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.
F-1 Proposed AASHTO Recommended Practice for Conducting Plant Aging Studies A P P E N D I X F Standard Recommended Practice for Measuring the Effects of Asphalt Plant Mixing and Processing on Binder Absorption by Aggregate and Asphalt Mixture Characteristics AASHTO Designation: X xx-xx INTRODUCTION The effects of plant production on asphalt material properties have been researched for many years resulting in a number of changes in mixture design procedures to produce laboratory specimens with the same degree of aging as mixtures processed through a batch or drum mixing plant. The importance of being able to simulate plant aging in the laboratory is high for mixture volumetric considerations such as asphalt absorption, and it is absolutely essential for performance testing of mixtures for moisture, rutting, and fatigue susceptibility. National Cooperative Highway Research Program (NCHRP) Project 9-52 provided guidance on laboratory short-term oven aging (STOA) techniques for warm mix asphalt (WMA) and hot mix asphalt (HMA) that were able to simulate plant aging over a wide range of variables, including WMA technology, aggregate asphalt absorption, recycled material content, and asphalt source. Recent and potential future changes in the production and formulations of asphalt mixtures such as the use of new WMA additives, higher recycled material content, and the use of asphalt recycling agents may require evaluations of their effects upon short-term aging. 1. SCOPE 1.1. This recommended practice provides a structure for conducting studies of mixture plant aging effects on asphalt mixtures. It is intended as a general guidance document that may need to be adapted for specific variables and conditions. It reflects the experience of the research team in NCHRP Project 9-52. An overall flowchart for a short-term aging study is shown in Figure 1.
F-2 Figure 1. Overall Flow Chart for Conducting Short-Term Aging Studies of Asphalt Mixtures 1.2. This standard may involve hazardous materials, operations, and equipment. This standard does not propose to address all safety problems associated with its usage. It is the duty and responsibility of the user of this standard to establish appropriate safety and health practices and determine the applicability of regulatory limitations prior to use. Note 1âThe values stated in SI units are to be regarded as the standard. 2. REFERENCED DOCUMENTS 2.1. Newcomb, D., A. Epps Martin, F. Yin, E. Arambula, E.S. Park, A. Chowdhury, R. Brown, C. Rodezno, N. Tran, E. Coleri, D. Jones, J.T. Harvey, and J.M. Signore. NCHRP Report 815: Short-Term Laboratory Conditioning of Asphalt Mixtures, Transportation Research Board, 2015. 2.2. Hogg, R.V. and J. Ledolter. Engineering Statistics, Macmillan, 1987. 3. EXPERIMENTAL DESIGN 3.1. The most important part of conducting an aging study is to define the goals and desired outcomes in terms of what questions are to be answered. Variables that may be of interest could include the effect of: Aggregate sources Asphalt sources WMA technologies Presence and concentration of recycled binder replacement Asphalt grades Experiment Design Planning LMLC Samples Sample Stockpiles and Tanks PMPC Samples Monitor Construction Conditions Obtain Samples Fabricate Samples IAW AASHTO M 323 â Use NCHRP 9-49/9-52 STOA Execution Compact at or near Plant Volumetric and Performance Testing Volumetric and Performance Testing Data Analysis Aggregate gradations
F-3 3.2. If a relatively few factors and levels of factors are included in the study, a full-factorial experiment will give the most definitive results. Each factor and level will comprise one cell in the experiment design. The number of replicate samples to be prepared for each cell should be no less than three, and should be increased for larger variability. It should be noted that the addition of a factor with two levels doubles the number of cells in the experiment design matrix and a full-factorial experiment design will soon become unwieldy. If there are a moderate to high number of cells, then greater cost and time efficiency can be gained by a partial factorial experiment wherein only some fraction of the cells in the factorial are populated. 3.3. The types of samples to be obtained should also be considered in the experiment design. It is possible that, for the laboratory portion of the study [laboratory-mixed, laboratory- compacted (LMLC) samples], more factors and levels may be desirable to obtain a more detailed, controlled set of data on the effects of factors at various levels. In fact, a preliminary laboratory study may highlight certain variables of interest and discount others so that a more robust field study may be conducted. 3.4. The field portion of the study should be designed to obtain the needed comparison between factors and levels as efficiently as possible. Thus, trends in the laboratory data from LMLC samples should be studied before a given factor or levels of a factor are considered for production in an asphalt plant. Samples taken during production are referred to as plant- mixed, plant-compacted (PMPC) specimens and are taken as near as practical to the point of production. The factors and levels for the plant portion of the study should be selected to be as efficient as possible to limit costs and the number of plant changes required in production. 3.5. An example of a simple experiment design to study the differences in short-term aging between WMA and HMA from two different asphalt binder sources (S1 and S2) are shown in Figure 2. Here there are only two factorsâasphalt mix type and asphalt binder sourceâ with two levels each. For the sake of illustration, three replicates (R1, R2, and R3) represent the individual tests in each cell. The number of replicates in the cells should be established based upon the reproducibility of the test (within-laboratory variability). The greater the variability, the greater the number of samples required. Figure 2. An Experimental Design for Asphalt Mix Type and Binder Source 3.6. The data analysis of the matrix is performed through an analysis of variance (ANOVA) to determine if the means of the cells are significantly different from each other. Following this, a pair-wise comparison of means [such as Tukeyâs honest significant difference (HSD) method] may be done to rank and group the levels of factors into those that are the same and those that are different. A good reference on experiment design is Hogg and Ledolter (1987). 4. PLANNING 4.1. Once the experiment design has been established, planning the schedule for sampling, Factor Asphalt Mix Type WMA HMA Binder Source S1 R1, R2, R3 R1, R2, R3 S2 R1, R2, R3 R1, R2, R3 sample preparation, and testing may begin. This process ensures that the proper amount of material for LMLC and PMPC samples are obtained from the specified points of sampling.
F-4 4.2. A checklist should be made to ensure that all the needed laboratory equipment and supplies are ready. This checklist should include all the equipment and supplies for volumetric testing and any desired performance testing. 4.3. Sampling for LMLC specimens should include aggregate and reclaimed asphalt pavement (RAP) stockpiles to be used in the mixtures as well as the virgin asphalt binder and any additives. Sampling and testing these materials should be as close as possible to the time of production. This will help ensure that the materials sampled will be those used in production. 4.4. Sampling asphalt mixtures at the plant should be done as close to the point of loadout from the silo as possible. Ideally, the plant asphalt mix for PMPC should be taken from a sampling platform prior to trucks exiting the plant site. A platform may not always be available, especially in the case of a portable plant. In these instances, the sample may be taken from the pavement site, preferably from a windrow, provided the job site is no more than a 20-minute drive from the laboratory where the PMPC sample will be compacted. 4.5. For each cell in the experiment design, the engineer or technician will need to determine the amount of material needed in each cell. Normally, a conservative approach of ordering 20 to 30 percent over the amount required for each cell works best to account for potential errors in fabrication and handling of specimens. 5. LABORATORY-MIXED, LABORATORY-COMPACTED SPECIMENS 5.1. Preparation: 5.1.1. As stated above, materials need to be sampled from aggregate stockpiles, RAP stockpiles, asphalt binder tanks, and additive bins or tanks in quantities sufficient to prepare enough LMLC samples for the study. The mixture design procedure in AASHTO M 323 should be followed in the preparation of laboratory samples with the exception of the STOA protocol for both volumetric and performance testing specimens. 5.1.2. NCHRP Projects 9-49 and 9-52 have shown that using an STOA process of storing the loose mix in an oven at 240°F (116°C) for WMA and at 275°F (135°C) for HMA at two hours prior to compaction produces volumetric, stiffness, and rutting resistance values which match plant-produced mixtures very closely. This differs from AASHTO M 323 which requires an STOA of 275°F (135°C) for two hours for volumetric testing and four hours for performance testing regardless of the mixture type. Thus, for LMLC specimens, the STOA of 240°F (116°C) for WMA and at 275°F (135°C) for HMA at two hours should be used for both volumetric and performance testing. 5.1.3. Compaction of the specimens should follow the number of required gyrations for 96 percent of maximum density for volumetric testing and 93 percent of maximum density for performance testing. Once the LMLC specimens have been compacted, they should be tested for volumetric mix parameters and appropriately shaped for performance testing. 5.2. Testing: 5.2.1. Volumetric and performance testing should follow sample preparation as closely as possible. Normally, testing should be accomplished within 48 hours of sample fabrication if stored in a 77°F (25°C) room or within two weeks if stored in a 60°F (15°C) chamber.
F-5 5.2.2. Volumetric parameters that may be of interest to compare LMLC short-term aged with PMPC specimens include maximum specific gravity (Gmm), effective binder content (Pbe), percent binder absorbed (Pba), and film thickness (FT). As shown in Figures 2 and 3, volumetric properties tracked very well between LMLC specimens and PMPC specimens with the exception of one mixture with a very highly absorptive aggregate. However, this procedure can be followed if there is a suspicion that the volumetrics between mixture design and production are significantly different. 5.2.3. Performance testing may include stiffness measurements such as resilient modulus or dynamic modulus; a rutting test such as Hamburg wheel-tracking test, asphalt pavement analyzer, or flow number; or a cracking test such as the semi-circular beam, disk compact- shaped tension, or the overlay tester. In NCHRP 9-52 it was noted that stiffness measurements were the best discriminators of performance between different mix parameters and plant characteristics. Although cracking evaluation was not a part of NCHRP 9-52, a complete field study is being design for cracking in NCHRP 9-57. 6. PLANT-MIXED, PLANT-COMPACTED SPECIMENS 6.1. Plant Operations: The field portion of the aging study may be conducted using either a drum mix plant or a batch plant. NCHRP Project 9-52 confirmed that both types of plants are capable of producing similar mixtures. Ensure that the asphalt plant is operating well and the proper materials are being introduced in the quantities determined from the mixture design. A minimum run of 500 tons of asphalt mix is required to be sure that the plant is operating normally before sampling the material for each condition to be studied. So, for the experiment design presented earlier, at least 500 tons of mix would need to be produced at each combination of asphalt mixture type and asphalt binder source. 6.2. Sampling and Preparation: 6.2.1. The plant-produced mixture should be sampled out of the back of a dump truck from a sampling platform positioned just after the trucks pull off the weigh scale at the plant. If circumstances prevent sampling from the truck, loose mix samples may be taken from the paving site and transported to the field lab for compaction provided that the time between production and sampling is no longer than 20 minutes and that the travel time from the paving site to the laboratory is no longer than 20 minutes. 6.2.2. After sampling the plant-produced mixture, it should be immediately transported to the field testing laboratory. Once at the laboratory, individual samples for compaction should be weighed into metal bowls and placed in an oven until the material reaches the field compaction temperature. The samples should then be left in the oven until they have reached the compaction temperature. Once the mixture has achieved the field compaction temperature, the samples should be compacted in a Superpave gyratory compactor to 96 percent of Gmm for volumetric testing and 93 percent of Gmm for performance testing. 6.3. Testing: 6.3.1. As with the LMLC mixtures, volumetric and performance testing should follow sample preparation as closely as possible. Normally, testing should be accomplished within 48 hours of sample fabrication if stored in a 77°F (25°C) room or within two weeks if stored in a 60°F (15°C) chamber. 6.3.2. The same volumetric parameters measured for the LMLC short-term aged specimens should be measured for the PMPC samples including Gmm, Pbe, Pba, and FT. A comparison of these
F-6 parameters with the LMLC volumetric measurements will verify that the plant-produced materials are sufficiently similar to those produced in mixture design so that the comparison performance testing results will not be affected by extraneous variables. Performance testing should include the same stiffness measurements, rutting tests, and cracking tests that were used in evaluating the LMLC material. 7. DATA ANALYSIS 7.1. The first step in data analysis should be to compare the results of LMLC volumetric and performance testing against that of the PMPC. A cell-by-cell comparison can be made by plotting the means of the cells for the two sample types against one another along a 45° line of equality as shown in Figure 3. This will show whether there is a bias toward either the lab-produced or field-produced material. If the data points fall close to the line of equality, then it is likely that there are minimal differences between the two types of preparation. Figure 3. Comparison of overall LMLC and PMPC results. 7.2. Next, a factor analysis can be done by comparing the results by using ANOVA as previously described to compare the means of different levels of factors. This will establish whether a factor influences the measured parameter. For instance, Figure 4 shows a hypothetical outcome in which WMA is shown to be less stiff than HMA. Figure 4. Stiffness Comparison for HMA versus WMA PM PC A ir Vo id s LMLC Air Voids W M A St iff ne ss HMA Stiffness
F-7 7.3. Once the ANOVA has been completed, an additional step of ranking the data can be accomplished using Tukeyâs method, for instance. This type of ranking is very beneficial when there are a number of levels of each factor. 7.4. It is strongly suggested that, as more factors and levels are considered, the advice of a statistics expert should be sought at the experiment design and data analysis stages.
