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.
111 APPENDIX A: REVIEW OF EXISTING BINDER FATIGUE TESTS AND SELECTION FOR FURTHER EVALUATION AS PART OF NCHRP 9-59 INTRODUCTION The purpose of this appendix is to present a discussion of how binder tests were selected for in depth evaluation as potential fatigue specification tests as part of NCHRP 9-59. REVIEW OF CURRENT BINDER FATIGUE TESTS The paragraphs below briefly describe a number of promising binder tests that have been proposed as fatigue specification tests. This is not intended to be a complete list but provides several examples of potential asphalt binder fatigue tests. In executing the project, the research team intends to thoroughly evaluate a wide range of tests, and select several of the most promising. As many different binder tests will be included in the final work plan as possible given the budget and schedule for NCHRP 9-59. The direct tension test (AASHTO T 314) has been used as an alternative specification test for many years to characterize the low temperature fracture properties of asphalt binders. Although this test was not widely used and is no longer supported by equipment manufacturers, AASHTO and ASTM standards exist for the test, and equipment manufacturers could probably easily gear up for renewed production and technical support for this test if the demand were there. In the 2012 FHWA ALF study, direct tension strain at failure exhibited a reasonably good correlation to laboratory mixture fatigue tests and to fatigue performance in the ALF lanes (Gibson et al., 2012). It is possible that this test could be modified for use in a fatigue specification by specifying the failure strain at the BBR critical temperature. This would in effect control the inherent ductility of the binder, independent of temperature, which might be an effective approach to controlling mixture fatigue performance. Figure A-1 shows an example of the potential effectiveness of this approach to controlling fatigue performance. This is a plot of cycles to 25-meter total crack lengths versus estimated failure strain from the direct tension test at the BBR critical stiffness temperature for the 2012 FHWA ALF study. The plot shows test results for which data was available without significant extrapolation; the R2 values are quite high for both the 100-mm and 150-mm pavement thickness. The NCHRP 9-59 research team originally considered the direct tension test to be a promising one for further evaluation in Phase II of the project. However, discussions with laboratory personnel and panel members familiar with the procedure indicated that it is in fact a very difficult and time-consuming test to perform and has never been widely used among state high departments or commercial laboratories. Furthermore, the direct tension equipment is no longer produced commercially. For these reasons, the SENB test (described below) has been proposed as an alternative to direct tension for evaluation in Phase II of NCHRP 9-59.
112 Figure A-1. Plot of Cycles to 25-Meter Total Crack Length as a Function of Estimated DT Failure Strain at BBR Critical Stiffness Temperature, for 2012 FHWA ALF Study (data as reported by Gibson et al., 2012). The yield energy test is a binder strength test performed using the DSR. In this test, the energy required to cause yielding in an asphalt binder specimen is measured. The test uses 8-mm parallel plates with a 2-mm gap and a strain rate of 1% per second (Johnson et al., 2009). This test correlated very well to the observed fatigue performance of mixes in the 2012 FHWA ALF experiment (Gibson et al., 2012). In the ALF study, the test temperature was selected to be the same as the temperature used in loading the pavements (19Â°C), but the test can probably be performed over a wider range of intermediate temperatures. In addition to the observed correlation to mixture fatigue performance, this test is promising because it uses the DSR and standard test fixtures, and it could be easily implemented. Non-uniform distribution of stresses and strains is a drawback of this test, as well as relatively slow loading, which is important and is a problem with many strength tests. Traffic associated loading is relatively fastâthe loads in a pavement occur in a matter of hundredths of a second; typical mixture modulus measurements are made at 10 Hz in part because this approximates the typical condition of traffic loading. The yield energy test (and many other strength tests) typically involves times to failure of perhaps 30 seconds, which are several orders of magnitude slower than traffic loading. Figure A-2 is a plot of shear stress as a function of shear strain as is used to determine yield energy (Gibson et al., 2012). In recent private conversations with one of the original developers of this procedure (Johnson, January 2015), it was stated that further research has cast doubts upon the effectiveness of this test and those involved with its development now believe that the linear amplitude sweep (LAS) procedure (described below) is more promising as a test for evaluating the fatigue resistance of asphalt binder. Therefore, this test is not considered a good candidate for inclusion in NCHRP 9-59. RÂ² = 83% RÂ² = 88% 0.0E+00 5.0E+06 1.0E+07 1.5E+07 0 100,000 200,000 300,000 1.0 2.0 3.0 Cy cle s t o 25 m C ra ck in g, 1 50 m m Cy cle s t o 25 m C ra ck in g, 1 00 m m Estimated DT Failure Strain (%) at BBR Critical Stiffness Temperature 100 mm 150 mm
113 Figure A-2. Plot of Shear Stress vs. Shear Strain Measured Using DSR to Determine Yield Energy (Gibson et al., 2012). There are a few fatigue-type tests that have recently been developed for evaluating asphalt binders using the DSR. There are at least two versions of this test. The first version is called the time sweep test and involves fatigue testing of an asphalt binder specimen using the parallel plate geometry in the DSR (Martono and Bahia, 2008). Like many fatigue tests, this test can be time consuming. Another potential drawback is that the shape of the specimen edge could deteriorate during such testing, and/or the specimen could partially detach from one or both plates (FHWA Binder ETG, 2012). Furthermore, although similar to the uniaxial testing used to characterize the fatigue response of asphalt mixtures, the stresses and strains in the parallel plate geometry are not constant but vary linearly across the specimen. A rigorous analysis is therefore difficult or impossible. The second version is called the stress sweep test and uses exponentially increasing stresses to reduce the time of testing. Figure A-3 shows a diagram illustrating the difference in the stress sweep and time sweep binder fatigue tests (Gibson, 2012). Another procedure, called the linear amplitude sweep (LAS) test, is similar to the stress sweep but the strain rather than the stress is gradually increased throughout the test. The LAS test has been standardized on a provisional basis in AASHTO TP-101-12-UL.
114 Figure A-3. Schematic of Stress Sweep and Time Sweep Binder Fatigue Tests (Gibson et al., 2012). Figure A-4 compares the LAS test results determined at 2.5% strain on extracted binders according to AASHTO TP 101 and the bending beam fatigue (BBF) test results determined at 400 microstrain and the field cracking performance of five mixes, including control virgin HMA, 50% RAP HMA, 50% RAP WMA, WMA foam, WMA additive, evaluated at the NCAT Pavement Test Track. The comparison shows a good correlation between the binder and mixture fatigue test results and a reasonable correlation between the binder fatigue test results and the field cracking performance of the five mixtures. The correlation between lab and field data can be improved when coupled with structural pavement analysis, as illustrated later in this work plan. Recently, an AASHTO provisional specification for the LAS test has been approved, TP 101-14: Estimating Damage Tolerance of Asphalt Binders Using the Linear Amplitude Sweep. This describes in detail the procedure for performing the LAS test and a standard procedure for estimating SVECD parameters from the resulting test data. However, this is a relatively new test procedure and the procedure recommended by researchers involved in developing the test is frequently revised. The procedure to be used in NCHRP 9-59 will be determined in consultation with researchers at the North Carolina State University and the University of Wisconsin at Madison. Another important finding in recent research on the LAS test, as presented at the April 2014 FHWA Binder Expert Task Group Meeting, is that the modulus range over which the test is valid is limited; at a frequency of 10 rad/s (1.6 Hz), the modulus should be between 0.5 and 35 MPa, while at 62 rad/s (10 Hz), the modulus should fall between 1.5 and 70 MPa. These limits help prevent delamination of the specimen from the rheometer plates at higher modulus values, and excessive bulging and plastic flow at low modulus values (Tabatabaee et al., 2014).
115 (a) LAS @ 2.5% strain vs. BBF @ 400 ms b) LAS @ 2.5% strain vs. %Area Cracked Figure A-4. Comparison of LAS Test Results with Bending Beam Fatigue Results and Field Cracking Performance The single edge notched beam (SENB) test is a fracture test that can be run on a modification of the BBR device (Velasquez et al., 2011). In this test, a standard-sized BBR specimen is notched and then tested at a constant strain rate until failure. Various fracture parameters including failure load, strain at failure deformation, failure energy and fracture toughness can be determined from this test. Velasquez and his associates showed that SENB-BBR test data correlated with both standard BBR test data and cracking temperature determined using the asphalt binder cracking device (ABCD) procedure (Velasquez et al., 2011). This test has not been standardized by any highway agency and has not been widely used by engineers and researchers. It is however soundly based in engineering principles and should provide a good measurement of failure properties at low temperatures, which should correlate to fatigue performance. The NCHRP 9-59 research team believes that a failure test should be included among the procedures evaluated in Phase II of the project, and the only other test that provides such information is the direct tension test, which is a very difficult test to run. The SENB should be relatively easy to implement using a modification of current BBR technology. The double edge-notched tension (DENT) test has recently been used successfully to characterize the fracture properties of both asphalt binders and HMA mixes (Andriescu et al., 2006; Andriescu and Hesp, 2009). In both mixture and binder tests, a rectangular test coupon is prepared with 45Â° notches on both sides at the center of the specimen. The specimen is loaded at a constant deformation rate until failure, and the load and energy to failure are measured. Typically, three sets of specimens are prepared with different ligament lengths (the ligament length is the distance between the notches at the midpoint of the specimen). Total energy to y = -0.9911x + 390117 RÂ² = 0.8542 0 50 100 150 200 250 300 350 400 0 100 200 300 LA S Cy cle s t o Fa ilu re (x 10 00 ) BBF Cycles to Failure (x1000) @ 400 microstrain y = -6698.9x + 287528 RÂ² = 0.5834 0 50 100 150 200 250 300 350 0 5 10 15 20 LA S Cy cle s t o Fa ilu re (x 10 00 ) Cracking (% of Lane Area)
116 failure is plotted as a function of ligament length; the intercept gives essential work of fracture, an indication of the inherent fracture toughness of the material. The slope of the plot gives the plastic work of fracture. The critical crack tip opening displacement (CTOD) is calculated by dividing the essential work of fracture by the maximum stress. Anderson discusses early versions of this test as used on metals to determine the crucial CTOD (Anderson, 1995). He points out that there are several distinct advantages to this procedure, including that it is one of the few fracture test methods that is applicable near the brittle-ductile transition, and that it is a relatively simple test to perform and analyze (Anderson, 1995). This test has been standardized by the Ontario Ministry of Transportation, as test method LS-299, and has also been used by a wide range of engineers and researchers. This test is considered a primary candidate for use in NCHRP 9-59. Figure A-5 is a photograph of the DENT test for asphalt binder, as reported by Gibson et al. (2012). Figure A-5. DENT Test Setup, for Asphalt Binder (Gibson et al., 2012). Although the DENT test as specified in LS-299 and an FHWA standard (FHWA-HRT-11- 045) is in many ways a very promising test for characterizing the fatigue performance potential of asphalt binders, there are two significant shortcomings in this procedure: (1) the complexity and time required in testing six different specimens; and (2) the amount of materials required for preparing these specimens. The large amount of material required for the DENT test in its current form is a serious problem considering that the test will likely be implemented using binder aged using an extended protocol, for example, RTFOT aging followed by 40-hour PAV conditioning. For these reasons, the NCHRP research team is proposing to use a simplified version of the DENT test, using only two specimens at one notch depth/ligament length. Results of testing at the FHWA on ALF binders clearly shows that extension to failure of a single test correlates very highly to both CTOD determined using the full procedure and to field performance of the ALF binders. Table A-1 shows the correlation (r2 values) among different DENT test parameters and FHWA ALF 2 cracking data (Gibson et al., 2012). Included in this
117 data were the results for five binders: PG 70-22 (control); air blown; SBS-LG; CR-TB (crumb rubber); and terpolymer. Included in this table are the following DENT parameters/ALF performance indicators: â DENT extension to failure (mm) at 25Â°C, 80 mm/min, 10 mm ligament â DENT CTOD (mm) at 25Â°C, 80 mm/min, three ligament lengths â DENT CTOD (mm) at 25Â°C, 100 mm/min, three ligament lengths â ALF cycles to first crack â ALF cycles to 25 m cracking The end point for extension to failure was calculated as the extension where the load fell to 20 % of the maximum value. This was done to try to avoid highly variable data when the extension becomes extremely large. The simple extension data shows r2 values of 99 % with both CTOD values, and the r2 values with the ALF cracking data are essentially identical to those for the CTOD data. Figure A-6 is a plot showing extension to failure (10 mm ligament length) and CTOD for the two loading rates. It appears that DENT extension to failure can be used as a simple but very effective surrogate for CTOD. Although this simplification might at first seem somewhat empirical, the NCHRP research team believes that DENT extension to failure is a good indication of crack tip blunting, which is directly related to fracture toughness. Besides providing for a quicker and less expensive test using one-third of the binder, the simplified approach to the DENT test will probably also provide more robust and more repeatable results since the reported parameter (extension at failure) does not rely on the calculation of slope and intercept from multiple tests. Table A-1. Correlation Among DENT Test Parameters and ALF Cracking Data (Gibson et al., 2012). Data Correlation (r2 value, %) DENT Extension to Failure, 10 mm Ligament (mm) DENT CTOD, 80 mm/min (mm) DENT CTOD, 100 mm/min (mm) CTOD, 80 mm/min 99 --- 100 CTOD, 100 mm/min 99 100 --- ALF cycles to surface cracking 97 97 98 ALF cycles to 25 m cracking 92 92 92
118 Figure A-6. Relationship Between Extension to Failure for 10 mm ligament at 80 mm/min and CTOD at 80 mm/min and 100 mm/min, from DENT Testing of Five FHWA ALF2 Binders at 25Â°C (Gibson et al., 2012). In the current Ontario LS-299 standard for the DENT test, testing conditions are given as 15Â°C and 50 mm/min. However, in the Appendix to FHWA-HRT-11-045, a test method for the DENT test gives test conditions of 25Â°C and 100 mm/min. At this time, it is anticipated that DENT testing carried out as part of NCHRP 9-59 will use a loading rate of 50 mm/min, and two test temperatures: 25Â°C and one other temperature that will depend on the consistency of the binder tested. The slower loading rate of 50 mm/min will be used because the FHWA specified rate of 100 mm/min cannot be obtained using a standard ductilometer, which has a maximum loading rate of 50 mm/min. There is no reason to expect that using the 50 mm/min as compared to 100 mm/min will produce results that differ significantly in their correlation to mixture fatigue performance, and the lower rate will make the test much easier to implement since it will then be possible to perform the test on commercially available equipment. The two test temperature selected will allow a comparison of the DENT failure properties with temperature and other binder properties such as modulus, and it will also allow for the estimation of DENT properties at some standard reference condition. Originally it was anticipated that the DENT test would be conducted at three temperatures, but discussions with engineers familiar with the DENT test indicated that this test provides meaningful data over a relatively narrow range of conditions, and testing over a wide range of temperatures would not be productive. Various rheological parameters have been used to characterize binder fatigue susceptibility, the two most important of these being the R-value from the CA model (discussed previously) and the Glover-Rowe parameter. These parameters, along with the Dahlquist criterion (based on a maximum storage modulus of 100 kPa), and possibly other similar parameters not yet identified RÂ² = 0.9879 RÂ² = 0.9895 0 20 40 60 80 0 50 100 150 200 CT OD , m m Extension to Failure, mm 80 mm/min 100 mm/min
119 by the research team will be evaluated as potential specification âtestsâ related to mixture fatigue performance. The values will be calculated from standard DSR testing to be performed on all binders included in the study, and the analysis will be essentially identical to that to be performed on other candidate mixture test data. Because once rheological data is collected using the DSR parameters such as these can easily be calculated for a wide range of conditions, it is likely that many rheological parameters such as the Glover-Rowe parameter will be evaluated as part of NCHRP 9-59. Binder Specifications and Potential Fatigue Tests in Other Countries The primary recent changes in asphalt specifications in Europe have dealt with the issue of âharmonization,â meaning developing a common framework for specifications, if not consistent specific values within those specifications. Current specifications, as put forth in EN 12591, rely on penetration and softening point tests, in an unaged state and after conditioning using the RTFOT (Austroads, 2013). The framework standard for polymer-modified asphalt, EN14023 adds force ductility and elastic recovery to the battery of tests used to specific asphalt binders. Although these tests probably address fatigue and fracture properties, there are currently no tests directly addressing these properties in the European binder specifications. This is perhaps because at least some researchers in Europe, to quote Herve DiBennedetto in a 2013 forum discussion, believe ââ¦that the fatigue characteristics and crack approaches that are used in practice for design are not really good toolsâ¦The physical phenomena describing fatigue and cracking are not well understoodâ¦â (AAPT, 2013). Europe is currently moving towards more performance-based asphalt specifications, under the leadership of CEN Technical Committee 336: Bituminous Binders. At this time, several fracture-based tests are under consideration for inclusion in the next generation of European asphalt specifications. These are the Fraass breaking point test EN 12593), the force ductility test and tension test (EN 13589 and EN 13587), the direct tension test (as once included in the U.S. Superpave specifications), and a relatively new fracture toughness test specification, CEN/TS 15963, which involves a simplified fracture mechanics based test on asphalt cement (Austroads, 2013; Planche, 2013). In this test, a composite beam specimen is formed with end pieces of aluminum and a central section of asphalt binder with a small notch. The specimen is tested in flexure until failure. In its most simple application, the test is used to control low temperature fracture properties, by determining a critical temperature as that at which the composite specimen fails at a deflection of 0.3 mm. More complicated analyses of the test are however possible (Planche, 2013). In Australia, asphalt binders are currently classified according to viscosity at 60Â°C. Other tests including in the specification (AS 2008 â 1997) are penetration at 25Â°C, flashpoint, toluene insoluble, and the hardening ratio of viscosity after RTFOT conditioning. There are no tests related to fracture and/or fatigue properties. Specifications in New Zealand (TNZ M/1: 2007) involve grading the asphalt according to penetration at 25Â°C, but include viscosity at 60Â°C, retained penetration after RTFOT conditioning, and ductility at 25Â°C, also after RTFOT
120 conditioning (Austroads, 2013). Thus, in these two countries, the only fracture/fatigue related test in use in asphalt specification is ductility. Asphalt cement in South Africa is graded on the basis of ductility. Other tests included in the South African specification (SANS 307-2005) include softening point and viscosity at both 60Â°C and 135Â°C. Penetration, softening point and viscosity at 60Â°C are also specified for the RTFOT residue. Ductility was included in South African asphalt specifications until 2002. Currently, no fatigue or fracture related tests are included in the South African specification for asphalt cement. In 2012, an International Forum on asphalt specifications in Latin America was held at the Annual Meeting of the Association of Asphalt Paving Technologists (AAPT). In general, asphalt technology in this region is behind that in the U.S., Canada and Europe. In Chile, asphalt specifications largely involve traditional, empirical tests such as penetration, softening point and ductility. In this case, ductility can be thought of as the fatigue/fracture specification test, much as it was in the U.S. during the 1950s and 1960s. For polymer-modified asphalts, elastic recovery is also used as a specification test which can also be considered to at least partly address fracture and fatigue properties. As in Chile, Columbia uses traditional empirical consistency tests in their specification, including ductility, although there are no separate requirements for polymer- modified binders. The Canadian provinces have all adopted the PG grading system as used in the U.S., but with a variety of significant variations. The following review largely is based on information provided in the Asphalt Institutes Specification Database, available online at www.asphaltinstitute.org/specification-databases/canadian-province-binder-spec-database/. In general, Canadian PG specifications, as should be expected, are geared towards grades with substantially lower critical temperatures, including such grades as PG 52-34, PG 58-34, PG 64- 28 and PG 76-28. Alberta includes requirements for direct tensionâa minimum 1.0 % failure strain at the critical low temperature after PAV aging. New Brunswick and Nova Scotia, on the other hand, have no requirements for direct tension. Quebec also has no requirements for direct tension but does include requirements for elastic recovery at 10Â°C for certain binders, generally ones with extended service temperature ranges that would tend to require polymer modification. Quebec includes requirements for reporting MSCR test data, but currently has no specific limits for this test. Unlike most other provinces, Saskatchewan does not use PG grading but instead still relies on penetration grading for asphalt cements. Saskatchewanâs specification includes requirements for viscosity at 60Â°C and ductility at either 15 or 25Â°C. There are also requirements for retained penetration and viscosity after thin-film oven test (TFOT) aging. For these provinces, fracture/fatigue related asphalt tests are limited to direct tension, elastic recovery and ductility. As mentioned in several other places within this report, Ontario has recently implemented several tests intended to address premature failure of asphalt pavements. These failures manifest in a variety of ways, including raveling and both load and non-load associated cracking. A few reports and papers have suggested that the use of recycled engine oil bottoms (REOBs) have
121 contributed significantly to this problem. The two tests that Ontario has put into their PG grading system to address these failures are the double-edge notched tension (DENT) test, and the extended bending beam rheometer (EBBR) test. The DENT test is described earlier in this report and is a notched-tension test devised to provide information on the fracture toughness of asphalt binders. The extended BBR test is similar to the standard method but involves testing after conditioning at the test temperature for extended periods of timeâ24 and 72 hours. This long conditioning time allows for significant physical hardening to occur in binders susceptible to this phenomenon. Comparison with field data suggests that the EBBR test provides a significantly better correlation to field performance compared to the standard BBR procedure (Marks, 2015). The DENT test has been previously identified as a promising test for evaluating the fatigue and fracture properties of asphalt cement. In Canadian studies, the EBBR test also appears promising, although it is not clear if this test would should similar good correlations to performance in more moderate U.S. climates. Also, the test is obviously time consuming because of the long conditioning time required. An important question is whether the substantial physical hardening observed for some binders in the EBBR test can be related to other, simpler tests. For instance, it is possible that the degree of physical hardening in the EBBR test might be related to the R value in the CA model. This would allow a specification to be devised that would be simpler and quicker, but that would also identify binders prone to excessive physical hardening. In summary, review of asphalt specifications and potential changes in these specifications in other countries does not suggest that there exist recently developed tests that might be effective in ensuring the fatigue performance of asphalt concrete mixtures, with the exception of the DENT test as implemented in Ontario. The EBBR test, also implemented in Ontario is an interesting test and appears to be effective in reducing thermal cracking but does not specifically address fatigue performance, although the results of this test might relate indirectly to fatigue performance. Other tests used in various countries that potentially address fatigue performance are existing tests, including ductility and the direct tension test. RATINGS OF CANDIDATE BINDER FATIGUE TESTS Based upon the literature review presented above, six binder tests/parameters have been identified as candidates for further evaluation: 1. The linear amplitude sweep (LAS) 2. The double-edge notched tension (DENT) test 3. The Glover-Rowe parameter (GRP) and other similar rheological parameters that can be calculated from DSR data and have been related to performance, such as the R-value and ÎTc. 4. The direct tension test 5. The single edge notched bending (SENB) test 6. The ductility test
122 The ductility test has been included in the ratings because of its historical significance, because it is still widely used in other countries and because there is substantial data in the literature linking ductility to asphalt concrete pavement durability and fatigue performance. Five criteria were selected for evaluating the tests: 1. Additional cost. This means additional cost required to run the test. Determining the GRP, R-value, and other parameters, for example, involves no additional cost because pavement testing laboratories should already have the equipment needed to produce the required data. Performing the DENT test would require significant equipment investment. 2. Active time requirement. This is the actual working time required to run the test. In the original criteria and submitted to the panel there was also a criterion for total time requirement, but in the final ratings it was found that these two time requirements were directly related, and so a single criterion is included in this final rating. 3. Correlation with performance. This is based on published reports and papers showing correlation between the proposed test and field performance or laboratory performance tests. 4. Engineering soundness. This is a subjective evaluation of how strongly the test is based on engineering principles, as opposed to being empirical. 5. Technical difficulty and ease of implementation. This is also a subjective evaluation of how difficult the test is to perform and the potential hurdles that would be encountered in implementing the test. Table A-2 is a summary of the resulting ranking of the candidate tests based on these criteria and their weights. These criteria and weights were submitted to the panel; the results are based in part upon the resulting response. Tables A-3 through A-7 explain how points were awarded for the five criteria. It should be emphasized that the rating in Table A-2 is not meant to be the only means for selecting the final tests including for evaluation in the laboratory testing phase of NCHRP 9-59. Instead, this table is meant to provide guidance to the researchers and panel in making the final selection. The highest rated tests were the LAS, GRP and related rheological parameters, and the DENT test. Although the ductility test scored relatively highly in this rating, the research team feels that there would be little support and significant opposition to reconsideration of this procedure, and it represents a significant step backwards from rational, engineering based tests and specifications for paving materials. As mentioned above, it has been included partly for historical reasons and partly because it is still in wide use in other parts of the world. The two remaining testsâdirect tension and SENBâare similar in that they provide information on the strain capacity of asphalt binders at low temperature but using different geometries. The rating of the direct tension test in Table A-2 has changed in this version of the Interim Report, as the score for technical difficulty/ease of implementation has been reduced from 3 to 1. This is largely because of clear indications from laboratory personnel within AAT and from several NCHRP 9-59 panel members that this test is in fact very difficult and time
123 consuming to perform and many people in the industry would react negatively to having it become a specification test. Table A-2. Ratings of Candidate Binder Fatigue Tests. Criteria Weight LAS DENT Glover- Rowe, Etc. Direct Tension SENB Ductility Equipment cost (additional/new) 10 5 3 5 2 4 2 Active time requirement 20 3 1 3 2 3 2 Correlation with performance 50 5 5 5 2 1 4 Engineering soundness 5 3 3 5 5 5 1 Technical difficulty/ease of implementation 15 5 1 5 1 3 1 Total 100 4.5 3.3 4.6 2.0 2.2 2.8 Table A-3. Description of Cost Criteria. Points CostâAdditional, New Equipment or Modifications Only 5 < $500 4 $500 to < $2,000 3 $2,000 to < $8,000 2 $8,000 to < $32,000 1 â¥ $32,000 Table A-4. Description of Total Time Criteria. Points Estimated Active Time to Complete Test, Hours not including conditioning 5 < 1 4 1 to < 2 3 2 to < 4 2 4 to < 8 1 â¥ 8
124 Table A-5. Description of Criteria for Correlation to Performance. Points Description of Existing Documentation Relating Test to Performance 5 Multiple independent studies showing good correlation to both laboratory performance tests and to field performance. 4 Multiple independent studies showing good correlation to laboratory performance tests with one study showing at least moderate correlation to field performance. 3 One study showing correlation to laboratory performance tests and one study showing correlation to field performance. 2 One study showing correlation to either laboratory performance tests or field performance. 1 No studies showing correlation to laboratory performance tests or field performance, but theory suggests that such correlations should exist. Table A-6. Description of Criteria for Engineering Soundness. Points Description of Engineering Basis of Test 5 Test is soundly based in engineering fundamentals. 3 Test has some basis in engineering fundamentals. 1 Test is entirely or almost entirely empirical. Table A-7. Description of Criteria for Technical Difficulty/Ease of Implementation. Points Description of Technical Difficulty/Implementation Issues 5 Test can be easily run by technician with no more than six months experience. Test (or a nearly identical test) is already a standard asphalt binder test. No unusually difficult procedures are involved. Example: DSR variants such as LAS. 3 Test can be easily run by technicians with one-year experience. Test (or a very similar test) is a commonly used standard method in other fields/geographic regions or was a standard/optional asphalt binder test in the past. Some parts of the procedure can be challenging. Example: direct tension test. 1 Test requires technicians with several years of experience. Not a standard procedure. Involves very challenging techniques to perform properly. Example: fracture mechanics-type tests, DENT test. Also, tests for which strong opposition to adoption exists in the pavement technology community, such as ductility.
125 Selection of Final Asphalt Binder Tests The NCHRP 9-59 research team has emphasized from the beginning of the projectâeven during writing the initial proposalâthat the funding and time allotted for the project was not adequate for the development, refinement and validation of a completely new test procedure for evaluating the fatigue performance of asphalt binders. Therefore, our approach has always emphasized identifying and evaluating tests that have already been developed and can be widely implemented quickly and easily. It is also essential that potential tests have shown promise by exhibiting good correlation to mixture fatigue performanceâeither in the laboratory or the field, preferably both. Based upon these considerations and the ratings summarized in Table A-2, the following tests have been selected for detailed evaluation in Phase II of NCHRP 9-59: â The LAS test â The simplified DENT test â Various rheological parameters, including the GRP, loss modulus, storage modulus and phase angle These tests all met the criteria developed by the research team for candidate binder tests to be included in Phase II of NCHRP 9-59: the have gone through initial development; they have been correlated to field performance; and they can be realistically implemented as specification tests. The laboratory testing and data analysis conducted during NCHRP 9-59 was largely based around these tests and was intended to provide a data set that could be used to do a thorough evaluation of the ability of these tests to predict binder fatigue and fracture performance.