EVALUATION OF THE USE OF RECLAIMED ASPHALT PAVEMENT IN STONE MATRIX ASPHALT MIXTURES Except where referenced is made to the work of others, the work described in this thesis is my own or was done in collaboration with my advisory committee. This thesis does not include proprietary or classified information. Adriana Vargas-Nordcbeck Certificate of Approval: David Timm Assistant Professor Civil Engineering Elton Ray Brown, Chair Director National Center for Asphalt Technology Randy West Assistant Director National Center for Asphalt Technology George T. Flowers Interim Dean Graduate School EVALUATION OF THE USE OF RECLAIMED ASPHALT PAVEMENT IN STONE MATRIX ASPHALT MIXTURES Adriana Vargas-Nordcbeck A Thesis Submitted to the Graduate Faculty of Auburn University in Partial Fulfillment of the Requirements for the Degree of Master of Science Auburn University December 17, 2007 iii EVALUATION OF THE USE OF RECLAIMED ASPHALT PAVEMENT IN STONE MATRIX ASPHALT MIXTURES Adriana Vargas-Nordcbeck Permission is granted to Auburn University to make copies of this thesis at its discretion, upon request of individuals or institutions and at their expense. The author reserves all publication rights. Signature of Author Date of Graduation iv VITA Adriana Vargas-Nordcbeck, daughter of Mario Vargas and Shirley Nordcbeck, was born November 26, 1977 in San Jose, Costa Rica. She graduated from the University of Costa Rica with a Bachelor of Science degree in Civil Engineering in August, 2003. She also attended the State University at Distance in San Jose, Costa Rica, and graduated with a Masters Degree in Business Administration in July, 2005. She began her studies as a graduate student at Auburn University in September, 2005. She married Fabricio Leiva- Villacorta on May 19, 2007. v THESIS ABSTRACT EVALUATION OF THE USE OF RECLAIMED ASPHALT PAVEMENT IN STONE MATRIX ASPHALT MIXTURES Adriana Vargas-Nordcbeck Master of Science, December 17, 2007 (MBA, UNED-Costa Rica, 2005) (B.S., University of Costa Rica, 2003) 171 Typed Pages Directed by E. Ray Brown Mixtures that contain reclaimed asphalt pavement (RAP) can typically perform as well or better than conventional HMA mixes. However, use of RAP has generally not been extended to stone matrix asphalt (SMA) production. This study evaluated the effect of RAP on combined aggregate properties, asphalt binder properties, and overall performance of SMA mixtures. The effect of type and size of RAP, as well as aggregate source on overall mix performance was also evaluated. Four types of RAP were combined at four levels (0%, 10%, 20% and 30%) with four aggregate sources. One source of virgin asphalt cement (PG76-22) was used in this study. vi Testing was performed to evaluate LA abrasion and flat and elongated particle content of the virgin and recycled aggregate blends. The effect of RAP addition on the rheological properties and performance grades of the combined binder blends was also evaluated. Finally, testing was performed to determine potential binder effect on resistance to moisture susceptibility, resistance to rutting, thermal cracking potential and fatigue life of the recycled mixtures. Results showed that only fatigue life of the mixes decreased significantly with the addition of RAP, but damage can be minimized by limiting the use of recycled SMA mixes to the top layers of the pavement and ensuring a good bond with the underlying layer. Overall, up to 20% RAP could be used without significantly affecting the performance of the mixes. vii ACKNOWLEDGMENTS The author would like to thank Dr. E. Ray Brown and Donald Watson for all their guidance and support in this endeavor. The author also acknowledges the advisory committee including Dr. David Timm and Dr. Randy West for all of their time and assistance during this project. Thanks are also due to the staff at the National Center for Asphalt Technology for all their assistance. Special thanks are due to her parents Mario and Shirley and to her sister Marcela, for all of their love and support throughout all academic endeavors. Finally to her husband Fabricio, for all of his encouragement and support in every step of the way. viii Style manual used: Proceedings, Association of Asphalt Paving Technologists Computer software used: Microsoft Word, Microsoft Excel, Minitab ix TABLE OF CONTENTS LIST OF TABLES??????????????????????????......xi LIST OF FIGURES??????????????????????????...xv CHAPTER 1. INTRODUCTION?????????????????????....1 1.1 BACKGROUND AND PROBLEM STATEMENT........................................... 1 1.2 OBJECTIVES..................................................................................................... 2 1.3 SCOPE OF STUDY ............................................................................................ 2 CHAPTER 2. LITERATURE REVIEW ............................................................................ 4 2.1 INTRODUCTION .............................................................................................. 4 2.2 REVIEWS........................................................................................................... 5 2.3 SUMMARY...................................................................................................... 36 CHAPTER 3. RESEARCH TEST PLAN ........................................................................ 38 3.1. PART 1 ? EVALUATION OF MATERIALS .................................................. 39 3.1.1 Evaluation of Aggregate Properties.............................................................. 39 3.1.2 Evaluation of Asphalt Binder Properties ...................................................... 47 3.2 PART 2 ? MIX DESIGNS ............................................................................... 53 3.3 PART 3 ? PERFORMANCE TESTS................................................................ 59 3.3.1 Moisture Susceptibility ................................................................................. 59 3.3.2 Rutting Susceptibility.................................................................................... 60 x 3.3.3 Creep Compliance....................................................................................... 61 3.3.4 Flexural Beam Fatigue................................................................................ 65 CHAPTER 4. TEST RESULTS AND ANALYSIS......................................................... 67 4.1 MATERIAL PROPERTIES ............................................................................. 67 4.1.1 Aggregates .................................................................................................... 67 4.1.2 Asphalt Binder .............................................................................................. 69 4.2 MIX DESIGNS ................................................................................................ 83 4.3 PERFORMANCE TESTS ................................................................................ 87 4.3.1 Moisture Susceptibility ................................................................................. 87 4.3.2 Rutting Susceptibility.................................................................................... 96 4.3.3 Indirect Tensile Creep Compliance ............................................................ 100 4.3.4 Flexural Beam Fatigue................................................................................ 107 4.3.5 Summary..................................................................................................... 119 CHAPTER 5. CONCLUSIONS AND RECOMMENDATIONS.................................. 125 CHAPTER 6. REFERENCES ........................................................................................ 128 APPENDIX A................................................................................................................. 133 APPENDIX B ................................................................................................................. 148 xi LIST OF TABLES Table 2.1. Rutting of Fine SMA vs. Standard Hot Mix: I-85 Test Section (2)?????5 Table 2.2. Friction Values for Fine SMA Mix: I-85 Test Section (2). ............................... 6 Table 2.3. Procedures for the Design of Mixtures Containing RAP (9)........................... 20 Table 2.4. Mix Design Parameters (14)............................................................................ 25 Table 2.5. Strength, Stiffness and Critical Pavement Temperature of the Mixes (16).... 27 Table 2.6. Change of IDT Properties for Long-Term Aged Mixture (17)........................ 29 Table 2.7. Summary Results of Fatigue Constants and Allowable Number of Loads (18). ................................................................................................................................... 34 Table 2.8. Analysis of Moisture Damage Potential (18). ................................................. 34 Table 3.1. Test Matrix for Mix Variables......................................................................... 39 Table 3.2. Properties of Virgin Aggregates. ..................................................................... 40 Table 3.3. Gradations for Mt. View Aggregates............................................................... 40 Table 3.4. Gradations for Lithia Springs Aggregates. ...................................................... 41 Table 3.5. Gradations for Camak Aggregates................................................................... 41 Table 3.6. Gradations for Ruby Aggregates. .................................................................... 41 Table 3.7. RAP Gradations. .............................................................................................. 45 Table 3.8. Asphalt Contents in RAP................................................................................. 46 Table 3.9. Gradations for Control Mixes.......................................................................... 54 Table 3.10. Gradations of Recycled SMA Mix Using DG1 RAP. ................................... 54 xii Table 3.11. Gradations of Recycled SMA Mix Using DG2 RAP. ................................... 55 Table 3.12. Gradations of Recycled SMA Mix Using -4 RAP......................................... 55 Table 3.13. Gradations of Recycled SMA Mix Using +4 RAP........................................ 55 Table 4.1. Aggregate Properties for Combined Blends. ................................................... 68 Table 4.2. Analysis of Variance for Aggregate Properties. .............................................. 68 Table 4.3. Aggregate Properties for RAP Material........................................................... 68 Table 4.4. Average Results for Aggregate Sources. ......................................................... 69 Table 4.5. Average Results for RAP Contents. ................................................................ 69 Table 4.6. Critical Temperatures and Performance Grades of Virgin and Recovered RAP Binders. ..................................................................................................................... 70 Table 4.7. Measured Binder Properties of +4 RAP Blends. ............................................. 71 Table 4.8. Measured Binder Properties of -4 RAP Blends............................................... 71 Table 4.9. Measured Binder Properties of DG1 RAP Blends........................................... 72 Table 4.10. Measured Binder Properties of DG2 RAP Blends......................................... 72 Table 4.11. Performance Grades of RAP Blends. ............................................................ 76 Table 4.12. Rate of Increase in G*/sin? for Unaged Blends. ........................................... 78 Table 4.13. Rate of Increase in G*/sin? for RTFO-Aged Blends..................................... 79 Table 4.14. Rate of Increase in G*sin? for RTFO and PAV-Aged Blends...................... 80 Table 4.15. Rate of Increase in Creep Stiffness................................................................ 82 Table 4.16. Rate of Decrease in Creep Rate..................................................................... 83 Table 4.17. Volumetric Properties of RAP Mixtures. ...................................................... 84 Table 4.18. Virgin and RAP Binder Contents for SMA Mixes........................................ 85 Table 4.19. Savings in Virgin Binder (%) for RAP Mixtures. ......................................... 87 xiii Table 4.20. Tensile Strengths for SMA Mixtures............................................................. 88 Table 4.21. Analysis of Variance for Tensile Strengths. .................................................. 89 Table 4.22. Tensile Strength Comparisons for SMA Mixes with Various RAP Contents. ................................................................................................................................... 90 Table 4.23. Tensile Strength Comparisons for SMA Mixes with Various RAP Types. .. 91 Table 4.24. Unconditioned Tensile Strengths Comparisons for Aggregate Source ? RAP Type Interaction........................................................................................................ 92 Table 4.25. Conditioned Tensile Strengths Comparisons for Aggregate Source ? RAP Type Interaction........................................................................................................ 92 Table 4.26. Moisture Susceptibility Results for Control Mixes. ...................................... 93 Table 4.27. Moisture Susceptibility Results for SMA Mixes Using +4 RAP. ................. 93 Table 4.28. Moisture Susceptibility Results for SMA Mises Using -4 RAP.................... 94 Table 4.29. Moisture Susceptibility Results for SMA Mixes Using DG1 RAP............... 94 Table 4.30. Moisture Susceptibility Results for SMA Mixes Using DG2 RAP............... 94 Table 4.31. Analysis of Variance for TSR........................................................................ 95 Table 4.32. Average TSR Values for Various RAP Contents. ......................................... 95 Table 4.33. Tensile Strength Ratios Comparisons for RAP Contents. ............................. 95 Table 4.34. Rutting Susceptibility Results for RAP Mixtures.......................................... 96 Table 4.35. Analysis of Variance for Rut Depths............................................................. 97 Table 4.36. Average Rut Depths for Various RAP Contents. .......................................... 97 Table 4.37. Differences in Rut Depth for Aggregate Sources. ......................................... 97 Table 4.38. Rut Depth Comparisons for RAP Types........................................................ 98 Table 4.39. Rut Depths for RAP Content ? RAP Type Interaction................................ 100 xiv Table 4.40. m-values for SMA Mixes. ........................................................................... 104 Table 4.41. Analysis of Variance for m-value................................................................ 105 Table 4.42. m-value Comparisons for Aggregate Sources. ............................................ 106 Table 4.43. m-value Comparisons for RAP Types......................................................... 107 Table 4.44. Test Results for High Strain Beams (800 ??).............................................. 108 Table 4.45. N f Comparisons for RAP Contents (800??). ............................................... 109 Table 4.46. Effect of RAP Binder on Fatigue Life (800 ??). ........................................ 109 Table 4.47. N f Comparisons for Aggregate Sources (800??)......................................... 110 Table 4.48. N f Comparisons for RAP Content ? Aggregate Source Interaction ............ 111 Table 4.49. Initial Stiffness Comparisons for Recycled SMA with Various RAP Contents (800??).................................................................................................................... 112 Table 4.50. Initial Dissipated Energy Comparisons for RAP Contents (800??)............ 113 Table 4.51. Average Test Results for Low Strain Beams............................................... 114 Table 4.52. N f and Percent Drop Comparisons for RAP Contents (400 ??). ................. 115 Table 4.53. Initial Stiffness Comparisons for Aggregate Sources (400 ??). .................. 117 Table 4.54. Initial Stiffness Comparisons for RAP Content ? Aggregate Source Interaction (400 ??). ............................................................................................... 118 Table 4.55. Fatigue Life Comparisons for Strain Levels................................................ 119 xv LIST OF FIGURES Figure 2.1. Distribution of Rut Measurements on SMA Pavements (3)???????..7 Figure 2.2. Los Angeles abrasion loss versus change in percent passing 4.75 mm sieve (5).............................................................................................................................. 10 Figure 2.3. F/E particle content versus change in percent passing 4.75 mm sieve (5). .... 11 Figure 2.4. Percent passing 4.75 mm sieve versus VMA (5). .......................................... 12 Figure 2.5. Percent passing 4.75 mm sieve versus VCA (5). ........................................... 13 Figure 2.6. Example of Superpave high temperature sweep blending charts (8). ............ 17 Figure 2.7. Load Vs. Number of Cycles to Failure in SCB Fatigue Test for PG64-22 Mixtures (17). ........................................................................................................... 30 Figure 2.8. Number of Cycles to Failure in Strain Controlled Beam Fatigue Test (17)..31 Figure 2.9. Effect of Test Temperature on Indirect Tensile Strength (18). ...................... 33 Figure 3.1. Gradations for Mt. View Aggregates. ............................................................ 42 Figure 3.2. Gradations for Lithia Springs Aggregates...................................................... 42 Figure 3.3. Gradations for Camak Aggregates. ................................................................ 43 Figure 3.4. Gradations for Ruby Aggregates.................................................................... 43 Figure 3.5. RAP Gradations.............................................................................................. 46 Figure 3.6. Rolling Thin Film Oven (20).......................................................................... 48 Figure 3.7. Pressure Aging Vessel (20). ........................................................................... 49 Figure 3.8. Basics of Dynamic Shear Rheometer (20). .................................................... 50 xvi Figure 3.9. Schematic of Bending Beam Rheometer (20)................................................ 52 Figure 3.10. Superpave Binder Specification Example (20). ........................................... 53 Figure 3.11. Gradations for Control Mixes....................................................................... 56 Figure 3.12. Gradations of Recycled SMA Mix Using DG1 RAP................................... 56 Figure 3.13. Gradations of Recycled SMA Mix Using DG2 RAP................................... 57 Figure 3.14. Gradations of Recycled SMA Mix Using -4 RAP. ...................................... 57 Figure 3.15. Gradations of Recycled SMA Mix Using +4 RAP. ..................................... 58 Figure 3.16. Indirect Tension Test Creep Compliance Curves (23)................................. 63 Figure 3.17. Prony Series Fit to Master Creep Compliance Curve (23).......................... 64 Figure 3.18. Determination of m, the Slope of the Log Creep Compliance Curve (23)... 65 Figure 4.1. Critical High Temperatures for Binder Blends. ............................................. 74 Figure 4.2. Critical Intermediate Temperatures for Binder Blends. ................................. 74 Figure 4.3. Critical Low Temperatures for Binder Blends. .............................................. 75 Figure 4.4. G*/sin? Trends for Unaged RAP Blends. ...................................................... 77 Figure 4.5. G*/sin? Trends for RTFO-Aged RAP Blends................................................ 79 Figure 4.6. G*sin? Trends for RTFO+PAV-Aged RAP Blends. ..................................... 80 Figure 4.7. Creep Stiffness Trends for RAP Blends......................................................... 81 Figure 4.8. Creep Rate Trends for RAP Blends................................................................ 83 Figure 4.9. Old to New Asphalt Ratio vs RAP Content. .................................................. 85 Figure 4.10. Asphalt Contents for SMA Mixtures............................................................ 87 Figure 4.11. Strength Values from Moisture Susceptibility Test. .................................... 89 Figure 4.12. Effect of RAP Percentage on Tensile Strength. ........................................... 90 Figure 4.13. Strength Values for RAP Types. .................................................................. 91 xvii Figure 4.14. Effect of Aggregate Source on Rut Depth.................................................... 98 Figure 4.15. Average Rut Depths for Various RAP Types............................................... 99 Figure 4.16. Effect of RAP Content on Creep Compliance for Recycled Mixes Using +4 RAP......................................................................................................................... 101 Figure 4.17. Effect of RAP Content on Creep Compliance for Recycled Mixes Using -4 RAP ........................................................................................................................ 101 Figure 4.18. Effect of RAP Content on Creep Compliance for Recycled Mixes Using DG1 RAP................................................................................................................ 102 Figure 4.19. Effect of RAP Content on Creep Compliance for Recycled Mixes Using DG2 RAP................................................................................................................ 102 Figure 4.20. Creep Compliance Master Curves for RAP Mixtures................................ 103 Figure 4.21. Average m-values for RAP Mixtures......................................................... 105 Figure 4.22. Effect of Aggregate Source on m-value. .................................................... 106 Figure 4.23. Average m-values for RAP Types.............................................................. 107 Figure 4.24. Number of Cycles to Failure for RAP Mixtures (800??)........................... 109 Figure 4.25. Effect of Aggregate Source on N f (800??)................................................. 110 Figure 4.26. Relationship between Initial Stiffness and Number of Cycles to Failure (800 ??). .......................................................................................................................... 112 Figure 4.27. Relationship between Drop in Initial Stiffness at 1,000,000 Cycles and Estimated N f (400 ??)............................................................................................. 115 Figure 4.28. Relationship between Initial Stiffness and N f (400 ??).............................. 116 Figure 4.29. Effect of Aggregate Source on Initial Stiffness (400 ??)........................... 117 Figure 4.30. Number of Cycles to Failure for High and Low Strain Levels. ................. 119 1 CHAPTER 1. INTRODUCTION 1.1 BACKGROUND AND PROBLEM STATEMENT Economic and environmental considerations have prompted the use of reclaimed asphalt pavement (RAP) in new asphalt mixes. Asphalt pavement is the most recycled product in the United States, both in terms of tonnage (73 million tons, more than any other material) and in terms of percentage (80 percent of reclaimed asphalt pavement is recycled, a higher percentage than any other substance) (1). RAP is used HMA pavement that has been milled up or crushed. It can be used as a constituent in new mixtures, with characteristics similar to those of virgin HMA mixtures. Benefits of using recycled HMA include lower costs, reduced waste and conservation of natural resources. Although RAP has been successfully incorporated in HMA applications, its use by many agencies has not been extended to the production of open-graded friction courses and stone matrix asphalt (SMA) mixtures. When SMA technology was first implemented in the United States in 1991, there was no experience with the use of RAP in this specialty mixture. Its effect on special requirements for SMA mixes, such as more cubical aggregate and use of polymer-modified asphalt and fiber stabilizers was uncertain, and therefore, its use in SMA mixtures has generally not been allowed. Based on the success obtained with the incorporation of RAP in conventional mixtures, the use of RAP in SMA mixtures needed to be evaluated. This research 2 evaluated the effect of RAP on aggregate, asphalt binder and combined mixture properties. 1.2 OBJECTIVES The objectives of this study were to: 1) Evaluate the effect of various RAP contents and sources on combined aggregate properties such as toughness/abrasion, and flat and elongated particles. 2) Evaluate the effect of RAP on asphalt binder properties such as dynamic shear and fatigue. 3) Determine the feasibility of using SMA mixtures as future RAP sources. 4) Evaluate the performance of SMA mixtures containing fractionated RAP and the potential economical benefits of using this type of material. 5) Evaluate the effect of various RAP sources of different gradation, asphalt content and aggregate properties on the overall performance of SMA mixtures. 1.3 SCOPE OF STUDY To accomplish the aforementioned objectives, this study started with a literature search and review of the information pertaining to the design of SMA mixtures and mixtures containing RAP and their performance. Based on the results of the literature study, a research plan was developed involving extensive laboratory testing, which included performing mix designs for different aggregate sources, RAP types and RAP proportions. For each blend, aggregate properties were determined, as well as optimum asphalt 3 content and volumetric properties. Performance tests were conducted to evaluate the mixtures at different RAP levels. A blend with no RAP was used as a baseline for the study for comparisons of mix performance. 4 CHAPTER 2. LITERATURE REVIEW 2.1 INTRODUCTION Several projects have studied the use of SMA mixes in the United States. Results have shown that the same benefits found in European mixes can be obtained with local materials and procedures. However, many specifications for material properties, gradation and volumetrics needed to be modified, and in some cases new requirements were developed. Use of RAP in HMA applications has also been widely investigated. Inclusion of RAP in HMA mixes has been shown to have not only economic and environmental benefits, but also in some cases it has improved performance. Combining RAP with virgin materials may affect mixture properties, and therefore it has been necessary to develop guidelines for the design of mixtures containing RAP. Performance tests conducted on recycled mixtures indicate that for the most part, they have been found to perform as well as virgin mixtures if properly accounted for in the mix design. 5 2.2 REVIEWS 1. Summary of Georgia?s Experience with Stone Matrix Asphalt Mixes by GDOT (2) This report summarizes the results of various research projects conducted by GDOT to assess the viability of using SMA mixes on the Georgia road system. Research Project No. 9102 evaluated the performance of SMA asphalt under stresses of heavy truck loadings and compared it to the performance of conventional GDOT mixes. Research Project No. 9202 evaluated the use of SMA as an overlay for Portland cement concrete (PCC) pavements. Both projects used the 50-blow Marshall Mix Design procedure, which is used in the design of European SMA. In Project No. 9102, coarse and fine SMA mixes were designed for use as intermediate layers and wearing courses, respectively. The mixes were placed in a 2.5- mile, high traffic volume test section on Interstate 85 northeast of Atlanta. Following the construction of the I-85 test section in 1991, rutting measurements were conducted between 1993 and 1995 to monitor rutting in the fine SMA and conventional mixes. The results indicated that SMA mixes exhibited significantly less rutting than conventional mixes (Table 2.1). Table 2.1. Rutting of Fine SMA vs. Standard Hot Mix: I-85 Test Section (2). Year SMA (mm) Standard (mm) 1993 0 3.0 1994 2.3 5.3 1995 2.5 6.8 6 The test section on I-85 was also used to monitor the friction provided by SMA mixes. Results indicated that the thick asphalt film in SMA mixes did not affect frictional properties, since the thicker film wears quickly at the surface (Table 2.2). Table 2.2. Friction Values for Fine SMA Mix: I-85 Test Section (2). Friction Number Date Number 11/91 42 2/92 50 1/96 50 A mix optimization research study was conducted in a joint study with Georgia Tech to learn more about methods of enhancing SMA performance. Findings from this study showed that GDOT fine SMA mixes undergo at least 30% to 40% less rutting than a typical GDOT dense-graded surface mix, and these fine SMA mixes typically have a fatigue life of 3 to 5 times that of a conventional surface mix. The study also indicated that by relaxing the aggregate quality requirements for SMA mixes important production cost savings could be realized without significantly reducing the performance of the mixes. In Europe, aggregate quality requirements for SMA mixes are typically very rigorous. Based on this research, GDOT implemented use of aggregates which have no more than 45% abrasion loss and which have no more than 20% flat and elongated particles when measured at the 3:1 ratio. Based on the combination of GDOT and European experience, SMA has proven to have the following intrinsic benefits: ? 30-40% less rutting than standard mixes ? 3 to 5 times greater fatigue life in laboratory experiments ? 30-40% longer service life ? Lower annualized cost 2. Performance of Stone Matrix Asphalt (SMA) Mixtures in the United States by Brown et al. (3) This report provides a summary of mix design and performance data obtained between 1994 and 1996 from 86 SMA projects involving a total of 140 test sections in 19 different states. All mixtures were designed using the 50-blow Marshall procedure and used a stabilizer (or special asphalt binder) to prevent draindown of the asphalt cement. In most cases, a fiber (cellulose or mineral) or a polymer was used as the stabilizer. The various SMA mixtures were inspected to determine performance. The study indicated that over 90% of the projects had rutting measurements less than 4 mm. Approximately 25% of the projects had no measurable rutting (Figure 2.1). Figure 2.1. Distribution of Rut Measurements on SMA Pavements (3). Cracking (thermal and reflective) did not represent a significant problem. SMA mixtures appeared to be more resistant to cracking than dense mixtures, most likely due 7 8 to the relatively high asphalt content and its resulting high film thickness. There was no evidence of raveling, and the biggest performance problem was the occurrence of fat spots, which is caused by segregation, draindown, high asphalt content or improper type or amount of stabilizer. The study concluded that SMA mixtures provided good performance in high traffic volume areas and that the increased benefits should compensate for the extra cost of construction. 3. Updated Review of Stone Matrix Asphalt and Superpave Projects by Watson (4). A second study (4) was conducted in September 2001 to evaluate long-term performance on some of the same SMA projects studied by Brown et al. The survey found that SMA mixtures had given exceptional rut-resistant performance, even when placed on high- traffic volume routes. Only one out of the 11 projects visited exhibited rutting in excess of 6 mm. Only one project had significant block-type cracking, believed to be caused by the stiff binder. The biggest long-term performance problem was transverse, reflective cracking. However, this problem appeared to be related to the use of SMA as a thin-lift overlay of PCC pavements. Comparisons between SMA mixes and conventional sections indicated that SMA mixtures may significantly reduce the rate of crack propagation when used as an overlay for concrete pavements. The fat spots, noted as the major performance problem in the original study (2) had been worn off by traffic over time and were not noticeable during the 2001 review. In 9 general, several projects were still in excellent condition after 9 years of service and based on an overall project condition rating, SMA mixes can be expected to last up to 25% longer than conventional mixes. 4. Development of a Mixture Design Procedure for Stone Matrix Asphalt (SMA) by Brown et al. (5). This study developed a mixture design procedure for SMA and evaluated material and mixture criteria for these mixes. Data were collected from a laboratory study conducted with various types of aggregates, fillers, asphalt binders and stabilizing additives. Parameters evaluated included aggregate toughness, flat and elongated particles, aggregate gradation, volumetric mix properties, asphalt binder content, compactive effort and asphalt binder draindown. Results indicated that there was a good correlation between aggregate breakdown and aggregate toughness as measured by the Los Angeles abrasion test for both Marshall (R 2 = 0.62) and SGC (R 2 = 0.84) compaction, as seen in Figure 2.2. To evaluate the effect of flat and elongated particles, mixtures were prepared with 0%, 25%, 50%, 75% and 100% flat and elongated aggregate. Samples were compacted with 50 blows of the Marshall hammer and aggregate breakdown was measured. Figure 2.3 shows that increased F/E particle content increases aggregate breakdown (R 2 =0.89). 50 blow Marshall Compaction Compaction with 100 revolutions of SGC Figure 2.2. Los Angeles abrasion loss versus change in percent passing 4.75 mm sieve (5). 10 Figure 2.3. F/E particle content versus change in percent passing 4.75 mm sieve (5). Increased aggregate breakdown resulted in lower VMA. High Los Angeles abrasion values (40% or higher) make meeting the VMA requirements and ensuring a reasonable high asphalt content more difficult. Figure 2.4 shows the change in VMA with change in percent passing the 4.75 mm sieve. As the percent passing the 4.75 mm sieve decreases, the VMA remains nearly constant, and then begins to increase once the percent passing the 4.75 mm sieve reaches 30-40 percent. The point at which the VMA begins to increase defines the condition at which stone-on-stone contact begins to develop. To ensure the formation of stone-on-stone contact, the percent passing the 4.75 mm sieve should be kept below 30 percent. 11 The presence of an adequate aggregate skeleton can also be verified by measuring the voids in the coarse aggregate (VCA) of the mix. Figure 2.5 shows that as the percent passing the 4.75 mm sieve decreases, the VCA of the mix also decreases. At approximately 30 percent passing the 4.75 mm sieve, the slope of the curve begins to decrease slightly, setting the point at which stone-on-stone contact begins to develop. The design air void range should be kept between 3 and 4 percent. To minimize fat spots and rutting, the air voids in warmer climates should be designed closer to 4 percent. Also, use of polymer modified asphalt produced better rut resistant mixes, while fiber stabilizers were superior in preventing draindown. A combination of stabilizers may provide the best properties in SMA mixes. Figure 2.4. Percent passing 4.75 mm sieve versus VMA (5). 12 Figure 2.5. Percent passing 4.75 mm sieve versus VCA (5). 5. NCHRP Project 9-12 (6). Research for Project 9-12, Incorporation of Reclaimed Asphalt Pavement in the Superpave System, was conducted in three separate, but related, studies: Black Rock Study The objective of this study was to determine whether RAP acts like a black rock or whether some blending occurs between the old and new binders. Three cases simulating possible interactions between the old and new binders were studied to investigate the behavior of RAP blends. Black Rock (BR) samples were made using virgin and recovered RAP aggregate with virgin binder (no RAP binder). Actual Practice (AP) samples were made using virgin binder and aggregate, mixed with RAP with its binder film intact. Total Blending (TB) samples were made using virgin and recovered RAP 13 14 aggregate. RAP binder was recovered, then blended with virgin binder in the specified percentages before mixing. All the samples were prepared on the basis of an equal volume of total binder. Three different RAPs, two different virgin binders, and two RAP contents (10 and 40 percent) were investigated in this primary phase of the project. The different cases of blending were evaluated through the use of various Superpave shear tests at high temperatures and of the indirect tensile creep and strength tests at low temperatures. Results indicate that even though there is no significant difference at low RAP contents, RAP does not act like a black rock, and blending of the old and new binders occurs to a significant extent. This means that at high RAP contents the hardened RAP binder must be accounted for in the virgin binder selection. Binder Effects Study This study investigated the effects of RAP content and stiffness on the blended binder properties. The same three RAPs and two virgin binders were evaluated in this phase of the project at RAP binder contents of 0, 10, 20, 40, and 100 percent. The blended binders were tested according to the AASHTO MP1 binder tests. The response variables for the experiment were the individual test results and critical temperatures determined at high and intermediate temperatures from the Dynamic Shear Rheometer (DSR) tests and at low temperatures from the BBR tests. The specific parameters studied were complex shear modulus (G*) and phase angle (?) from the DSR and stiffness and m-value from the BBR. It was found that at low RAP contents, the effects of the RAP binder are negligible. At intermediate RAP contents, these effects can be compensated for by using 15 a virgin binder that is one grade softer on both the high- and low- temperature grades. Higher RAP contents require the use of blending charts to determine the appropriate virgin binder grade. Mixture Effects Study This study investigated the effects of RAP on total mixture properties. Shear tests and indirect tensile tests were conducted to assess the effects of RAP on mixture stiffness at high, intermediate, and low temperatures. Beam fatigue testing was also conducted at intermediate temperatures. RAP contents of 0, 10, 20, and 40 percent were evaluated. The tests indicated that high RAP contents increase the mixture stiffness, and therefore, a softer virgin binder must be used to improve the fatigue and low-temperature cracking resistance of the mixture. 6. Laboratory Investigation of Mixing Hot-Mix Asphalt with Reclaimed Asphalt Pavement by Huang et al. (7). This study analyzed the blending process of RAP with virgin mixture. One type of screened RAP consisting only of ?No. 4 particles was blended with virgin coarse aggregate at different percentages, and binder rheological tests were performed to characterize properties of binders at different layers of aggregate particle coating. An extreme case was evaluated by mixing RAP and virgin aggregates without any new asphalt binder. The objective was to find out to what extent the aged asphalt will ?get away? from the RAP particles under pure mechanical mixing. Results indicated that only a small proportion of the aged binder would be available to blend with the virgin binder. 16 A blended mixture containing 20% RAP and PG 64-22 binder was used to simulate actual plant mixing. The mixture was subjected to staged extraction and recovery by soaking it in trichloroethylene solution for 3 minutes and then decanting the solution, repeating the process several times. This process allowed the formation of different layers of asphalt around the RAP particles. Results showed that the influence of RAP on the virgin binder was very limited. Only a small portion of RAP asphalt participated in the remixing process; other portions formed a stiff coating around RAP aggregates, and RAP acted as a ?black rock?. 7. Designing Recycled Hot-Mix Asphalt Mixtures Using Superpave Technology by Kandhal and Foo (8). This project developed a procedure for selecting the performance grade (PG) of virgin asphalt binder in a recycled HMA mixture based on the Superpave PG grading system. Blending charts were constructed and evaluated based on test parameters obtained from the dynamic shear rheometer (DSR) and therefore, only high and intermediate test temperatures were considered. Two blending charts were used to determine the high temperature value of recycled asphalt binder. The first high temperature sweep blending chart determined the temperature at which G*/sin? of the unconditioned recycled asphalt binder is 1.0 kPa. The second high temperature sweep blending chart determined the temperature at which G*/sin? of RTFO residue of the recycled asphalt binder is 2.2 kPa. The high temperature value of the recycled asphalt binder is defined as the lower temperature value given by these two high temperature sweep blending charts. The intermediate temperature sweep blending chart determined the temperature at which G*sin? of RTFO+PAV residue of the recycled asphalt binder is 5 MPa. These charts indicated a linear relationship between the logarithm of binder shear stiffness (expressed as G*/sin?) and percent of virgin asphalt in a virgin and RAP binder blend, as shown in Figure 2.6. Figure 2.6. Example of Superpave high temperature sweep blending charts (8). 17 18 The following recommendations were made for proper selection of PG asphalt binder: ? High temperature value of the recycled asphalt binder performance grade can be determined by using only one high temperature sweep blending chart. High temperature sweep blending chart ?G*/sin ?= 1.0 kPa? is recommended over high temperature sweep blending chart ?G*/sin ?=2.2 kPa? because it does not require running the RTFO test. ? Although the intermediate temperature sweep blending chart ?G*sin?=5 MPa? was expected to determine the maximum amount of RAP, it allowed unusually high percentages of RAP, which are inconsistent with the field experience with recycled HMA. Use of the intermediate temperature sweep blending chart is not recommended at the present time. ? A three-tier system of selecting the PG grade of the virgin asphalt binder was recommended for recycled mixes: Tier 1: If the amount of RAP in the HMA mix is equal to or less than 15%, the selected PG grade of the virgin asphalt binder should be the same as the Superpave specified PG grade. Tier 2: If the amount of RAP in the HMA mix is more than 15% but equal to or less than 25%, the selected PG grade of the virgin asphalt binder should be one grade below (both high and low temperature grade) the Superpave specified PG grade. The use of a specific grade blending chart to select the high temperature grade of the virgin asphalt binder is optional. 19 Tier 3: If the amount of RAP in the HMA mix is more than 25%, use the specific grade blending chart to select the high temperature grade of the virgin asphalt binder. The low temperature grade should be at least one grade lower than the binder grade specified by Superpave. 8. Guidelines for the Design of Superpave Mixtures Containing Reclaimed Asphalt Pavement (RAP) by Bukowski (9). This guideline was developed by the FHWA Superpave Mixtures Expert Task Group and outlines the proper means for incorporating RAP in Superpave mixtures. It suggests that aggregate and asphalt binder in the RAP should be considered as part of the aggregate and asphalt binder contents of the total mix, respectively. Also, all aggregate requirements for the aggregate blend must be satisfied. Asphalt binder grade must be adjusted depending upon the amount of RAP included in the mixture, according to the following three categories: Tier 1: Up to 15% RAP by weight of total mixture Tier 2: 16% to 25% RAP by weight of total mixture Tier 3: Above 25% RAP by weight of total mixture Tier 1 does not require any modification of the mix design process, and the selection of the grade of virgin asphalt binder is based on typical requirements for climatic conditions and predicted traffic. Determination of asphalt binder content in RAP is left to the discretion of the agency. Tier 2 requires determination of the asphalt binder content in the RAP. For Tier 3, the grade of virgin asphalt binder is either set to one grade lower than that usually selected for given climatic conditions, or selected from a blending 20 chart. Table 2.3 summarizes the tests required on the RAP and selection of asphalt binder grade. Table 2.3. Procedures for the Design of Mixtures Containing RAP (9). Tier Determine RAP AC Content Measure RAP Gradation Measure RAP AC Stiffness Measure Agg Blend Properties PG Grade Change 1 (a) x no x none 2 x x no (b) x one grade lower (c) 3 x x yes x use blending chart (a) At the discretion of the agency (b) Unless blending chart is used (c) Or use blending chart 9. Effect of Reclaimed Asphalt Pavement on Binder Properties Using the Superpave System by Kennedy et al. (10). In this study, rheological properties were measured for different combinations and percentages of aged asphalts and virgin asphalts. It includes test results from Superpave binder tests conducted on unaged binders at the high-temperature range, as well as test results on blends aged using the rolling thin film oven test (RTFOT) and pressure aging vessel (PAV) conducted at high-, low-, and intermediate-temperature ranges. Six asphalts were chosen from the Material Reference Library (MRL) for this experiment. Two of the binders were chosen arbitrarily to be aged to simulate RAP binder and then combined with the four virgin binders at different percentages (0, 15, 25, 55 and 100%). Engineering characteristics of the virgin-RAP blends were determined with the aid of a dynamic shear rheometer (DSR) and a bending beam rheometer (BBR). 21 The result of this study is a methodology for determining the effect of RAP on rheological properties of PG binders in the Superpave system. Specific conclusions drawn from this study include: ? The stiffness of the binder is higher at higher percentages of RAP binder. ? The rate of change of stiffness (G*/sin ?, G*sin ?, or creep stiffness) is either constant from 0?100% RAP binder or increases with lower temperatures. ? The rate of change of stiffness is either constant from 0?100% RAP or increases at higher percentages of RAP in the blend. 10. Determination of Recycled Asphalt Pavement (RAP) Content in Asphalt Mixes Based on Expected Mixture Durability by Abdulshafi et al. (11). This study developed a method to evaluate the effects of RAP content on long-term durability of a bituminous concrete mixture, which may be used to select an optimum RAP content. The procedure includes preparation of test specimens at different levels of RAP addition. Each set of specimens is divided into two subsets to be tested for indirect tensile strength; one is tested in dry condition and the other is subjected to vacuum saturation, followed by a freeze cycle and warm water soaking prior to testing. During the testing, load and deformation data are continuously collected, and the resultant energy needed to fail a specimen is calculated. Numerical indices of absorbed energy are computed from the test data obtained for the dry and conditioned subsets of specimens. The mix that has the greatest index of absorbed energy is selected as having the optimum RAP content. 22 11. Recycled Asphalt Pavement (RAP) Effects on Binder and Mixture Quality by Li et al. (12). This study investigated the effect of various types and percentages of RAP on asphalt binder and asphalt mixture properties. Ten mixtures were prepared using two asphalt binders (PG58-28 and PG58-34) and two RAP sources, identified as follows: ? Millings ? RAP from a single source, milled up from I-494 in Maple Grove ? RAP ? RAP combined from a number of sources and crushed at the HMA plant. In addition to the control mixtures, asphalt mixtures were prepared with 20% and 40% of each of the RAP sources. The dynamic modulus proposed by the recent AASHTO design guide was used to determine the effect of various percentages of RAP on mixture properties. Stiffness and moisture susceptibility results were also used to determine the effect of RAP on the asphalt mixture properties. From the complex modulus test results, it was observed that addition of RAP to a mixture generally increased the complex modulus and mixture stiffness. However, this does not always occur at low temperatures. Asphalt binder grade and RAP source had a significant effect on mixture stiffness. The complex modulus for the mixtures made with PG 58-28 asphalt binder was always higher than that from the mixtures made with a softer PG 58-34 asphalt binder. Also, addition of the millings led to a larger increase in stiffness than the similar addition of RAP. Mixtures containing RAP showed significant variability and the variability increased with the increase in RAP content. The IDT creep test was performed at temperatures of -18?C and -24?C. Results indicated that generally stiffness increases as the percentage of RAP or millings increases. The mixtures with PG 58-34 binder were softer than the mixtures with PG 58- 23 28 binder. For the mixtures with PG 58-28 binder, as the percentage of RAP or millings increased, the IDT strength increased, while the mixtures with PG 58-34 binder did not show the same trend. Moisture susceptibility test data indicated that as the percentage of RAP increased the strength also increased, while the tensile strength ratio decreased. Binder tests showed that the addition of RAP improved the binder grade in terms of high temperature performance, while the low temperature performance did not change significantly except for the case when 40% RAP was added, meaning that the resulting binder blends would be more resistant to rutting and equally resistant to thermal cracking compared to virgin binders. The tests on the binders indicated that using 20% RAP in asphalt mixtures does not significantly affect the performance. RAP amounts of 40% have a significant effect on the performance of the mixtures. 12. Use of Reclaimed Asphalt Pavement in Superpave Hot-Mix Asphalt Applications by Stroup-Gardiner and Wagner (13). This research evaluated the effectiveness of screening RAP stockpiles into coarse and fine fractions. This practice was found to maximize the use of RAP and produce a range of HMA mixtures that meet Superpave requirements. The coarser fraction was used in a typical 12.5 mm below-the-restricted-zone Superpave gradation, while the finer fraction was used in a 12.5 mm above-the-restricted- zone gradation. Screening the RAP increased uniformity in coarser aggregate fractions and allowed up to 40 percent of this material to be used and still meet below-the- restricted-zone Superpave gradation requirements by reducing the amount of finer 24 aggregate fractions, especially the minus 0.075 mm material. The use of RAP in these mixtures reduced neat asphalt requirements by 18 to 33 percent. The use of the finer RAP fraction in above-the-restricted-zone Superpave gradations resulted in a reduction in neat asphalt of about 25 percent. Addition of this material decreased rutting potential and temperature susceptibility. However, the amount of material to be used was limited to a maximum of 15 percent in order to meet above- the-restricted-zone gradation requirements. 13. Mechanistic and Volumetric Properties of Asphalt Mixtures with Recycled Asphalt Pavement by Daniel and Lachance (14). This research examined how the addition of RAP changes the volumetric and mechanistic properties of asphalt mixtures. Two RAP sources, a processed RAP and an unprocessed RAP (grindings), were used to study the change in volumetric properties and one RAP source was used for dynamic modulus and creep testing. A control mixture containing only virgin materials (0% RAP) was tested along with mixtures containing 15, 25 and 40% RAP. The volumetric properties of the different mixes are shown in Table 2.4. For the processed RAP mixtures, the VMA and VFA values for the 25% and 40% RAP contents were higher than those for the control and 15% mixtures. For the grindings RAP mixtures, the VMA values increase with RAP content and the VFA values for all mixtures are higher than for the control mix. It is hypothesized that this difference is due to the extent of blending of the RAP material with the virgin materials. 25 Table 2.4. Mix Design Parameters (14). Processed Grindings Control 15% RAP 25% RAP 40% RAP 15% RAP 25% RAP 40% RAP % AC 4.8 5.1 5.4 4.9 4.9 5.2 5.2 Gmm 2.451 2.483 2.445 2.466 2.452 2.460 2.475 VMA 13.1 13.3 16.3 15.2 13.8 14.3 14.7 VFA 69.4 69.9 75.4 73.6 71.8 71.0 73.0 DP 1.14 1.10 0.88 1.02 0.91 0.75 0.75 % AC = asphalt content; Gmm = maximum theoretical specific gravity; DP = dust proportion. The study also indicated that there is an optimal preheating time for RAP to allow the particles to soften, break down, and blend with the virgin materials. At 15% RAP, the stiffness of the mixture increased and the compliance decreased, which indicates that the mixture will be more resistant to permanent deformation and less resistant to fatigue and thermal cracking, due to the addition of aged binder contained in the RAP. However, mixtures containing 25 and 40% RAP did not follow the expected trends and behaved similar to the control mixture. A combination of gradation, asphalt content and volumetric properties is likely the cause of these trends. 14. Five Year Experience of Low-Temperature Performance of Recycled Hot Mix by Tam et al. (15). This project investigated the relative resistance of recycled hot mixes to thermal cracking, as compared to conventional mixes. Two criteria were used: limiting stiffness and fracture temperature (FT). Materials were selected from five recycling contracts covering different regions, virgin asphalt cements, and recycling ratios. Direct tension tests were performed at different temperatures to determine the tensile strengths, strains, and stiffnesses of the different mixtures. Results indicated that when using the limiting stiffness approach, the recycled mixes had higher stiffness values 26 than the conventional mix, which would translate into higher thermal cracking susceptibility. Thermal contraction was used to estimate the induced strain due to thermal shrinkage under a restrained condition. In this case, only one mixture had a fracture temperature below the FT of the virgin mix, confirming the findings from the direct tension tests. Results from field data and laboratory tests revealed that mixes with low RAP content or high penetration virgin asphalt cement had better performance than those with high RAP content or using low penetration virgin asphalt cement. It was also found that fracture temperature, stiffness and viscosity increased with aging of the pavement, reducing its resistance to low temperature cracking. 15. Investigation of Properties of Plant-Produced RAP Mixtures by McDaniel et al. (16). This experiment examined the influence of RAP content in the mixture and recovered binder properties of plant-produced hot mix asphalt. For low temperature properties the plant-produced mixtures were tested for creep compliance and tensile strength. For high temperature properties the mixtures were tested for dynamic modulus (|E*|). The virgin and recovered RAP binders were also tested for complex shear modulus (G*). Three percentages of RAP were added (15%, 25% and 40%) using two binder grades (PG64-22 and PG58-28). Indirect tensile strength results showed that, in general, mixtures with higher strength also showed higher stiffness values. Mixtures with lower stiffness values have a better ability to relax the thermal stresses that develop as the pavement cools. In addition, 27 high strength is also required to resist cracking by traffic loads. Table 2.5 shows that the mixture with the highest RAP content (Mixture D) had the highest strength and stiffness; hence, the warmest critical temperature (T c ). Mixture E with the lowest strength also had the lowest stiffness and a low Tc value. Mixtures with the softer binder (PG58-28) showed lower strengths at a given RAP content than the corresponding mixtures with PG64-22, as expected. Table 2.5. Strength, Stiffness and Critical Pavement Temperature of the Mixes (16). RAP Strength (kPa) Stiffness T c Mixture % Binder Rep1 Rep2 Rep3 Avg. (GPa) (?C) A 0 PG64-22 3284 3393 2785 3154 14.7 -28.9 B 15 PG64-22 3359 3525 2831 3238 17.3 -23.3 C 25 PG64-22 3498 3245 3150 3298 17.7 -25.6 D 40 PG64-22 4056 4165 3390 3870 19.2 -22.8 E 25 PG58-28 3153 3143 2413 2903 13.1 -27.2 F 40 PG58-28 3272 3370 2988 3210 16.1 -23.9 16. Laboratory Study of Fatigue Characteristics of HMA Mixtures Containing RAP by Huang et al. (17). This project evaluated the laboratory fatigue characteristics of asphalt mixtures containing RAP. A typical surface mixture commonly used in the state of Tennessee was evaluated at 0, 10, 20 and 30 percent of No. 4 sieve screened RAP materials. One type of aggregate (limestone) and two types of asphalt binders (PG64-22 and PG76-22) were considered in this study. Fatigue characteristics of mixtures were evaluated through indirect tensile strength (IDT), beam fatigue, and semi-circular fatigue tests (SCB). These three tests represented three different test modes: indirect tensile at monotonic loading, SCB at cyclic constant stress, and third-point beam at cyclic constant strain. Half of the specimens were subjected to laboratory long-term aging prior to performance tests. 28 The indirect tensile stress (ITS) and strain test was used to determine the tensile strength and strain of the mixtures. This test was conducted at 25 o C and a 2 inch/min deformation rate. The toughness index (TI), a parameter describing the toughening characteristics in the post-peak region, was also calculated from the indirect tensile test results. The TI compares the performance of a specimen with that of an elastic perfectly plastic reference material, for which the TI remains a constant of 1. For an ideal brittle material with no post-peak load carrying capacity, the value of TI equals zero. In this study, the values of indirect tensile toughness index were calculated up to tensile strain of one percent. Results from the IDT test revealed that increasing the percentage of screened RAP materials generally increased the tensile strengths, and decreased toughness indices for both un-aged and aged mixes. Increasing RAP percentages had significantly different effects in IDT properties for mixtures with PG64-22 than those with PG76-22, especially for the mixtures subjected to laboratory long-term aging. As shown in Table 2.6, the increase of RAP had more tensile strength gains (about 5 to 10% greater for PG64-22 mixtures), no (or less) tensile strain loss at failure (1% smaller for the PG64-22 mixture at 30% RAP content), and less decrease in post-failure toughness index (9.8 to 24.3% less for PG64-22 mixtures), suggesting that the recycled mixes would have an increased fatigue life. 29 Table 2.6. Change of IDT Properties for Long-Term Aged Mixture (17). IDT Properties Tensile Strength Change, % Strain at Failure Change, % Toughness Index Change, % AC %RAP PG76-22 PG64-22 PG76-22 PG64-22 PG76-22 PG64-22 10 4.94 10.9 2.14 9.41 -13.6 -3.82 20 12.2 17.1 -9.57 4.38 -34.5 -11.9 30 18.82 28.9 -12.3 -11.3 -45.0 -20.7 Note: The values in the table indicated the increase or decrease or properties relative to the control mix (with 0% RAP). In the SCB fatigue test, the inclusion of RAP generally increased the fatigue life of the mixtures in this study, as well as the total dissipated energy. Long-term aging also increased fatigue life. For mixes subjected to long-term aging, the slope of fatigue curves in load versus log(N f ) increased significantly when the RAP increased to 30 percent, which indicated potential lower fatigue life for these mixes at lower stress levels (Figure 2.7). Failure Cycles Vs. Applied Loads, Unaged Mixes: PG 64-22 0 500 1000 1500 2000 2500 3000 3500 4000 1 10 100 1000 10000 100000 1000000 Cycles, N f Lo ad , l b s . 0% RAP, unaged 10% RAP, unaged 20% RAP, unaged 30% RAP, unaged Failure Cycles Vs. Applied Loads, Aged Mixes: PG 64-22 0 500 1000 1500 2000 2500 3000 3500 4000 1 10 100 1000 10000 100000 1000000 Cycles, N f L o a d , lbs 0% RAP, aged 10% RAP, aged 20% RAP, aged 30% RAP, aged Figure 2.7. Load Vs. Number of Cycles to Failure in SCB Fatigue Test for PG64-22 Mixtures (17). Results from beam fatigue tests indicated that the inclusion of RAP generally increased the flexural stiffness of the mixtures. Fatigue life as defined by AASHTO TP8- 94 generally increased with the increase of RAP percentages. The percentage of increase in fatigue life is more significant for long-term aged mixtures with PG64-22 asphalt (up to 1.8 higher than the virgin mix) than those with PG76-22 (up to 0.6 times higher than 30 the virgin mix). For mixtures with PG76-22 asphalt, without long-term aging, the fatigue life decreased with the inclusion of RAP (Figure 2.8). Number of Cycles to Failure: PG 64-22 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 unaged long-term aged C ycles, Nf 0% RAP 10% RAP 20% RAP 30% RAP Number of Cycles to Failure: PG 76-22 0 50,000 100,000 150,000 200,000 250,000 300,000 unaged long-term aged Cy cl e s , Nf 0% RAP 10% RAP 20% RAP 30% RAP Figure 2.8. Number of Cycles to Failure in Strain Controlled Beam Fatigue Test (17). In summary, the results from this study indicated that the inclusion of RAP generally increased the stiffness, indirect tensile strength and laboratory fatigue resistance 31 32 for the mixtures studied. Mixture properties changed significantly at 30% RAP content as compared to those with 10 and 20 percent. 17. A Comparison of the Predicted Performance of Virgin and Recycled Mixes by Puttagunta et al. (18). This study compared the fatigue and moisture damage potential of virgin and recycled mixes through the use of indirect tensile strength and resilient modulus tests, as stipulated by the Asphalt Aggregate Mixture Analysis System (AAMAS) guidelines, which recommend the use of indirect tensile strength and resilient modulus to assess both fatigue and moisture damage potential. One source of RAP was used to prepare mixes with recycling ratios of 25% and 50%. Results for the indirect tensile strength test indicated that the virgin mix had a tensile strength about 1,000 kPa higher at low temperatures as compared to the recycled mixes (Figure 2.9). The tensile strength of all mixes decreased as temperature increased, but the rate of decrease was higher for the virgin mix (about 5 kPa/?C higher). The difference in the tensile strengths of the 25% and 50% recycled materials was small at all temperatures (less than 100 kPa). Figure 2.9. Effect of Test Temperature on Indirect Tensile Strength (18). From the results of the resilient modulus test it was concluded that the virgin mix had a higher resilient modulus than the recycled mixes at all temperatures (from 238 MPa at 40?C to 1,667MPa at 5?C for mixtures with 25% RAP and from 255 MPa at 40?C to 1,548MPa at 5?C for mixtures with 50% RAP). The rate of decrease of resilient modulus with test temperature for the virgin mix was 68 to 73 MPa/?C higher than for the recycled mixes between 5 and 22?C but was almost equal between 22 and 40?C (only 8 to 10 MPa/?C higher). Again, at all temperatures the difference between the results of the 25% and 50% recycled mixes was small (not more than 120 MPa). The fatigue analysis showed that in general, the virgin mix had higher resistance to fatigue cracking than the recycled mixes (up to 7.86x10 8 repetitions higher at 5?C). The AAMAS procedure utilizes the resilient moduli and failure strain from the indirect tensile strength test to calculate the fatigue coefficients K 1 and K 2 . The fatigue performances of the 25% and 50% recycled mixes were relatively similar at all temperatures, with a maximum difference of 2.8x10 7 repetitions at 5?C (Table 2.7). 33 34 Table 2.7. Summary Results of Fatigue Constants and Allowable Number of Loads (18). Mixture Temperature (?C) Virgin Mix 25% RAP 50% RAP Resilient modulus (MPa) 5 22 40 6,928 2,020 1,259 5,261 1,594 1,021 5,380 1,621 1,004 Fatigue constant, K 1 5 22 40 4.83x10 -8 6.68x10 -6 4.43x10 -7 1.45x10 -7 1.72x10 -5 1.0x10 -4 1.33x10 -7 1.61x10 -5 1.0x10 -4 Fatigue constant, K 2 5 22 40 3.59 3.05 2.85 3.47 2.95 2.76 3.48 2.96 2.75 Allowable number of loads, N f 5 22 40 8.34x10 8 7.64x10 6 1.99x10 5 4.83x10 7 2.35x10 6 1.42x10 5 7.61x10 7 2.61x10 6 1.52x10 5 Moisture damage analysis based on retained stability as defined by the Marshall design method indicated that the virgin and recycled mixes offered good resistance to moisture damage (over 90% retained stability). However, the AAMAS procedure, which involves the use of the indirect tensile strength and resilient modulus, predicted a resistance to moisture damage that falls below the minimum criteria. The AAMAS procedure also predicted resistance to moisture damage that increased with increasing recycling ratios (up to 0.32 higher, as shown in Table 2.8). This may be attributed to the fact that recycled aggregates allow a better coating with new asphalt as compared to virgin aggregates. Table 2.8. Analysis of Moisture Damage Potential (18). Reclaimed Asphalt Pavement Virgin mix 25% 50% AAMAS criterion Tensile strength ratio 0.59 0.81 0.91 >0.80 Modulus of resilient ratio 0.68 0.85 0.90 >0.80 35 18. Behavior of Recycled Asphalt Pavements at Low Temperatures by Sargious and Mushule (19). This study was conducted to evaluate the behavior of recycled asphalt pavements with respect to low-temperature cracking. A recycled mix consisting of 45% RAP and 55% virgin materials as well as a virgin control mix were used. Using mix properties that were determined experimentally in the laboratory, thermal stresses resulted from drop in temperature and the expected cracking temperatures were determined for both mixes. An experimental analysis based on laboratory tests that consider the pavement properties only, as well as a more complete theoretical analysis based on a finite element computer program were included. The mix properties that were determined experimentally included density, resilient modulus, tensile strength, coefficient of thermal contraction, thermal conductivity, and specific heat. The data required by the program are the ambient air temperature, the cross-section geometry, the thermal and elastic properties of pavement and subgrade, and the surface thermal characteristics. The results for both experimental and laboratory-based experimental analyses indicated that the performance of recycled asphalt pavements with respect to low- temperature cracking is superior to that of virgin asphalt pavements of comparable initial properties. Recycled mixtures had lower crack temperatures (-27?C for the virgin and -31.5?C for the recycled materials), which may be due to factors such as the use of a soft asphalt in the recycled mix as a modifier. Recycled mixtures also had higher coefficient of thermal conductivity (0.37 to 0.50 W/(m?C) higher), higher tensile strength (360 to 1260 kPa higher) and lower coefficient of thermal contraction (0.12x10 5 /?C to 36 0.19x10 5 /?C lower) than those of virgin mixtures. The theoretical work showed that pavement thickness and subgrade type play an important role in low-temperature cracking for both virgin and recycled asphalt pavements. 2.3 SUMMARY The literature review revealed that studies conducted on SMA mixture design and performance (2, 3, 4, 5) concluded that SMA mixes had benefits such as reduced rutting, greater fatigue life and longer service life. Certain modifications in the requirements have been made to adapt the mixtures to the material characteristics and conditions found in the United States. Research on the use of recycled asphalt pavement has found that RAP does not act like a black rock and partial blending occurs (6, 7, 8, 9). Guidelines have been developed to incorporate RAP in conventional HMA mixtures, establishing that at low RAP contents (up to 15%) the binder effects are negligible and no modification is required in the design process. At intermediate RAP contents (16% to 25%), these effects can be compensated for by using a virgin binder that is one grade softer on both the high- and low- temperature grades. Higher RAP contents (over 25%) require the use of blending charts to determine the appropriate virgin binder grade. It has also been found that addition of RAP increases the binder stiffness (6, 10), and hence, the mixture stiffness. This may affect low temperature performance (14, 15) and fatigue life (17, 18). On the other hand, increase in mix stiffness resulted in higher indirect tensile strength (17, 18), which improved rutting and moisture resistance (12, 14, 18). 37 Fractionated RAP has been successfully used in Superpave mixtures (13). RAP material passing the 1.18 mm sieve represents the fine aggregates bound in small conglomerates by the RAP asphalt, which cannot be separated during milling or sieving operations. RAP sources studied had similar gradations at and above the 1.18 mm size, but showed significant differences in the amount of material passing the 1.18 mm sieve. By removing this material, the uniformity between the RAP sources in the coarser fractions could be increased. The fine RAP fractions generally have a higher asphalt content than the coarse fractions due to the higher surface area per unit weight associated with fine aggregate gradations. This higher binder content may reduce the required virgin binder content noticeably while using a lower percentage of RAP material. It has also been observed that using the fine RAP fraction increases the mixture stiffness which reduces rutting potential. The information collected suggests that use of RAP in SMA mixtures could produce important benefits in terms of performance. The effect of increased stiffness must be carefully studied, since SMA mixes could be especially vulnerable to distresses associated with this property, such as thermal and fatigue cracking. 38 CHAPTER 3. RESEARCH TEST PLAN The research approach was divided into three parts as they relate to the objectives of the study: evaluation of materials, mix designs and performance tests. The experiment was planned as a 4x4x4 factorial design, with three factors (aggregate source, RAP content and RAP type) at four levels each. This allowed studying the contributions that each of the factors make individually to the response, as well as the effect of the interaction of treatment factors. The full factorial design would require 64 treatment combinations, but due to time constraints and a need to keep research costs in a reasonable range, a one-fourth fraction was selected so that the number of mix designs to be evaluated could be limited without sacrificing the integrity of the experiment. Table 3.1 shows the test matrix for the fractional factorial design. 39 Table 3.1. Test Matrix for Mix Variables. RAP content, % Aggregate source RAP source 0 10 20 30 Regular X SMA X Fine-graded X Mountain View Coarse-graded X Regular X SMA X Fine-graded X Lithia Springs Coarse-graded X Regular X SMA X Fine-graded X Camak Coarse-graded X Regular X SMA X Fine-graded X Ruby Coarse-graded X 3.1. PART 1 ? EVALUATION OF MATERIALS This study involved evaluating material properties of aggregates, asphalt binder and the combined blend of virgin materials and RAP. 3.1.1 Evaluation of Aggregate Properties Four aggregate sources were used in this study: Florida Rock at Mt. View, Martin- Marietta at Ruby, Martin-Marietta at Camak, and Vulcan at Lithia Springs. These sources were chosen because they have been widely used in SMA production in Georgia with positive results. Their properties are shown in Table 3.2. Tables 3.3 through 3.6 and Figures 3.1 through 3.4 show the aggregate gradations for each source. The ?M? 40 denomination on some of the aggregates means that they are manufactured screenings, while the ?W? denominations correspond to washed screenings, which have a lower dust content. Table 3.2. Properties of Virgin Aggregates. Specific Gravities Aggregate Source General Character of Material Bulk SSD App. Absorption, % Mt. View Granite Gneiss/ Amphibolite 2.640 2.659 2.691 0.72 Lithia Springs Granite Gneiss 2.591 2.608 2.635 0.62 Camak Granite Gneiss 2.638 2.655 2.682 0.62 Ruby Gneiss/ Amphibolite 2.734 2.746 2.767 0.43 Table 3.3. Gradations for Mt. View Aggregates. Percent Passing Sieve Size 007 089 W10 1" 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 1/2" 97.0 100.0 100.0 3/8" 48.0 100.0 100.0 #4 3.0 22.0 99.0 #8 3.0 4.0 83.0 #16 2.0 2.0 66.0 #30 2.0 2.0 53.0 #50 2.0 1.0 37.0 #100 1.0 1.0 18.0 #200 1.0 1.0 6.0 41 Table 3.4. Gradations for Lithia Springs Aggregates. Percent Passing Sieve Size 007 089 810 1" 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 1/2" 85.0 100.0 100.0 3/8" 50.0 100.0 100.0 #4 6.0 30.0 84.0 #8 2.0 2.0 62.0 #16 1.0 2.0 50.0 #30 1.0 1.0 41.0 #50 1.0 1.0 28.0 #100 1.0 1.0 21.0 #200 1.0 1.0 10.0 Table 3.5. Gradations for Camak Aggregates. Percent Passing Sieve Size 007 M10 1" 100.0 100.0 3/4" 100.0 100.0 1/2" 94.0 100.0 3/8" 56.0 100.0 #4 10.0 98.0 #8 3.0 82.0 #16 3.0 62.0 #30 2.0 50.0 #50 1.0 36.0 #100 1.0 25.0 #200 1.0 12.0 Table 3.6. Gradations for Ruby Aggregates. Percent Passing Sieve Size 007 M10 1" 100.0 100.0 3/4" 100.0 100.0 1/2" 96.0 100.0 3/8" 55.0 100.0 #4 2.0 99.0 #8 1.0 82.0 #16 1.0 62.0 #30 1.0 49.0 #50 1.0 37.0 #100 1.0 27.0 #200 1.0 18.0 0.45 Power Gradation Chart #200 #100 #50 #30 #16 #8 #4 3/8" 1/2" 3/4" 1" 0 10 20 30 40 50 60 70 80 90 100 Sieve Size Per cen t Pa ssi n g 007 089 W10 Figure 3.1. Gradations for Mt. View Aggregates. 0.45 Power Gradation Chart #200 #100 #50 #30 #16 #8 #4 3/8" 1/2" 3/4" 1" 0 10 20 30 40 50 60 70 80 90 100 Sieve Size Per cen t Pa ssi n g 007 089 810 Figure 3.2. Gradations for Lithia Springs Aggregates. 42 0.45 Power Gradation Chart #200 #100 #50 #30 #16 #8 #4 3/8" 1/2" 3/4" 1" 0 10 20 30 40 50 60 70 80 90 100 Sieve Size Per cen t Pa ssi n g 007 M10 Figure 3.3. Gradations for Camak Aggregates. 0.45 Power Gradation Chart #200 #100 #50 #30 #16 #8 #4 3/8" 1/2" 3/4" 1" 0 10 20 30 40 50 60 70 80 90 100 Sieve Size Per cen t Pa ssi n g 007 M10 Figure 3.4. Gradations for Ruby Aggregates. 43 44 Four RAP sources were used in this study: conventional RAP, RAP from reclaimed SMA, fine-graded RAP (-4 RAP), and coarse-graded RAP (+4 RAP). Table 3.7 and Figure 3.5 show the RAP gradations used in this study and Table 3.8 shows the asphalt cement content of each RAP source. SMA RAP was included in this study to evaluate the possibility of recycling SMA material back into an SMA mixture. As the first SMA projects reach the end of their service life, it is important to determine if the stiff mastic in an SMA mix might prevent it from being recycled or if the proportion of RAP may have to be reduced. However, when the SMA RAP received from GDOT was tested it was found that its gradation and asphalt content did not match those of an SMA mix. SMA mixes generally have about 25 percent of the material passing the No. 4 sieve, and in this case, that amount was 77 percent. It is known that this RAP material was crushed to achieve the 12.5 mm NMAS and this may have affected its gradation. The asphalt content of the SMA RAP was unusually low with an average of only 4.4 percent based on weight of total mix (the usual asphalt content being about 6 percent). Based on this result, it is likely that the RAP from the SMA project also included a portion of the underlying 19 mm Superpave mixture as a result of the milling process. Overall, there was no significant difference between the gradations of conventional and SMA RAP, and from this point on they will be treated and referred to as dense-graded RAP 1 (DG1 RAP) and dense-graded RAP 2 (DG2 RAP), respectively. The use of fractionated RAP material into coarse and fine-graded stockpiles was also considered in this study. +4 RAP was used as a substitute of a portion of the No. 7 stone, typically used in high quantities in SMA production. This option would be 45 beneficial in the event that quarries were faced with a supply shortage of No. 7 stone due to its high demand in other HMA and concrete mix applications. -4 RAP was used as a substitute for a portion of the asphalt content, since its higher surface area makes it richer in asphalt cement. This would represent an advantage because asphalt cement is typically the most expensive component of a mixture. Table 3.7. RAP Gradations. Percent Passing Sieve Size DG1 RAP DG2 RAP -4 RAP +4 RAP 1" 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 99.0 1/2" 99.0 100.0 100.0 96.0 3/8" 93.0 95.0 100.0 84.0 #4 73.0 77.0 100.0 37.0 #8 58.0 61.0 81.0 25.0 #16 47.0 50.0 65.0 21.0 #30 38.0 42.0 53.0 18.0 #50 29.0 32.0 40.0 15.0 #100 19.0 20.0 25.0 10.0 #200 11.2 12.0 15.0 6.2 0.45 Power Gradation Chart #200 #100 #50 #30 #16 #8 #4 3/8" 1/2" 3/4" 1" 0 10 20 30 40 50 60 70 80 90 100 Sieve Size Per cen t Pa ssi n g DG1 DG2 -4 +4 Figure 3.5. RAP Gradations. Table 3.8. Asphalt Contents in RAP RAP Source Asphalt Content, % +4 4.5 -4 6.2 DG1 5.6 DG2 4.4 Aggregate toughness was determined by the Los Angeles abrasion test (ASTM C131), which measures the resistance of coarse aggregates to degradation by abrasion and impact. The aggregate is placed in a metal drum along with a charge of steel balls, and the drum is rotated 500 times at a speed of 30 - 33 revolutions per minute (RPM). The inside of the drum is equipped with an angle iron which runs longitudinally. This causes the charge of aggregate and balls to fall with a heavy impact once during each revolution, breaking the aggregate particles into smaller particles. At the completion of 46 47 the test, the aggregate is shaken over a No. 12 sieve and the amount which passes through the sieve, expressed as a percentage of the total charge, is the Los Angeles abrasion value designated "percent loss". Aggregates must be tough in order to prevent crushing and abrasive wear during manufacturing, placing and compaction of HMA. This aggregate property is especially critical in gap-graded mixtures such as SMA because excessive aggregate breakdown will fill void spaces within the mixture and thereby reduce the amount of asphalt cement that would otherwise be needed. As the asphalt content is reduced, the durability of the mixture suffers and results in premature aging and deterioration. The flat and elongated property was determined by GDT-129. This characteristic is defined as the percentage by weight of coarse aggregates that have a length in excess of three times its average thickness, in accordance with the test procedure. This test was performed to ensure that the aggregate contained cubical particles capable of distributing traffic loads through the stone-on-stone coarse aggregate skeleton of an SMA mix. This also contributes to the improved rutting resistance of SMA mixes as compared to conventional mixtures. 3.1.2 Evaluation of Asphalt Binder Properties The binder from RAP materials was recovered through Abson recovery tests (ASTM D1856) and its properties were evaluated by means of the Dynamic Shear Rheometer (DSR) and Bending Beam Rheometer (BBR). Asphalt cement from samples of the proposed blends was also extracted and analyzed for rheological properties and performance grade. Short-term aging of the blended binders was achieved using the Rolling Thin Film Oven (RTFO) procedure according to AASHTO T240, which simulates aging during construction. In this test, a moving film of asphalt binder is heated in an oven for 85 minutes at 163? C. The moving film is created by placing the asphalt binder sample in a small jar then placing the jar in a circular metal carriage that rotates within the oven (Figure 3.6). This rotation is used to continually expose fresh films of the binder to hot air. Figure 3.6. Rolling Thin Film Oven (20). Long-term aging, which simulates several years of exposure to the environment, was achieved using the Pressure Aging Vessel (PAV) in accordance to AASHTO PP1. The PAV is an oven-pressure vessel combination that takes RTFO aged samples and 48 exposes them to high air pressure (2070 kPa) and temperature (90?C, 100?C or 110?C, depending upon expected climatic conditions) for 20 hours. Figure 3.7. Pressure Aging Vessel (20). Engineering properties of the blended binders were obtained through Dynamic Shear Rheometer (DSR) and Bending Beam Rheometer (BBR) testing. The DSR test was used in accordance with AASHTO TP 5 to measure the complex shear modulus (G*) and phase angle (?) of the blended binders at high and intermediate temperatures. The test uses a thin asphalt binder sample placed between two plates. The lower plate is fixed while the upper plate oscillates back and forth across the sample at 1.59 Hz to create a shearing action. These oscillations at 1.59 Hz (10 radians/sec) are meant to simulate the shearing action corresponding to a traffic speed of about 90 km/hr (55 mph). 49 Figure 3.8. Basics of Dynamic Shear Rheometer (20). The physical properties measured with the DSR allow obtaining the rutting and fatigue parameters, which are used to quantify the asphalt binder?s contribution in resisting those types of distresses. Rutting is considered a stress controlled, cyclic loading phenomenon. Each traffic loading cycle does work that contributes to deform the HMA pavement surface. A part of this work is recovered by elastic rebound of the surface while some is dissipated in the form of permanent deformation and heat. The work dissipated per loading cycle at a constant stress can be expressed as: ? ? ? ? ? ? ?= ? ? ?? sin/ 1 2 G W oc Equation 3.1 (21) Where: W c = work dissipated per load cycle ? o = stress applied during the load cycle G * = complex modulus ? = phase angle 50 The amount of work dissipated per loading cycle is inversely proportional to G * /sin?, called the rutting parameter. In order to minimize permanent deformation, W c must be minimized as well. This indicates that higher values of G * /sin? correspond to binders with better rutting resistance In the case of fatigue cracking, this distress is considered a strain controlled phenomenon. The work dissipated per loading cycle at a constant strain can be expressed as: [ ]?? sin 2 ? ??= GW oc Equation 3.2 (21) where ? is the strain and the other variables are as previously described. Fatigue cracking is minimized by decreasing the term G * sin? (fatigue parameter). The BBR test was performed according to AASHTO TP 1 to determine the binder?s propensity to thermal cracking. The BBR basically subjects a simple asphalt beam to a small (1,000 mN) load over 240 seconds. Then, using basic beam theory, the BBR calculates the flexural creep stiffness (S) and logarithmic creep rate (m) of the asphalt binder. The creep stiffness of the asphalt binder beam at 60 seconds loading time is given by: () ()tbh PL tS ? 3 3 4 = Equation 3.3 (21) Where: S(t) = creep stiffness at time, t = 60 seconds P = applied constant load, 100 g L = distance between beam supports, 102 mm b = beam width, 12.5 mm 51 h = beam thickness, 6.25 mm ?(t) = deflection at time, t = 60 seconds The m-value is the rate of change of the stiffness, S(t), with loading time and is used to describe how the asphalt binder relaxes under load. Figure 3.9. Schematic of Bending Beam Rheometer (20). Creep stiffness is related to thermal stresses in an HMA pavement due to shrinking while the m-value is related to the ability of an HMA pavement to relieve these stresses. Therefore, asphalt binders with minimum creep stiffness and maximum creep rate are desired in order to resist thermal cracking. The Superpave asphalt binder specification (AASHTO MP1) is intended to control permanent deformation, low temperature cracking and fatigue cracking in asphalt pavements. The specification accomplishes this by controlling the various physical 52 properties described previously (G*/sin?, G*sin?, S(t) and m-value). The physical properties remain constant for all performance grades (PG), but the temperatures at which these properties must be achieved vary depending on the climate in which the binder is expected to serve (Figure 3.10). Figure 3.10. Superpave Binder Specification Example (20). 3.2 PART 2 ? MIX DESIGNS RAP material, virgin asphalt and virgin aggregate were proportioned to produce 12.5 mm SMA mix designs. The 50-blow Marshall procedure, which is used by GDOT, was used for asphalt mixture compaction and PG 76-22 was used as the standard performance grade asphalt. RAP was blended at four proportions (0%, 10%, 20%, and 30%) to determine the effect of RAP over the ranges of anticipated use. A blend with no RAP was 53 54 used as a baseline for the study for comparisons of mix performance. A RAP content of 10% represented the least amount that can feasibly be utilized, and a maximum RAP content of 30% was used because it is improbable that blends with greater contents of RAP would be able to meet gradation and volumetric requirements of the mix design. The gradations for the control and recycled mixes are shown in Tables 3.9 through 3.13 and Figures 3.11 through 3.15. Table 3.9. Gradations for Control Mixes. Percent Passing Sieve Size Mt. View Lithia Springs Camak Ruby 1" 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 1/2" 98.0 90.0 95.0 96.9 3/8" 64.6 66.5 55.2 65.4 #4 24.6 25.8 18.5 24.4 #8 20.3 16.7 16.0 20.5 #16 17.2 14.4 14.0 16.9 #30 15.5 13.1 12.8 14.6 #50 13.3 11.4 11.4 12.4 #100 10.0 10.4 10.2 10.6 #200 8.0 8.5 8.4 8.6 Table 3.10. Gradations of Recycled SMA Mix Using DG1 RAP. Percent Passing Sieve Size 10% RAP 20% RAP 30% RAP 1" 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 1/2" 89.1 97.0 95.8 3/8" 63.4 67.1 62.5 #4 27.7 25.9 27.6 #8 20.6 21.4 22.4 #16 17.2 18.1 19.1 #30 15.2 15.7 16.4 #50 12.6 13.3 13.7 #100 10.7 10.7 10.6 #200 8.1 8.4 8.0 55 Table 3.11. Gradations of Recycled SMA Mix Using DG2 RAP. Percent Passing Sieve Size 10% RAP 20% RAP 30% RAP 1" 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 1/2" 95.6 98.0 90.1 3/8" 60.1 64.1 65.6 #4 25.9 24.9 31.2 #8 21.8 20.4 23.7 #16 18.5 17.3 19.7 #30 16.3 15.7 17.3 #50 13.8 13.6 14.2 #100 11.2 10.4 10.5 #200 8.6 8.4 7.8 Table 3.12. Gradations of Recycled SMA Mix Using -4 RAP. Percent Passing Sieve Size 10% RAP 20% RAP 30% RAP 1" 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 1/2" 97.7 95.5 97.3 3/8" 60.0 59.4 70.0 #4 25.2 26.4 34.6 #8 22.2 21.9 28.3 #16 18.7 18.7 23.5 #30 16.6 16.3 19.9 #50 14.1 13.7 16.0 #100 10.4 10.6 11.5 #200 8.2 8.3 8.3 Table 3.13. Gradations of Recycled SMA Mix Using +4 RAP. Percent Passing Sieve Size 10% RAP 20% RAP 30% RAP 1" 100.0 100.0 100.0 3/4" 99.9 99.8 99.7 1/2" 96.8 89.9 97.0 3/8" 66.9 65.9 64.6 #4 24.9 27.2 23.8 #8 20.4 19.7 19.3 #16 17.0 16.8 16.7 #30 14.8 15.1 15.1 #50 12.7 12.9 13.4 #100 10.6 11.0 10.2 #200 8.5 8.5 8.1 0.45 Power Gradation Chart #200 #100 #50 #30 #16 #8 #4 3/8" 1/2" 3/4" 1" 0 10 20 30 40 50 60 70 80 90 100 Sieve Size Per cen t Pa ssi n g Mt. View Lithia Springs Ruby Camak Figure 3.11. Gradations for Control Mixes. 0.45 Power Gradation Chart #200 #100 #50 #30 #16 #8 #4 3/8" 1/2" 3/4" 1" 0 10 20 30 40 50 60 70 80 90 100 Sieve Size Per cen t Pa ssi n g 10% 20% 30% Figure 3.12. Gradations of Recycled SMA Mix Using DG1 RAP. 56 0.45 Power Gradation Chart #200 #100 #50 #30 #16 #8 #4 3/8" 1/2" 3/4" 1" 0 10 20 30 40 50 60 70 80 90 100 Sieve Size Per cen t Pa ssi n g 10% 20% 30% Figure 3.13. Gradations of Recycled SMA Mix Using DG2 RAP. 0.45 Power Gradation Chart #200 #100 #50 #30 #16 #8 #4 3/8" 1/2" 3/4" 1" 0 10 20 30 40 50 60 70 80 90 100 Sieve Size Per cen t Pa ssi n g 10% 20% 30% Figure 3.14. Gradations of Recycled SMA Mix Using -4 RAP. 57 0.45 Power Gradation Chart #200 #100 #50 #30 #16 #8 #4 3/8" 1/2" 3/4" 1" 0 10 20 30 40 50 60 70 80 90 100 Sieve Size Per cen t Pa ssi n g 10% 20% 30% Figure 3.15. Gradations of Recycled SMA Mix Using +4 RAP. In the mix design, the gradations of the SMA mixtures were kept as close as possible to each other to provide a proper comparison among the different aggregate sources and RAP types. The gradation charts in Figures 3.11 through 3.15 show that there is little variation in the percent passing each sieve size among the different mixtures (in most cases under 5%). Fiber stabilizing additive (cellulose fiber) and mineral filler (marble dust) were included in the mixture as specified by GDOT. Once the blends for each type of aggregate were determined, an initial asphalt content was estimated for each mixture. Replicate samples prepared for each blend were mixed at three different asphalt contents and conditioned in accordance with AASHTO R30. Specimens were compacted using a Marshall hammer, following procedures in AASHTO T-245. The bulk specific gravity of each specimen was determined by 58 59 AASHTO T 166. The theoretical maximum specific gravity of the loose HMA mix samples was measured in accordance with AASHTO T 209. Percent of air voids in the mix, voids in mineral aggregate (VMA) and voids filled with asphalt (VFA) were calculated for each mixture. 3.3 PART 3 ? PERFORMANCE TESTS The performance of the bituminous concrete test mixes was evaluated by subjecting specimens to diametral tensile strength, moisture susceptibility, flexural beam fatigue, APA rutting, and indirect tensile creep compliance tests. 3.3.1 Moisture Susceptibility The effect of RAP addition on the moisture susceptibility of the mixtures was evaluated by determining the diametral tensile strength on dry and wet specimens according to GDT-66, Evaluating the Moisture Susceptibility of Bituminous Mixtures by Diametral Tensile Splitting. In this test, internal water pressures in the mixtures are produced by vacuum saturation followed by a freeze and a warm-water soaking cycle. Six Marshall specimens were prepared using optimum asphalt content and compacted to 7.0 ? 1.0 percent air voids. A subset of three specimens remained unconditioned and was used as the control group. The other subset was partially vacuum saturated with water for 30 minutes and then subjected to a freeze (-18?C for 15 hours) and thaw (60?C for 24 hours) cycle. Both subsets were then tested for indirect tensile strength at a load rate of 0.065 in/minute. The diametral tensile strength of each specimen was determined by Equation 3.4. tD P S ? 2 = Equation 3.4 Where: S = tensile strength, psi (kPa) P = maximum load, pounds (N) t = specimen height immediately before tensile test, inches (millimeters) D = specimen diameter, inches (millimeters) The percentage of retained strength (TSR) was calculated by comparing the properties of dry specimens with water-conditioned specimens. 100?= c a S S TSR Equation 3.5 Where: TSR = percent retained strength S a = average tensile strength of conditioned subset, psi (kPa) S c = average tensile strength of control subset, psi (kPa) 3.3.2 Rutting Susceptibility Rutting susceptibility of the mixtures was tested with the Asphalt Pavement Analyzer (APA), according to GDT-115, Determining Rutting Susceptibility Using the Loaded Wheel Tester. The APA is a modification of the Georgia Loaded Wheel Tester (GLWT), 60 and it follows a similar rut-testing procedure. A wheel is loaded onto a pressurized linear hose and tracked back and forth over a testing sample to induce rutting. Six samples for each mix type were compacted with a gyratory compactor to 5.0 ? 1 percent air voids and tested at 64?C using a vertical load of 100 lbs. and hose pressure of 100 psi for 8,000 cycles. 3.3.3 Creep Compliance The creep compliance of the mixtures was evaluated according to AASHTO T 322, Determining the Creep Compliance and Strength of Hot-Mix Asphalt Using the Indirect Tensile Test Device, in order to determine if the addition of RAP affected the resistance of the mixtures to thermal cracking. Three replicate specimens for each mixture were compacted with a gyratory compactor to approximately 7% air voids and cut to dimensions of 150 mm diameter by 50 mm height. The tensile creep compliance was determined by applying a static compressive load of fixed magnitude along the diametral axis of each specimen for 100 s. Each specimen was tested at temperatures of -20, -10 and 0?C. The horizontal and vertical deformations measured near the center of the specimen were used to calculate the tensile creep compliance as a function of time, given by the following relationship: () ( ) 0 ? ? t tD = Equation 3.5 where ?(t) is the strain and ? 0 is the stress. 61 Compliance is a way of characterizing the stiffness of a material. Another term frequently used is creep stiffness, S(t), which is the inverse of creep compliance as determined from a creep test: () () ()ttD tS ? ? 0 1 == Equation 3.6 An example of creep compliance curves measured at multiple temperatures using the indirect tensile test at low temperature is presented in Figure 3.16. A nonlinear regression routine is used to determine the master creep compliance curve from the creep compliance curves measured at multiple temperatures. The regression is performed in two steps. First, a regression is performed to simultaneously determine the temperature shift factors (a t ) and the parameters for the following Prony series (Maxwell model) representation of the master creep compliance curve: () () () v N i i i eDDD ? ? ? ?? +?+= ? =1 / 10 Equation 3.7 Where: D (?) = creep compliance at reduced time ? ? = reduced time (= t/a t ) a t = temperature shift factor D(0), D i, ? I , ? v = Prony series parameters 62 Figure 3.16. Indirect Tension Test Creep Compliance Curves (23). In essence, the regression finds the best shift factors and Prony series parameters to fit the measured data based upon a least-squares criterion. One of the temperatures is selected as the reference temperature for the master curve (typically -20?C), and thus the creep compliance curve at this temperature is fixed in time (a t = 1). The regression determines the amount of time (horizontal) shift required for the curves at the remaining temperatures to result in a smooth master curve. Each of these remaining creep compliance curves will thus have a shift factor (a t ) associated with it. Figure 3.17 shows the shifted creep compliance data. 63 Figure 3.17. Prony Series Fit to Master Creep Compliance Curve (23). The second step in the regression routine is to fit a second functional form to the master creep compliance information. This second functional form is the following power law: ( ) ( ) m DDD ?? 1 0 += Equation 3.8 where D(?) and ? are as defined previously, and D(0), D 1 , and m are the coefficients of the functional form. The primary purpose for fitting this functional form is to determine the parameter m. This parameter is essentially the slope of the linear portion of the master creep compliance curve on a log-log plot (Figure 3.18). It has been found to be an important parameter in distinguishing between the thermal cracking performance of different materials. 64 Figure 3.18. Determination of m, the Slope of the Log Creep Compliance Curve (23). 3.3.4 Flexural Beam Fatigue Fatigue tests were conducted according to AASHTO TP 8, Determining the Fatigue Life of Compacted Hot-Mix Asphalt Subjected to Repeated Flexural Bending, to evaluate the stiffening effect of RAP on the mixture and its impact on the long-term fatigue life of the pavement. Three replicate beams were compacted with a kneading compactor to 6.0 ? 1.0 percent air voids and cut to dimensions of 380 mm long by 50 mm thick by 63 mm wide. The beams were placed in four-point loading and subjected to repeated haversine loads. The deflection caused by the load was measured at the center of the beam. The tests were performed under a constant-strain condition, at strain levels of 400 and 800 micro-strain and at a temperature of 20?C. The variables measured include the number of cycles to failure, initial and final stiffnesses, and dissipated energy. The flexural stiffness is defined as: 65 t t S ? ?? = 1000 Equation 3.9 Where: S = flexural stiffness, MPa ? t = maximum tensile stress, kPa ? t = maximum tensile microstrain The initial stiffness is defined as the measured flexural stiffness after 50 cycles. The number of cycles to failure (N f ) is the load cycle at which the specimen exhibits a 50 percent reduction in stiffness relative to the initial stiffness. The dissipated energy is calculated by determining the area within the stress-strain hysteresis loop for each captured data pulse. The cumulative dissipated energy is the summation of the dissipated energy per cycle. 66 67 CHAPTER 4. TEST RESULTS AND ANALYSIS 4.1 MATERIAL PROPERTIES 4.1.1. Aggregates Properties of the virgin and recycled SMA mixes used in Part 1 of the study are shown in Table 4.1. These data were used to perform an analysis of variance (shown in Table 4.2), which indicated that combined blend properties such as LA abrasion and flat and elongated particle content are mainly influenced by the aggregate source (p-values < 0.001). At 95% confidence level, RAP content and RAP type did not have a significant effect on percent loss or F/E particle content. As Table 4.3 shows, there is little variation in the aggregate properties of the RAP materials (3.0% difference for % loss and 0.8% difference for F/E particle content), which is the reason why RAP type is not significant in this data. 68 Table 4.1. Aggregate Properties for Combined Blends. Aggregate % RAP RAP Type LA Abrasion, % loss F/E particles, % (3:1 ratio) 0 DG1 48.0 8.6 10 -4 48.0* 4.0 20 DG2 46.7 5.4 Mt. View 30 +4 47.7 4.0 0 -4 39.6 11.3 10 DG1 40.4 11.8 20 +4 41.1 13.1 Lithia Springs 30 DG2 41.0 17.5 0 +4 37.3 9.3 10 DG2 38.7 7.9 20 -4 37.3* 9.3 Camak 30 DG1 40.3 12.9 0 DG2 21.1 6.4 10 +4 23.8 4.7 20 DG1 26.4 5.7 Ruby 30 -4 21.1* 6.4 *Same as virgin blend because coarse recycled material was not added. Table 4.2. Analysis of Variance for Aggregate Properties. LA Abrasion F/E Particles Factor F-statistic p-value F-statistic p-value Agg. Source 763.25 0.000 25.05 0.000 RAP Content 2.55 0.082 2.89 0.058 RAP Type 7.24 0.001 1.85 0.167 Table 4.3. Aggregate Properties for RAP Material. RAP Type LA Abrasion, % loss F/E particles, % (3:1 ratio) DG1 47.2 6.8 DG2 44.2 6.0 -4 N/A N/A +4 47.2 6.8 The effect of RAP addition on aggregate properties depended on the quality of the virgin and recycled materials contained in the blend. As mentioned above, the differences among aggregate sources were significant (Table 4.4), and they determined the results for 69 the combined blend. If a virgin aggregate is combined with RAP aggregates that have higher percent loss values, increasing the RAP content will increase the percent loss of the blend. Likewise, if a virgin aggregate is combined with RAP aggregates that have lower percent loss values, increasing the RAP content will decrease the percent loss of the blend. However, the differences produced by the increase in RAP content were very small (less than 1.5%, as Table 4.5 shows) and were not significant for these data. The results for the F/E particles had a maximum difference of 3.1% among RAP contents, which was not significant. Table 4.4. Average Results for Aggregate Sources. Aggregate Source LA Abrasion, % Loss F/E Particles, % (3:1 ratio) Mt. View 47.6 5.5 Lithia Springs 40.5 13.4 Camak 38.2 9.8 Ruby 23.1 5.8 Table 4.5. Average Results for RAP Contents. RAP Content, % LA Abrasion, % Loss F/E Particles, % (3:1 ratio) 0 36.5 8.9 10 37.5 7.1 20 37.9 8.4 30 37.5 10.2 4.1.2 Asphalt Binder The single type of virgin asphalt cement was PG 76-22, which is a polymer-modified asphalt. The results for the virgin and recovered RAP asphalts are presented in Table 4.6. 70 The critical high temperatures for the extracted RAP binders were obtained by testing the recovered RAP binder as if it had been RTFO aged. Table 4.6. Critical Temperatures and Performance Grades of Virgin and Recovered RAP Binders. Recovered RAP Binders Aging Property Virgin Binder +4 -4 DG1 DG2 Original G*/sin ?, kPa 78.9 --- --- --- --- RTFO G*/sin ?, kPa 79.2 87.4 89.0 89.0 94.2 RTFO+PAV G* sin ?, kPa BBR S, MPa BBR m-value 21.1 -27.2 -24.4 26.0 -27.9 -25.5 26.5 -28.1 -20.1 27.6 -30.2 -23.9 28.5 -25.1 -18.4 PG Actual MP1 78.9-24.4 76-22 87.4-25.5 82-22 89.0-20.1 88-16 89.0-23.9 88-22 94.2-18.4 94-16 The actual binder properties of the blends are shown in Tables 4.7 through 4.10. It can be observed that among RAP types, blends that contain DG2 RAP had higher values of G*/sin? at a given RAP content for both original (1.96 kPa and higher at passing temperatures) and RFTO aged samples (3.5 kPa and higher at passing temperatures), which indicates better resistance of the resulting binder blends to rutting. Blends containing DG2 RAP also had lower values of G*sin? than the corresponding blends containing other RAP binder types (maximum 3,847 kPa at passing temperature). This is indicative of a higher fatigue cracking resistance for these binder blends. Finally, the properties obtained with the BBR test (creep stiffness and creep rate) also had more favorable results for blends using DG2 RAP binder. These blends had lower stiffness (maximum 151 MPa at passing temperature) and higher creep rate (0.322 and higher at passing temperatures) than the corresponding blends containing other RAP binder types. Low creep stiffness values are desired in order to minimize thermal stresses, while the m-value must be high to maximize the ability of the HMA pavement to relieve 71 those stresses; therefore, mixtures that contain DG2 RAP have a binder blend that is more resistant to thermal cracking. Table 4.7. Measured Binder Properties of +4 RAP Blends. +4 RAP Aging Property Critical Property Temp. ?C 10% 20% 30% Original G*/sin ?, kPa ? 1.00 kPa 76 82 1.290 0.762 1.634 0.967 2.521 1.354 RTFO G*/sin ?, kPa ? 2.20 kPa 76 82 2.695 1.570 2.802 1.559 3.312 1.968 RTFO+PAV G* sin ?, kPa BBR S, MPa BBR m-value ? 5,000 kPa ? 300 MPa ? 0.300 25 22 -12 -18 -12 -18 3,427 4,915 148 273 0.332 0.276 3,614 5,190 157 270 0.324 0.223 4,413 6,233 167 264 0.304 0.266 Table 4.8. Measured Binder Properties of -4 RAP Blends. -4 RAP Aging Property Critical Property Temp. ?C 10% 20% 30% Original G*/sin ?, kPa ? 1.00 kPa 76 82 1.578 0.884 1.657 0.956 1.849 1.036 RTFO G*/sin ?, kPa ? 2.20 kPa 76 82 3.018 1.760 3.318 1.818 3.964 2.087 RTFO+PAV G* sin ?, kPa BBR S, MPa BBR m-value ? 5,000 kPa ? 300 MPa ? 0.300 25 22 -12 -18 -12 -18 4,370 6,227 176 287 0.304 0.261 4,153 5,915 179 297 0.306 0.263 4,690 6,520 182 297 0.291 0.266 72 Table 4.9. Measured Binder Properties of DG1 RAP Blends. DG1 RAP Aging Property Critical Property Temp. ?C 10% 20% 30% Original G*/sin ?, kPa ? 1.00 kPa 76 82 88 1.518 0.877 1.593 0.885 1.112 0.603 RTFO G*/sin ?, kPa ? 2.20 kPa 76 82 88 3.032 1.736 3.183 1.843 2.243 1.217 RTFO+PAV G* sin ?, kPa BBR S, MPa BBR m-value ? 5,000 kPa ? 300 MPa ? 0.300 25 22 -12 -18 -12 -18 3,688 5,297 164 308 0.313 0.326 3,997 5,715 164 344 0.311 0.254 4,149 5,854 168 345 0.304 0.271 Table 4.10. Measured Binder Properties of DG2 RAP Blends. DG2 RAP Aging Property Critical Property Temp. ?C 10% 20% 30% Original G*/sin ?, kPa ? 1.00 kPa 76 82 88 2.039 1.150 1.965 1.101 2.676 0.652 RTFO G*/sin ?, kPa ? 2.20 kPa 76 82 3.488 2.042 4.038 2.274 4.213 2.676 RTFO+PAV G* sin ?, kPa BBR S, MPa BBR m-value ? 5,000 kPa ? 300 MPa ? 0.300 25 22 -12 -18 -12 -18 3,030 4,374 142 317 0.334 0.281 3,416 4,876 127 299 0.329 0.271 3,847 5,438 151 317 0.322 0.260 Table 4.6 showed that the critical high temperatures of the RAP binders are higher than that of the virgin binder, which suggests that the combined blends of the recycled mixture should be more resistant to rutting. The intermediate temperatures were also higher for the RAP binders, meaning that the fatigue resistance of the asphalt blends may be affected by the addition of RAP binder. The ability of the combined blends to resist thermal cracking could also be affected, since three of the RAP binders had critical low 73 temperatures higher than that of the virgin binder and this could cause the combined blends to have higher critical low temperatures as well. DG2 RAP binder has a higher critical high temperature (Table 4.6), which caused the DG2 RAP binder blends to have critical high temperatures at least 1.7?C higher than the other blends (Figure 4.1). It can also be observed that blends containing -4 RAP always had a higher critical high temperature than blends containing +4 RAP. This could be due to the higher asphalt content in -4 RAP, which results in a lower demand of virgin asphalt in the mix. The resulting asphalt blends are therefore stiffer and have higher values of critical high temperatures. Even though the recovered DG2 RAP binder had the highest temperatures in both cases, in general, the combined DG2 RAP blends had critical intermediate temperatures between 0.5?C and 3.0?C lower than the other blends, as shown in Figure 4.2. The DG2 RAP blends also had most critical low temperatures between 0.1?C and 4.3?C lower than the other blends, and only in one case was the critical temperature exceeded by that of the blend containing +4 RAP (Figure 4.3). 75 76 77 78 79 80 81 82 83 84 85 10 20 30 % RAP Blend C r i t ic al H i gh Tem p er ature (? C) +4 -4 DG1 DG2 Figure 4.1. Critical High Temperatures for Binder Blends. 19 20 21 22 23 24 25 10 20 30 % RAP Blend Crit ic a l I n t e rm ed ia t e T e m p e r at u r e ( ? C) +4 -4 DG1 DG2 Figure 4.2. Critical Intermediate Temperatures for Binder Blends. 74 -30.0 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 10 20 30 % RAP Blend Cr iti cal Low Te mper atur e (? C) +4 -4 DG1 DG2 Figure 4.3. Critical Low Temperatures for Binder Blends. Table 4.11 shows the results for the performance grades of the binder blends. The addition of 10% RAP binder did not change the performance grade of the binder blends. Increasing the RAP binder content to 20% only affected the DG2 blend by raising the high-temperature grade by one grade. The low temperature performance grade remained the same as the virgin asphalt binder. Finally, the use of 30% RAP binder reduced the low-temperature grade of the -4 blend by one grade, raised the high-temperature grade of the DG1 blend by one grade, and had no further effect on the DG2 and +4 blends. These results support the ETG design guidelines (8), where no PG grade change in the virgin binder is necessary for mixtures containing less than 15% RAP; mixtures containing 16 to 25% RAP require the new asphalt binder to be one grade lower than the 75 76 grade required for a virgin asphalt binder, and mixtures with over 25% RAP need to select the new asphalt binder using a blending chart. Table 4.11. Performance Grades of RAP Blends. Performance Grade RAP Source % RAP blend Actual MP 1 +4 0 10 20 30 PG 78.9-24.4 PG 78.3-26.4 PG 78.5-25.0 PG 80.7-22.6 PG 76-22 PG 76-22 PG 76-22 PG 76-22 -4 0 10 20 30 PG 78.9-24-4 PG 79.5-22.6 PG 80.1-22.8 PG 81.5-19.8 PG 76-22 PG 76-22 PG 76-22 PG 76-16 DG1 0 10 20 30 PG 78.9-24.4 PG 79.5-23.5 PG 80.1-23.5 PG 82.2-22.7 PG 76-22 PG 76-22 PG 76-22 PG 82-22 DG2 0 10 20 30 PG 78.9-24.4 PG 81.2-25.8 PG 82.6-25.1 PG 83.8-24.1 PG 76-22 PG 76-22 PG 82-22 PG 82-22 The trends of binder properties obtained with the DSR and BBR tests in the range of 0% to 30% RAP binder in the blend were analyzed for the different mixtures in this study. The rates of change of the properties in that range were calculated to assess the impact of the RAP asphalt content in the blends. These rates were computed as the change in the binder property divided by the change in RAP content, for the entire range studied. The results are discussed below. DSR Results Results for the rutting and fatigue parameters were analyzed for original (unaged) blends, RTFO-aged and RTFO+PAV aged blends at failing and passing temperatures. Estimated and actual critical temperatures were compared for original and aged blends. Figure 4.4 shows that, as expected, the rutting parameter G*/sin? in the original blends was higher at the low temperature and increased with the addition of RAP binder because the old binder makes the resulting blends stiffer. The rates of increase were also higher at the low temperature (Table 4.12), with the biggest rates being that of +4 and DG2 RAP mixtures, but they were not significant for the range of RAP binder percentages used in this study (0-30%). +4 RAP 0.0 0.5 1.0 1.5 2.0 2.5 3.0 102030 % RAP Blend G*/sin ? (kP a ) 76 ?C 82 ?C -4 RAP 0.0 0.4 0.8 1.2 1.6 2.0 10203 % RAP Blend G* / s in ? (k P a ) 0 76 ?C 82 ?C DG1 RAP 0.0 0.4 0.8 1.2 1.6 2.0 2.4 0102030 DG2 RAP 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 01020 % RAP Blend G*/sin ? (kPa) 30 76 ?C 82 ?C % RAP Blend G * /s in ? (k Pa) 76 ?C 82 ?C Figure 4.4. G*/sin? Trends for Unaged RAP Blends. 77 78 Table 4.12. Rate of Increase in G*/sin? for Unaged Blends. Increase Rate (kPa/%RAP) RAP Type 76?C 82?C +4 0.041 0.020 -4 0.018 0.010 DG1 0.021 0.012 DG2 0.086 0.064 The results for RFTO-aged blends, shown in Figure 4.5 were similar to those of unaged blends. Lower temperature and higher RAP binder percentages increased G*/sin?, but not at a high rate, as shown in Table 4.13 (increase rates smaller than 0.05kPa/%RAP). This suggests that even though the rutting parameter increases with the addition of RAP for both original and RTFO-aged blends, the rates of change are so small that the rutting resistance of the mixes is not likely to be significantly improved. +4 RAP 0.0 1.0 2.0 3.0 4.0 0102030 % RAP Blend G*/sin ? (kPa) 76 ?C 82 ?C -4 RAP 0.0 1.0 2.0 3.0 4.0 0102030 % RAP Blend G*/sin ? (kPa) 76 ?C 82 ?C DG1 RAP 0.0 1.0 2.0 3.0 4.0 0 10203 % RAP Blend G*/sin ? (kPa ) 0 76 ?C 82 ?C DG2 RAP 0.0 1.0 2.0 3.0 4.0 5.0 0102030 % RAP Blend G*/sin ? (kP a ) 76 ?C 82 ?C Figure 4.5. G*/sin? Trends for RTFO-Aged RAP Blends. Table 4.13. Rate of Increase in G*/sin? for RTFO-Aged Blends. Increase Rate (kPa/%RAP) RAP Type 76?C 82?C +4 0.016 0.007 -4 0.037 0.011 DG1 0.027 0.016 DG2 0.046 0.031 For the RFTO and PAV residues, the fatigue parameter G*sin? increased with the addition of RAP binder (Figure 4.6). The results were considerably higher the lower temperature. Unlike the trends for the unaged and RTFO-aged blends, the rate of increase was much more significant in this case (over 25 kPa/%RAP), especially for the +4 and -4 RAP blends, where the rate of increase is as high as 68.4 kPa/%RAP (Table 4.14). 79 These results indicate that the fatigue resistance of the binder blends is highly influenced by temperature and RAP content. Addition of RAP binder may result in mixes more susceptible to fatigue cracking, especially at high temperatures. +4 RAP 2,500 3,500 4,500 5,500 6,500 102030 % RAP Blend G*s i n ? (k Pa) 25 ?C 22 ?C -4 RAP 2,000 3,000 4,000 5,000 6,000 7,000 0 10203 % RAP Blend G*sin ? (k Pa) 0 25 ?C 22 ?C DG1 RAP 2,500 3,500 4,500 5,500 6,500 102030 DG2 RAP 2,000 3,000 4,000 5,000 6,000 10203 % RAP Blend G*s i n ? (k Pa) 0 25 ?C 22 ?C % RAP Blend G*sin ? (kP a ) 25 ?C 22 ?C Figure 4.6. G*sin? Trends for RTFO+PAV-Aged RAP Blends. Table 4.14. Rate of Increase in G*sin? for RTFO and PAV-Aged Blends. Increase Rate (kPa/%RAP) RAP Type 22?C 25?C +4 58.9 43.9 -4 68.4 53.1 DG1 46.2 35.1 DG2 32.4 25.0 80 BBR Results Results for creep stiffness and creep rate obtained with the bending beam rheometer were analyzed for the asphalt blends at failing and passing temperatures. Figure 4.7 shows that the creep stiffness of the blends increased with lower temperature and addition of RAP binder, although it did it at a slow rate, as seen in Table 4.15 (maximum 2.85 MPa/%RAP). In most cases, the increase rate was higher at the lower temperature (Table 4.15). The results suggest that addition of RAP is not highly influential for the creep stiffness of the binder blend at the temperatures studied (low rates of increase), but could become more significant at lower temperatures. +4 RAP 100 150 200 250 300 102030 % RAP Blend Cree p St i ffne ss (MPa) -12 ?C -18 ?C -4 RAP 100 150 200 250 300 10203 % RAP Blend Cree p St i ffne ss (MPa) 0 -12 ?C -18 ?C DG1 RAP 100 150 200 250 300 350 0102030 DG2 RAP 100 150 200 250 300 350 01020 % RAP Blend C r eep Stiffness (MP a ) 30 -12 ?C -18 ?C % RAP Blend C r eep Stiffness (MP a ) -12 ?C -18 ?C Figure 4.7. Creep Stiffness Trends for RAP Blends. 81 82 Table 4.15. Rate of Increase in Creep Stiffness. Increase Rate (MPa/%RAP) RAP Type -18?C -12?C +4 0.150 0.733 -4 1.250 1.233 DG1 2.850 0.767 DG2 1.917 0.200 Figure 4.8 shows that, as expected, higher RAP binder percentages and low temperature resulted in lower creep rates. The decrease rate for these values was very small (less than 0.0011 MPa/%RAP) and had a maximum variation of 0.008 MPa/%RAP between temperatures (Table 4.16). As with creep stiffness, the small rates of change suggest that increasing RAP content does not decrease the creep rate significantly, and that the thermal cracking resistance of the recycled binder blends will not be affected. +4 RAP 0.20 0.25 0.30 0.35 0.40 0102030 % RAP Blend Cree p r a te -12 ?C -18 ?C -4 RAP 0.20 0.25 0.30 0.35 0.40 01020 % RAP Blend Cre e p rate 30 -12 ?C -18 ?C DG1 RAP 0.20 0.25 0.30 0.35 0.40 0102030 DG2 RAP 0.20 0.25 0.30 0.35 0.40 01020 % RAP Blend Cree p r a te 30 -12 ?C -18 ?C % RAP Blend Cree p r a te -12 ?C -18 ?C Figure 4.8. Creep Rate Trends for RAP Blends. Table 4.16. Rate of Decrease in Creep Rate. Decrease Rate (MPa/%RAP) RAP Type -18?C -12?C +4 -0.0007 -0.0007 -4 -0.0007 -0.0011 DG1 -0.0006 -0.0007 DG2 -0.0009 -0.0001 4.2 MIX DESIGNS Table 4.17 shows the virgin asphalt content, total asphalt content (virgin binder plus RAP binder), voids in the mineral aggregate (VMA), and voids filled with asphalt (VFA). The mix design information for all mixtures is presented in Appendix A. The VMA values 83 84 were calculated using the effective specific gravity (G se ) of the aggregate blends, as specified by GDOT. It can be observed that the results for VMA and VFA do not change significantly with the addition of RAP or with the types of RAP used. Table 4.17. Volumetric Properties of RAP Mixtures. Aggregate Source % RAP RAP Type Total AC, % VMA, % VFA, % 0 +4 6.2 18.3 74.6 10 DG2 6.1 18.2 77.2 20 -4 6.8 19.3 80.3 Mt. View 30 DG1 6.3 18.3 78.9 0 -4 6.2 18.1 76.5 10 DG1 6.0 17.6 76.8 20 +4 6.4 18.0 79.4 Lithia Springs 30 DG2 6.0 17.3 77.1 0 DG1 6.9 19.6 78.8 10 -4 6.4 18.9 76.4 20 DG2 6.9 19.4 79.0 Camak 30 +4 7.3 19.8 82.8 0 DG2 6.2 18.5 76.7 10 +4 5.9 18.0 76.4 20 DG1 6.6 18.9 79.8 Ruby 30 -4 5.8 17.6 76.5 One parameter that can be useful to evaluate the impact of RAP addition on mixture performance is the ratio of old binder to virgin binder, shown in Table 4.18. Figure 4.9 shows that as RAP content increases, the ratio of old binder to virgin binder increases as well. When grouped by RAP type, mixes that contained -4 RAP had the highest ratio (0.19) due to the higher asphalt content present in -4 RAP (6.2%). For aggregate sources, mixes that contained Ruby aggregates had the highest ratio (0.18). Figure 4.10 shows the virgin and RAP asphalt binder contents for each mixture. Table 4.18. Virgin and RAP Binder Contents for SMA Mixes. RAP Type % RAP Agg. Source Virgin AC, % Old AC, % Old AC/ Virgin AC 0 Camak 6.9 0.0 0.00 10 Ruby 5.5 0.4 0.08 20 Lithia Springs 5.5 0.9 0.16 +4 30 Mt. View 5.0 1.3 0.26 0 Lithia Springs 6.2 0.0 0.00 10 Mt. View 5.5 0.6 0.11 20 Camak 5.7 1.2 0.21 -4 30 Ruby 4.0 1.8 0.45 0 Mt. View 6.2 0.0 0.00 10 Lithia Springs 5.5 0.5 0.10 20 Ruby 5.5 1.1 0.19 DG1 30 Camak 5.7 1.6 0.28 0 Ruby 6.2 0.0 0.00 10 Camak 6.0 0.4 0.07 20 Mt. View 6.0 0.8 0.14 DG2 30 Lithia Springs 4.7 1.3 0.27 y = 0.0099x R 2 = 0.8652 0.00 0.10 0.20 0.30 0.40 0.50 0102030 % RAP R A P B i nde r / V i r g i n B i n der Figure 4.9. Old to New Asphalt Ratio vs RAP Content. 85 86 Table 4.19 illustrates the savings in virgin binder content for all RAP types. These savings were calculated as a percent of reduction on virgin binder compared to the control mix. Since there are four different mixes with 0% RAP (one for each aggregate source), these percentages were computed by matching the recycled SMA mix with the control mix that contained the same aggregate source. For example, the mix that contains 10% -4 RAP and Camak aggregates was compared to the control mix that contains Camak aggregates; while the mix that contains 20% -4 RAP and Mt. View aggregates was compared to the control mix that contains Mt. View aggregates, and so on. At 10% and 20% RAP, the reduction in the required virgin binder is very similar (averages of 11.7% and 10.8%, respectively). Normally, the required amount of virgin binder will decrease with RAP content, and these results can be attributed to variability in mix design. At 30% RAP the savings increase dramatically to an average of 24.1%, which would represent an important economical benefit, since asphalt cement is the most expensive component of an HMA mix. This benefit is particularly important for mixtures containing -4 RAP, where the virgin binder required is reduced in up to 35.5%. 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 0102030010203001020300102030 +4 -4 DG1 DG2 RAP Content, % % Bi nder Virgin AC RAP AC Figure 4.10. Asphalt Contents for SMA Mixtures. Table 4.19. Savings in Virgin Binder (%) for RAP Mixtures. RAP Type % RAP DG1 DG2 -4 +4 Average 10 11.3 13.0 11.3 11.3 11.7 20 11.3 3.2 17.4 11.3 10.8 30 17.4 24.2 35.5 19.4 24.1 4.3 PERFORMANCE TESTS 4.3.1 Moisture Susceptibility Table 4.20 and Figure 4.11 show the wet (conditioned) and dry (unconditioned) strength values for the SMA mixes with recycled material. An analysis of variance (shown in Table 4.21) indicated that at 95% confidence level, the amount of RAP significantly influences the tensile strength (p-values < 0.001), while type of RAP has a significant effect (p-value = 0.015 for dry strength and 0.019 for wet strength). This was expected 87 88 because increasing the RAP content increases the amount of old binder (binder from RAP) and makes the mixture stiffer, which has an effect on the tensile strength and bond to the aggregates. The analysis of variance also indicated that the interaction between aggregate source and RAP type is significant for both conditioned and unconditioned tensile strengths (p-values < 0.001). Table 4.20. Tensile Strengths for SMA Mixtures. RAP Type % RAP Agg. Source Unconditioned Tensile Strength (psi) Conditioned Tensile Strength (psi) 0 Camak 87.7 78.7 10 Ruby 103.6 92.3 20 Lithia Springs 87.0 90.8 +4 30 Mt. View 105.7 99.8 0 Lithia Springs 89.0 70.9 10 Mt. View 87.2 88.2 20 Camak 98.5 85.7 -4 30 Ruby 144.7 141.7 0 Mt. View 71.6 71.6 10 Lithia Springs 83.6 79.7 20 Ruby 83.1 83.6 DG1 30 Camak 124.8 118.1 0 Ruby 69.7 70.2 10 Camak 85.1 72.3 20 Mt. View 101.9 91.9 DG2 30 Lithia Springs 94.8 95.8 0 20 40 60 80 100 120 140 160 0 102030 0 102030 0 102030 0 102030 +4 -4 DG1 DG2 % RAP S t reng th , psi Wet Strength Dry Strength Figure 4.11. Strength Values from Moisture Susceptibility Test. Table 4.21. Analysis of Variance for Tensile Strengths. Unconditioned Conditioned Factor F-statistic p-value F-statistic p-value Agg. Source 2.24 0.099 3.18 0.035 RAP Content 18.21 0.000 33.61 0.000 RAP Type 3.96 0.015 3.73 0.019 Agg. Source x RAP Type 16.33 0.000 33.90 0.000 Figure 4.12 shows that strength increased as the percentage of RAP increased. This is not surprising since recycled SMA mixtures contain RAP binder that would be expected to have an effect on the mixture properties of the samples, such as increased stiffness. Table 4.22 shows the differences in tensile strengths for various RAP contents and confirms that the higher the RAP content, the more significant those differences become. This can be attributed to the higher old to new binder ratio for mixtures with higher RAP contents. 89 y = 1.1648x + 77.427 R 2 = 0.8746 y = 1.2789x + 70.266 R 2 = 0.8956 0 20 40 60 80 100 120 140 102030 % RAP Tensile S t re n g th, psi Unconditioned Conditioned Unconditioned Conditioned Figure 4.12. Effect of RAP Percentage on Tensile Strength. Table 4.22. Tensile Strength Comparisons for SMA Mixes with Various RAP Contents. Unconditioned Conditioned % RAP Difference of Means (psi) Are differences significant at 95% confidence level? Difference of Means (psi) Are differences significant at 95% confidence level? 10.30 13.03 37.91 No No Yes 10.28 15.16 41.01 No Yes Yes 2.74 27.62 No Yes 4.88 30.72 No Yes 0 ? 0 ? 0 ? 10 ? 10 ? 20 ? 10 20 30 20 30 30 24.88 Yes 25.85 Yes The pairwise comparisons for tensile strengths among RAP types are shown in Table 4.23, where it can be observed that the differences among RAP types are not significant, except between mixes containing -4 RAP and mixes containing DG2 RAP. Figure 4.13 shows that recycled mixtures using -4 RAP had the highest tensile strengths (96.6 psi for conditioned and 104.9 psi for unconditioned specimens), which was 90 expected because as mentioned earlier, these mixtures have the highest old to new binder ratio (an average of 0.19), and this increases stiffness. Recycled mixtures using DG2 RAP, have the lowest old to new binder ratio (an average of 0.12) which produces a softer binder blend, and this caused them to have the lowest tensile strengths (82.6 psi for conditioned and 87.9 psi for unconditioned specimens). Table 4.23. Tensile Strength Comparisons for SMA Mixes with Various RAP Types. Unconditioned Conditioned RAP Type Difference of Means (psi) Are differences significant at 95% confidence level? Difference of Means (psi) Are differences significant at 95% confidence level? -2.89 14.18 5.26 No No No -5.69 8.38 2.14 No No No 17.07 8.16 Yes No 14.07 7.83 Yes No DG1 ? DG1 ? DG1 ? DG2 ? DG2 ? -4 ? DG2 -4 +4 -4 +4 +4 -8.92 No -6.24 No 0 20 40 60 80 100 120 140 DG1 DG2 -4 +4 RAP Type Strength, psi Conditioned Unconditioned Figure 4.13. Strength Values for RAP Types. 91 92 Tables 4.24 and 4.25 show the difference of mean tensile strengths for the interaction term for conditioned and unconditioned specimens. The results are consistent with the trend observed for the individual factors: mixtures with higher old to new binder ratio had higher tensile strengths, such as mixtures containing Ruby aggregates and -4 RAP, which had an average ratio of 0.45 and tensile strengths up to 75.0 psi higher for unconditioned and 71.5 psi higher for conditioned specimens. Table 4.24. Unconditioned Tensile Strengths Comparisons for Aggregate Source ? RAP Type Interaction. Difference of Means (psi) RAP Type Mt. View Lithia Spr. Camak Ruby 30.3 15.6 34.2 11.2 5.5 3.5 -39.7 -26.3 -37.1 -13.4 61.6 20.5 -14.7 3.8 -5.4 -7.7 13.3 2.6 75.0 33.9 DG1 ? DG1 ? DG1 ? DG2 ? DG2 ? -4 ? DG2 -4 +4 -4 +4 +4 18.5 -2.4 -10.8 -41.4 Significant differences at 95% confidence level are in bold. Table 4.25. Conditioned Tensile Strengths Comparisons for Aggregate Source ? RAP Type Interaction. Difference of Means (psi) RAP Type Mt. View Lithia Spr. Camak Ruby 20.3 16.6 28.2 16.1 -8.8 11.1 -45.8 -32.4 -39.4 -13.4 58.1 8.7 -3.7 7.9 -24.9 -5.0 13.4 6.3 71.5 22.1 DG1 ? DG1 ? DG1 ? DG2 ? DG2 ? -4 ? DG2 -4 +4 -4 +4 +4 11.7 19.9 -7.0 -49.4 Significant differences at 95% confidence level are in bold. Tables 4.26 through 4.30 summarize the moisture susceptibility results. All mixtures contained between 0.8 and 1.0 percent of lime by total weight of mix. A minimum TSR of 0.80 is generally required by GDOT for SMA mixtures. However, a 93 TSR of 0.7 may be acceptable so long as all individual test values exceed 100 psi. All mixtures met or exceeded the 0.8 minimum retained strength. Table 4.26. Moisture Susceptibility Results for Control Mixes. Measurement Mt. View Lithia Springs Camak Ruby Unconditioned Samples % Air voids 7.2 7.1 7.1 6.5 Load, lbs 1,124 1,397 1,377 1,094 Dry S, psi 71.6 89.4 87.7 69.7 Conditioned Samples % Air voids 7.2 7.7 7.7 6.9 Load, lbs 1,124 1,114 1,236 1,103 Wet S, psi 71.6 70.9 78.7 70.2 % Saturation 77.4 75.0 54.2 64.4 TSR 1.00 0.80 0.90 1.01 Table 4.27. Moisture Susceptibility Results for SMA Mixes Using +4 RAP. +4 RAP Measurement 10% Ruby 20% Lithia Springs 30% Mt. View Unconditioned Samples % Air voids 6.8 7.3 6.9 Load, lbs 1,628 1,367 1,661 Dry S, psi 103.6 87.0 105.7 Conditioned Samples % Air voids 6.8 7.3 6.8 Load, lbs 1,450 1,426 1,568 Wet S, psi 92.3 90.8 99.8 % Saturation 85.6 86.0 71.7 TSR 0.89 1.04 0.94 94 Table 4.28. Moisture Susceptibility Results for SMA Mises Using -4 RAP. -4 RAP Measurement 10% Mt. View 20% Camak 30% Ruby Unconditioned Samples % Air voids 6.9 6.6 6.5 Load, lbs 1,370 1,547 2,273 Dry S, psi 87.2 98.5 144.7 Conditioned Samples % Air voids 6.9 7.0 6.4 Load, lbs 1,385 1,346 2,226 Wet S, psi 88.2 85.7 141.7 % Saturation 78.4 76.0 82.6 TSR 1.01 0.87 0.98 Table 4.29. Moisture Susceptibility Results for SMA Mixes Using DG1 RAP. DG1 RAP Measurement 10% Lithia Springs 20% Ruby 30% Camak Unconditioned Samples % Air voids 7.1 6.2 6.6 Load, lbs 1,313 1,305 1,960 Dry S, psi 83.6 83.1 124.8 Conditioned Samples % Air voids 7.1 6.2 6.6 Load, lbs 1,252 1,314 1,855 Wet S, psi 79.7 83.6 118.1 % Saturation 75.1 90.0 61.4 TSR 0.95 1.01 0.95 Table 4.30. Moisture Susceptibility Results for SMA Mixes Using DG2 RAP. DG2 RAP Measurement 10% Camak 20% Mt.View 30% Lithia Springs Unconditioned Samples % Air voids 6.9 6.7 6.9 Load, lbs 1,337 1,601 1,488 Dry S, psi 85.1 101.9 94.8 Conditioned Samples % Air voids 6.9 6.8 6.9 Load, lbs 1,136 1,443 1,504 Wet S, psi 72.3 91.9 95.8 % Saturation 90.8 75.2 79.9 TSR 0.85 0.90 1.01 95 The results for the analysis of variance shown in Table 4.31 indicated that the TSR values do not change significantly with the variations in RAP content (p-value = 0.682). The analysis of variance indicated that none of the main factors had a significant effect on TSR. Table 4.32 shows that the average TSR values range from 0.93 for control mixes to 0.97 for mixes with 30% RAP content, all well above the minimum specified by GDOT. The differences in TSR are shown in Table 4.33, where it can be seen that the increase is not significant (less than 0.05). Table 4.31. Analysis of Variance for TSR. Factor F-statistic p-value Agg. Source 1.38 0.265 RAP Content 0.50 0.682 RAP Type 0.62 0.605 R 2 = 16.49% Table 4.32. Average TSR Values for Various RAP Contents. RAP Content, % TSR 0 0.93 10 0.93 20 0.96 30 0.97 Table 4.33. Tensile Strength Ratios Comparisons for RAP Contents. RAP Content, % Difference of Means Are differences significant at 95% confidence level? 0.002 0.033 0.044 No No No 0.031 0.042 No No 0 ? 0 ? 0 ? 10 ? 10 ? 20 ? 10 20 30 20 30 30 0.011 96 4.3.2 Rutting Susceptibility Table 4.34 shows the APA results. An analysis of variance (Table 4.35) indicated that aggregate source, RAP type and the interaction between RAP content and RAP type were significant factors, while RAP content did not have an effect on rutting susceptibility (p- value = 0.720). Table 4.36 shows that the average rut depths for different RAP contents range from 3.1 mm to 3.4 mm, all below the maximum 5 mm criteria specified by GDOT. It is probable that because the rut depths were already low, RAP content did not have a significant impact on rutting performance. Table 4.34. Rutting Susceptibility Results for RAP Mixtures. Aggregate Source RAP Type % RAP Rut depth, mm DG1 0 3.11 -4 10 3.23 DG2 20 4.41 Mt. View +4 30 1.70 -4 0 2.37 DG1 10 2.00 +4 20 2.44 Lithia Springs DG2 30 4.50 +4 0 3.67 DG2 10 1.96 -4 20 1.48 Camak DG1 30 3.25 DG2 0 3.58 +4 10 5.16 DG1 20 5.38 Ruby -4 30 3.85 97 Table 4.35. Analysis of Variance for Rut Depths. Factor F-statistic p-value Agg. Source 14.36 0.000 RAP Content 0.45 0.720 RAP Type 2.86 0.038 RAP Content x RAP Type 15.24 0.000 Table 4.36. Average Rut Depths for Various RAP Contents. RAP Content, % Rut Depth, mm 0 3.18 10 3.09 20 3.43 30 3.32 The analysis of variance (Table 4.35) suggests that rutting is highly affected by the aggregate source and the interaction between RAP content and RAP type (p-values < 0.001). RAP type also has a significant impact on rutting performance (p-value = 0.038). Figure 4.14 and Table 4.37 show that mixtures that contained Ruby aggregates had higher rut depths than the rest of the mixtures (up to 1.9 mm higher), but these results still met the design criteria and therefore there is not a practical difference in terms of rutting performance. Table 4.37. Differences in Rut Depth for Aggregate Sources. Aggregate Source Difference of Means (mm) Are differences significant at 95% confidence level? -0.285 -0.522 1.381 No No Yes -0.237 1.665 No Yes Mt. View ? Mt. View ? Mt. View ? Lithia Springs ? Lithia Springs ? Camak ? Lithia Springs Camak Ruby Camak Ruby Ruby 1.902 Yes 0.0 1.0 2.0 3.0 4.0 5.0 6.0 Mt. View Lithia Springs Camak Ruby Aggregate Source Ru t Dep t h , mm Figure 4.14. Effect of Aggregate Source on Rut Depth. Figure 4.15 shows the average rut depths for various RAP types. There is no significant difference in the results among RAP types, except between mixtures with -4 RAP and mixtures with DG2 RAP (Table 4.38), which would be expected because of the difference in old to new binder ratio (0.07 higher for mixes containing -4 RAP). However, all mixtures had average rut depths below 5 mm, which means that rutting performance was not really affected by changing the RAP type. Table 4.38. Rut Depth Comparisons for RAP Types. RAP Type Difference of Means (mm) Are differences significant at 95% confidence level? 0.178 -0.702 -0.192 No No No -0.880 -0.370 Yes No DG1 ? DG1 ? DG1 ? DG2 ? DG2 ? -4 ? DG2 -4 +4 -4 +4 +4 0.510 No 98 0.0 1.0 2.0 3.0 4.0 5.0 6.0 DG1 DG2 -4 +4 RAP Type Ru t Dep t h , mm Figure 4.15. Average Rut Depths for Various RAP Types. Table 4.39 shows the differences in rut depth for the interaction between RAP content and RAP type. In general, specimens that contained RAP had no significant differences in rut depth when compared to the control mixtures (0% RAP). Two mixtures, one containing 20% DG1 RAP and one containing 10% +4 RAP, exhibited rut depths significantly higher (up to 3.5 mm higher than other mixtures) that also exceeded the 5 mm maximum criteria (Table 4.34). The fact that mixes with these RAP types only affected rutting performance at a particular RAP content can be attributed to test variability. 99 100 Table 4.39. Rut Depths for RAP Content ? RAP Type Interaction. Difference of Means (mm) % RAP DG1 DG2 -4 +4 -1.11 2.26 0.14 -1.62 0.82 0.91 0.86 -0.89 1.48 1.49 -1.23 -1.97 3.38 1.25 2.45 2.54 -1.75 0.62 -2.72 -3.46 0 ? 0 ? 0 ? 10 ? 10 ? 20 ? 10 20 30 20 30 30 -2.13 0.09 2.37 -0.74 Significant differences at 95% confidence level are in bold. 4.3.3 Indirect Tensile Creep Compliance As mentioned in Section 3.3.4, the creep compliance test is used to evaluate thermal cracking resistance of the mixtures. Figures 4.16 through 4.19 show the creep compliance results at 50 seconds. It is clear that, as expected, creep compliance increases with temperature. However, there is not a clear relationship between creep compliance and RAP content in these data. Addition of RAP does not clearly change the stiffness of the mix, as characterized by the IDT creep test, suggesting that the low temperature performance would not be affected by the higher RAP content. +4 RAP 0.0E+00 1.0E-06 2.0E-06 3.0E-06 4.0E-06 5.0E-06 6.0E-06 7.0E-06 8.0E-06 9.0E-06 0102030 % RAP C r e e p C o mpl i ance @ t = 50 sec , 1/ psi -20?C -10?C 0?C Figure 4.16. Effect of RAP Content on Creep Compliance for Recycled Mixes Using +4 RAP. -4 RAP 0.0E+00 5.0E-07 1.0E-06 1.5E-06 2.0E-06 2.5E-06 3.0E-06 3.5E-06 4.0E-06 4.5E-06 0102030 % RAP C r e e p C o mpl i ance @ t = 50 sec , 1/ psi -20?C -10?C 0?C Figure 4.17. Effect of RAP Content on Creep Compliance for Recycled Mixes Using -4 RAP . 101 DG1 RAP 0.0E+00 1.0E-06 2.0E-06 3.0E-06 4.0E-06 5.0E-06 6.0E-06 7.0E-06 8.0E-06 0102030 % RAP C r e e p C o mpl i ance @ t = 50 sec , 1/ psi -20?C -10?C 0?C Figure 4.18. Effect of RAP Content on Creep Compliance for Recycled Mixes Using DG1 RAP. DG2 RAP 0.0E+00 2.0E-06 4.0E-06 6.0E-06 8.0E-06 1.0E-05 1.2E-05 0102030 % RAP C r e e p C o mpl i ance @ t = 50 sec , 1/ psi -20?C -10?C 0?C Figure 4.19. Effect of RAP Content on Creep Compliance for Recycled Mixes Using DG2 RAP. 102 +4 RAP 1.0E-07 1.0E-06 1.0E-05 1.0E-04 1.0E-03 1.0E-02 1.0E-01 1.0E+00 1.0E+00 1.0E+02 1.0E+04 1.0E+06 1.0E+08 1.0E+10 Reduced Time, s Cre e p Compliance, 1/psi Control 10% RAP 20% RAP 30% RAP -4 RAP 1.0E-07 1.0E-06 1.0E-05 1.0E-04 1.0E-03 1.0E-02 1.0E-01 1.0E+00 1.0E+00 1.0E+02 1.0E+04 1.0E+06 1.0E+08 1.0E+10 Reduced Time, s Cre e p Compliance, 1/psi Control 10% RAP 20% RAP 30% RAP Regular RAP 1.0E-07 1.0E-06 1.0E-05 1.0E-04 1.0E-03 1.0E-02 1.0E-01 1.0E+00 1.0E+00 1.0E+02 1.0E+04 1.0E+06 1.0E+08 1.0E+10 Reduced Time, s C r eep Compliance , 1/psi Control 10% RAP 20% RAP 30% RAP SMA RAP 1.0E-07 1.0E-06 1.0E-05 1.0E-04 1.0E-03 1.0E-02 1.0E-01 1.0E+00 1.0E+00 1.0E+02 1.0E+04 1.0E+06 1.0E+08 1.0E+10 Reduced Time, s C r eep Compliance , 1/psi Control 10% RAP 20% RAP 30% RAP Figure 4.20. Creep Compliance Master Curves for RAP Mixtures. The other parameter that measures thermal cracking resistance is the m-value. Figure 4.20 shows the creep compliance master curves used to calculate the m-value for each mix (Table 4.40). The m-value is obtained by fitting a power law through the master compliance curve obtained from the indirect tensile creep tests, following the procedure described in Section 3.3.4. Mixtures with higher m-values tend to have greater resistance to thermal cracking. 103 104 Table 4.40. m-values for SMA Mixes. RAP Type % RAP Agg. Source m-value 0 Camak 0.440 10 Ruby 0.548 20 Lithia Springs 0.484 +4 30 Mt. View 0.372 0 Lithia Springs 0.339 10 Mt. View 0.410 20 Camak 0.518 -4 30 Ruby 0.331 0 Mt. View 0.574 10 Lithia Springs 0.320 20 Ruby 0.630 DG1 30 Camak 0.687 0 Ruby 0.598 10 Camak 0.599 20 Mt. View 0.386 DG2 30 Lithia Springs 0.432 Table 4.41 shows that the m-value is not influenced by RAP content (p-value = 0.552) but is significantly affected by the aggregate source and RAP type. The influence of aggregate source is somewhat unexpected because thermal cracking resistance is not really defined by the aggregate properties, and is instead dictated by the binder properties. This may be attributed to the difference in mean old to new binder ratios among mixtures with different aggregate sources. Figure 4.21 shows the m-values as a function of RAP content and confirms that there is not a strong relationship between the two (R 2 = 0.29). Even though the m-value appears to decrease with RAP content (higher old to new binder ratios), the poor correlation between the two variables is not sufficient to conclude that addition of RAP would affect thermal cracking potential significantly. Table 4.41. Analysis of Variance for m-value. Factor F-statistic p-value Agg. Source 5.16 0.004 RAP Content 0.71 0.552 RAP Type 3.49 0.025 y = -0.0013x + 0.493 R 2 = 0.2948 0.430 0.450 0.470 0.490 0.510 0102030 % RAP m-value Figure 4.21. Average m-values for RAP Mixtures. Tables 4.42 and 4.43 show all pairwise comparisons among levels for aggregate source and RAP type, and the effect of these variables on the m-value is shown in Figures 4.22 and 4.23. It can be observed that in general, Camak and Ruby mixtures had m- values up to 0.17 higher than Mt. View and Lithia Springs mixtures. Mixtures containing DG1 RAP performed better than only -4 RAP mixtures (Table 4.43). This may be attributed to the higher amount of RAP binder present in -4 RAP mixtures and its rheological properties, which as noted earlier, indicated these binder blends were more sensitive to thermal cracking. 105 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Mt. View Lithia Springs Camak Ruby Aggregate Source m- val ue Figure 4.22. Effect of Aggregate Source on m-value. Table 4.42. m-value Comparisons for Aggregate Sources. Aggregate Source Difference of Means Are differences significant at 95% confidence level? Mt. View ? Mt. View ? Mt. View ? Lithia Springs Camak Ruby -0.019 0.148 0.114 No Yes No Lithia Springs ? Lithia Springs ? Camak Ruby 0.167 0.133 Yes Yes Camak ? Ruby -0.034 No 106 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 DG1 DG2 -4 +4 RAP Type m- val ue Figure 4.23. Average m-values for RAP Types. Table 4.43. m-value Comparisons for RAP Types. RAP Type Difference of Means Are differences significant at 95% confidence level? -0.049 -0.153 -0.114 No Yes No -0.104 -0.065 No No DG1 ? DG1 ? DG1 ? DG2 ? DG2 ? -4 ? DG2 -4 +4 -4 +4 +4 0.039 No 4.3.4 Flexural Beam Fatigue As described in the previous chapter, beams were tested at two strain levels to simulate different pavement structures. The high strain level (800 ??) simulates a thin pavement with weak structure or poor subgrade, while the low strain level (400 ??) simulates a thick pavement with adequate subgrade (6). Because low strain beams typically do not reach the termination stiffness in a reasonable amount of time, it was necessary to 107 108 establish a cut off point of 1,000,000 cycles, which allowed the test to be completed in a maximum time of nearly 28 hours. High Strain Results Table 4.44 shows the results for the high strain beams. An analysis of variance indicated that the number of cycles to failure is affected by the aggregate source and RAP content (p-values < 0.001) and the interaction of the two (p-value = 0.023). Figure 4.24 shows that as the amount of RAP increases, the number of cycles to failure decreases. As shown in Table 4.45, the fatigue life of the mixes is reduced as soon as RAP is added. This was expected because as the old to new binder ratio increases mixes become stiffer and tend to fail sooner in a constant strain test (Table 4.46). Table 4.44. Test Results for High Strain Beams (800 ??). Aggregate Source % RAP RAP Type Cycles to Failure Initial Stiffness (MPa) Final Stiffness (MPa) Initial Dissipated Energy (kPa) 0 DG1 45,800 2,952 1,473 0.985 10 -4 50,143 3,169 1,577 1.069 20 DG2 31,013 4,325 2,148 1.105 Mt. View 30 +4 19,880 4,315 2,150 0.964 0 -4 58,703 3,090 1,532 1.039 10 DG1 44,877 3,411 1,690 0.937 20 +4 57,940 3,179 1,583 0.965 Lithia Springs 30 DG2 16,753 4,949 2,469 0.662 0 +4 92,070 3,220 1,607 0.985 10 DG2 40,947 3,221 1,600 1.055 20 -4 71,070 3,380 1,677 1.028 Camak 30 DG1 74,760 3,433 1,710 0.717 0 DG2 72,680 3,028 1,510 0.862 10 +4 40,933 3,141 1,566 1.034 20 DG1 21,123 3,348 1,671 0.894 Ruby 30 -4 4,273 4,798 2,382 0.542 y = -1141.3x + 63555 R 2 = 0.8695 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 102030 % RAP Cycl es to Fai l u r e Figure 4.24. Number of Cycles to Failure for RAP Mixtures (800??). Table 4.45. N f Comparisons for RAP Contents (800??). RAP Content (%) Difference of Means (Cycles) Are differences significant at 5% confidence level? -23,088 -22,027 -38,397 Yes Yes Yes 1,062 -15,308 No No 0 ? 0 ? 0 ? 10 ? 10 ? 20 ? 10 20 30 20 30 30 -16,370 Yes Table 4.46. Effect of RAP Binder on Fatigue Life (800 ??). % RAP Old binder/ New binder Cycles to Failure Initial Stiffness (MPa) 0 0.00 67,313 3,073 10 0.09 44,225 3,236 20 0.17 45,287 3,558 30 0.32 28,917 4,374 Figure 4.25 shows the effect of aggregate source on fatigue life. Table 4.47 indicates that Camak mixtures reached a higher number of cycles to failure than the rest. 109 Camak blends did not have aggregate properties that would seem to significantly improve the fatigue resistance of the mixture; however, they had a low old to new binder ratio (an average of 0.12) and the specimens tested had lower air voids than the other beams (0.6 percent lower on average). Both properties are desirable to obtain greater fatigue life in HMA mixtures. 0 20,000 40,000 60,000 80,000 100,000 Camak Lithia Springs Mt. View Ruby Aggregate Source Cyc l es to F a il u r e Figure 4.25. Effect of Aggregate Source on N f (800??). Table 4.47. N f Comparisons for Aggregate Sources (800??). Aggregate Source Difference of Means (Cycles) Are differences significant at 5% confidence level? 7,859 33,002 -1,957 No Yes No 25,143 -9,816 Yes No Mt. View ? Mt. View ? Mt. View ? Lithia Springs ? Lithia Springs ? Camak ? Lithia Springs Camak Ruby Camak Ruby Ruby -34,959 Yes 110 111 Table 4.48 shows the difference of mean N f for the interaction between aggregate source and RAP content. It is important to note that the number of cycles to failure only changes significantly with RAP content for Ruby mixtures, and only at the 30% RAP level, which is expected because these mixtures have the highest old to new binder ratio (0.45). Table 4.48. N f Comparisons for RAP Content ? Aggregate Source Interaction (800 ??). Difference of Means (Cycles) % RAP Mt. View Lithia Spr. Camak Ruby 4,343 -14,787 -25,920 -13,827 -763 -41,950 -51,123 -21,000 -17,310 -31,747 -51,557 -68,407 -19,130 -30,263 13,063 -28,123 30,123 33,813 -19,810 -36,660 0 ? 0 ? 0 ? 10 ? 10 ? 20 ? 10 20 30 20 30 30 -11,133 -41,187 3,690 -16,850 Significant differences at 95% confidence level are in bold. As mentioned earlier, in a controlled-strain test, stiffer mixes are expected to fail earlier (lower number of cycles to failure). It can be observed that the results in Figure 4.26 followed the expected trend, and that the higher initial stiffnesses are related to higher RAP contents and higher old to new asphalt ratios (up to 0.32 on average, as previously shown in Table 4.46). An analysis of variance confirmed that the initial stiffness of the mixtures is influenced by RAP content (p-value < 0.001). Table 4.49 shows that the increase in initial stiffness becomes more significant at higher RAP contents due to the higher amount of old binder present (up to 1.5 percent by total weight of mix higher). y = -535.04Ln(x) + 9174.5 R 2 = 0.4454 2,000 2,500 3,000 3,500 4,000 4,500 5,000 5,500 0 20,000 40,000 60,000 80,000 100,000 120,000 Cycles to Failure In itial St i f fn e s s (MP a ) 0% 10% 20% 30% Figure 4.26. Relationship between Initial Stiffness and Number of Cycles to Failure (800 ??). Table 4.49. Initial Stiffness Comparisons for Recycled SMA with Various RAP Contents (800??). RAP Content (%) Difference of Means (MPa) Are differences significant at 95% confidence level? 163.1 485.4 1301.6 No Yes Yes 322.3 1,138.5 Yes Yes 0 ? 0 ? 0 ? 10 ? 10 ? 20 ? 10 20 30 20 30 30 816.2 Yes The initial dissipated energy is the energy required at the beginning of the test to deflect the beam, and is an indication of the susceptibility of the mixture to fatigue damage. Mixtures that are more resistant to fatigue cracking would require a higher amount of energy to produce this damage. An analysis of variance showed that RAP content significantly affects the initial dissipated energy (p-value < 0.001). However, the 112 113 only significant difference was found for 30% RAP mixes, in which the dissipated energy was up to 0.37 MPa lower (Table 4.50). This is also related to the high old to new binder ratio (an average of 0.32) that stiffens the mix and makes it more easily damaged. Table 4.50. Initial Dissipated Energy Comparisons for RAP Contents (800??). RAP Content (%) Difference of Means (kPa) Are differences significant at 95% confidence level? 0.088 0.063 -0.282 No No Yes -0.026 -0.371 No Yes 0 ? 0 ? 0 ? 10 ? 10 ? 20 ? 10 20 30 20 30 30 -0.345 Yes Fatigue cracking is a distress that initiates at the bottom of the HMA layer where the tensile stress is the highest then propagates to the surface as one or more longitudinal cracks. Since SMA mixes are typically placed at or near the surface of the pavement, fatigue cracking may not be a major concern. Low Strain Results The results for the low strain beams are shown in Table 4.49. Most tests were stopped at the cutoff point of 1,000,000 cycles, and only a few specimens reached 50% of the initial stiffness before that. The number of cycles to failure shown in Table 4.51 corresponds to the extrapolated value from the best fit curve of the N f vs stiffness plot at 50% of the initial stiffness. Because data extrapolation can be a source of error, the percent drop in initial stiffness at the end of the test (1,000,000 cycles maximum) was also included in the results. Low percentages indicate that the specimen had experienced 114 less damage at the end of the test and it was likely to withstand a higher number of cycles before reaching the failure point. As shown in Figure 4.27, there is a good correlation between the percent drop in stiffness and estimated N f (R 2 = 0.66). Table 4.51. Average Test Results for Low Strain Beams. Aggregate Source % RAP RAP Type Cycles to Failure * Initial Stiffness (MPa) % Drop in Stiffness ** Initial Dissipated Energy (kPa) 0 DG1 3,601,543 3,369 38.4 0.271 10 -4 5,153,700 3,450 37.6 0.267 20 DG2 2,338,070 3,485 35.9 0.279 Mt. View 30 +4 1,332,883 4,715 46.6 0.199 0 -4 5,526,153 3,653 36.5 0.290 10 DG1 3,686,010 3,845 39.6 0.292 20 +4 4,791,923 3,547 38.6 0.279 Lithia Springs 30 DG2 1,532,050 4,166 45.2 0.485 0 +4 4,353,263 3,187 38.3 0.267 10 DG2 4,686,913 3,571 34.7 0.438 20 -4 2,689,257 3,139 41.8 0.245 Camak 30 DG1 3,857,159 3,285 33.3 0.251 0 DG2 5,405,080 3,502 38.6 0.223 10 +4 4,552,187 3,570 39.1 0.294 20 DG1 4,460,207 3,395 39.3 0.274 Ruby 30 -4 759,820 4,621 49.9 0.317 *Extrapolated values for specimens that did not reach 50% of the initial stiffness after 1,000,000 cycles. **Measured at 1,000,000 cycles. y = -5.6204Ln(x) + 123.23 R 2 = 0.6636 20 30 40 50 60 0.0.E+00 2.0.E+06 4.0.E+06 6.0.E+06 8.0.E+06 1.0.E+07 Cycles to Failure % Dr o p i n Stiffn es s 0% 10% 20% 30% Figure 4.27. Relationship between Drop in Initial Stiffness at 1,000,000 Cycles and Estimated N f (400 ??). An analysis of variance indicated the main factor that influenced the number of cycles to failure was RAP content (p-value = 0.002). Again, this is related to the old to new binder ratio that increases with RAP content. Fatigue life is more affected at high ratios (over 0.3), where the number of cycles to failure was reduced by up to 2.8 million cycles (Table 4.52). Similar trends were obtained when using the percent drop in initial stiffness instead of N f . Table 4.52. N f and Percent Drop Comparisons for RAP Contents (400 ??). RAP Content (%) Difference of Means (Cycles) Difference of Means (%) Are differences significant at 95% confidence level? -201,808 -1,151,646 -2,851,032 -0.22 0.94 5.75 No No Yes -949,838 -2,649,224 1.17 5.98 No Yes 0 ? 0 ? 0 ? 10 ? 10 ? 20 ? 10 20 30 20 30 30 -1,699,386 4.81 No 115 As with the high strain results, N f decreases with an increase in initial stiffness. Figure 4.28 shows that in general, mixtures with 30% RAP have higher initial stiffness and lower fatigue life. An analysis of variance confirmed that initial stiffness is highly influenced by RAP content (p-value < 0.001) and aggregate source (p-value = 0.001). y = -366.78Ln(x) + 9115 R 2 = 0.3238 2,000 2,500 3,000 3,500 4,000 4,500 5,000 5,500 6,000 0.E+00 2.E+06 4.E+06 6.E+06 8.E+06 1.E+07 Cycles to Failure In i t ia l Stiffn es s (MPa ) 0% 10% 20% 30% Figure 4.28. Relationship between Initial Stiffness and N f (400 ??). Figure 4.29 and Table 4.53 show that mixtures that contained Camak aggregates had the lower stiffness among aggregate sources. As mentioned earlier, these mixtures have more virgin binder content (old to new asphalt ratio = 0.13), which is likely the cause of this result. Table 4.54 shows the difference of means for the initial stiffness for the interaction between RAP content and aggregate source. It was found that specimens that contained Lithia Springs and Camak aggregates (both with old to new asphalt ratios 116 = 0.13, lowest among aggregate sources) did not change their initial stiffness significantly with an increase in RAP content. 0 600 1,200 1,800 2,400 3,000 3,600 4,200 4,800 Camak Lithia Springs Mt. View Ruby Aggregate Source In itia l S t iffn e s s (M Pa ) Figure 4.29. Effect of Aggregate Source on Initial Stiffness (400 ??). Table 4.53. Initial Stiffness Comparisons for Aggregate Sources (400 ??). Aggregate Source Difference of Means (MPa) Are differences significant at 5% confidence level? 47.6 -459.6 16.9 No Yes No -507.2 -30.7 Yes No Mt. View ? Mt. View ? Mt. View ? Lithia Springs ? Lithia Springs ? Camak ? Lithia Springs Camak Ruby Camak Ruby Ruby 476.5 Yes 117 118 Table 4.54. Initial Stiffness Comparisons for RAP Content ? Aggregate Source Interaction (400 ??). Difference of Means (MPa) % RAP Mt. View Lithia Spr. Camak Ruby 81 116 1,346 192 -106 514 383 -48 98 68 -107 1,119 35 1,265 -298 321 -432 -286 -175 1,050 0 ? 0 ? 0 ? 10 ? 10 ? 20 ? 10 20 30 20 30 30 1,230 620 146 1,226 Significant differences at 5% confidence level are in bold. For low strain specimens, the analysis of variance indicated that the dissipated energy was not influenced by any of the main factors or interaction terms at the 95% significance level. High Strain and Low Strain Comparison As expected, the average number of cycles to failure was higher for low strain samples than for high strain samples (in the order of millions of cycles higher, as shown in Figure 4.30). Mixtures with no RAP showed the best performance (N f up to 2.5 times higher than recycled mixtures, as seen in Table 4.55) because they only contain virgin binder that is less stiff. As the RAP content is increased and more old binder goes into the mix, the fatigue life of the specimens is significantly reduced for both strain levels. This is associated with an increase in initial stiffness that causes earlier failure in controlled- strain specimens. However, this increase in stiffness may not affect performance when the mixes are placed near the top of the pavement, since fatigue cracking originates at the bottom of the HMA layer. 200 400 600 800 1000 10,000 100,000 1,000,000 10,000,000 Cycles to Failure Strai n Le vel, ?? 0% 10% 20% 30% Figure 4.30. Number of Cycles to Failure for High and Low Strain Levels. Table 4.55. Fatigue Life Comparisons for Strain Levels. % RAP Old binder/ New binder N f (400 ??) N f (800 ??) 0 0.00 67,313 4,721,510 10 0.09 44,225 4,519,703 20 0.17 45,287 3,569,864 30 0.32 28,917 1,870,478 4.3.5 Summary The results of this study have shown that adding RAP to an SMA mix has an impact in the sense that a portion of the total binder content corresponds to old (aged) binder. As the RAP content increases, this portion of binder increases as well. The main implication is that the stiffness of the resulting asphalt blend is higher than that of the virgin binder, and therefore mixture stiffness increases with RAP content. For RAPs with high asphalt content, such as the fine-graded portion of screened RAP, the old to new binder ratio is 119 higher and the effect is greater. The magnitude of the effect also depends on the properties of the RAP binder compared to the virgin binder. For the performance tests conducted in this study, the increase in stiffness caused by higher old to new asphalt ratios did not have a significant effect in most cases. Figure 4.31 shows that there is a poor correlation between old to new binder ratio and TSR (R2 = 0.08). However, all mixtures were above the minimum requirement and the TSR values increased slightly as RAP content increased; therefore moisture susceptibility was not an issue for recycled SMA mixes. y = 0.1504x + 0.9229 R 2 = 0.0768 0.60 0.70 0.80 0.90 1.00 1.10 0.00 0.10 0.20 0.30 0.40 0.50 RAP Binder/Virgin Binder TSR GDOT Minumum Required Figure 4.31. Effect of RAP Binder on TSR. Figure 4.32 shows that the rut depths were not correlated to the old to new binder ratio either (R 2 = 0.01). Still, most mixtures had average rut depths below the maximum specified by GDOT. Two recycled mixtures had results that exceeded the design criteria by no more than 0.4 mm, which could be attributed to test variability. In general, 120 mixtures had good rutting performance that was not significantly changed by the presence of old binder. y = 0.878x + 3.1287 R 2 = 0.0087 1.0 2.0 3.0 4.0 5.0 6.0 7.0 0.00 0.10 0.20 0.30 0.40 0.50 RAP Binder/Virgin Binder Rut Depth, mm GDOT Maximum Limit Figure 4.32. Effect of RAP binder on Rutting Performance. Susceptibility of the mixes to thermal cracking was poorly correlated to the old to new binder ratio (R 2 = 0.02), as shown in Figure 4.33. This could be due to the fact that even though the old binder content increases, there was not a significant change in the combined binder blend properties that control thermal cracking (creep stiffness and creep rate, shown in Figure 4.34). 121 y = -0.1386x + 0.4993 R 2 = 0.0231 0.30 0.40 0.50 0.60 0.70 0.00 0.10 0.20 0.30 0.40 0.50 RAP Binder/Virgin Binder m - val u e (cre ep co mp l i a n ce test) Figure 4.33. Effect of RAP Binder on Thermal Cracking. y = 1.45x + 269.25 R 2 = 0.3239 y = 0.6525x + 146.78 R 2 = 0.2363 0 100 200 300 400 0102030 % RAP Blend Cree p Stiffne s s (M Pa) -12?C -18?C y = -0.0009x + 0.2823 R 2 = 0.3496 y = -0.0006x + 0.3258 R 2 = 0.31 0.20 0.24 0.28 0.32 0.36 0.40 0 102030 % RAP Blend Cr e e p Ra te -12?C -18?C Figure 4.34. Effect of RAP on Low Temperature Binder Properties. One result that was affected by the increase in the old to new binder ratio was the fatigue life of the mixes. Figure 4.34 shows that adding more RAP binder to the mixtures produces lower number of cycles to failure for specimens tested in controlled-strain mode. This occurs because the stiffness of the mix increases with higher old to new asphalt ratios. 122 It was observed that the fatigue life of the mixes was significantly reduced at high strain levels. However, recycled mixes are stiffer and will have less strain. Fatigue life may not be as affected unless the mixes are used in thin pavements. Also, SMA mixes are likely to be used as a surface layer, and because fatigue cracking originates at the bottom of the HMA pavement layers, it may not be a concern for this type of mixture. y = 72064e -4.4141x R 2 = 0.5453 1,000 10,000 100,000 0.00 0.10 0.20 0.30 0.40 0.50 RAP Binder/Virgin Binder C y c l es to Failur e (8 00 ?? ) y = 5E+06e -3.6289x R 2 = 0.6491 100,000 1,000,000 10,000,000 0.00 0.10 0.20 0.30 0.40 0.50 RAP Binder/Virgin Binder C y c l es to Failur e (4 00 ?? ) Figure 4.35. Effect of RAP Binder on Fatigue Life. It is important to mention that the type of RAP used in recycled mixes can also have an important influence in performance. When the fine-graded portion of the RAP was used, the amount of old binder was increased because this portion of the RAP typically has a higher asphalt content than dense-graded or coarse graded RAP. As already discussed, this results in higher stiffness of the mix. It is expected that mixes that contain fine-graded RAP will have good resistance to moisture susceptibility and permanent deformation, but low fatigue life. Thermal cracking may not be a major concern for the reasons discussed above. One benefit of using fine-graded RAP is that the virgin binder requirement can be considerably reduced, lowering the cost of the mix. 123 124 Replacing a percentage of the No. 7 stone with coarse-graded aggregate did not affect the performance of the recycled mixes significantly. The low asphalt content characteristic of coarse-graded RAP produces asphalt blends with a low old to new asphalt ratio that did not increase the stiffness of the mixes in a way that would influence performance. Moisture susceptibility, permanent deformation and thermal cracking were not a concern for mixtures containing coarse-graded RAP. Fatigue life may be reduced depending on the RAP content used. The advantage of substituting virgin material with RAP aggregate is that recycled aggregates could be used if quarries were faced with a critical supply shortage of No. 7 stone due to its high demand, and still obtain a mix with characteristics similar to those of a virgin SMA mix. The feasibility of using RAP from reclaimed SMA could not be evaluated conclusively. The SMA RAP received for this project did not match some of the characteristics of an SMA mix (asphalt content, percent passing the No. 4 sieve). This has been partially attributed to circumstances that occurred during the milling process and the fact that the material was crushed to have 100 percent passing the 12.5 mm sieve. The resulting RAP was more similar to a dense-graded mix with low asphalt content. No general conclusions can be made unless it is assured that the same conditions would be repeated as part of a standard procedure for this type of RAP. 125 CHAPTER 5. CONCLUSIONS AND RECOMMENDATIONS The following general conclusions and recommendations were obtained from this research: ? Tests on the aggregate properties of the combined blends indicated that addition of RAP changes the LA abrasion and F/E particle content depending on the properties of the RAP aggregates in relation to the virgin aggregates. However, for the aggregates and RAPs tested the change was not significant up to 30% RAP. ? Use of RAP changed the engineering properties of the resulting binder blends due to the increased old to new binder ratio. The stiffness of the binder blend (G*/sin?, G*sin? and creep stiffness) increases with RAP content, particularly increasing the fatigue cracking potential. ? The volumetric properties of the mix (air voids, VMA and VFA) were met with all of the RAP stockpiles and various RAP contents. ? RAP content influenced only the tensile strength and fatigue life (N f ) of the mixtures. Increasing RAP content resulted in higher tensile strengths (conditioned and unconditioned) and lower number of cycles to failure. It can be concluded that RAP content significantly affects only the fatigue performance of the mixtures, especially at high strain levels. 126 ? Separating the RAP into fine and coarse-graded fractions produced two stockpiles with different properties. The fine-graded portion had a high asphalt content and therefore, produced mixes with high old to new binder ratios. Additionally, fine- graded RAP contains more material passing the No. 200 sieve, which must be accounted for during mix design. The coarse-graded portion had lower asphalt content, which indicates that mixtures containing this material will have a lower amount of old binder and less increase in stiffness. ? Use of fine-graded RAP reduced the virgin binder requirements due its high asphalt content, which translates into increased economic benefits. However, mixes that contain fine-graded graded RAP are stiffer because they have higher old to new binder ratios and are more susceptible to fatigue cracking. ? Use of coarse-graded RAP allowed reducing the No. 7 stone requirement without affecting the performance of the mixes. This may be beneficial in the case that quarries were faced with a shortage due to the high demand of this material. ? It is uncertain whether SMA pavement material can be successfully recycled back into an SMA mixture. The reclaimed asphalt used in this study had a mix gradation resembled a dense-graded mix and it had low asphalt content. Unless the same conditions are always repeated as part of a standard procedure for this material, it can not be assured that mixes containing recycled SMA will perform similar to other recycled mixes. ? Because fatigue cracking is the main concern for recycled mixtures and this distress originates at the bottom of the HMA layer, it is recommended that SMA 127 mixes containing RAP be used primarily in the top layers of the pavement. Also, a good bond between the SMA layer and the underlying material must be provided. ? Adding RAP up to 30% had little effect on the low temperature PG properties. The low temperature grade of the combined binder blends was raised by one grade on only one of the cases. This may indicate that the grade of virgin binder does not have to be adjusted to provide the desired properties. 128 REFERENCES 1. Beyond Roads, Questions and Answers. Asphalt Education Partnership. http://www.beyondroads.com/index.cfm?fuseaction=page&filename=asphaltQan dA.html. Accessed March 15 th , 2006. 2. Summary of Georgia?s Experience with Stone Matrix Asphalt Mixes. Georgia Department of Transportation. http://www.dot.state.ga.us/dot/construction/materials-research/b- admin/research/onlinereports%5Cr-SMA2002.pdf. Accessed February 21 st , 2006. 3. Brown, E.R., R.B. Mallick, J. E. Haddock and J. Bukowski. Performance of Stone Matrix Asphalt (SMA) Mixtures in the United States. National Center for Asphalt Technology, Report No. 97-01, Jan. 1997. 4. Watson, D. E. Updated Review of Stone Matrix Asphalt and Superpave Projects. In Transportation Research Record 1832, TRB, National Research Council, Washington, D.C., 2003, pp. 217-223. 5. Brown, E.R., R.B. Mallick, J. E. Haddock and T.A. Lynn. Development of a Mixture Design Procedure for Stone Matrix Asphalt (SMA). National Center for Asphalt Technology, Report No. 97-03, March 1997. 6. McDaniel, R.S., H. Soleymani, R.M. Anderson, P. Turner, and R. Peterson. Recommended Use of Reclaimed Asphalt Pavement in the Superpave Mix Design 129 Method. Contractor?s Final Report. NCHRP Web Document 30 (Project D9-12). 2000. 7. Huang, B., G. Li, D. Vukosavljevic, X. Shu, and B.K. Egan. Laboratory Investigation of Mixing Hot-Mix Asphalt with Reclaimed Asphalt Pavement. In Transportation Research Record 1929, TRB, National Research Council, Washington, D.C., 2005, pp. 37-45. 8. Kandhal, P.S. and K.Y. Foo. Designing Recycled Hot Mix Asphalt Mixtures Using Superpave Technology. National Center for Asphalt Technology, Report No. 96- 05, Jan. 1997. 9. Bukowski, J. R., Guidelines for the Design of Superpave Mixtures Containing Reclaimed Asphalt Pavement (RAP), Memorandum, ETG Meeting, FHWA Superpave Mixtures Expert Task Group, San Antonio, Texas, March, 1997. 10. Kennedy T.W., W.O. Tam, and M. Solaimanian. Effect of Reclaimed Asphalt Pavement on Binder Properties Using the Superpave System. Center for Transportation Research, Bureau of Engineering Research, The University of Texas at Austin. Research Report 1250-1, September 1998. 11. Abdulshafi, O., B. Kedzierki, and M.G. Fitch. Determination of Recycled Asphalt Pavement (RAP) Content in Asphalt Mixes Based on Expected Mixture Durability. Ohio State University. December, 2000. 12. Li X., T.R. Clyne, and M.O. Marasteanu. Recycyled Asphalt Pavement (RAP) Effects on Binder and Mixture Quality. Minnesota Department of Transportation. July, 2004. 130 13. Stroup-Gardiner M. and C. Wagner. Use of RAP in Superpave HMA Applications. In Transportation Research Record 1681, TRB, National Research Council, Washington, D.C., 1999, pp. 1-9. 14. Daniel, J.S., and A. Lachance. Mechanistic and Volumetric Properties of Asphalt Mixtures with Recycled Asphalt Pavement. In Transportation Research Record 1929, TRB, National Research Council, Washington, D.C., 2005, pp. 28-36. 15. Tam, K.K., P. Joseph, and D.F. Lynch. Five Year Experience of Low- Temperature Performance of Recycled Hot Mix. In Transportation Research Record 1362, TRB, National Research Council, Washington, D.C., 1992, pp. 56- 65. 16. McDaniel, R.S., A. Sha, G.A. Huber, and V.L. Gallivan. Investigation of Properties of Plant-Produced RAP Mixtures. Transportation Research Board Annual Meeting 2007 Paper #07-2855. 17. Huang, B., W.R. Kingery, and Z. Zhang. Laboratory Study of Fatigue Characteristics of HMA Mixtures Containing RAP. Presented at the International Symposium on Design and Construction of Long Lasting Asphalt Pavements, Auburn, AL. June, 2004. 18. Puttagunta, R., S.Y. Oloo, and A.T. Bergan. A Comparison of the Predicted Performance of Virgin and Recycled Mixes. Canadian Journal of Civil Engineering. Vol. 24, National Research Council of Canada, Ottawa. 1997. pp. 115-121. 131 19. Sargious, M. and N. Mushule. Behaviour of Recycled Asphalt Pavements at Low Temperatures. Canadian Journal of Civil Engineering. Vol. 18, National Research Council of Canada, Ottawa. 1991. pp. 428-435. 20. McGennis, R.B., S. Shuler, and H.U. Bahia. Background of Superpave Asphalt Binder Test Methods. FHWA, Report No. FHWA-SA-94-069, July 1994. 21. Roberts, F.L., P.S. Kandhal, E.R. Brown, D.Y. Lee, T.W. Kennedy. Hot Mix Asphalt Materials, Mixture Design and Construction. National Asphalt Pavement Association Research and Education Foundation, 1996. 22. Kandhal, P.S. and F. Parker. Aggregate Tests Related to Asphalt Concrete Performance in Pavements. National Cooperative Highway Research Program NCHRP Report No. 405, 1998. 23. Strategic Highway Research Program Report SHRP-A-357. Development and Validation of Performance Prediction Models and Specifications for Asphalt Binders and Paving Mixes. Strategic Highway Research Program, National Research Council, Washington D.C., 1993. 24. Kandhal, P.S. Accelerated Laboratory Rutting Tests: Evaluation of the Asphalt Pavement Analyzer. National Cooperative Highway Research Program NCHRP Report No. 508, 2003. 25. Strategic Highway Research Program Report SHRP-A-404. Fatigue Response of Asphalt-Aggregate Mixes. Strategic Highway Research Program, National Research Council, Washington D.C., 1994. 132 26. Christensen, D. Analysis of Creep Data from Indirect Tension Test on Asphalt Concrete. Asphalt Paving Technology. Volume 67, Journal of the Association of Asphalt Paving Technologists, St. Paul, MN, 1998. pp. 458-492. 133 APPENDIX A Laboratory Mix Designs 134 A.1 Mix Designs for Mt. View Mixtures Mt. View 0% RAP Aggregate Components Sieve size 007 089 M10 Marble dust Lime Blend Proportions 68.0% 12.0% 13.0% 6.0% 1.0% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 100.0 1/2" 97.0 100.0 100.0 100.0 100.0 98.0 3/8" 48.0 100.0 100.0 100.0 100.0 64.6 #4 3.0 22.0 99.0 100.0 100.0 24.6 #8 3.0 4.0 83.0 100.0 100.0 20.3 #16 2.0 2.0 66.0 100.0 100.0 17.2 #30 2.0 2.0 53.0 100.0 100.0 15.5 #50 2.0 1.0 37.0 100.0 100.0 13.3 #100 1.0 1.0 18.0 98.0 100.0 10.0 #200 1.0 1.0 6.0 90.0 100.0 8.0 Series % AC VMA VFA 1 6.0 18.3 74.6 Mt. View 10% -4 RAP Aggregate Components Sieve size 007 089 M10 Marble dust Lime RAP Blend Proportions 77.0% 0.0% 7.0% 5.0% 1.0% 10.0% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 100.0 100.0 1/2" 97.0 100.0 100.0 100.0 100.0 100.0 98.0 3/8" 48.0 100.0 100.0 100.0 100.0 100.0 64.6 #4 3.0 22.0 99.0 100.0 100.0 100.0 24.6 #8 3.0 4.0 83.0 100.0 100.0 81.0 20.3 #16 2.0 2.0 66.0 100.0 100.0 65.0 17.2 #30 2.0 2.0 53.0 100.0 100.0 53.0 15.5 #50 2.0 1.0 37.0 100.0 100.0 40.0 13.3 #100 1.0 1.0 18.0 98.0 100.0 25.0 10.0 #200 1.0 1.0 6.0 90.0 100.0 15.0 8.0 Series % AC VMA VFA 1 4.5 17.3 66.8 2 5.0 17.1 74.7 3 5.5 18.2 77.2 3.0 3.5 4.0 4.5 5.0 5.5 6.0 4.0 4.5 5.0 5.5 6.0 %AC Ai r V o i d s 16.8 17.2 17.6 18.0 18.4 4.0 4.5 5.0 5.5 6.0 %AC VM A 64.0 68.0 72.0 76.0 80.0 4.0 4.5 5.0 5.5 6.0 %AC VF A Mt. View 20% DG2 RAP Aggregate Components Sieve size 007 089 M10 Marble dust Lime RAP Blend Proportions 67.0% 7.0% 0.0% 4.8% 0.9% 20.3% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 100.0 100.0 1/2" 97.0 100.0 100.0 100.0 100.0 100.0 98.0 3/8" 48.0 100.0 100.0 100.0 100.0 95.0 64.1 #4 3.0 22.0 99.0 100.0 100.0 77.0 24.9 #8 3.0 4.0 83.0 100.0 100.0 61.0 20.4 #16 2.0 2.0 66.0 100.0 100.0 50.0 17.3 #30 2.0 2.0 53.0 100.0 100.0 42.0 15.7 #50 2.0 1.0 37.0 100.0 100.0 32.0 13.6 #100 1.0 1.0 18.0 98.0 100.0 20.0 10.4 #200 1.0 1.0 6.0 90.0 100.0 12.0 8.4 Series % AC VMA VFA 1 5.5 20.4 69.9 2 6.0 19.3 80.3 3 6.5 19.7 84.1 135 2.0 3.0 4.0 5.0 6.0 7.0 5.0 5.5 6.0 6.5 7.0 %AC Ai r V o i d s 18.8 19.2 19.6 20.0 20.4 20.8 5.05.56.06.57.0 %AC VM A 66.0 70.0 74.0 78.0 82.0 86.0 90.0 5.0 5.5 6.0 6.5 7.0 %AC VF A Mt. View 30% +4 RAP Aggregate Components Sieve size 007 089 M10 Marble dust Lime RAP Blend Proportions 58.7% 0.0% 5% 5% 0.8% 30.5% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 99.0 99.7 1/2" 97.0 100.0 100.0 100.0 100.0 96.0 97.0 3/8" 48.0 100.0 100.0 100.0 100.0 84.0 64.6 #4 3.0 22.0 99.0 100.0 100.0 37.0 23.8 #8 3.0 4.0 83.0 100.0 100.0 25.0 19.3 #16 2.0 2.0 66.0 100.0 100.0 21.0 16.7 #30 2.0 2.0 53.0 100.0 100.0 18.0 15.1 #50 2.0 1.0 37.0 100.0 100.0 15.0 13.4 #100 1.0 1.0 18.0 98.0 100.0 10.0 10.2 #200 1.0 1.0 6.0 90.0 100.0 6.2 8.1 Series % AC VMA VFA 1 5.0 18.3 78.9 2 5.5 18.9 81.8 3 6.0 19.9 82.8 136 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.5 5.0 5.5 6.0 6.5 %AC Ai r V o i d s 18.0 18.4 18.8 19.2 19.6 20.0 4.55.05.56.06.5 %AC VM A 76.0 78.0 80.0 82.0 84.0 86.0 4.5 5.0 5.5 6.0 6.5 %AC VF A A.2 Mix Designs for Lithia Springs Mixtures Lithia Springs 0% RAP Aggregate Components Sieve size 007 089 810 Marble dust Lime Blend Proportions 67.0% 13.0% 13.0% 6.0% 1.0% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 100.0 1/2" 85.0 100.0 100.0 100.0 100.0 90.0 3/8" 50.0 100.0 100.0 100.0 100.0 66.5 #4 6.0 30.0 84.0 100.0 100.0 25.8 #8 2.0 2.0 62.0 100.0 100.0 16.7 #16 1.0 2.0 50.0 100.0 100.0 14.4 #30 1.0 1.0 41.0 100.0 100.0 13.1 #50 1.0 1.0 28.0 100.0 100.0 11.4 #100 1.0 1.0 21.0 98.0 100.0 10.4 #200 1.0 1.0 10.0 90.0 100.0 8.5 Series % AC VMA VFA 1 6.0 17.8 75.8 137 Lithia Springs 10% DG1 RAP Aggregate Components Sieve size 007 089 810 Marble dust Lime RAP Blend Proportions 71.9% 0.0% 12.6% 4.5% 1.0% 10.0% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 100.0 100.0 1/2" 85.0 100.0 100.0 100.0 100.0 99.0 89.1 3/8" 50.0 100.0 100.0 100.0 100.0 93.0 63.4 #4 6.0 30.0 84.0 100.0 100.0 73.0 27.7 #8 2.0 2.0 62.0 100.0 100.0 58.0 20.6 #16 1.0 2.0 50.0 100.0 100.0 47.0 17.2 #30 1.0 1.0 41.0 100.0 100.0 38.0 15.2 #50 1.0 1.0 28.0 100.0 100.0 29.0 12.6 #100 1.0 1.0 21.0 98.0 100.0 19.0 10.7 #200 1.0 1.0 10.0 90.0 100.0 11.2 8.1 Series % AC VMA VFA 1 5.0 17.2 72.0 2 5.5 17.6 76.8 3 6.0 18.5 78.9 3.5 3.8 4.1 4.4 4.7 5.0 4.5 5.0 5.5 6.0 6.5 %AC Ai r V o i d s 16.8 17.2 17.6 18.0 18.4 18.8 4.55.05.56.06.5 %AC VM A 70.0 72.0 74.0 76.0 78.0 80.0 82.0 4.5 5.0 5.5 6.0 6.5 %AC VF A 138 Lithia Springs 20% +4 RAP Aggregate Components Sieve size 007 089 810 Marble dust Lime RAP Blend Proportions 61.8% 0.0% 12.0% 5.0% 0.9% 20.3% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 99.0 99.8 1/2" 85.0 100.0 100.0 100.0 100.0 96.0 89.9 3/8" 50.0 100.0 100.0 100.0 100.0 84.0 65.9 #4 6.0 30.0 84.0 100.0 100.0 37.0 27.2 #8 2.0 2.0 62.0 100.0 100.0 25.0 19.7 #16 1.0 2.0 50.0 100.0 100.0 21.0 16.8 #30 1.0 1.0 41.0 100.0 100.0 18.0 15.1 #50 1.0 1.0 28.0 100.0 100.0 15.0 12.9 #100 1.0 1.0 21.0 98.0 100.0 10.0 11.0 #200 1.0 1.0 10.0 90.0 100.0 6.2 8.5 Series % AC VMA VFA 1 5.0 17.7 74.4 2 5.5 18.0 79.4 3 6.0 18.4 83.8 2.0 2.5 3.0 3.5 4.0 4.5 5.0 4.5 5.0 5.5 6.0 6.5 %AC Ai r V o i d s 17.2 17.4 17.6 17.8 18.0 18.2 18.4 18.6 4.55.05.56.06.5 %AC VM A 66.0 70.0 74.0 78.0 82.0 86.0 90.0 4.5 5.0 5.5 6.0 6.5 %AC VF A 139 Lithia Springs 30% DG2 RAP Aggregate Components Sieve size 007 089 810 Marble dust Lime RAP Blend Proportions 65.7% 0.0% 0.0% 3.0% 0.8% 30.5% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 100.0 100.0 1/2" 85.0 100.0 100.0 100.0 100.0 100.0 90.1 3/8" 50.0 100.0 100.0 100.0 100.0 95.0 65.6 #4 6.0 30.0 84.0 100.0 100.0 77.0 31.2 #8 2.0 2.0 62.0 100.0 100.0 61.0 23.7 #16 1.0 2.0 50.0 100.0 100.0 50.0 19.7 #30 1.0 1.0 41.0 100.0 100.0 42.0 17.3 #50 1.0 1.0 28.0 100.0 100.0 32.0 14.2 #100 1.0 1.0 21.0 98.0 100.0 20.0 10.5 #200 1.0 1.0 10.0 90.0 100.0 12.0 7.8 Series % AC VMA VFA 1 4.5 17.3 74.9 2 5.0 17.5 80.8 3 5.5 18.4 82.2 2.5 3.0 3.5 4.0 4.5 4.0 4.5 5.0 5.5 6.0 %AC Ai r V o i d s 16.8 17.2 17.6 18.0 18.4 18.8 4.0 4.5 5.0 5.5 6.0 %AC VM A 72.0 76.0 80.0 84.0 88.0 4.0 4.5 5.0 5.5 6.0 %AC VF A 140 141 A.3 Mix Designs for Camak Mixtures Camak 0% RAP Aggregate Components Sieve size 007 M10 Marble dust Lime Blend Proportions 83.0% 10.0% 6.0% 1.0% 100% 1" 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 1/2" 94.0 100.0 100.0 100.0 95.0 3/8" 46.0 100.0 100.0 100.0 55.2 #4 2.0 98.0 100.0 100.0 18.5 #8 1.0 82.0 100.0 100.0 16.0 #16 1.0 62.0 100.0 100.0 14.0 #30 1.0 50.0 100.0 100.0 12.8 #50 1.0 36.0 100.0 100.0 11.4 #100 1.0 25.0 98.0 100.0 10.2 #200 1.0 12.0 90.0 100.0 8.4 Series % AC VMA VFA 1 6.0 21.1 73.3 Camak 10% DG2 RAP Aggregate Components Sieve size 007 M10 Marble dust Lime RAP Blend Proportions 73.0% 11.0% 4.9% 0.9% 10.2% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 100.0 1/2" 94.0 100.0 100.0 100.0 100.0 95.6 3/8" 46.0 100.0 100.0 100.0 95.0 60.1 #4 2.0 98.0 100.0 100.0 77.0 25.9 #8 1.0 82.0 100.0 100.0 61.0 21.8 #16 1.0 62.0 100.0 100.0 50.0 18.5 #30 1.0 50.0 100.0 100.0 42.0 16.3 #50 1.0 36.0 100.0 100.0 32.0 13.8 #100 1.0 25.0 98.0 100.0 20.0 11.2 #200 1.0 12.0 90.0 100.0 12.0 8.6 Series % AC VMA VFA 1 5.5 18.4 72.5 2 6.0 18.9 76.4 3 6.5 19.2 81.1 3.0 3.5 4.0 4.5 5.0 5.5 5.0 5.5 6.0 6.5 7.0 %AC Ai r V o i d s 18.2 18.4 18.6 18.8 19.0 19.2 19.4 5.0 5.5 6.0 6.5 7.0 %AC VM A 70.0 74.0 78.0 82.0 86.0 5.0 5.5 6.0 6.5 7.0 %AC VF A Camak 20% -4 RAP Aggregate Components Sieve size 007 M10 Marble dust Lime RAP Blend Proportions 75.1% 0.0% 4.0% 0.9% 20.0% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 100.0 1/2" 94.0 100.0 100.0 100.0 100.0 95.5 3/8" 46.0 100.0 100.0 100.0 100.0 59.4 #4 2.0 98.0 100.0 100.0 100.0 26.4 #8 1.0 82.0 100.0 100.0 81.0 21.9 #16 1.0 62.0 100.0 100.0 65.0 18.7 #30 1.0 50.0 100.0 100.0 53.0 16.3 #50 1.0 36.0 100.0 100.0 40.0 13.7 #100 1.0 25.0 98.0 100.0 25.0 10.6 #200 1.0 12.0 90.0 100.0 15.0 8.3 Series % AC VMA VFA 1 5.0 18.7 74.2 2 5.5 19.2 77.8 3 6.0 19.8 81.1 142 3.0 3.5 4.0 4.5 5.0 4.5 5.0 5.5 6.0 6.5 %AC Ai r V o i d s 18.0 18.4 18.8 19.2 19.6 20.0 4.55.05.56.06.5 %AC VM A 70.0 74.0 78.0 82.0 86.0 4.5 5.0 5.5 6.0 6.5 %AC VF A Camak 30% DG1 RAP Aggregate Components Sieve size 007 M10 Marble dust Lime RAP Blend Proportions 65.6% 0.0% 3.5% 0.8% 30.1% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 100.0 1/2" 94.0 100.0 100.0 100.0 99.0 95.8 3/8" 46.0 100.0 100.0 100.0 93.0 62.5 #4 2.0 98.0 100.0 100.0 73.0 27.6 #8 1.0 82.0 100.0 100.0 58.0 22.4 #16 1.0 62.0 100.0 100.0 47.0 19.1 #30 1.0 50.0 100.0 100.0 38.0 16.4 #50 1.0 36.0 100.0 100.0 29.0 13.7 #100 1.0 25.0 98.0 100.0 19.0 10.6 #200 1.0 12.0 90.0 100.0 11.2 8.0 Series % AC VMA VFA 1 5.5 19.6 81.3 2 6.0 20.1 85.1 3 6.5 20.6 88.0 143 2.0 2.4 2.8 3.2 3.6 4.0 5.0 5.5 6.0 6.5 7.0 %AC Ai r V o i d s 19.2 19.6 20.0 20.4 20.8 5.05.56.06.57.0 %AC VM A 78.0 80.0 82.0 84.0 86.0 88.0 90.0 5.0 5.5 6.0 6.5 7.0 %AC VF A A.4 Mix Designs for Ruby Mixtures Ruby 0% RAP Aggregate Components Sieve size 007 M10 Marble dust Lime Blend Proportions 77.0% 18.0% 4.0% 1.0% 100% 1" 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 1/2" 96.0 100.0 100.0 100.0 96.9 3/8" 55.0 100.0 100.0 100.0 65.4 #4 2.0 99.0 100.0 100.0 24.4 #8 1.0 82.0 100.0 100.0 20.5 #16 1.0 62.0 100.0 100.0 16.9 #30 1.0 49.0 100.0 100.0 14.6 #50 1.0 37.0 100.0 100.0 12.4 #100 1.0 27.0 98.0 100.0 10.6 #200 1.0 18.0 90.0 100.0 8.6 Series % AC VMA VFA 1 6.5 19.4 76.7 2 7.0 18.8 86.5 144 Ruby 10% +4 RAP Aggregate Components Sieve size 007 M10 Marble dust Lime RAP Blend Proportions 69.9% 15.0% 4.0% 0.9% 10.2% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 99.0 99.9 1/2" 96.0 100.0 100.0 100.0 96.0 96.8 3/8" 55.0 100.0 100.0 100.0 84.0 66.9 #4 2.0 99.0 100.0 100.0 37.0 24.9 #8 1.0 82.0 100.0 100.0 25.0 20.4 #16 1.0 62.0 100.0 100.0 21.0 17.0 #30 1.0 49.0 100.0 100.0 18.0 14.8 #50 1.0 37.0 100.0 100.0 15.0 12.7 #100 1.0 27.0 98.0 100.0 10.0 10.6 #200 1.0 18.0 90.0 100.0 6.2 8.5 Series % AC VMA VFA 1 5.0 18.1 68.9 2 5.5 18.0 76.4 3 6.0 18.5 80.2 3.0 3.5 4.0 4.5 5.0 5.5 6.0 4.5 5.0 5.5 6.0 6.5 %AC Ai r V o i d s 17.6 17.8 18.0 18.2 18.4 18.6 4.55.05.56.06.5 %AC VM A 66.0 70.0 74.0 78.0 82.0 4.5 5.0 5.5 6.0 6.5 %AC VF A 145 Ruby 20% DG1 RAP Aggregate Components Sieve size 007 M10 Marble dust Lime RAP Blend Proportions 70.0% 5.0% 4.0% 0.9% 20.1% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 100.0 1/2" 96.0 100.0 100.0 100.0 99.0 97.0 3/8" 55.0 100.0 100.0 100.0 93.0 67.1 #4 2.0 99.0 100.0 100.0 73.0 25.9 #8 1.0 82.0 100.0 100.0 58.0 21.4 #16 1.0 62.0 100.0 100.0 47.0 18.1 #30 1.0 49.0 100.0 100.0 38.0 15.7 #50 1.0 37.0 100.0 100.0 29.0 13.3 #100 1.0 27.0 98.0 100.0 19.0 10.7 #200 1.0 18.0 90.0 100.0 11.2 8.4 Series % AC VMA VFA 1 5.0 18.4 75.6 2 5.5 18.9 79.8 3 6.0 19.7 81.9 3.0 3.5 4.0 4.5 5.0 4.5 5.0 5.5 6.0 6.5 %AC Ai r V o i d s 17.6 18.0 18.4 18.8 19.2 19.6 20.0 4.55.05.56.06.5 %AC VM A 70.0 74.0 78.0 82.0 86.0 4.5 5.0 5.5 6.0 6.5 %AC VF A 146 Ruby 30% -4 RAP Aggregate Components Sieve size 007 M10 Marble dust Lime RAP Blend Proportions 66.7% 0.0% 2.5% 0.9% 29.9% 100% 1" 100.0 100.0 100.0 100.0 100.0 100.0 3/4" 100.0 100.0 100.0 100.0 100.0 100.0 1/2" 96.0 100.0 100.0 100.0 100.0 97.3 3/8" 55.0 100.0 100.0 100.0 100.0 70.0 #4 2.0 99.0 100.0 100.0 100.0 34.6 #8 1.0 82.0 100.0 100.0 81.0 28.3 #16 1.0 62.0 100.0 100.0 65.0 23.5 #30 1.0 49.0 100.0 100.0 53.0 19.9 #50 1.0 37.0 100.0 100.0 40.0 16.0 #100 1.0 27.0 98.0 100.0 25.0 11.5 #200 1.0 18.0 90.0 100.0 15.0 8.3 Series % AC VMA VFA 1 4.0 17.6 76.5 2 4.5 17.5 83.9 3 5.0 18.4 85.4 2.0 2.5 3.0 3.5 4.0 4.5 3.5 4.0 4.5 5.0 5.5 %AC Ai r V o i d s 16.8 17.2 17.6 18.0 18.4 18.8 3.54.04.55.05. %AC VM A 74.0 78.0 82.0 86.0 90.0 3.5 4.0 4.5 5.0 5.5 %AC VF A 147 148 APPENDIX B Individual Test Results 149 Table B.1. Results from Moisture Susceptibility Test. Agg. Source % RAP RAP Type % Air Voids Wet Strength, psi Dry Strength, psi TSR Mt. View 0 DG1 7.2 69.90 77.86 0.90 Mt. View 0 DG1 7.2 77.09 64.94 1.19 Mt. View 0 DG1 7.1 67.67 71.94 0.94 Mt. View 10 -4 6.7 90.91 100.65 0.90 Mt. View 10 -4 6.9 86.45 83.21 1.04 Mt. View 10 -4 7.1 87.09 77.79 1.12 Mt. View 20 DG2 7.3 95.75 107.80 1.12 Mt. View 20 DG2 9.6 84.03 100.60 0.89 Mt. View 20 DG2 6.6 95.87 97.30 0.84 Mt. View 30 +4 6.7 101.99 107.84 0.95 Mt. View 30 +4 6.3 110.14 105.62 1.04 Mt. View 30 +4 7.3 87.28 103.77 0.84 Lithia Spr. 0 -4 7.6 67.55 91.80 0.74 Lithia Spr. 0 -4 7.9 74.36 97.08 0.77 Lithia Spr. 0 -4 7.5 70.79 79.26 0.89 Lithia Spr. 10 DG1 7.1 84.93 85.56 0.99 Lithia Spr. 10 DG1 6.8 73.78 79.96 0.92 Lithia Spr. 10 DG1 7.4 80.41 85.18 0.94 Lithia Spr. 20 +4 7.4 88.87 95.68 0.93 Lithia Spr. 20 +4 7.0 98.74 77.35 1.28 Lithia Spr. 20 +4 7.4 84.67 88.04 0.96 Lithia Spr. 30 DG2 7.3 98.99 98.55 1.00 Lithia Spr. 30 DG2 6.8 92.37 85.94 1.07 Lithia Spr. 30 DG2 6.5 95.94 99.76 0.96 Camak 0 +4 7.5 77.70 87.54 0.96 Camak 0 +4 8.0 78.80 90.34 0.89 Camak 0 +4 7.5 79.50 85.18 0.87 Camak 10 DG2 6.6 76.90 100.39 0.77 Camak 10 DG2 6.9 68.05 79.26 0.86 Camak 10 DG2 7.2 72.00 75.69 0.95 Camak 20 -4 6.8 96.58 108.54 0.89 Camak 20 -4 6.7 76.52 100.20 0.76 Camak 20 -4 7.8 84.03 86.64 0.97 Camak 30 DG1 6.4 127.39 136.81 0.93 Camak 30 DG1 7.1 114.78 124.08 0.93 Camak 30 DG1 6.2 112.05 113.51 0.99 Ruby 0 DG2 7.0 72.10 70.00 1.03 Ruby 0 DG2 6.0 71.60 70.00 1.02 Ruby 0 DG2 6.7 67.00 69.10 0.97 Ruby 10 +4 6.7 93.65 102.24 0.92 Ruby 10 +4 6.7 89.89 102.56 0.88 Ruby 10 +4 6.9 93.39 106.12 0.88 Ruby 20 DG1 6.1 81.42 84.54 0.96 Ruby 20 DG1 6.4 76.78 85.63 0.90 Ruby 20 DG1 6.1 92.69 79.13 1.17 Ruby 30 -4 6.1 154.89 168.96 0.92 Ruby 30 -4 6.9 134.77 139.99 0.96 Ruby 30 -4 6.3 135.54 125.16 1.08 150 Table B.2. Results from Rutting Susceptibility Test. Agg. Source % RAP RAP Type % Air Voids Rut Depth, mm Mt. View 0 DG1 5.4 3.13 Mt. View 0 DG1 5.0 2.94 Mt. View 0 DG1 5.2 1.76 Mt. View 0 DG1 5.3 3.39 Mt. View 0 DG1 5.0 0.95 Mt. View 0 DG1 5.1 6.50 Mt. View 10 -4 5.0 4.42 Mt. View 10 -4 5.1 3.18 Mt. View 10 -4 4.8 1.40 Mt. View 10 -4 4.9 3.11 Mt. View 10 -4 5.1 5.86 Mt. View 10 -4 5.1 1.41 Mt. View 20 DG2 5.1 4.23 Mt. View 20 DG2 4.7 6.01 Mt. View 20 DG2 4.9 3.34 Mt. View 20 DG2 4.8 2.05 Mt. View 20 DG2 4.4 8.33 Mt. View 20 DG2 5.0 2.50 Mt. View 30 +4 5.2 1.25 Mt. View 30 +4 4.9 2.35 Mt. View 30 +4 4.9 1.89 Mt. View 30 +4 4.9 1.23 Mt. View 30 +4 4.9 2.49 Mt. View 30 +4 4.8 0.99 Lithia Spr. 0 -4 4.7 3.14 Lithia Spr. 0 -4 4.2 1.86 Lithia Spr. 0 -4 4.3 1.89 Lithia Spr. 0 -4 4.7 3.59 Lithia Spr. 0 -4 4.7 1.98 Lithia Spr. 0 -4 4.3 1.76 Lithia Spr. 10 DG1 4.6 2.73 Lithia Spr. 10 DG1 4.7 2.37 Lithia Spr. 10 DG1 4.8 1.31 Lithia Spr. 10 DG1 4.6 2.15 Lithia Spr. 10 DG1 4.7 1.25 Lithia Spr. 10 DG1 4.7 2.19 Lithia Spr. 20 +4 4.7 2.69 Lithia Spr. 20 +4 4.8 2.48 Lithia Spr. 20 +4 4.7 1.77 Lithia Spr. 20 +4 4.9 0.93 Lithia Spr. 20 +4 4.4 3.33 Lithia Spr. 20 +4 4.2 3.45 Lithia Spr. 30 DG2 5.6 2.77 Lithia Spr. 30 DG2 5.2 5.75 Lithia Spr. 30 DG2 5.8 5.15 Lithia Spr. 30 DG2 5.3 2.88 Lithia Spr. 30 DG2 5.6 5.69 Lithia Spr. 30 DG2 5.4 4.76 151 Table B.2 (cont.). Results from Rutting Susceptibility Test. Agg. Source % RAP RAP Type % Air Voids Rut Depth, mm Camak 0 +4 4.9 6.83 Cama +4 4.8 4.27 Camak 0 +4 4.7 2.39 Cama +4 4.5 4.42 Camak 0 +4 4.2 3.45 Cama +4 4.3 0.68 Camak 10 DG2 4.7 1.62 Cama 10 DG 4.6 2.14 Camak 10 DG2 4.4 1.97 Cama 10 DG 4.9 1.58 Camak 10 DG2 4.5 2.56 Cama 10 DG 4.6 1.89 Camak 20 -4 4.3 1.40 Cama 20 -4 4.4 1.83 Camak 20 -4 4.1 1.93 Cama 20 -4 4.8 0.95 Camak 20 -4 4.2 1.87 Cama 20 -4 5. 0.91 Camak 30 DG1 4.7 4.04 Cama 30 DG 4.8 2.95 Camak 30 DG1 4.7 2.53 Cama 30 DG 5.0 4.20 Camak 30 DG1 4.3 2.75 Cama 30 DG 4.1 3.04 Ruby 0 DG2 5. 5.89 Ruby DG 5.4 1.73 Ruby 0 DG2 4.5 3.85 Ruby DG 4.2 4.05 Ruby 0 DG2 4.8 2.70 Ruby DG 4.4 3.30 Ruby 10 +4 5. 4.40 Ruby 10 +4 5.2 5.60 Ruby 10 +4 5.3 6.17 Ruby 10 +4 5.6 5.83 Ruby 10 +4 5.0 5.03 Ruby 10 +4 5. 3.92 Ruby 20 DG1 4.8 5.74 Ruby 20 DG 4.7 4.56 Ruby 20 DG1 4.8 5.99 Ruby 20 DG 5.0 6.47 Ruby 20 DG1 5.2 5.84 Ruby 20 DG 4.5 3.67 Ruby 30 -4 5.4 4.95 Ruby 30 -4 5.2 3.08 Ruby 30 -4 5.8 4.03 Ruby 30 -4 5.2 4.84 Ruby 30 -4 5.3 3.25 Ruby 30 -4 4.4 2.97 152 153 Table B.4. Results from Fatigue Test (400 ??). Agg. Source % RAP RAP Type % Air Voids N f Initial Stiffness, MPa Diss. Energy (kPa) Mt. View 0 DG1 6.8 1,202,320 3,102 0.247 Mt. View 0 DG1 6.5 6,365,620 3,374 0.277 Mt. View 0 DG1 6.9 3,236,690 3,632 0.288 Mt. View 10 -4 6.6 5,319,730 3,194 0.246 Mt. View 10 -4 6.3 4,017,630 3,236 0.254 Mt. View 10 -4 6.3 6,123,740 3,921 0.302 Mt. View 20 DG2 5.3 1,755,400 3,660 0.272 Mt. View 20 DG2 5.5 2,890,920 3,366 0.277 Mt. View 20 DG2 5.7 2,367,890 3,430 0.288 Mt. View 30 +4 7.4 2,450,210 4,314 0.197 Mt. View 30 +4 6.9 342,580 5,572 0.197 Mt. View 30 +4 6.5 1,205,860 4,260 0.203 Lithia Spr. 0 -4 6.4 6,098,480 3,564 0.277 Lithia Spr. 0 -4 5.9 4,917,110 3,738 0.298 Lithia Spr. 0 -4 5.4 5,562,870 3,656 0.295 Lithia Spr. 10 DG1 5.4 5,434,130 3,972 0.298 Lithia Spr. 10 DG1 5.6 3,055,130 3,896 0.299 Lithia Spr. 10 DG1 6.0 2,568,770 3,667 0.278 Lithia Spr. 20 +4 6.1 6,648,340 3,472 0.278 Lithia Spr. 20 +4 6.1 2,593,300 3,674 0.285 Lithia Spr. 20 +4 6.7 5,134,130 3,494 0.275 Lithia Spr. 30 DG2 5.8 2,274,170 4,160 0.296 Lithia Spr. 30 DG2 5.2 2,053,980 4,126 0.265 Lithia Spr. 30 DG2 5.1 268,000 4,213 0.894 Camak 0 +4 5.8 3,062,380 3,480 0.277 Camak 0 +4 6.0 3,576,160 3,252 0.273 Camak 0 +4 6.4 6,421,250 2,830 0.251 Camak 10 DG2 6.0 3,149,900 3,605 0.513 Camak 10 DG2 5.8 4,033,030 3,463 0.513 Camak 10 DG2 5.3 6,877,810 3,644 0.288 Camak 20 -4 5.2 3,163,080 3,542 0.265 Camak 20 -4 7.0 4,225,320 3,077 0.247 Camak 20 -4 6.8 679,370 2,798 0.224 Camak 30 DG1 5.7 3,771,280 2,858 0.238 Camak 30 DG1 5.0 4,144,106 3,267 0.258 Camak 30 DG1 5.1 3,656,092 3,730 0.258 Ruby 0 DG2 6.7 5,528,870 3,636 0.221 Ruby 0 DG2 6.2 4,984,710 3,568 0.224 Ruby 0 DG2 6.2 5,701,660 3,302 0.224 Ruby 10 +4 5.6 5,364,710 3,726 0.311 Ruby 10 +4 6.3 2,020,360 3,466 0.296 Ruby 10 +4 5.4 6,271,490 3,519 0.275 Ruby 20 DG1 6.2 8,368,160 3,074 0.257 Ruby 20 DG1 5.7 3,557,870 3,488 0.289 Ruby 20 DG1 5.5 1,454,590 3,623 0.275 Ruby 30 -4 6.9 543,000 4,411 0.308 Ruby 30 -4 6.4 1,091,240 4,863 0.321 Ruby 30 -4 6.2 645,220 4,588 0.321 154 Table B.5. Results from Fatigue Test (800 ??). Agg. Source % RAP RAP Type % Air Voids N f Ini. Stiffness, MPa Diss. Energy (kPa) Mt. View 0 DG1 6.7 78,940 2,929 0.989 Mt. View 0 DG1 6.1 35,440 3,030 1.004 Mt. View 0 DG1 6.6 23,020 2,898 0.963 Mt. View 10 -4 6.7 62,700 3,156 0.995 Mt. View 10 -4 5.2 57,900 3,599 0.900 Mt. View 10 -4 6.0 29,830 2,752 0.916 Mt. View 20 DG2 5.0 37,600 4,480 0.898 Mt. View 20 DG2 6.3 37,420 4,277 0.910 Mt. View 20 DG2 5.4 18,020 4,219 0.873 Mt. View 30 +4 6.9 22,860 4,372 0.701 Mt. View 30 +4 7.2 20,520 4,331 0.670 Mt. View 30 +4 6.7 16,260 4,243 0.698 Lithia Spr. 0 -4 7.0 49,280 3,003 1.017 Lithia Spr. 0 -4 5.9 73,420 3,325 1.119 Lithia Spr. 0 -4 7.0 53,410 2,941 0.980 Lithia Spr. 10 DG1 5.9 44,850 3,506 1.095 Lithia Spr. 10 DG1 5.6 32,990 3,480 1.087 Lithia Spr. 10 DG1 5.5 56,790 3,248 1.024 Lithia Spr. 20 +4 6.3 43,640 3,226 1.041 Lithia Spr. 20 +4 6.4 80,150 3,064 0.997 Lithia Spr. 20 +4 6.1 50,030 3,246 1.045 Lithia Spr. 30 DG2 5.2 25,850 5,029 0.526 Lithia Spr. 30 DG2 5.4 11,150 4,569 0.586 Lithia Spr. 30 DG2 5.0 13,260 5,250 0.513 Camak 0 +4 5.7 109,140 3,323 0.894 Camak 0 +4 5.0 109,470 3,273 0.867 Camak 0 +4 5.9 57,600 3,065 0.801 Camak 10 DG2 5.5 58,090 3,265 1.075 Camak 10 DG2 6.5 21,880 3,277 1.025 Camak 10 DG2 5.6 42,870 3,121 1.001 Camak 20 -4 5.0 82,650 3,556 1.084 Camak 20 -4 6.9 33,500 3,253 1.012 Camak 20 -4 6.1 97,060 3,331 0.799 Camak 30 DG1 5.0 93,550 3,679 1.128 Camak 30 DG1 5.0 65,730 3,578 0.799 Camak 30 DG1 6.1 65,000 3,043 1.007 Ruby 0 DG2 6.8 78,700 3,096 0.896 Ruby 0 DG2 5.2 50,660 3,133 0.885 Ruby 0 DG2 6.5 88,680 2,854 0.806 Ruby 10 +4 6.8 59,420 3,192 1.065 Ruby 10 +4 7.0 26,920 2,969 1.013 Ruby 10 +4 5.6 36,460 3,262 1.087 Ruby 20 DG1 6.4 37,170 2,978 0.978 Ruby 20 DG1 5.8 11,390 3,550 1.199 Ruby 20 DG1 5.6 14,810 3,515 1.138 Ruby 30 -4 6.6 5,180 4,933 0.722 Ruby 30 -4 6.7 4,340 4,709 0.638 Ruby 30 -4 6.7 3,300 4,753 0.627