWD, SSG, DCP, shear, and raveling tests were performed The research team established three adjacent test blocks in each test block as soon as the test section was available in a random location along the length of the test section following compaction. After all testing was complete, the and in the center of the lane. Each of the three laborato- roadway was reopened to traffic. From the collected data, ries was randomly assigned to one of the test blocks. The the single-operator and multi-laboratory precision values location of each test within the test block was arranged in were calculated.