SPR 396-9A JUNE 2015 Performance Evaluation of Arizona’s LTPP SPS-9 Project: Strategic Study of Flexible Pavement Binder Factors Arizona Department of Transportation Research Center Performance Evaluation of Arizona’s LTPP SPS‐9 Project: Strategic Study of Rehabilitation of Flexible Pavement Binder Factors SPR‐396‐9A June 2015 Prepared by: Jason Puccinelli Nichols Consulting Engineers 1885 S. Arlington Avenue, Suite 111 Reno, NV 89509‐3370 Steven M. Karamihas The University of Michigan Transportation Research Institute 2901 Baxter Road Ann Arbor, MI 48109 Kathleen T. Hall Kathleen T. Hall Consulting 1271 Huntington Drive South Mundeleine, IL 60060 Jonathan Minassian Kevin Senn Nichols Consulting Engineers 1885 S. Arlington Avenue, Suite 111 Reno, NV 89509‐3370 Published by: Arizona Department of Transportation 206 S. 17th Avenue Phoenix, AZ 85007 in cooperation with US Department of Transportation Federal Highway Administration This report was funded in part through grants from the Federal Highway Administration, U.S. Department of Transportation. The contents of this report reflect the views of the authors, who are responsible for the facts and the accuracy of the data, and for the use or adaptation of previously published material, presented herein. The contents do not necessarily reflect the official views or policies of the Arizona Department of Transportation or the Federal Highway Administration, U.S. Department of Transportation. This report does not constitute a standard, specification, or regulation. Trade or manufacturers’ names that may appear herein are cited only because they are considered essential to the objectives of the report. The U.S. government and the State of Arizona do not endorse products or manufacturers. 1. Report No. FHWA-AZ-15-396(9A) 2. Government Accession No. 3. Recipient's Catalog No. 4. Title and Subtitle 5. Report Date Performance Evaluation of Arizona’s LTPP SPS‐9 Project: Strategic Study of Flexible Pavement Binder Factors June 2015 6. Performing Organization Code 7. Author(s) 8. Performing Organization Report No. Jason Puccinelli, Steven Karamihas, Kathleen T. Hall, Jonathan Minassian, Kevin Senn 9. Performing Organization Name and Address 10. Work Unit No. (TRAIS) Nichols Consulting Engineers 1885 South Arlington Avenue Suite 111 Reno, Nevada 89509‐3370 11. Contract or Grant No. SPR 000‐147 (396) 9A The University of Michigan Transportation Research Institute 2901 Baxter Road Ann Arbor, Michigan 48109 12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered Arizona Department of Transportation 206 S. 17th Avenue Phoenix, Arizona 85007 14. Sponsoring Agency Code 15. Supplementary Notes Prepared in cooperation with the US Department of Transportation, Federal Highway Administration 16. Abstract As part of the Long Term Pavement Performance (LTPP) Program, the Arizona Department of Transportation (ADOT) constructed eight Specific Pavement Studies 9 (SPS‐9) test sections on Interstate 10 near Phoenix (04B900). SPS‐9A 04B900 is an overlay project and is accordingly given independent analysis and documentation in this report separate from Arizona SPS‐9B projects (040900 and 04A900) located on US 93, which were new construction and are documented in a separate report. The SPS‐9A project studied the effect of asphalt specification and mix designs on flexible pavements, specifically comparing Superpave binders with commonly used agency binders. Opened to traffic in 1995, the project was monitored at regular intervals until it was rehabilitated in 2005. Surface distress, profile, and deflection data collected throughout the life of the pavement were used to evaluate the performance of various flexible pavement design features, layer configurations, and thickness. This report documents the analyses conducted as well as practical findings and lessons learned that will be of interest to ADOT. 17. Key Words 18. Distribution Statement LTPP, pavement performance, profile, distress, FWD, flexible, AC, deflections, roughness, backcalculation Document is available to the U.S. public through the National Technical Information Service, Springfield, VA 22161 19. Security Classification 20. Security Classification 21. No. of Pages 82 Unclassified Unclassified 22. Price 23. Registrant's Seal SI* (MODERN METRIC) CONVERSION FACTORS APPROXIMATE CONVERSIONS TO SI UNITS Symbol When You Know Multiply By LENGTH in ft yd mi inches feet yards miles in 2 ft 2 yd ac 2 mi 2 square square square acres square fl oz gal 3 ft 3 yd fluid ounces gallons cubic feet cubic yards oz lb T ounces pounds short tons (2000 lb) o Fahrenheit fc fl foot-candles foot-Lamberts lbf 2 lbf/in poundforce poundforce per square inch Symbol When You Know mm m m km millimeters meters meters kilometers 25.4 0.305 0.914 1.61 To Find Symbol millimeters meters meters kilometers mm m m km square millimeters square meters square meters hectares square kilometers mm 2 m 2 m ha 2 km milliliters liters cubic meters cubic meters 3 shown in m mL L 3 m 3 m grams kilograms megagrams (or "metric ton") g kg Mg (or "t") AREA inches feet yard 645.2 0.093 0.836 0.405 2.59 miles 2 VOLUME 29.57 3.785 0.028 0.765 NOTE: volumes greater than 1000 L shall be MASS 28.35 0.454 0.907 TEMPERATURE (exact degrees) F 5 (F-32)/9 or (F-32)/1.8 Celsius o lux 2 candela/m lx 2 cd/m C ILLUMINATION 10.76 3.426 FORCE and PRESSURE or STRESS 4.45 6.89 newtons kilopascals N kPa APPROXIMATE CONVERSIONS FROM SI UNITS Multiply By LENGTH 0.039 3.28 1.09 0.621 To Find Symbol inches feet yards miles in ft yd mi AREA 2 mm 2 m 2 m ha 2 km square millimeters square meters square meters hectares square kilometers 0.0016 10.764 1.195 2.47 0.386 square square square acres square inches feet yards miles 2 in 2 ft 2 yd ac 2 mi VOLUME mL L 3 m m3 milliliters liters cubic meters cubic meters 0.034 0.264 35.314 1.307 g kg Mg (or "t") grams kilograms megagrams (or "metric ton") o Celsius fluid ounces gallons cubic feet cubic yards fl oz gal 3 ft yd3 ounces pounds short tons (2000 lb) oz lb T MASS 0.035 2.202 1.103 TEMPERATURE (exact degrees) C 1.8C+32 Fahrenheit o foot-candles foot-Lamberts fc fl F ILLUMINATION lx 2 cd/m lux 2 candela/m N kPa newtons kilopascals 0.0929 0.2919 FORCE and PRESSURE or STRESS 0.225 0.145 poundforce poundforce per square inch lbf 2 lbf/in *SI is the symbol for the International System of Units. Appropriate rounding should be made to comply with Section 4 of ASTM E380. (Revised March 2003) Contents EXECUTIVE SUMMARY ............................................................................................................ 1 CHAPTER 1. INTRODUCTION .................................................................................................... 3 CHAPTER 2. SPS‐9A DEFLECTION ANALYSIS ............................................................................ 15 ANALYSIS OF DEFLECTION DATA ............................................................................................... 15 MAXIMUM DEFLECTION, MINIMUM DEFLECTION, AND AREA VALUE..................................... 15 BACKCALCULATION PROCEDURE .............................................................................................. 19 BACKCALCULATION USING EVERCALC SOFTWARE ................................................................... 23 KEY FINDINGS FROM THE SPS‐9A DEFLECTION ANALYSIS ........................................................ 26 CHAPTER 3. SPS‐9A DISTRESS ANALYSIS ................................................................................. 29 AC DISTRESS TYPES.................................................................................................................... 29 RESEARCH APPROACH............................................................................................................... 30 OVERALL PERFORMANCE TREND OBSERVATIONS.................................................................... 32 KEY FINDINGS FROM THE SPS‐9A DISTRESS ANALYSIS ............................................................. 39 CHAPTER 4. SPS‐9A ROUGHNESS ANALYSIS ............................................................................ 41 PROFILE DATA SYNCHRONIZATION ........................................................................................... 41 DATA EXTRACTION .................................................................................................................... 42 CROSS CORRELATION ................................................................................................................ 42 SYNCHRONIZATION ................................................................................................................... 43 DATA QUALITY SCREENING ....................................................................................................... 43 SUMMARY ROUGHNESS VALUES .............................................................................................. 47 PROFILE ANALYSIS TOOLS ......................................................................................................... 52 DETAILED OBSERVATIONS......................................................................................................... 54 SUMMARY ................................................................................................................................. 64 CHAPTER 5. CONCLUSIONS AND RECOMMENDATIONS .......................................................... 69 REFERENCES........................................................................................................................... 71 APPENDIX: ROUGHNESS VALUES ........................................................................................... 73 v List of Figures Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure 10. Figure 11. Figure 12. Figure 13. Figure 14. Figure 15. Figure 16. Figure 17. Figure 18. Figure 19. Figure 20. Figure 21. Figure 22. Figure 23. Figure 24. Figure 25. Figure 26. Existing Pavement Structure for the SPS‐9A Project .................................................... 4 Pavement Structure for the SPS‐9A Project After Construction ................................. 4 Layout of the SPS‐9A Test Sections .............................................................................. 5 SPS‐9A Test Section Layout .......................................................................................... 6 Location of SPS‐9A 04B900 Test Sections .................................................................... 7 Gradations of Superpave Aggregate and Agency Standard Aggregate ...................... 12 Average Normalized Dmax by Test Section .................................................................. 16 Average Normalized Dmin by Test Section .................................................................. 17 AREA Values by Test Section ...................................................................................... 19 Backcalculated AC Modulus by Test Section .............................................................. 25 Backcalculated MR by Test Section ............................................................................ 26 Structural Damage Trends for SPS‐9A Test Sections .................................................. 33 Environmental Damage Trends for SPS‐9A Test Sections .......................................... 34 Structural Damage Index and Pavement Structure Summary ................................... 35 Environmental Damage Index and Pavement Structure Summary ........................... 36 Rutting Index and Pavement Structure Summary ...................................................... 37 IRI Progression of Section 04B901.............................................................................. 48 IRI Progression of Section 04B902.............................................................................. 48 IRI Progression of Section 04B903.............................................................................. 49 IRI Progression of Section 04B959.............................................................................. 49 IRI Progression of Section 04B960.............................................................................. 50 IRI Progression of Section 04B961.............................................................................. 50 IRI Progression of Section 04B962.............................................................................. 51 IRI Progression of Section 04B964.............................................................................. 52 Summary of IRI Ranges ............................................................................................... 66 Comparison of HRI to MRI .......................................................................................... 74 vi List of Tables Table 1. Table 2. Table 3 Table 4. Table 5. Table 6. Table 7. Table 8. Table 9. Table 10. Table 11. Table 12. Table 13. Table 14. Table 15. Table 16. Table 17. Table 18. Table 19. Table 20. Table 21. SPS‐9A Site Structural Factors ..................................................................................... 3 SPS‐9A Mix Properties (As Designed) ........................................................................... 9 SPS‐9A Mix Properties (As Constructed) .................................................................... 10 Dynamic Modulus (E*) for Test Section 041007/04B962 .......................................... 11 Climatic Information for SPS‐9A ................................................................................. 12 SPS‐9A Traffic‐Loading Summary ............................................................................... 13 General Trends of D0 and Area Values ....................................................................... 18 Structural Parameter Statistics for SPS‐9A ................................................................. 22 Backcalculation Seed Value and Modulus Range ....................................................... 23 Backcalculation Moduli Statistics for SPS‐9A Test Sections ....................................... 24 Flexible Pavement Distress Types and Failure Mechanisms ...................................... 30 Profile Measurement Visits of the SPS‐9A Site .......................................................... 41 Selected Repeats of Section 04B901 .......................................................................... 44 Selected Repeats of Section 04B902 .......................................................................... 44 Selected Repeats of Section 04B903 .......................................................................... 44 Selected Repeats of Section 04B959 .......................................................................... 45 Selected Repeats of Section 04B960 .......................................................................... 45 Selected Repeats of Section 04B961 .......................................................................... 45 Selected Repeats of Section 04B962 .......................................................................... 45 Selected Repeats of Section 04B964 .......................................................................... 46 Roughness Values ....................................................................................................... 74 vii List of Acronyms and Abbreviations AASHTO AB AC ACFC ADOT ARAC COV Dmax Dmin E* EP ESAL FWD HRI I‐10 IRI ksi lbf LTPP MR MP MRI NWP PCC PSD psi RN SMA SN SNeff SP‐I SP‐III SPS WP American Association of State Highway and Transportation Officials aggregate base asphalt concrete asphalt concrete friction course Arizona Department of Transportation asphalt rubber asphalt concrete coefficient of variation maximum deflection minimum deflection dynamic modulus effective pavement modulus equivalent single axle load falling weight deflectometer Half‐car Roughness Index Interstate 10 International Roughness Index kips per square inch pound force Long Term Pavement Performance resilient modulus milepost Mean Roughness Index non‐wheelpath portland cement concrete power spectral density pounds per square inch Ride Number stone mastic asphalt structural number effective structural number Superpave, Level I Superpave, Level III Specific Pavement Studies wheelpath viii Acknowledgments The project team would like to acknowledge the Arizona Department of Transportation (ADOT) for sponsoring this project. In addition, the authors thank the ADOT Research Center and the Technical Advisory Committee for their input as well as the leadership of Christ Dimitroplos. Larry Scofield’s contribution to the report is also greatly appreciated. The comprehensive information stored in the Long Term Pavement Performance database allowed for this research to be conducted. ix x EXECUTIVE SUMMARY As part of the Long Term Pavement Performance (LTPP) Program, the Arizona Department of Transportation (ADOT) constructed eight Specific Pavement Studies 9 (SPS‐9) test sections on Interstate 10 near Phoenix, identified herein as SPS‐9A. The SPS‐9A project studied the effect of asphalt specification and mix designs on flexible pavements, specifically comparing Superpave binders with commonly used agency binders. Each of the eight SPS‐9A (04B900) sections received the same basic rehabilitation using different materials as part of the standard experiment. These sections had the same structural properties before receiving the mill and overlay treatment. Construction of all eight sections occurred in March 1995, and they were placed out of study in February 2005 when the sections were milled and overlaid. This report provides general information about the project location, including climate, traffic, and subgrade conditions, as well as details about the mix designs of each test section. All eight of the SPS‐9A test sections were constructed consecutively and exposed to the same traffic‐loading, climate, and subgrade conditions, which allowed for direct comparisons between mix design performance without the confounding effects introduced by different in situ conditions. Most sections had a clear increase in magnitude of environmental distress approximately 10 years after construction. Where fatigue cracking was very prevalent, it was difficult to match individual cracks to roughness within the measured profile. However, in a few cases, features in the profiles that affected the roughness were found that correspond directly to the location of transverse cracks noted in the distress survey. From a roughness perspective, the stone mastic asphalt (SMA) cellulose and asphalt rubber asphalt concrete sections outperformed the Superpave mixes. Considering all distresses, the SMA cellulose significantly outperformed the other sections of this project. The vast majority of sections showed significant growth in longitudinal and, consequently, fatigue cracking. This significant growth in cracking was observed in the final distress survey, which implies that the growth occurred in between the last two surveys, seven to 10 years after the sections were constructed, with the rate of crack growth slowly increasing until the sections were placed out of study. All sections performed well with regard to rut resistance. Rutting would not have triggered a rehabilitation event for any section. 1 2 CHAPTER 1. INTRODUCTION Understanding how design features contribute to long‐term pavement performance can be extremely valuable to pavement designers looking to optimize resources and improve overall performance. This study’s objectives were to document the overall performance trends of the Specific Pavement Studies 9 (SPS‐9) project, identify key differences in performance between the various asphalt specifications and mix designs, and document key findings that would be useful to the Arizona Department of Transportation (ADOT). This report provides the results of surface distress, deflection, and profile analyses for the Long Term Pavement Performance (LTPP) SPS‐9 project on Interstate 10 (I‐10) near Phoenix. SPS‐9 sites were designed to study the effect of asphalt specification and mix designs on flexible pavements, specifically comparing Superpave binders with commonly used agency binders. The SPS‐9A site (04B900) consisted of eight existing sections that each received the same basic rehabilitation using different materials. These sections had the same structural properties before receiving the mill and overlay treatment. Table 1 summarizes the structural design of the test sections. All test sections had approximately the same thickness; the LTPP construction report (FHWA 1998) provides more detail about the layout and structural properties of the site. Table 1. SPS‐9A Site Structural Factors Experimental Asphalt Concrete 04B901 04B902 Existing Asphalt Concrete Layer Thickness (inches) 2.8 3.4 Standard Asphalt Concrete Layer Thickness (inches) 4.4 4.4 04B903 04B959 3.0 2.5 4.4 4.5 4.0 3.0 04B960 04B961 04B964 2.5 2.5 2.5 4.5 4.5 4.5 3.0 3.0 3.0 041007/ 04B962 3.6 4.3 2.9 Section 3 Layer Thickness (inches) 3.5 3.0 Type Agency standard Superpave, Level I (SP‐1) (PG 76‐10) SP‐1 (PG 70‐10) Stone mastic asphalt (SMA) polymer with asphalt concrete friction course SMA polymer SMA cellulose Asphalt rubber asphalt concrete Superpave, Level III r in March 1995.. The resurfaccing treatmen nt consisted o of milling 4 inches Each test section was resurfaced e pavement, placingg a 4‐inch standard (1.5‐innch maximum m aggregate) aasphalt concrete from the existing (AC) layerr, and then ovverlaying with h 3 inches of the t experimeental surfacess, as shown in n Figures 1 an nd 2. The test sections s were e located entirrely on a shallow fill of nattive material.. The subgrad de and embankm ment material are coarse‐grained silty saands with graavel and cobb bles. Figure F 1. Existting Pavemen nt Structure ffor the SPS‐99A Project Figure e 2. Pavement Structure fo or the SPS‐9A A Project Afteer Constructio on 4 e from m milepost (M MP) 112.81 to o MP 122.29 oon westbound d I‐10. The sitte is located aat The site extended latitude 33°27’45” and d longitude ‐112°28’10”, with w an approxximate elevattion of 1059 fft. The terrain n surroundiing the test se ection is sligh htly rolling and the roadwaay is straight. Figures 3 thrrough 5 illustrrate the test se ections locate ed within the SPS‐9A proje ect. ARIZONA SPS-9A S 09B90 00 I-10 02/02/ /95 o the SPS‐9A A Test Section ns Figurre 3. Layout of 5 Figure 4. SPS‐9A Test Section Layout 6 Figure 5. 5 Location of SPS‐9A 04B9900 Test Sections (Courtesyy of Google M Maps) ntal sections, three are spe ecifically for tthe SPS‐9A exxperiment: 04 4B901, 04B90 02, Of the eigght experimen and 04B903. 04B901 iss a standard agency a mix de esign using thhe Marshall 75 blow mix design. 04B902 and e, Level 1 (SP‐1) mixes thatt used PG 76‐ 10 and AC‐400 (PG 70‐10) b binders, 04B903 are Superpave respective ely. When usiing the LTPP Bind B 3.1 softw ware, the recoommended b binder for thiss project site was PG 76‐10.. The followin ng inputs were used in the program:  Laatitude: 33.46 6 degrees.  Lo owest yearly air temperatu ure: ‐4.5° C.  Ye early degree‐‐days greater than 10° C: 5387. 5  Lo ow air temperature standaard deviation: 2.2° C.  Desired reliabiility: 98 perce ent.  Depth of layer: 0 mm.  Trraffic speed: Fast.  Trraffic loading: Up to 3 million equivalen nt single axle loads (ESALs)). 7 The SPS‐9A experiment also included five supplemental test sections: 04B959, 04B960, 04B961, 04B964, and 041007/04B962. Sections 04B959, 04B960, and 04B961 were stone mastic asphalt (SMA) pavements. Section 04B959 was an SMA polymer with an asphalt concrete friction course (ACFC); 04B960 was an SMA polymer pavement; and 04B961 was an SMA cellulose pavement. Section 04B964 was an asphalt rubber asphalt concrete (ARAC) pavement. All mixtures included Type II portland cement concrete (PCC) as an admixture. Although eight experimental sections were constructed, time and budget constraints did not allow for detailed information to be recorded on the supplemental sections. The aggregate properties of Superpave sections 04B902 and 04B903 follow:  Bulk specific gravity of aggregate: 2.672.  Bulk specific gravity of admixture: 3.140.  Bulk specific gravity of the total gradation: 2.677.  Effective specific gravity of the total gradation: 2.712. Table 2 provides the design mix properties and Table 3 provides the construction mix properties for the SPS‐9A test sections. 8 Table 2. SPS‐9A Mix Properties (As Designed) Layer Mix type Maximum specific gravity Bulk specific gravity Asphalt content (%) Air voids (%) Mineral aggregate air voids (%) Effective asphalt content (%) Number of blows Asphalt grade PG high temperature (°C) PG low temperature (°C) Maximum particle size (mm) Bulk density (kg/m3) Rice density (kg/m3) Avg. (MR) at 5° C Avg. (MR) at 25° C Avg. (MR) at 40° C 04B901 04B902 04B903 04B960 04B961 Agency mix SP‐1 (PG 76‐10) SP‐1 (PG 70‐10) SMA polymer SMA cellulose Marshall Superpave Superpave N/A N/A All Sections Standard AC below experi‐ mental layer N/A N/A 2.532 2.532 N/A N/A N/A 2.368 2.425 2.426 N/A N/A N/A 3.8 6 4.3 4.2 4.3 4.2 5.2 N/A 5.3 N/A 3.4 N/A N/A 13.3 13.3 N/A N/A N/A N/A 3.8 3.8 N/A N/A N/A 75 N/A N/A N/A N/A AC‐40 N/A AC‐40 N/A N/A N/A N/A 76 76 70 N/A N/A N/A 10 10 10 N/A N/A N/A N/A 25.4 25.4 N/A N/A N/A 2368 N/A N/A N/A N/A N/A 2483 N/A N/A N/A N/A 15.663 9.27 2.743 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A 19.134 11.068 3.52 N/A: Not available. 9 Table 3. SPS‐9A Mix Properties (As Constructed) 04B901 Layer Mix type Bulk specific gravity (mean) Bulk specific gravity (minimum) Bulk specific gravity (maximum) Asphalt content (mean) Asphalt content (minimum) Asphalt content (maximum) Air voids (mean) (%) Air voids (minimum) (%) Air voids (maximum) (%) 04B902 04B903 04B961 Standard AC layer Agency mix Standard AC layer SP‐1 (PG 76‐10) Standard AC layer Standard AC layer SMA cellulose N/A N/A N/A Superpave N/A N/A N/A N/A N/A 2.449 N/A N/A N/A N/A N/A N/A 2.265 N/A N/A N/A N/A N/A N/A 2.505 N/A N/A N/A N/A 3.6 3.7 3.4 4.1 3.3 3.6 5.1 3.6 3.65 3.2 4.0 3.2 3.5 5.1 3.6 3.75 3.7 4.1 3.5 3.8 5.1 3.8 6.2 3.1 3.8 2.4 2.9 3.6 3.7 5.55 2.1 3.0 N/A 2.6 3.4 3.8 7.15 4.1 5.1 N/A 3.1 3.8 N/A: Not available. The dynamic modulus (E*) was calculated for Section 041007/04B962. The E* values provided in Table 4 are estimates based on the resilient modulus Artificial Neural Network model developed in 2011 (Kim et al. 2011). 10 Table 4. Dynamic Modulus (E*) for Test Section 041007/04B962 Layer Existing AC Standard AC SP‐3 SP‐1 (PG 76‐10) Temperature (°C) 14 40 70 100 130 14 40 70 100 130 14 40 70 100 130 14 40 70 100 130 Sample Age (Days) 44 44 44 44 44 135 135 135 135 135 135 135 135 135 135 58 58 58 58 58 Frequency 0.1 0.5 1 5 10 25 3694828 1967348 532003 115027 35623 4330092 2935557 1127095 282964 79880 4321909 2919646 1126592 283947 78523 4359022 2954525 1155592 295975 81971 4074048 2497805 834205 196541 56534 4576202 3421560 1602026 474572 134431 4573729 3404883 1597170 476537 133794 4612543 3439803 1629823 494355 139875 4216398 2724167 994522 247269 70009 4664152 3611171 1826265 585666 169391 4664235 3595022 1819156 587732 169201 4703856 3630074 1852780 608223 176830 4500353 3225375 1430188 414452 117823 4832633 3999767 2367267 917717 289834 4838556 3986572 2355060 918662 290872 4880106 4022336 2389785 945083 303140 4603928 3425915 1640226 511889 148342 4891801 4144552 2600349 1092434 363395 4900125 4133229 2586341 1092199 364931 4942501 4169478 2621197 1120821 379601 4725184 3673586 1931954 667432 201351 4959504 4315779 2900745 1349704 485354 4970844 4307322 2884986 1347284 487343 5014285 4344357 2919867 1378316 505439 The gradations for the Superpave and standard agency aggregate are shown in Figure 6. As previously mentioned, some mixture and other data were not available for all test sections. By LTPP definitions, the SPS‐9A project site is a dry, no‐freeze environment (Table 5). The temperature and precipitation information in Table 5 represents 40 years of recorded data collected at nearby weather stations. The solar radiation and humidity data were summarized from 15 years of weather station data from the nearby SPS‐2 project. 11 Control points Restricted zone Figure 6. Gradations of Superpave Aggregate and Agency Standard Aggregate (FHWA 1998) Table 5. Climatic Information for SPS‐9A Annual average daily mean temperature (°F) Annual average daily maximum temperature (°F) Annual average daily minimum temperature (°F) Absolute maximum annual temperature (°F) Absolute minimum annual temperature (°F) Number of days per year above 32 °F Number of days per year below 32 °F Annual average freezing index (°F‐days) Annual average precipitation (inches) Annual average daily mean solar radiation (W/ft2) Annual average daily maxi‐ mum relative humidity (%) Annual average daily minimum relative humidity (%) 40‐Year Average 40‐Year Maximum 40‐Year Minimum 72 74 69 88 90 85 55 60 51 116 123 111 25 31 17 176 196 133 17 43 1 0 0 0 7.7 15.2 1.8 22.7 36.8 1.65 53 64 43 17 22 13 12 Table 6 summarizes the total ESALs computed from traffic‐loading information collected at the SPS‐9 site. The ESAL values for 1993 and 1994 are ADOT estimates; no monitoring traffic data were available for this period. Table 6. SPS‐9A Traffic‐Loading Summary Year 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 ESALs 1,400,000* 1,100,000* 1,283,553 1,253,915 1,289,820 1,374,457 954,526 2,786,163 1,702,068 2,581,494 3,062,289 2004 1,446,194 2005 1,229,188 *ADOT traffic estimates. No monitoring data available. After experiment surfaces were placed in 1995, the following maintenance activities were performed:  Section 04B901 (agency mix, PG 76‐10): No rehabilitation or maintenance conducted.  Section 04B902 (SP‐1, PG 76‐10): Pothole patching in 2003.  Section 04B903 (SP‐1, PG 70‐10): No rehabilitation or maintenance conducted.  Section 04B959 (SMA polymer with ACFC): No rehabilitation or maintenance conducted.  Section 04B960 (SMA polymer): No rehabilitation or maintenance conducted.  Section 04B961 (SMA cellulose): No rehabilitation or maintenance conducted.  Section 04B964 (ARAC): No rehabilitation or maintenance conducted.  Section 041007/04B962 (SP‐3): Pothole patching in 2003. 13 All test sections were placed out of study because of reconstruction in the summer of 2005 except for Section 041007/04B962, which was placed out of study in 2007. Three analyses were conducted on the SPS‐9A project to evaluate pavement performance: deflection, distress, and profile. The following sections address each analysis, including a description of the research approach along with performance comparisons between test sections, overall trends, a summary of the results, and key findings. 14 CHAPTER 2. SPS‐9A DEFLECTION ANALYSIS Falling weight deflectometer (FWD) data provide information about the overall strength (i.e., stiffness) of the pavement structure and individual layers. At the SPS‐9A site, researchers used this information to evaluate changes with time or, as in the case of the asphalt‐bound layers, temperature, and they performed additional analyses to understand how various design features affect structural performance. ANALYSIS OF DEFLECTION DATA Using the nondestructive FWD deflection testing data, researchers can identify the structural condition of the sections over their service life. In this chapter, three levels of analysis are presented. First, researchers produced the deflection profile plots of maximum deflection (D0), minimum deflection (D7/ D8), and AREA value for all the sections to identify changes in the pavement and subgrade over time. Next, they backcalculated subgrade resilient modulus (MR), effective pavement modulus (EP), and effective structural number (SNeff) as outlined in the AASHTO Guide for Design of Pavement Structures (AASHTO 1993). Finally, they backcalculated AC modulus and MR using industry standard software. MAXIMUM DEFLECTION, MINIMUM DEFLECTION, AND AREA VALUE Maximum Deflections The normalized average maximum deflection (Dmax) (D0, measured at the center of the FWD load plate, normalized to a load level of 9000 pounds and an AC mix temperature of 68 °F) typically indicates the total stiffness of the pavement structure (surface and base) and the underlying subgrade. Increases in the normalized average maximum deflection (or Dmax) observed over time may be due to weakening of the pavement structure or weakening of the subgrade. Figure 7 shows Dmax results for each test section from the first round of testing to the last. Except for Section 041007, the first round of testing for all sections was performed in 1995. When this testing was performed, the top lift of the AC layer had not been placed, resulting in the relatively high deflections that were observed throughout the project. The second test was performed in 1997—21 months after construction. As expected, Dmax reduced significantly compared to the first round of testing. 15 04B901 04B902 0 04B B903 04B959 04B9600 04B961 04B964 04 41007 Figure 7. Average No ormalized Dmaax by Test Secction m Deflection Minimum The minim mum deflectio on (Dmin) was observed in the t sensor fa rthest from tthe loading pllate, which fo or LTPP can be either sen nsor No. 7 or sensor s No. 8, depending oon the configu uration used. The Dmin read dings malized to stan ndard 9000 pounds, p but no temperaturre correction factor was applied. The Dmin also norm readings were w indicativve of the subggrade charactteristics. Figuure 8 shows th he Dmin measurement from m the first round d of testing to o the last round of testing. In general, ssubgrade mod dulus did nott change much throughout the test sittes. Section 04B964 had th he strongest ssubgrade and d Section 04B959 had the weakest subgrade. s Inte erestingly, subgrade strength increasedd with time in n Section 041007. 16 04B901 04B902 04 4B903 04B9 959 04B96 0 04B961 04B964 04 41007 Figure 8. Average No ormalized Dmiin by Test Secction AREA Value The AREA A parameter iss commonly used u as a meaans of quantiffying the relaative stiffnesss of a pavemeent section. The T equation for f the AREA value is: A  6D0  D1  D2  D3  / D0 Where (Eq. 1 1) A = AREAA value D0 = surfaace deflectionn at center off test load D1 = surfaace deflectionn at 12 inchess D2 = surfface deflectioon at 24 inchees D3 = surfaace deflectionn at 36 inchess 17 The AREA value is the normalized area of a slice taken through any deflection basin between the center of the loaded area and 36 inches. This area is said to be normalized because it is divided by the maximum deflection, D0. The maximum value of the AREA parameter is 36 inches, which would result from testing an extremely rigid section of pavement, and it occurs when all four deflections are equal. The minimum AREA is 11.02 inches, which would result from deflection measurements on a one‐layer system of homogeneous material. This would imply that the pavement structure is of the same stiffness as the underlying soil. The state of Washington suggested that general trends of pavement condition can be concluded from the combination of AREA value and maximum deflection (Table 7) (Mahoney 1995). Table 7. General Trends of D0 and Area Values (Mahoney 1995) FWD‐Based Parameter AREA Generalized Conclusions Dmax Low Low Weak structure, strong subgrade Low High Weak structure, weak subgrade High Low Strong structure, strong subgrade High High Strong structure, weak subgrade Figure 9 shows the average AREA value of the SPS‐9A test sections from the first round of testing to the last round of testing. As shown in the figure, a significant increase in AREA value between first and second round of testing was observed, which coincides with the Dmax observation. 18 04B901 04B902 04 4B903 04B959 04B9660 04B961 04B964 04 41007 Figgure 9. AREA A Values by Teest Section CULATION PR ROCEDURE BACKCALC The AASH HTO Guide forr Design of Pa avement Strucctures (1993) outlined a prrocedure for calculating MR, the effecttive modulus of all paveme ent layers abo ove the subgrrade, and SNeeff using meassured deflectiion data. The deflections, which w are me easured at a distance d at le ast 0.7 times the radius off the stress bu ulb at nt interface, are a considere ed to reflect t he deformatiion of the sub bgrade layer o only the subgrade‐pavemen b used to com mpute MR. Th he backcalculaated MR can bbe calculated d as: and can be 1    2 MR Where  P rD DR (Eq. 2 2) M R = bacckcalculated subgrade s resilient moduluss  = Poisson’s ratio (  = 0.5 was asssumed in thee analysis) P = applieed load (lbf) r = distannce from centter of load plaate to DR (innches) DR = pavement surfacce deflection at distance r from the cennter of the looad plate (inchhes) 19 The radius of the stress bulb can be determined from the following equation:  EP   ae  a 2   D 3  M R   Where (Eq. 3) ae = radius of stress bulb at the subgrade‐pavement interface (inches) a = FWD load plate radius (inches) D = total thickness of pavement layers (inches) EP = effective pavement modulus M R = backcalculated subgrade resilient modulus To obtain EP in this equation, the researchers used an equation linking the FWD deflection at the center plate (Dmax), EP, and MR: 1  1  2 D  1    1 a d 0  1.5 Pa   Ep   D Ep    3 M 1   R  a MR      Where          (Eq. 4) d0 = deflection at the pavement surface and adjusted to a standard temperature of 68 °F (inches) P = contact pressure under the loading plate (psi) a = load plate radius (inches) M R = subgrade resilient modulus (psi) D = actual pavement structure thickness (inches) EP = effective modulus of the pavement structure (psi) 20 Once EP was determined, SNeff could be calculated: SN eff  0.0045 D E p  0.33 Where (Eq. 5)  SN eff  effective structural number D = total thickness of pavement structure above the subgrade (inches) EP = effective modulus of the pavement structure above the subgrade (psi) To accommodate the large quantity of data, the researchers developed a spreadsheet to calculate MR, EP, and SNeff for each test section. Table 8 presents the statistics of these structural parameters. For most of the sections, MR remained fairly constant over the monitoring period. The only exception was observed in Section 041007, where the backcalculated MR jumped from 33 ksi in 1994 to 69 ksi in 1995 and decreased to 33 ksi in 1997. This fluctuation could be due to the construction on the section in 1995. In terms of pavement structure, all test sections showed a similar trend in that EP and SNeff remained fairly constant in the 1997, 1999, and 2002 testing. However, researchers observed a decreased EP and SNeff in the last round of testing. 21 Table 8. Structural Parameter Statistics for SPS‐9A MR (psi) Section Date 04B901 04B901 04B901 04B901 04B901 04B902 04B902 04B902 04B902 04B902 04B903 04B903 04B903 04B903 04B903 04B959 04B959 04B959 04B959 04B959 04B960 04B960 04B960 04B960 04B960 04B961 04B961 04B961 04B961 04B961 04B964 04B964 04B964 04B964 04B964 041007 041007 041007 041007 041007 041007 041007 041007 1995 1997 1999 2002 2005 1995 1997 1999 2002 2005 1995 1997 1999 2002 2005 1995 1997 1999 2002 2005 1995 1997 1999 2002 2005 1995 1997 1999 2002 2005 1995 1997 1999 2002 2005 1989 1991 1994 1995 1997 1999 2002 2005 Ep (psi) Average Maximum Minimum 28,949 30,949 32,871 37,831 28,529 29,358 32,687 33,542 35,680 29,868 26,580 32,357 31,638 43,805 35,165 19,017 25,928 27,160 25,788 23,765 23,862 30,122 31,379 30,024 29,167 28,151 33,420 34,192 33,709 32,191 38,767 41,652 43,940 46,528 41,183 49,092 38,798 33,301 69,973 33,205 34,584 35,338 31,516 40,648 39,980 41,180 63,421 37,779 32,949 34,782 35,486 43,918 37,235 28,533 41,607 35,479 92,041 53,471 21,774 29,983 31,151 29,404 27,492 25,904 35,505 36,306 33,951 34,731 30,832 36,819 37,360 36,783 38,099 47,576 49,051 52,980 59,964 49,657 65,536 48,721 40,727 86,868 36,522 38,738 56,764 61,815 17,693 21,703 23,327 25,439 19,614 25,105 30,628 30,959 28,849 23,749 24,874 28,577 28,464 29,890 24,310 17,176 21,428 23,189 22,016 20,718 21,126 27,881 29,211 27,723 25,585 24,254 30,685 30,934 31,509 28,905 28,941 30,434 32,478 33,516 29,079 36,758 30,569 26,524 55,787 27,632 28,995 26,784 24,346 COV (%) 19.4 16.4 15.7 27.7 16.3 6.2 4.2 4.5 12.7 12.6 4.1 12.8 5.8 50.4 26.9 6.3 9.2 8.2 8.5 9.0 4.8 6.3 5.2 4.6 7.0 5.5 5.1 4.8 4.7 8.1 14.4 11.4 12.3 14.4 13.7 14.4 10.0 9.9 11.9 6.1 6.1 16.6 18.5 22 Average Maximum Minimum 136,723 265,736 274,896 306,204 183,396 134,326 295,456 315,002 332,304 250,433 155,837 286,703 301,070 383,312 279,985 123,133 326,017 315,924 381,757 311,894 171,883 352,891 353,936 413,589 356,083 158,711 372,434 387,084 422,313 374,253 184,452 325,854 319,282 363,137 347,947 133,433 164,005 186,302 116,089 338,247 399,774 353,617 239,315 183,080 366,474 397,821 460,182 311,676 186,140 364,630 393,299 477,059 336,611 230,668 381,018 390,280 631,204 384,813 217,788 462,314 421,827 546,338 436,190 282,444 494,825 498,019 590,113 508,556 236,806 460,651 482,755 528,367 454,893 323,261 425,091 424,569 499,699 455,412 194,227 263,434 289,728 178,055 461,604 529,702 537,917 461,379 98,503 214,224 169,551 156,372 74,500 90,208 240,600 251,270 237,039 182,012 105,480 210,870 234,847 196,096 161,391 68,275 208,958 45,251 246,210 191,329 100,725 279,765 45,357 323,200 248,675 99,114 276,907 287,473 316,203 290,402 108,978 219,385 215,871 251,992 197,200 67,490 78,824 81,070 77,353 184,745 204,192 149,727 94,113 COV (%) 15.0 14.9 19.2 29.2 33.9 17.9 11.5 12.1 20.8 18.1 18.3 13.2 12.6 29.9 23.9 30.5 19.5 24.9 20.6 20.2 24.9 14.8 23.2 15.7 18.0 24.3 13.0 13.6 14.9 13.0 29.0 16.7 18.1 19.8 17.9 28.5 30.5 34.9 25.3 20.3 20.2 28.5 34.2 SNeff 3.98 5.86 5.87 5.81 4.78 4.00 6.22 6.37 6.24 5.75 4.19 6.33 6.43 6.52 5.91 3.53 6.07 6.10 6.24 5.96 4.03 6.32 6.41 6.61 6.26 3.96 6.44 6.52 6.60 6.40 4.10 6.03 5.99 6.25 6.22 3.48 3.79 3.86 4.30 6.20 6.64 5.93 5.26 BACKCALCULATION USING EVERCALC SOFTWARE The FWD data were also processed through the backcalculation software Evercalc developed by Washington State Department of Transportation. One set of FWD data at each station was selected for backcalculation using the representative thickness of each test section obtained from the LTPP database to determine MR of each layer. Table 9 shows the seed value and modulus range used for backcalculation. The pavement structure was first assumed as a four‐layer system for analysis: AC, aggregate base (AB), subgrade, and bedrock. However, after running several initial analyses, the researchers found that the base layer was not producing reasonable moduli values. Consequently, instead of calculating each individual layer moduli, the base layer was combined into the subgrade layer for consideration, and the backcalculation analysis was repeated. The results of this approach produced more reasonable moduli values. Table 9. Backcalculation Seed Value and Modulus Range Layer Description AC AB Subgrade Seed Modulus (ksi) 400 25 15 Poisson’s Ratio 0.35 0.3 0.4 Minimum Modulus (ksi) 100 10 5 Maximum Modulus (ksi) 2100 150 50 Table 10 provides the statistics of backcalculated moduli for the test sections. A similar pavement response was observed on the three LTPP test sections (04B901, 04B902, and 04B903). A rapid increment of the AC moduli was observed in the second round of testing (1997). The AC moduli dropped significantly in 2002. In general, Sections 04B902 and 04B903 (SP‐1) showed a higher AC moduli value compared to Section 04B901 (agency standard mix). No significant difference was found between Section 04B902 (PG 76‐10) and Section 04B903 (PG 70‐10). All three SMA sections (04B959, 04B960, and 04B961) showed superior AC moduli among the eight test sections. Section 04B964 (ARAC) and Section 041007 (SP‐3) showed the lowest AC moduli values. 23 Table 10. Backcalculation Moduli Statistics for SPS‐9A Test Sections Section Date Backcalculated AC Modulus (ksi) 04B901 04B901 04B901 04B901 04B901 04B902 04B902 04B902 04B902 04B902 04B903 04B903 04B903 04B903 04B903 04B959 04B959 04B959 04B959 04B959 04B960 04B960 04B960 04B960 04B960 04B961 04B961 04B961 04B961 04B961 04B964 04B964 04B964 04B964 04B964 041007 041007 041007 041007 041007 041007 041007 041007 1995 1997 1999 2002 2005 1995 1997 1999 2002 2005 1995 1997 1999 2002 2005 1995 1997 1999 2002 2005 1995 1997 1999 2002 2005 1995 1997 1999 2002 2005 1995 1997 1999 2002 2005 1989 1991 1994 1995 1997 1999 2002 2005 389.7 693.5 729.9 340.8 321.7 320.3 793.8 831.8 428.1 468.8 487.8 774 806.4 554.5 554.2 509.4 1257.7 1469.8 833.1 1112.4 303.5 1730 1514.6 723 1309.4 505.4 1593.7 1563.6 1343.8 1273.4 448.8 703.5 590 353.5 626.3 373.8 304.1 412.2 244.5 906.2 845.9 457.7 409.8 24 Backcalculated Subgrade Modulus (ksi) 25.7 33.5 35.8 39.4 29.1 23.7 36.8 36.8 38.5 31.1 22.5 34.9 32.8 42.5 35.4 15.9 25.1 27.5 24.2 22.5 19.4 32.5 33.5 30.5 29.8 24 35.9 35.5 36.3 34.9 31.1 48.2 48.3 50 45.5 30.1 29.4 24.6 24.2 32.9 34.2 34.1 30.1 Root‐Mean Square Error (%) 6.07 5 3.72 8.48 3.12 8.2 5.78 5.58 9.9 6.62 4.93 5.18 4.99 8.17 6.23 2.63 1.34 1.93 1.78 1.77 8.09 3.33 3.8 5.14 2.85 2.23 2.52 2.28 3.06 2.48 18.87 11.27 12.47 18.84 10.25 10.1 3.66 2.17 7.18 2.71 3.43 4.39 2.17 In general, backcalculaated AC modu ulus decrease ed as the paveement aged (Figure 10). In n some cases the AC modullus increased with time, which w could be e due to agingg of the asphaalt binder. Th he decrease in n AC moduli vaalue is probab bly due to the e appearance of distress annd the resultiing weakening in overall pavementt strength. Figure 11 shows the baackcalculated d subgrade mo odulus valuess. The variatio on in the Evercalc backcalcu ulated subgrad de modulus within w each se ection is simillar to the preevious backcalculated subggrade modulus using u the American Associiation of State e Highway annd Transportaation Officialss (AASHTO) procedure e except Section 041007. In I general, un niform subgraade modulus was observed d at the test ssite during the e test period. The average backcalculatted subgrade modulus of aall sections is 32 ksi. The he highest co oefficient of variation v (COV V) of the subggrade moduluus was 21 perrcent in Sectio on 04B903; th lowest CO OV was 14 percent in Section 041007. 04B9 901 04B902 2 04B903 04B959 044B960 04B9961 04B964 4 041007 Figure 10. Backcalculatted AC Modu ulus by Test Section 25 04B9 901 04B902 2 04B903 04B959 044B960 04B9961 04B964 4 041007 Figure e 11. Backcalculated MR b by Test Sectio on KEY FINDINGS FROM THE T SPS‐9A DEFLECTION D ANALYSIS A The deflecction at the center c plate shows an imprrovement in ppavement strrength betweeen testing conducted d in 1995 and d 1997 as a re esult of adding the structu ral layer in 19995 after the first round of testing waas completed d. All of the se ections remained structuraally sound during the 1997 7, 1999, and 2 2002 testing. Th he researchers observed a decline in paavement struucture in the llast round of testing in 200 05, mainly du ue to the weakening in the e AC layer, sin nce the subgraade didn’t change significaantly over tim me. The MR, EP, and SNeff re esults after ussing the AASH HTO Guide forr Design of Paavement Stru uctures (1993)) procedure e follow:  Th he average EP showed an increasing i tre end in the seccond, third, a nd fourth rou und of testingg and decreased in the last round d of testing. These results ccan be explained by age haardening of the AC t last roundd of testing because signifficant pavemeent laayer. A decreaase of EP was observed in the distresses deve eloped.  5 construction n, the average e backcalcula ted SNeff did not change significantly w within After the 1995 o the monitoring period d. eaach section over 26 The results of MR and the AC moduli after using the industrial standard backcalculation software Evercalc are summarized below:  After construction in 1995, little variation in the AC moduli was observed in most sections. The first significant drop was observed in 2002 in all sections, however, the AC moduli bounced back in the 2005 testing. A similar trend was observed among the three core LTPP test sections. The Superpave mix design sections showed higher AC moduli values compared to the agency standard mix section. No significant difference was found between the PG 76‐10 and PG 70‐10 sections. The three SMA sections showed the highest AC moduli value among all test sections.  The trend of the Evercalc backcalculated subgrade moduli was similar to the trend observed in the subgrade moduli using the AASHTO procedure. In general, uniform subgrade modulus was observed at the test site during the test period. The average backcalculated subgrade modulus of all sections was 32 ksi. The COV of subgrade modulus within each test section ranged from 14 percent to 21 percent. 27 28 CHAPTER 3. SPS‐9A DISTRESS ANALYSIS This chapter includes analyses and results from evaluating distress data collected from the SPS‐9A site using LTPP manual survey techniques (Miller and Bellinger 2003). Surface distress provides powerful information about the nature and extent of pavement deterioration, which can be used to quantify performance trends as well as to investigate how design features affect service life. All of the flexible SPS‐9A test sections were constructed consecutively and exposed to the same traffic‐ loading, climate, and subgrade conditions, allowing for direct comparisons between layer configurations and design features without the confounding effects introduced by different in situ conditions. AC DISTRESS TYPES Surface deterioration is composed of multiple distress types. The raw distress data for each section are not included in this report but are available for download from LTPP Products Online (http://www.infopave.com/Data/StandardDataRelease/). Distress type definitions follow (Huang 1993):  Fatigue cracking: A series of interconnecting cracks caused by repeated traffic loading. Cracking initiates at the bottom of the asphalt layer where tensile stress is the highest under the wheel load. With repeated loading, the cracks propagate to the surface.  Longitudinal wheelpath (WP) cracking: Cracking parallel to the centerline occurring in the WP. This cracking can be the early stages of fatigue cracking or can initiate from construction‐related issues such as paving seams and segregation of the mix during paving. In the latter case, cracking is typically very straight (no meandering).  Longitudinal non‐wheelpath (NWP) cracking: Cracking parallel to the centerline occurring outside the WP. This cracking is not load‐related and can initiate from paving seams or where segregation issues occurred during paving. Cracking can also be caused by tensile forces experienced during temperature changes. Pavements with oxidized or hardened asphalt are more prone to this type of cracking.  Transverse cracking: Cracking that is predominantly perpendicular to the pavement centerline. Cracking starts from tensile forces experienced during temperature changes. Pavements with oxidized or hardened asphalt are more prone to this type of cracking.  Block cracking: Cracking that forms a block pattern and divides the surface into approximately rectangular pieces. Cracking initiates from tensile forces experienced during temperature changes. This distress type indicates that the AC has significantly oxidized or hardened.  Raveling: Wearing away of the surface caused by dislodging of aggregate particles and loss of asphalt binder. Raveling is caused by moisture stripping and asphalt hardening. 29  Bleeding: Excessive bituminous binder on the surface that can lead to loss of surface texture or a shiny, glass‐like, reflective surface. Bleeding is a result of high asphalt content or low air void content in the mix.  Rutting: A surface depression in the WPs. Rutting can result from consolidation or lateral movement of material due to traffic loads. It can also signify plastic movement of the asphalt mix because of inadequate compaction, excessive asphalt, or a binder that is too soft given the climatic conditions. These distress types can be grouped into two general categories based on cause of failure mechanism: structural or environmental factors. Table 11 summarizes the flexible pavement distress types and their associated failure mechanisms. Table 11. Flexible Pavement Distress Types and Failure Mechanisms Failure Mechanism Traffic/Load Climate/Materials Related Related X X X X X X X X X Distress Type Fatigue cracking Longitudinal WP cracking Longitudinal NWP cracking Transverse cracking Block cracking Raveling Bleeding Rutting RESEARCH APPROACH Investigators began this analysis with a cursory review of all distress data collected at each test section to identify suspect or inconsistent information. Team members used photos and distress maps to verify quantities reported in the database. Because of the subjective nature of the data collection technique (raters must select distress type and severity based on a set of rules), variation is expected in distress data. The SPS‐9A data set was well within the acceptable range of variability. Distress data are reported at three severity levels: low, moderate, and high. Inconsistencies between severity levels within a distress type create one of the largest sources of variability in distress data (Rada et al. 1999). In addition, conducting analyses on three separate severity levels for each distress type becomes increasingly complex with results that are difficult to interpret. To reduce variability and to 30 consolidate the information for analyses, the researchers summed the quantities from the three severity levels into one composite value. As shown in Table 11, pavement deterioration (when not directly attributable to mix problems or construction deficiencies) can be attributed to structural or environmental factors. Structural factors are the result of traffic loading relative to the structural capacity of the pavement section. Environmental factors represent the influence of climate on pavement deterioration. Therefore, structural and environmental indices were developed to focus the analyses on overall structural and environmental damage, which are more consistent and provide a better avenue for comparison, rather than on individual types of distress, which vary from section to section and year to year. The structural damage index consists of those distresses generally manifesting from the portion of the pavement that experiences loading (i.e., WPs). Therefore, the structural damage index was presented as the percentage of WP damage and included fatigue and longitudinal WP cracking. To normalize fatigue and longitudinal cracking, the structural damage index took the form of the following expression: S Where F  1 ft  Clwp 2Wwp Ls (Eq. 6)  S = structural damage index F = area of fatigue (ft2) Clwp = length of longitudinal WP cracking (ft) Wwp = width of WP = 3.28 (ft) Ls = length of test section (ft) The environmental damage index is a composite of distresses that generally result from climatic effects. The entire pavement surface is subject to environmental distress; therefore, the environmental damage index was characterized as the percentage of total pavement area damaged. Typically, transverse cracking, longitudinal cracking (outside of the WPs), and block cracking are specific to environmental damage. To normalize the environmental distress for the total area, the environmental damage index was expressed as: E B Cnwp Ct   Atot Ls Ls (Eq. 7)  31 Where E = environmental damage index B = area of block cracking (ft2) C nwp = length of NWP cracking (ft) Ct = length of transverse cracking (ft) Atot = total area of test section (ft2) Ls = length of test section (ft) Although the structural and environmental distress factors clearly affected the SPS‐9A project’s structural and functional service life, rutting, patching, and other surface defects (such as potholes, bleeding, and raveling) also affected performance. Rutting data reported in this study were generated using a 6 ft straightedge reference (Simpson 2001). Replicate data were not collected for the SPS‐9A project. Therefore, standard statistical comparisons (i.e., t tests) to determine the significance of findings could not be conducted. Instead, the evaluation consisted of graphical comparisons between test sections from data collected at the same points in time. OVERALL PERFORMANCE TREND OBSERVATIONS While gathering pavement distress data, researchers became aware of a few significant trends affecting the overall pavement performance of the project. These observations were clearly driving issues for this project and were intrinsically important pieces of the distress performance. Before receiving the experimental AC overlays, each section received the same rehabilitation treatment and existing AC was milled. Manual distress surveys were performed on each section after milling, but before the experimental layers were added. These surveys showed similar conditions throughout all the sections. No global preventive maintenance or rehabilitation was performed on any of the test sections. Sections 04B961 and 04B964 exhibited raveling in the 2005 and 2002 surveys, respectively. Sections 04B902 and 041007 exhibited significant pumping; Sections 04B903 and 04B959 experienced minimal pumping. Figure 12 shows the structural damage trends for each section. The performance trends are relatively consistent and within the expected range of variation. All sections (except 04B959 and 04B961) showed a rapid accumulation of structurally related distresses approximately 10 years after construction. 32 The Marshall mix (Section 04B901) contained a smaller percentage of asphalt binder than the Superpave mixes (Sections 04B902 and 04B903). Higher asphalt binder contents typically produce better resistance to fatigue cracking; however, all three sections accumulated similar amounts of fatigue cracking. Compared to the rest of the SPS‐9A project, Section 04B959 (SMA polymer with ACFC) and Section 04B961 (SMA cellulose) exhibited significantly smaller amounts of structural damage accumulation. Manual Survey Distress Data 160% 140% 04B901 04B902 04B903 04B959 04B960 04B961 04B964 041007/B962 Structural Damage Index 120% 100% 80% 60% 40% 20% Ja n06 Ja n05 Ja n04 Ja n03 Ja n02 Ja n01 Ja n00 Ja n99 Ja n98 Ja n97 Ja n96 Ja n95 Ja n94 0% Date Figure 12. Structural Damage Trends for SPS‐9A Test Sections Figure 13 shows the overall environmental damage trends for each section. The performance trends are relatively consistent and within the expected range of variation. 33 Manual Survey Distress Data 350% 300% 04B901 04B902 04B903 04B959 04B960 04B961 04B964 041007/B962 Environmental Damage Index 250% 200% 150% 100% 50% Ja n06 Ja n05 Ja n04 Ja n03 Ja n02 Ja n01 Ja n00 Ja n99 Ja n98 Ja n97 Ja n96 Ja n95 Ja n94 0% Date Figure 13. Environmental Damage Trends for SPS‐9A Test Sections Performance Comparisons In‐depth analyses and comparisons were conducted for all of the SPS‐9A test sections. Figure 14 summarizes the structural damage index and pavement structure for each section; Figure 15 summarizes the environmental damage index and pavement structure. Both damage indices reported are based on the data collected in January 2005 (just before going out of study). 34 2005 Manual Distress Data (B900) 160% 140% 140% 120% 111% Structural Damage Index 108% 106% 102% 100% 100% 80% 60% 40% 20% 15% 3% 0% B901 04B901 AS PG 76-10 (19mm) B902 04B902 SP-1 PG 76-10 (19mm) B903 04B903 SP-1 PG 70-10 (19mm) B959 04B959 SMA Polymer W/ACFC B960 04B960 SMA Polymer B961 04B961 SMA Cellulose B964 04B964 1007/B962 041007/04B962 ARAC SP-3 Figure 14. Structural Damage Index and Pavement Structure Summary 35 2005 Manual Distress Data (B900) 350% 302% 300% Environmental Damage Index 250% 214% 200% 150% 136% 126% 106% 100% 50% 38% 22% 7% 0% 04B901 B901 AS PG 76-10 (19mm) 04B902 04B903 B902 B903 SP-1 PG 76-10 SP-1 PG 70-10 (19mm) (19mm) 04B959 B959 SMA Polymer W/ACFC 04B960 B960 SMA Polymer 04B961 B961 SMA Cellulose 04B964 B964 041007/04B962 1007/B962 ARAC SP-3 Figure 15. Environmental Damage Index and Pavement Structure Summary Figure 16 summarizes the rutting and pavement structure for each section. Significant variation in rutting performance between the sections did not exist. All sections exhibited less than 7 mm of rutting after over seven years in service, which is well below the level required to trigger improvements in most pavement management systems. Therefore, rutting was not the driving factor in the overall condition of the pavement. 36 2005 Rutting Index (B900) 7 6.4 6.2 6.0 6 5.5 Rutting Index (mm) 5 4.6 4.1 3.8 4 3.6 3 2 1 0 04B901 B901 AS PG 76-10 (19mm) 04B902 B902 SP-1 PG 76-10 (19mm) 04B903 B903 SP-1 PG 70-10 (19mm) 04B959 B959 SMA Polymer W/ACFC 04B960 B960 SMA Polymer 04B961 B961 SMA Cellulose 04B964 B964 041007/04B962 1007/B962 ARAC SP-3 Figure 16. Rutting Index and Pavement Structure Summary Following is a synopsis of the findings and performance of each section, including structural deterioration, environmental deterioration, rutting, and other unique circumstances. Section 04B901 (Standard Agency Mix, Marshall 75, PG 76‐10) This section exhibited average cracking performance, both structurally and environmentally. The structural damage index was 108 percent. Five of eight test sections experienced similar quantities of structural damage. The environmental damage index was 126 percent. Three of eight test sections experienced similar quantities of structural damage. The rate of structural and environmental deterioration increased from 2002 to 2005. Section 04B902 (SP‐1, PG 76‐10) This section exhibited the most structural damage of all the test sections; however it performed very well against environmental damage. The structural damage index was 140 percent, while most sections experienced quantities near 100 percent. The environmental damage index was 38 percent—much 37 lower than the majority of the test sections, which experienced quantities from 100 percent to 300 percent. Section 04B902 also exhibited the highest amount of pumping among the test sections. The rate of structural deterioration increased from 2002 to 2005. Section 04B903 (SP‐1, AC‐40 [PG 70‐10]) Structural deterioration was average in this section, but the environmental deterioration was well below the average for the SPS‐9A project. In fact, Section 04B903 exhibited the highest resistance to environmental damage. The rate of structural deterioration increased from 2002 to 2005. Section 04B959 (SMA Polymer with ACFC) This section exhibited the highest resistance against structural damage for the SPS‐9A project. After 10 years, the structural damage index was at 3 percent, while most sections had a structural damage index over 100 percent. However, this section also exhibited the second largest amount of environmental damage of the entire SPS‐9A project as a result of NWP longitudinal cracking along the edge of the lanes throughout the entire section. The rate of environmental deterioration increased from 2002 to 2005. Section 04B960 (SMA Polymer) Like Section 04B901, this section exhibited average cracking performance, both structurally and environmentally. The structural damage index was 102 percent. Five of eight test sections experienced similar quantities of structural damage. The environmental damage index was 106 percent. Three of eight test sections experienced similar quantities of structural damage. The rate of structural and environmental deterioration increased from 2002 to 2005. Section 04B961 (SMA Cellulose) This section was the best‐performing pavement in the entire SPS‐9A project. It was significantly better at resisting structural and environmental damage throughout the duration of its life when compared to the other test sections. The environmental and structural damage indices were each approximately 20 percent, which was significantly lower than the majority of the other test sections. Sections 04B961 and 04B964 experienced the highest amount of pavement raveling. Section 041007/04B962 (SP‐3) Like Sections 04B901 and 04B960, this section exhibited average cracking performance, both structurally and environmentally. The structural damage index was 106 percent. Five of eight test sections experienced similar quantities of structural damage. The environmental damage index was 136 percent. Three of eight test sections experienced similar quantities of structural damage. Like Section 04B902, this section exhibited a significant amount of pumping. The rate of structural and environmental deterioration increased from 2002 to 2005. 38 Section 04B964 (ARAC) This section performed average against structural damage, but accumulated the highest amount of environmental damage. The structural damage index was 100 percent. Five of eight test sections experienced similar quantities of structural damage. However, the environmental damage index was 302 percent, which was significantly higher than all the other test sections. Sections 04B964 and 04B961 experienced the highest amount of pavement raveling. KEY FINDINGS FROM THE SPS‐9A DISTRESS ANALYSIS The distress data captured at the SPS‐9A project provide valuable insight into pavement performance, design, management, and construction. Highlights from the SPS‐9A distress analysis follow:  Most every section (except Sections 04B959 and 04B961) showed significant growth in fatigue and longitudinal cracking 10 years after construction.  Construction quality can play a major role in performance. The construction observations documented in the LTPP construction report were limited to Sections 04B901, 04B902, and 04B903. However, the report showed that all three sections were free of any construction issues.  Half the sections (04B901, 04B959, 04B960, and 041007/04B962) had reasonable patterns of environmental distress growth with a clear increase in magnitude approximately 10 years after construction. Section 04B964 had little environmental distress growth for seven years but it rapidly increased after 10 years.  For the SP‐1 mix designs (Sections 04B902 and 04B903), the PG 70‐10 binder (Section 04B903) performed significantly better structurally and environmentally than the PG 76‐10 binder (Section 04B902). It also performed significantly better than the standard agency mix at resisting environmental deterioration.  SMA with cellulose fibers was, by far, the best pavement mix at resisting both structural and environmental deterioration.  All sections performed well with regard to rut resistance. Rutting would not have triggered a rehabilitation event for any section.  With no replicate sections, there is limited ability to assess potential variability independent of actual performance.  Three sections (Sections 04B959, 04B902, and 041007/04B962) received patching at some point. 39 40 CHAPTER 4. SPS‐9A ROUGHNESS ANALYSIS This chapter characterizes the surface roughness of these sections throughout their service life and links the observations to records of pavement distress and its development. Investigators collected road profile measurements on this site about once per year starting with the winter after the site was opened to traffic. This study analyzed the profiles in detail by calculating their roughness values, examining the spatial distribution of roughness within them, viewing them with post‐processing filters, and examining their spectral properties. These analyses provided details about the roughness characteristics of the road and provided a basis for quantifying and explaining the changes in roughness with time. PROFILE DATA SYNCHRONIZATION Profile data were collected from the entire SPS‐9A site on eight dates, from January 29, 1997, through December 17, 2004 (Table 12). Raw profile data were available for all visits. Each visit produced a minimum of five repeat profile measurements. Table 12. Profile Measurement Visits of the SPS‐9A Site Visit Date 01 Jan. 29, 1997 Feb. 2, 1997 Dec. 5, 1997 Dec. 7, 1998 Nov. 10, 1999 Nov. 12, 2001 Feb. 09, 2004 Dec. 17, 2004 02 03 04 05 06 07 08 Repeats by Section 04B901 04B902 04B903 04B959 04B960 04B961 04B962 04B964 7 7 7 7 7 7 7 7 — — — — — — 9 9 7 7 7 7 7 7 7 7 5 7 5 5 5 5 7 7 7 7 7 7 7 7 7 7 7 9 7 9 7 9 7 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 In visit 01, all eight sections were covered together in each profile. Visit 02 only covered Sections 04B962 and 04B964. Visit 03 covered Sections 04B962 and 04B964 together in one set of passes, and all of the other sections together in another. Each section was measured individually in visits 04 through 08. 41 Researchers used the raw data to synchronize all of the profiles to each other through their entire history. Three clues were available for this purpose: (1) the site layout from the construction report, (2) event markers in the raw profiles from the start and end of each section, and (3) automated searching for the longitudinal offset between repeat measurements. DATA EXTRACTION Researchers extracted profiles of individual test sections directly from the raw measurements for two reasons: First, profiles were collected in visits 01 through 06 at a 0.98‐inch sample interval and in visits 07 and 08 at about a 0.77‐inch sample interval. These data appeared in the database after the application of an 11.8‐inch moving average and decimation to a sample interval of 5.91 inches. The raw data contained the more detailed profiles. Second, this study depended on consistency of the profile starting and ending points with the construction layout and consistency of the section limits with time. In particular, a previous quality check revealed some profiles were shifted (Evans and Eltahan 2000). CROSS CORRELATION Searching for the longitudinal offset between repeat profile measurements that provides the best agreement between them is a helpful way to refine their synchronization. This can be done by inspecting filtered profile plots, but it is very time‐consuming. Visual assessment is also somewhat subjective when two profiles do not agree well, which is often the case when measurements are made several years apart. In this study, investigators used an automated procedure rather than visual inspection to find the longitudinal offset between measurements. In this procedure, which is based on a customized version of cross correlation (Karamihas 2004), a basis measurement is designated that is considered to have the correct longitudinal positioning. A candidate profile is then searched for the longitudinal offset that provides the highest cross correlation to the basis measurement. A high level of cross correlation requires a good match of profile shape, the location of isolated rough spots, and overall roughness level. Therefore, the correlation level is often only high when the two measurements are synchronized. When the optimal offset is found, a profile is extracted from the candidate measurement with the proper overall length and endpoint positions. For the remainder of this discussion, this procedure will be referred to as automated synchronization. For this application, investigators performed cross correlation after the International Roughness Index (IRI) filter was applied to the profiles rather than using the unfiltered profiles, which helped assign the proper weighting to relevant profile features. In particular, it increased the weighting of short‐ wavelength roughness that may be linked to pavement distress. This enhanced the effectiveness of the automated synchronization procedure. The long‐wavelength content within the IRI output helped ensure that the longitudinal positioning was nearly correct, and the short‐wavelength content was able to leverage profile features at isolated rough spots to fine‐tune the positioning. 42 SYNCHRONIZATION To extract profiles of individual test sections from the raw measurements, investigators: 1. Established a basis measurement for each section from visit 08. The first repeat measurement was used for this purpose. All of the sections were 500 ft long. 2. Automatically synchronized the other eight repeats from visit 08 to the basis set. 3. Automatically synchronized the measurements from the previous visit to the current basis set. 4. Designated the previous visit as the current visit. 5. Replaced the basis set with a new set of synchronized measurements from the first repeat of the current visit. 6. Repeated steps 3 through 5 until visit 01 was complete. DATA QUALITY SCREENING Investigators performed data quality screening to select five repeat profile measurements from each visit of each section. Among the group of available runs, investigators selected the five measurements that exhibited the best agreement with each other. In this case, agreement between any two profile measurements was judged by cross‐correlating them after applying the IRI filter (Karamihas 2004). In this method, the IRI filter is applied to the profiles; then the output signals are compared rather than the overall index. High correlation by this method requires that the overall roughness as well as the details of the profile shape that affect the IRI agree. The IRI filter was applied before correlation in this case for several reasons: • Direct correlation of unfiltered profiles places a premium on very long‐wavelength content, but ignores much of the contribution of short‐wavelength content. • Correlation of IRI filter output emphasizes profile features in (approximate) proportion to their effect on the overall roughness. • Correlation of IRI filter output provides a good trade‐off between emphasizing localized rough features at distressed areas in the pavement and placing too much weight on the very short‐ duration, narrow features (spikes) that are not likely to agree between measurements because the IRI filter amplifies short‐wavelength content, but attenuates macrotexture, megatexture, and spikes. • A relationship has been demonstrated between the cross‐correlation level of IRI filter output and the expected agreement in overall IRI (Karamihas 2004). 43 Note: This method was performed with a special provision for correcting modest longitudinal distance measurement errors. Each comparison between profiles produced a single value that summarized their level of agreement. When nine repeat profile measurements were available, they produced 36 correlation values. Any subgroup of five measurements could be summarized by averaging the relevant 10 correlation values. Researchers selected the subgroup that produced the highest average and excluded the other repeats from most of the analyses discussed in the remainder of this report. Since the number of available profiles ranged from six to nine, the number of measurements that were excluded ranged from one to four. Tables 13 through 20 list the selected repeats for each visit of each section and the composite correlation level produced by them. Visit 01 03 04 05 06 07 08 Visit 01 03 04 05 06 07 08 Visit 01 03 04 05 06 07 08 1 1 1 2 2 1 2 Table 13. Selected Repeats of Section 04B901 Repeat Numbers Composite Correlation 2 3 5 6 0.901 3 4 5 7 0.898 2 3 4 5 0.916 4 5 6 7 0.912 3 4 6 7 0.870 3 5 7 9 0.659 3 5 7 8 0.722 2 1 1 2 5 3 1 Table 14. Selected Repeats of Section 04B902 Repeat Numbers Composite Correlation 3 4 5 7 0.935 3 4 5 7 0.896 2 4 5 6 0.931 3 4 5 6 0.871 6 7 8 9 0.827 6 7 8 9 0.639 5 7 8 9 0.768 2 1 1 3 1 1 1 Table 15. Selected Repeats of Section 04B903 Repeat Numbers Composite Correlation 3 5 6 7 0.929 2 4 6 7 0.914 2 3 4 5 0.918 4 5 6 7 0.943 3 4 6 7 0.894 3 5 6 7 0.853 2 3 8 9 0.588 44 Visit 01 03 04 05 06 07 08 Visit 01 03 04 05 06 07 08 Visit 01 03 04 05 06 07 08 Visit 01 02 03 04 05 06 07 08 1 1 1 2 2 1 3 Table 16. Selected Repeats of Section 04B959 Repeat Numbers Composite Correlation 3 4 6 7 0.932 2 4 5 7 0.923 2 3 4 5 0.921 3 4 5 6 0.922 5 7 8 9 0.944 3 5 6 7 0.905 5 6 8 9 0.870 1 1 1 1 1 1 5 Table 17. Selected Repeats of Section 04B960 Repeat Numbers Composite Correlation 2 3 5 7 0.925 2 4 6 7 0.915 2 3 4 5 0.903 2 3 4 6 0.924 2 3 4 7 0.949 2 3 4 5 0.853 6 7 8 9 0.864 1 2 1 1 1 2 1 Table 18. Selected Repeats of Section 04B961 Repeat Numbers Composite Correlation 2 3 4 7 0.943 4 5 6 7 0.948 2 3 4 5 0.915 2 4 5 7 0.942 2 3 4 6 0.953 3 5 6 9 0.901 6 7 8 9 0.826 2 2 1 1 1 1 3 1 Table 19. Selected Repeats of Section 04B962 Repeat Numbers Composite Correlation 3 4 6 7 0.873 4 5 7 9 0.910 2 4 5 7 0.876 2 5 6 7 0.914 2 3 4 6 0.901 2 3 4 6 0.793 5 6 8 9 0.735 3 4 6 7 0.651 45 Visit 01 02 03 04 05 06 07 08 1 2 1 1 1 2 1 1 Table 20. Selected Repeats of Section 04B964 Repeat Numbers Composite Correlation 2 4 5 6 0.914 5 7 8 9 0.934 2 5 6 7 0.878 2 4 5 6 0.888 2 5 6 7 0.883 3 6 7 9 0.937 5 7 8 9 0.833 2 4 6 9 0.728 The process for selecting five repeat measurements from a larger group is similar to the practice within LTPP except that it is based on composite agreement in profile rather than the overall index value. The correlation levels listed in Tables 13 through 20 provide an appraisal of the agreement between profile measurements for each visit of each section. When two profiles produce a correlation level above 0.82, their IRI values are expected to agree within 10 percent most (95 percent) of the time. Above this threshold, the agreement between profiles is usually acceptable for studying the influence of distresses on profile. When two profiles produce a correlation level above 0.92, they are expected to agree within 5 percent most of the time. Above this threshold, the agreement between profiles is good. Correlation above 0.92 often depends on consistent lateral tracking of the profiler and may be very difficult to achieve on highly distressed surfaces. The IRI values provided in this report will be the average of five observations, which will tighten the tolerance even further. Overall, most of the groups of measurements in Tables 13 through 20 exhibited good or better correlation, and most exhibited acceptable correlation. Any group of repeat measurements that produced a composite correlation level below 0.82 was investigated using filtered plots; these results follow:  Section 04B901, visit 07 and 08: Narrow downward spikes throughout the profiles significantly diminished the correlation.  Section 04B902, visit 07: Narrow downward spikes throughout the left side profiles significantly diminished the correlation. These appeared in the same location in some, but rarely all, repeat measurements.  Section 04B902, visit 08: Narrow downward spikes in isolated locations within the profiles significantly diminished the correlation.  Section 04B903, visit 08: Narrow downward spikes in the left side profiles significantly diminished the correlation. These appeared in the same location in some, but rarely all, repeat measurements. 46  Section 04B961, visit 08: Correlation was diminished by short‐wavelength “noise” throughout the section on both the left and right side and several extraneous spikes in the left side profiles.  Section 04B962, visits 06 through 08: Narrow downward spikes that rarely appeared in the same location in more than two repeat measurements significantly diminished the correlation.  Section 04B964, visits 07 and 08: Correlation was diminished by short‐wavelength “noise” throughout the section on both the left and right side. SUMMARY ROUGHNESS VALUES Figures 17 through 24 show the left and right IRI values for each pavement section over its monitoring period. For most of the sections, this includes 14 summary IRI values: two per visit over seven visits. The figures show the IRI values versus time in years. In this case, “years” refers to the number of years between the measurement date and April 1, 1995. (All of the test sections were constructed in the second half of March 1995.) Fractions of a year are estimated to the nearest day. To supplement the plots, the appendix to this report lists the IRI, Half‐car Roughness Index (HRI), and Ride Number (RN) of each section for each visit. These roughness values are the average of the five repeat measurements selected in the data quality screening. These are not necessarily the same five repeat measurements selected for the LTPP database. The appendix also provides the standard deviation of IRI over the five repeat measurements to help identify erratic roughness values that result from transverse variations in profile caused by surface distresses. Figures 17 through 24 provide a snapshot of the roughness history of each pavement section. The remainder of this chapter characterizes the profile content that made up the roughness and explains the profile features that contributed to roughness progression. 47 IRI (in/mi) IRI (inches/mile) 150 100 50 Left Right Section B901 0 0 2 4 6 8 10 Years Figure 17. IRI Progression of Section 04B901 IRI (in/mi) IRI (inches/mile) 150 100 50 Left Right Section B902 0 0 2 4 6 Years Figure 18. IRI Progression of Section 04B902 48 8 10 IRI (in/mi) IRI (inches/mile) 150 100 50 Left Right Section B903 0 0 2 4 6 8 10 Years Figure 19. IRI Progression of Section 04B903 IRI (in/mi) IRI (inches/mile) 150 100 50 Left Right Section B959 0 0 2 4 6 Years Figure 20. IRI Progression of Section 04B959 49 8 10 IRI (in/mi) IRI (inches/mile) 150 100 50 Left Right Section B960 0 0 2 4 6 8 10 Years Figure 21. IRI Progression of Section 04B960 IRI (in/mi) IRI (inches/mile) 150 100 50 Left Right Section B961 0 0 2 4 6 Years Figure 22. IRI Progression of Section 04B961 50 8 10 (in/mi) IRI (inches/mile) 300 250 200 150 100 50 Left Right Section B962 0 0 2 4 6 Years Figure 23. IRI Progression of Section 04B962 51 8 10 (in/mi) IRI (inches/mile) 150 100 50 Left Right Section B964 0 0 2 4 6 8 10 Years Figure 24. IRI Progression of Section 04B964 PROFILE ANALYSIS TOOLS Investigators used various analytical techniques to study the profile characteristics of each pavement section and their change with time. These tools help study roughness, roughness distribution, and roughness progression of each test section, including concentrated roughness that may be linked to pavement distress. The discussion of each analysis and plotting method is rather brief; Sayers and Karamihas (1996b) provide more details about these methods. Roughness Values Investigators calculated left IRI, right IRI, Mean Roughness Index (MRI), HRI, and RN values. The appendix provides the average value of each index for each visit of each section. The discussion of roughness in this analysis emphasizes the left and right IRI. Nevertheless, comparing the progression of HRI and RN to the MRI provides additional information about the type of roughness that is changing. For example, a low HRI value relative to MRI indicates roughness that exists on only one side of the lane. Further, aggressive degradation of RN without a commensurate growth in MRI signifies that the developing roughness is biased toward short‐wavelength content. 52 Elevation Profile Plots A simple way to learn about the type of roughness that exists within a profile is to view the trace. However, certain key details of the profile are often not as obvious in a raw profile trace as they may be after the profile is filtered. Three types of filtered plots were inspected for every visit of every section:  Long wavelength: A profile smoothed with a base length of 25 ft and anti‐smoothed with a base length of 125 ft.  Medium wavelength: A profile smoothed with a base length of 5 ft and anti‐smoothed with a base length of 25 ft.  Short wavelength: A profile smoothed with a base length of 1 ft and anti‐smoothed with a base length of 5 ft. These filters were used to screen the profiles for changes with time and special features of interest. The terms “long,” “medium,” and “short” are relative, and in this case pertain to the relevant portions of the waveband that affects the IRI. The long‐wavelength portion of the profile was typically very stable with time. However, the long‐wavelength profile plots of every section changed somewhat between visit 06 and 07—not by a change in the surface characteristics of the section, but by a change in profiler make and the associated change in filtering practices. The medium‐wavelength plots provided a view of the features in a profile that were likely to have a strong effect on the IRI and may change with time. The short‐wavelength elevation plots also typically progressed with time, but only affected the IRI through localized roughness or major changes in content with time. However, the short‐wavelength elevation plots helped identify and track the progression of narrow dips and other short‐duration features that may have been linked to distress. In addition to filtered plots, every profile was viewed in its raw form. This helped reveal noteworthy features that did not necessarily affect the IRI, but helped establish a link between surface distress and profile properties. Two examples of this were narrow downward spikes in the profiles caused by cracking and deep, short‐duration dips at potholes. Roughness Profile A roughness profile provides a continuous report of road roughness using a given segment length (Sayers 1990). Instead of summarizing the roughness by providing the IRI for an entire pavement section, the roughness profile shows how IRI varies with distance along the section by using a sliding window to display the IRI of every possible segment of given base length along the pavement. A roughness profile displays the spatial distribution of roughness within a pavement section. As such, it can be used to distinguish road sections with uniform roughness from sections with roughness levels 53 that change over their length. Further, the roughness profile can pinpoint locations with concentrated roughness and estimate the contribution of a given road disturbance to the overall IRI. In this work, roughness profiles were generated and viewed using a base length of 25 ft, that is, every point in the plot shows the IRI of a 25‐ft segment of road, starting 12.5 ft upstream and ending 12.5 ft downstream. Any location where a peak occurs in the roughness profile that is greater than or equal to 2.5 times the average IRI for the entire section is considered an area of localized roughness. All areas of localized roughness are discussed in the detailed observations by identifying them, listing their severity, and describing the underlying profile features that caused them. Power Spectral Density Plots A power spectral density (PSD) plot of an elevation profile shows the distribution of its content within each waveband. An elevation profile PSD is displayed as mean square elevation versus wave number, which is the inverse of wavelength. PSD plots were calculated from the slope profile rather than the elevation profile, which aided in the interpretation of the plots because the content of a slope PSD typically covers fewer orders of magnitude than an elevation PSD. A PSD plot is generated by performing a Fourier transform on a profile (or in this case, a slope profile). The PSD’s value in each waveband is derived from the Fourier coefficients and represents the contribution to the overall mean square of the profile in that band. The slope PSD plots provided a very useful breakdown of the content within a profile. In particular, the plots reveal (1) cases in which significant roughness is concentrated within a given waveband; (2) the type of content that dominates the profile (e.g., long, medium, or short wavelength); (3) the type of roughness that increases with time; and (4) the type of roughness that is stable with time. For this project, the PSDs rarely provided much value beyond what was learned using filtered elevation plots and roughness profiles. However, any valuable observations that could be made from a PSD plot are discussed in the following section. Distress Surveys Once the analysis and plotting were completed, all of the observations were compared to the manual distress surveys performed on each section. Manual distress survey results were available for each section on at least six dates over the monitoring history, starting in February 1995. These were performed using LTPP protocols by technicians certified to perform distress surveys. The surveys provided a means of relating profile features to known distresses. DETAILED OBSERVATIONS This section reports key observations from the roughness index progression, filtered elevation profile plots, roughness profiles, PSD plots, and distress surveys. In many cases, similar behavior was noted for multiple sections. These observations are repeated under the heading of every section where 54 appropriate. However, changes in profile properties with time that were caused by changes in profiler make or model are summarized at the end of the report. Section 04B901 Roughness The IRI of the left side increased steadily from 66 to 90 inches/mi over the monitoring period, with the exception of a value of 128 inches/mi in visit 07. The IRI of the right side increased from 65 to 103 inches/mi over the monitoring period, but most of the increase occurred in the last two visits. The HRI was 11 percent to 14 percent below the MRI in visits 01 through 06, and was 19 percent to 21 percent below the MRI in visits 07 and 08. Elevation Profile Plots The left side profiles were very consistent with each other over the first five visits. Visit 06 profiles were also consistent with previous visits in the medium‐ and long‐wavelength range. However, the agreement in the short‐wavelength range was not as good because of narrow downward spikes in some locations within the visit 06 profiles. By visit 07, a very high concentration of narrow downward spikes up to 0.5 inch deep appeared in the left side profiles over the entire length of the section. These sometimes appeared at the same location in more than one repeat measurement, but rarely appeared in all five. In visit 08, very few narrow downward spikes appeared in the profiles with the exception of a patch of spikes about 0.5 inch deep that occurred 5 to 10 ft and 37 to 39 ft from the start of the section. The right side profiles were consistent with each other over the first six visits. (A minor exception was repeat 5 from visit 01, which was different from the others in the long‐wavelength range.) Visit 07 profiles were very similar to those of previous visits except they included several narrow downward spikes that were rarely in the same location in more than one repeat measurement. The spikes appeared with the highest density in the last 140 ft of the section. Profiles from visit 08 included a much higher number of downward spikes over the entire section, and they were most prevalent in the first 100 ft of the section. Two bumps stood out in the right side profiles from all seven visits. The first was about 115 ft from the start of the section and was 4 ft long and up to 0.15 inch high. The second was about 200 ft from the start of the section and was about 10 ft long and 0.2 inch high. Roughness Profiles The left side roughness profiles were very consistent in visits 01 through 05. No localized roughness appeared in these visits, although the roughness was not particularly evenly distributed. The roughness was highest about 125 ft from the start of the section. This is near a series of four bumps up to 5 ft long and more than 0.1 inch high. The final bump in the series was about 131 ft from the start of the section. It was 5 ft long and over 0.2 inch high. By visit 08, the peak value of the roughness profiles in this location was 195 to 215 inches/mi. 55 In visits 06 and 07, the roughness profiles were not consistent with each other because of the hit‐or‐ miss nature of the downward spikes within the elevation profiles. The right side roughness profiles were fairly consistent with each other in visits 01 through 06. In visits 07 and 08, additional roughness appeared in every repeat measurement, but rarely in the same locations in any two repeats. Localized roughness appeared in all visits within the last 25 ft of the section. This area had peak roughness values of over 160 inches/mi in all visits. This roughness was caused by a series of disturbances about 5 ft long. (It is not clear if these are bumps or dips.) In visit 07 and one repeat measurement from visit 08, the peak roughness was much higher in this location because a narrow dip up to 1 inch deep appeared 480.5 ft from the start of the section in these profiles. Distress Surveys WP cracking was first recorded in January 1999. This covered more than half of the left WP. By April 2002 cracking had covered all of the left WP, and longitudinal cracking covered about half of the right WP. In January 2005, areas of cracking covered both WPs over the entire length of the section. Transverse cracking was also recorded on both sides in multiple locations. (Ten feet of longitudinal distance without at least one transverse crack was rare.) Nothing in the distress surveys explain the localized roughness on the left side described above or the reduction in roughness on the left side between visits 07 and 08. The narrow dip 480.5 ft from the start of the section on the right side is in about the same location as a transverse crack recorded in April 2002 and January 2005. Section 04B902 Roughness The IRI of the left side did not follow a clear trend, but increased overall from 58 inches/mi in visit 01 to 79 inches/mi in visit 08. In visit 07, the left side IRI averaged 143 inches/mi, with a standard deviation of 28.1 inches/mi. These values were much higher than in any other visit. The IRI of the right side held steady between 59 and 63 inches/mi in visits 01 through 06, then increased to 89 inches/mi by visit 08. The HRI was 14 percent to 22 percent below the MRI. Elevation Profile Plots The right side profiles were consistent in visits 01 through 06, and similar in visit 07. (A minor exception was repeat 5 from visit 01, which was different from the rest in the long‐wavelength range.) A bump appears in the profiles of all visits from 195 to 203 ft that is about 0.2 inch high. The profiles from visit 08 include narrow downward spikes throughout the section. These often appear in only one repeat measurement in a given specific location, but some areas have a higher density of spikes than others. The left side profiles were consistent in visits 01 through 05. (A minor exception was repeat 5 from visit 01, which was different from the rest in the long‐wavelength range.) A patch of densely spaced narrow dips, up to 0.3 inch deep, appeared 120 to 145 ft from the start of the section in visit 05. Three new areas of densely spaced narrow dips appeared in the visit 06 profiles, from 215 to 235 ft, 305 to 330 ft, and 405 to 435 ft from the start of the section. 56 All visit 07 profiles included narrow downward spikes throughout the entire length of the section. These spikes were typically 0.4 inch or more deep and were poorly correlated among the five repeat measurements. Visit 07 profiles also included four bumps about a foot long: (1) about 120 ft from the start of the section, 0.5 inch high; (2) about 134 ft from the start of the section, 0.35 inch high; (3) about 304 ft from the start of the section, 0.2 inch high, detected in two repeats; and (4) about 307 ft from the start of the section, 0.3 inch high, detected in four repeats. Visit 08 profiles included narrow dips in the same areas as visit 06 along with a few narrow downward spikes in other areas. None of the four bumps listed for visit 07 appeared in the visit 08 profiles. The patches of narrow dips in visit 08 were much more repeatable, isolated, and in some cases more severe than in previous visits. Roughness Profiles An area of localized roughness appeared in the right side profiles about 197 ft from the start of the section in all visits. This was caused by the 8‐ft‐long bump described above. A peak also appeared in visits 01 through 07 of up to 140 inches/mi that is 120 ft from the start of the section. It was caused by a bump about 0.1 inch high from 113 to 117 ft from the start of the section. The right side roughness profiles from visit 08 were not very consistent because of the lack of repeatability in the appearance of the narrow downward spikes. The roughness profiles from the left side included localized roughness in four of the five repeat measurements about 135 ft from the start of the section in visit 05. In visit 06, and particularly in repeat 6, the roughness profiles showed a significant increase over those of visit 05 in the areas where the narrow dips appeared. Visit 07 was much rougher than visit 06 in several areas, again because of the narrow dips. The roughness was highest in the first 150 ft of the section and in the last 100 ft (in repeat 7 only). Visit 08 roughness profiles were much more repeatable on the left side than in visits 06 and 07. Areas of high roughness were centered about 125, 220, 315, and 420 ft from the start of the section. Distress Surveys The patch of narrow dips in the left side profiles in visit 05 corresponds to an area of cracking recorded in the left WP in January 1999. This was the only cracking recorded in that survey. In April 2002, the distress survey recorded areas of cracking over most of the left WP and longitudinal cracking over much of the length of the section in the right WP. Three of the four bumps listed for the left side profiles in visit 07 appear in locations where potholes were recorded in January 2005. The only pothole that was not in the vicinity of a corresponding bump was about 235 ft from the start of the section. In January 2005, potholes were detected in the left WP 121, 235, 305, 307, and 312 ft from the start of the section. These were locations where narrow dips were the most prevalent in visit 08 profiles, and all of these areas corresponded to peaks in the roughness profiles from visit 08. 57 Section 04B903 Roughness The IRI of the left side increased from 59 to 66 inches/mi in visits 01 through 06, then increased to 143 inches/mi by visit 08. The IRI of the right side held steady between 67 and 71 inches/mi over the monitoring period. The HRI was 15 percent to 20 percent below the MRI. Elevation Profile Plots In visit 06, narrow dips began to emerge in the left side profiles. They were up to 0.15 inch deep and appeared in the same location in more than one profile in some cases. Visit 07 profiles included more narrow dips on the left side. The dips were up to 0.5 inch deep, and the majority of them did not appear in the same location in multiple repeat measurements. Visit 08 included a much higher density of narrow dips. Many of the dips under 0.3 inch deep were not well repeated, but several deeper dips appeared at the same location in more than one repeat measurement. Similar dips appeared in the right side profiles from visits 06 through 08, but not nearly as many, and they were much less severe. A dip 1 ft long and at least 2.4 inches deep appeared in the left side profiles from visit 07 about 122 ft from the start of the section. A dip 1 ft long and 2 inches deep appeared in three of the five left side profiles from visit 08 about 333 ft from the start of the section. Roughness Profiles The roughness was very uniform along the section on the left side in visits 03 and 04. In visit 06, the roughness was somewhat uniform along the section, but not very consistent among the repeat measurements. Severe localized roughness with a peak value of 540 inches/mi appeared in the location of the deep dip in visit 07. This feature alone added more than 25 inches/mi to the overall IRI of the left side. Severe localized roughness with a peak value of 630 to 695 inches/mi appeared about 333 ft from the start of the section in three of the five repeat measurements from visit 08. This accounted for about 30 inches/mi of additional roughness in those repeat measurements. It also explains the high (15.4 inches/mi) standard deviation among the left side IRI values. No localized roughness was found in the right side profiles; they were very consistent with time in visits 01 through 07. Distress Surveys No distress was recorded until April 2002 when longitudinal cracking and large areas of WP cracking were recorded. In January 2005, both WPs were covered with cracking over the entire length of the section. Nothing in the distress survey explains the deep dips in visits 07 and 08. 58 Section 04B959 Roughness The IRI of the left side increased from 41 inches/mi in visit 01 to 47 inches/mi in visit 08 with a peak value of 50 inches/mi in visit 07. The IRI of the right side increased steadily from 51 to 73 inches/mi. The HRI was 13 percent to 16 percent below the MRI. Elevation Profile Plots Profiles were very consistent over the monitoring period on the left side. The only feature that stood out was a dip at the start of the section 0.1‐0.2 inch deep and about 10 ft long. The deepest part of the dip was only a few feet from the start of the section. Its shape was only fully developed in profiles that included lead‐in, such as visits 07 and 08. The right side profiles included narrow dips 188.5 ft from the start of the section in some of the repeat measurements from visits 04 and 05. But the profiles were fairly consistent in visits 01 through 07. In visit 08, several narrow dips up to 0.1 inch deep appeared, particularly in the first third of the section. Roughness Profiles The roughness of the left side was uniform along the section with the exception of localized roughness at the very start of the section, which was caused by the long dip described above and roughness that appeared just ahead of the section start. The localized roughness only appeared in visits 07 and 08 because they included profiles upstream of the section start. In the early visits, the roughness of the right side was uniform along the section. In later visits, and particularly in visit 08, most of the increase in roughness occurred in the first third of the section. Distress Surveys Nothing in the distress surveys explains the narrow dip in visits 04 and 05. Very little distress was recorded before January 2005. By January 2005, significant cracking had appeared. Much of the cracking was found in the first third of the section along the right WP. This included transverse cracks and large areas of cracking within the WP. Narrow dips appeared in the visit 08 profiles in the position of most of the transverse cracks. Section 04B960 Roughness The IRI of the left side held steady between 45 and 47 inches/mi in visits 01 through 06, then increased to 54 inches/mi by visit 08. The IRI of the right side only ranged from 45 to 48 inches/mi. The HRI was 13 percent to 18 percent below the MRI. 59 Elevation Profile Plots The right side profiles were very consistent with time. Two small disturbances stood out in the medium‐ wavelength (and raw) profiles that were about 190 and 380 ft from the start of the section. The same features appeared on the left side, but they did not stand out among the roughness from the rest of the section. The left side profiles were very consistent with time in visits 01 through 06. In visit 07, a sunken area appeared 441 to 447 ft from the start of the section that included two dips within it, each at least 0.1 inch deep. In visit 08, this entire area was about 0.1 inch below the surrounding pavement, and it included three dips within it, each at least another 0.2 inch deep. Roughness Profiles The right side roughness profiles were very consistent with time. In all visits, the roughness was about double the average near 190 and 380 ft from the start of the section. The left side roughness profiles were fairly consistent with time in visits 01 through 06. In visit 07, a peak of about 100 inches/mi appeared in the roughness profile about 445 to 450 ft from the start of the section. By visit 08, the same area showed peak roughness of up to 160 inches/mi. This localized roughness was caused by the sunken area with dips described above and was responsible for the majority of the increase in roughness of the entire section after visit 06. Distress Surveys The distress surveys recorded very little distress before April 2002. Some transverse and longitudinal cracking was recorded in April 2002. This included longitudinal cracking from 435 to 450 ft from the start of the section on the left side of the lane and transverse cracking across the left WP about 443 and 447 ft from the start of the section. This corresponds to the sunken area and the dips within it described above. The distress survey from January 2005 observed cracking over the entire section in both WPs, but several transverse cracks in the area from 440 to 450 ft were also specifically recorded. Section 04B961 Roughness The IRI increased from 52 to 59 inches/mi on the left side, but the progression was not steady. The IRI only ranged from 56 to 60 inches/mi on the right side. The HRI was 10 percent to 12 percent below the MRI. This was the lowest difference observed for the SPS‐9A project, and it is an unusually low difference for a full‐depth asphalt pavement. Elevation Profile Plots Profiles were very consistent over the monitoring period on both sides of the lane. They were unusually consistent with time in the short‐wavelength range. Very few features stood out. One minor exception: a series of short (about 0.05 inch), narrow bumps 60 to 100 ft from the section start on both sides. 60 Roughness Profiles No localized roughness was detected on the left side, although the area from 60 to 100 ft from the start of the section was rougher than the rest. On the right side, a peak value of up to 140 inches/mi appeared about 95 ft from the start of the section. This localized roughness was caused by the bumps described above. Distress Surveys Nothing in the distress surveys directly explains the series of narrow bumps observed in the profiles. Very little distress was observed before the final survey in January 2005. That survey recorded a significant amount of raveling, particularly on the left side of the lane. Section 04B962 Roughness The IRI of the left side held steady between 51 and 54 inches/mi in visits 01 through 05, then increased rapidly to 281 inches/mi in visit 08. The IRI of the left side held steady between 58 and 61 inches/mi in visits 01 through 06, then increased to 76 inches/mi by visit 08. The HRI was 13 percent to 17 percent below the MRI. Elevation Profile Plots Profiles from the left side were consistent over visits 01 through 05, and profiles from the right side were very consistent over visits 01 through 06. Those profiles included areas with small bumps 2 to 5 ft long from 180 to 210 ft and 360 to 420 ft from the start of the section on both sides. The right side profiles included several narrow downward spikes throughout the section in visits 07 and 08. Heavy concentrations of these spikes appeared 162 to 172 ft, 230 to 245 ft, 295 to 305 ft, 355 to 365 ft, and 475 ft from the start of the section. (These account for the additional roughness compared to visits 01 through 06.) In visit 06, the left side profiles included narrow downward spikes in the ranges from 60 to 80 ft, 150 to 180 ft and 410 to 440 ft from the start of the section. The dips often appeared in the same location in more than one repeat measurement, but rarely in all five. The deepest, and most well repeated, dip was less than 1 ft long, up to 1.4 inch deep, and appeared 430 ft from the start of the section. In visit 07, the left side profiles included narrow downward spikes up to 1 inch deep over the entire section. They usually appeared in the same location in more than one repeat measurement, but not in all repeats. The highest concentration of downward spikes appeared between 360 and 450 ft from the start of the section. In one repeat measurement, a dip 0.5 ft long and 0.8 inch deep appeared 170 ft from the start of the section. In two other repeat measurements, a dip appeared 80 ft from the start of the section that was 0.5 ft long and at least 0.7 inch deep. 61 The visit 07 profiles from the left side also included five longer disturbances that were detected in all five repeat measurements: (1) a 2 ft long, 0.5 inch deep dip centered 418 ft from the start of the section; (2) a 4 ft long, 0.5 inch deep dip centered 431 ft from the start of the section; (3) a 1.5 ft long, 0.4 inch high bump 471 ft from the start of the section; (4) a 2 ft long, 0.4 inch high bump 484 ft from the start of the section; and (5) a 1.5 ft long, up to 0.3 inch high bump 492 ft from the start of the section. All of these features appeared among a high concentration of narrow downward spikes. The visit 08 profiles from the left side included many areas with a high concentration of narrow downward spikes up to 2.5 inches deep, but the spikes rarely appeared in more than the same location in more than two repeat measurements. The highest concentration of spikes appeared in repeats 4 and 7. The profiles also included some bumps up to 2 ft long and 0.5 inch high in the last 40 ft of the section and a 4 ft long, 0.5 inch deep dip in the last 5 ft of the section. Roughness Profiles Roughness profiles for the right side were very consistent in visits 01 through 06. Although no localized roughness appeared in these profiles, the distribution of roughness along the section was not very uniform. The growth in roughness after visit 06 took place over the entire length of the section. No localized roughness was detected on the left side in visits 01 through 06, although the areas with small bumps described above were rougher than the surrounding pavement. In visits 06 and 07, localized roughness (or at least increased roughness) in the left side profiles appeared about 80 ft from the start of the section in one repeat measurement per visit and about 175 ft from the start of the section in one or two repeat measurements per visit. These are locations where deep, narrow dips were found in some, but not all, repeat measurements. Localized roughness also appeared in visit 06 about 430 ft from the start of the section and in the last 20 ft of the section. Both of these areas appeared as severe localized roughness in visit 07, with peak values of up to 550 inches/mi in the roughness profiles. In visit 08, the left side profiles were very rough over much of the section. (The exception was the area from 230 to 350 ft from the start of the section.) The most severe localized roughness appeared (1) at the very end of the section; (2) in one repeat measurement about 20 ft from the start of the section; and (3) in all but one repeat measurement about 80 ft from the start of the section. Distress Surveys The distress survey of January 1999 showed five areas of longitudinal cracking or WP cracking. Many of these areas correspond to the locations where narrow downward spikes appeared in the left side profiles from visit 06. In visits 07 and 08, cracking covered the entire left WP (and most of the right WP). This probably explains the downward spikes that appeared throughout the profiles and the hit‐or‐miss nature of their distribution. The distress survey in April 2002 recorded potholes about 427 ft and 434 ft from the start of the section. The distress survey in January 2005 recorded a large number of potholes and patches in the left WP over 62 the last 100 ft of the section. Their locations corresponded to the locations of longer bumps and dips in the visit 07 and 08 profiles. The deep narrow dip in the left side profiles from visits 06 through 08 did not correspond to anything in the distress surveys beyond the WP cracking found over the entire section. Section 04B964 Roughness The IRI increased fairly steadily from 41 to 52 inches/mi on the left side and from 42 to 49 inches/mi on the right side. The HRI was 11 percent below the MRI in visit 01. The gap between HRI and MRI grew steadily to 17 percent by visit 08. Elevation Profile Plots The profiles changed very little over the monitoring period with the exception of three rough features.  After visit 02, a dip appeared about 136 ft from the start of the section on the left side. It was about 1 ft long and grew in depth from 0.15 inch to 0.4 inch over visits 03 through 08. However, the dip was not present in every repeat measurement.  A dip up to 0.15 inch deep appeared 155 ft from the start of the section’s left side in all visits.  Starting in visit 04, a narrow dip appeared in the right side profiles about 160 ft from the start of the section. The dip rarely appeared in all five repeat measurements within a visit and ranged in depth up to 0.25 inch. Roughness Profiles Together, the first two dips described above caused an area of localized roughness that grew in severity to a peak value of 130 inches/mi by visit 08. The increased roughness at this area contributed significantly to the growth in roughness with time on the left side between visits 01 and 07. In the most extreme case (visit 04, repeat 01), the third dip described above caused a peak value in the roughness profile of 106 inches/mi on the right side. Other than the influence of the dip 160 ft from the start of the section, roughness was very evenly distributed along the section on the right side. PSD Plots Although the content was biased toward long wavelengths, some content was isolated around wavelengths of about 15 to 16 ft on the left side. 63 Distress Surveys Nothing in the distress surveys directly explains the three narrow dips described above. Very little distress was observed before April 2002. In April 2002 and January 2005, the distress surveys recorded raveling along the entire section in both WPs. SUMMARY This section summarizes important observations from each pavement section within the SPS‐9A project limits. Some observations within this report were common to more than one pavement section, as described below. This section of the report, in conjunction with the roughness progression plots (Figures 17 through 24), provides the essential information about each pavement section. Profiles from Sections 04B901, 04B902, 04B903, and 04B962 included narrow downward spikes dispersed throughout their length. The spikes usually appeared in only one or two repeat measurements at each location. These were caused by cracking, which covered both WPs of Sections 04B901, 04B902, 04B903, and 04B962 in the later visits. The spikes typically first appeared in profile measurements from November 2001 (visit 06), but were much more prevalent in February and December 2004 (visits 07 and 08). The spikes were usually more severe and numerous in the left side profiles. In Sections 04B901 and 04B902, the spikes were most severe in the left side profiles from visit 07. Although it is not clear why fewer spikes were detected in visit 08 on those two sections, it may be a consequence of a subtle difference in profiler positioning. The spikes account for most of the increase in the IRI values, particularly on the left side, in Sections 04B901, 04B902, 04B903, and 04B962 after November 2001. Sections 04B902 and 04B962 included several potholes in the left WP that were observed in the distress survey from January 2005. The profiles in the location of these potholes were rarely consistent among the five repeat measurements from the last two profiling visits. In Section 04B902, the narrow downward spikes found throughout the profiles were somewhat more likely to appear near the potholes than in other locations, but no direct relationship was observed. Bumps appeared near two of the potholes in two of the five repeat profile measurements in February 2004. This may have been caused by narrow patching that was only covered by the lateral placement of the profiler in two of the passes, but no patching was noted. In Section 04B962, two of the seven potholes caused dips about 0.5 ft long and up to 1 inch deep in the last profiling visit that were fairly well repeated. The final distress survey (January 2005) for Sections 04B901, 04B902, 04B903, 04B960, 04B962, and 04B964 all showed areas of cracking that cover both WPs over the entire section length. In these instances, fatigue cracking is so prevalent that it is difficult to match individual cracks to roughness within the measured profile. However, in a few cases, features in the profiles that affected the roughness were found that correspond directly to the location of transverse cracks noted in the distress survey. Several narrow dips appeared in Section 04B959 in the final profiling visit that corresponded to the locations of transverse cracks noted in January 2005. In Section 04B901, a narrow dip appeared in the right side profiles about 480 ft from the start of the section in the last two profiling visits. This is a location where transverse cracking was noted in April 2002. An area of localized roughness about 10 ft long appeared in the last two visits of Section 04B960 that included multiple transverse cracks. 64 Profiler Model The change in profiler in late 2002 affected the long‐wavelength content of the profiles on every test section because the newer profiler used a high‐pass filter that eliminated a little more of the profile content than the previous device. The change in high‐pass filtering methods had no probable effect on the measurement of localized roughness or the study of narrow bumps and dips caused by distresses. Another minor device effect within the profiles was peaks in the PSD plots with no pavement‐related explanation. In visits 01 through 06 (measured by the K.J. Law Engineers T‐6600 profiler), all profiles from the left and right side included a peak in their spectral content at a wavelength somewhere between 0.35 and 0.52 ft and another at a wavelength of double the first. Individual Test Sections The summaries below provide the most important observations made about each test section. To help provide context for these summary statements, Figure 25 shows the range of left and right IRI for each section. Note that the highest IRI value for some of the sections did not occur in the final visit. (See the appendix or Figures 17 through 24.) Section 04B901 Although this section started out moderately smooth, with an MRI value of 65 inches/mi, roughness grew significantly to a final MRI value of 97 inches/mi. While the roughness of the right side grew steadily, the roughness of the left side reached a peak IRI value of 128 inches/mi in February 2004 and then reduced to a final value of 90 inches/mi. In the last three profiling visits, the profiles included narrow downward spikes that often did not appear in the same location in more than one repeat measurement. The highest concentration of dips appeared in the left side profiles in February 2004, which accounts for much of the additional roughness over the previous and subsequent visits. Localized roughness appeared in the right side profiles over the last 25 ft of the section because of 5‐ft‐long bumps and dips. In 2004, the profiles in this area also often included a 1‐inch deep narrow dip at a transverse crack 480.5 ft from the start of the section. Section 04B902 The section was somewhat smooth until 2004. Starting with the profile measurement of November 2001, the profiles included an increasing number of narrow dips, particularly on the left side, that often did not appear in the same location in more than one repeat measurement. These dips were most severe in the left side profiles from February 2004, which had an average roughness value of 143 inches/mi, but were not at all consistent with each other. Although the spikes were less severe in November 2004, the profiles included patches of narrow dips in three areas where potholes were recorded in the distress surveys. 65 04B901 B901 L R 04B902 B902 L R B903 L 04B903 R B959 L 04B959 R B960 L 04B960 R B961 L 04B961 R B962 L 04B962 R B964 L 04B964 R 0 50 100 150 200 250 300 IRI (in/mi) (inches/mile) Figure 25. Summary of IRI Ranges Section 04B903 The right side of this section remained somewhat smooth throughout the monitoring history, but the left side IRI more than doubled between visits 06 and 08. Starting in visit 06, the profiles included an increasing number of narrow dips, particularly on the left side, that often did not appear in the same location in more than one repeat measurement. In visit 07, a deep (2.4‐inch) narrow dip appeared 122 ft from the start of the section on the left side. A deep (2‐inch) narrow dip appeared 333 ft from the start of the section in three of the five repeat measurements from visit 08 on the left side. Both dips caused severe localized roughness. Roughness at the dip in visit 08 accounted for the high (15.4 inches/mi) standard deviation in the left side IRI from visit 08. 66 Section 04B959 This section remained smooth throughout the monitoring period, although the MRI progressed by 13 inches/mi. An area of localized roughness appeared on the left side at the start of the section because of a long dip and some rough features just ahead of the section start. The roughness of the right side was uniform along the section, until the roughness progressed significantly in the first third of the section because of the influence of cracking. Section 04B960 The MRI increased from 46 to 51 inches/mi over the monitoring period. An area of localized roughness appeared on the left side about 445 ft from the start of the section by the end of the monitoring period. A sunken area of pavement, including three narrow dips about 0.2 inch deep at transverse cracks, caused the roughness. This area accounts for most of the roughness progression of the entire section. Section 04B961 The MRI ranged from 55 to 59 inches/mi over the entire monitoring period. An area of increased roughness appeared on the right side of the lane 60 to 100 ft from the start of the section because of a series of short, narrow bumps in the profiles. The bumps appeared on the left side of the lane also, but they did not increase the roughness as much there. The bumps were very consistently measured over the monitoring period. Section 04B962 This section remained smooth until 2004, when the final MRI value was 178 inches/mi, more than three times the initial value of 55 inches/mi. The sharp increase in roughness occurred primarily within the left side profiles. By the final visit, the left side profiles were strongly affected by roughness at potholes and patching, and included a high concentration of deep, narrow downward spikes caused by cracking in the WP. Most of the potholes and patches appeared in the last 100 ft of the section. Section 04B964 The MRI increased from 41 to 50 inches/mi over the monitoring period. The majority of the roughness progression on this section was caused by three narrow dips that grew in severity with time: (1) 136 ft from the start of the section on the left side, (2) 155 ft from the start of the section on the left side, and (3) 160 ft from the start of the section on the right side. 67 68 CHAPTER 5. CONCLUSIONS AND RECOMMENDATIONS ADOT initiated the SPS‐9A project to study the relative performance of the Superpave binders compared to the commonly used agency binders, which will provide a foundation for future design decisions. Surface distress, deflection, and profile data were used as the basis for performance evaluation and were analyzed as part of the study. The SPS‐9A project offers a unique opportunity to directly compare performance of various asphalt binders while reducing the confounding effect of other variables such as traffic‐loading, climate, and subgrade conditions. However, findings drawn from this evaluation must be considered carefully. The experiment design did not offer replicate treatments to verify findings. Conclusions drawn from this comparison are based on one set of in situ conditions; observations from other climate or loading scenarios may differ from those noted within this report. Therefore, findings reported may be unique to the conditions and construction of this site. Despite these issues, the data captured at the project provides valuable insight into pavement performance, design, management, and construction. Following is a summary of lessons learned from the performance data collected at the SPS‐9A site:  Roughness and roughness progression alone cannot be used to represent the health of a test section. Several test sections did not exhibit changes in roughness in proportion to the amount of fatigue cracking, and sections that had clearly reached the end of their service lives did not necessarily have roughness values that would trigger a rehabilitation event.  Most every section except Section 04B959 (SMA polymer with ACFC) and Section 04B961 (SMA polymer with cellulose) showed significant growth in fatigue and longitudinal cracking 10 years after construction.  Where fatigue cracking was very prevalent, it was difficult to match individual cracks to roughness within the measured profile. However, in a few cases, features in the profiles that affected the roughness corresponded directly to the location of transverse cracks noted in the distress survey.  Construction quality can play a major role in performance. The construction observations documented in the LTPP construction report were limited to Superpave sections (Sections 04B902 and 04B903) and the agency standard mix section (Section 04B901). However, it showed that all three sections were free of any construction issues.  Half of the sections (04B901, 04B959, 04B960, 041007/04B962) had reasonable patterns of environmental distress growth with a clear increase in magnitude approximately 10 years after construction. Section 04B964 had little environmental distress growth for seven years and then rapidly increased after 10 years. 69  With no replicate sections, researchers had limited ability to assess potential variability independent of actual performance.  For the SP‐1 mix designs (Sections 04B902 and 04B903), the PG 70‐10 binder (Section 04B903) performed significantly better, both structurally and environmentally, than the PG 76‐10 binder (Section 04B902). It also performed significantly better than the standard agency mix at resisting environmental deterioration.  All sections performed well with regard to rut resistance. Rutting would not have triggered a rehabilitation event for any section.  SMA with cellulose fibers was, by far, the best pavement mix at resisting both structural and environmental deterioration. From a roughness perspective, the SMA and ARAC sections outperformed the Superpave mixes. Based on these findings, the research team recommends the following:  Most of the pavement test sections appeared to have experienced top‐down cracking; however, this could not be confirmed. It is recommended that forensic analysis be performed at other locations throughout Arizona to learn about the factors contributing to top‐down cracking.  If the SMA with cellulose is sufficiently cost‐effective, ADOT should consider expanding its use in overlay applications. 70 REFERENCES American Association of State Highway and Transportation Officials (AASHTO). 1993. AASHTO Guide for Design of Pavement Structures. Washington, D.C.: American Association of State Highway and Transportation Officials. Evans, L. D. and A. Eltahan. 2000. LTPP Profile Variability. Publication FHWA‐RD‐00‐113. McLean, VA: Federal Highway Administration: 178. Federal Highway Administration (FHWA). 1998. “Arizona SPS‐9, I‐10 Westbound Milepost 112‐123, Construction Report on Site 04B900, Arizona Department of Transportation, Final.” Huang, Y. H. 1993. Pavement Analysis and Design. Englewood Cliffs, NJ: Prentice‐Hall. Karamihas, S. M. 2004. “Development of Cross Correlation for Objective Comparison of Profiles.” International Journal of Vehicle Design 36(2/3): 173–193. Karamihas, S. M., T. D. Gillespie, and S. M. Riley. 1995. “Axle Tramp Contribution to the Dynamic Wheel Loads of a Heavy Truck.” Proceedings of the 4th International Symposium on Heavy Vehicle Weights and Dimensions, Ann Arbor, MI. C. B. Winkler (ed.): 425–434. Kim, Y. Richard, B. Underwood, M. Sakhaei Far, N. Jackson, and J. Puccinelli. 2011. LTPP Computed Parameter: Dynamic Modulus. Publication FHWA‐HRT‐10‐035. Washington, D.C.: Federal Highway Administration. Mahoney, J. 1995. WSDOT Pavement Guide, Volume 2: Pavement Notes. Olympia: Washington State Department of Transportation. Miller, J. S. and W. Y. Bellinger. 2003. Distress Identification Manual for the Long‐Term Pavement Performance Program. Revised fourth edition. Publication FHWA‐RD‐03‐031. Washington, D.C.: Federal Highway Administration. Rada, G. R., C. L. Wu, R. K. Bhandari, A.R. Shekharan, G. E. Elkins, and J. S. Miller. 1999. Study of LTPP Distress Data Variability, Volume I. Publication FHWA‐RD‐99‐074. Washington, D.C.: Federal Highway Administration. Sayers, M. W. 1989. “Two Quarter‐Car Models for Defining Road Roughness: IRI and HRI.” Transportation Research Record: Journal of the Transportation Research Board 1215: 165–172. Sayers, M. W. 1990. “Profiles of Roughness.” Transportation Research Record: Journal of the Transportation Research Board 1260: 106–111. Sayers, M. W. and S. M. Karamihas. 1996a. “Estimation of Rideability by Analyzing Longitudinal Road Profile.” Transportation Research Record: Journal of the Transportation Research Board 1536: 110–116. 71 Sayers, M. W. and S. M. Karamihas. 1996b. Interpretation of Road Roughness Profile Data. Publication FHWA‐RD‐96‐101 (Washington, D.C.: Federal Highway Administration): 177. Simpson, A. L. 2001. Characterization of Transverse Profiles. Publication FHWA‐RD‐01‐024. Washington, D.C.: Federal Highway Administration. 72 APPENDIX: ROUGHNESS VALUES This appendix lists the left International Roughness Index (IRI), right IRI, mean roughness index (MRI), Half‐car Roughness Index (HRI), and Ride Number (RN) values for each visit of each section. The roughness values are the average for five repeat runs. The five runs were selected from a group of as many as nine by automated comparison of profiles, as described in the report. Values of standard deviation are also provided for left and right IRI to reveal cases of high variability among the five measurements. However, the screening procedure used to select five repeats usually helped reduce the level of scatter. The discussion of roughness in the report emphasizes the left and right IRI. Nevertheless, the other indexes do provide useful additional information. MRI is simply the average of the left and right IRI value. HRI is calculated by converting the IRI filter into a half‐car model (Sayers 1989) by collapsing the left and right profile into a single profile in which each point is the average of the corresponding left and right elevation. The IRI filter is then applied to the resulting signal. The HRI is very similar to the IRI except that side‐to‐side deviations in profile are eliminated. The result is that the HRI value for a pair of profiles will always be lower than the corresponding MRI value. Comparing the HRI and MRI value provides a crude indication of the significance of roll (i.e., side‐by‐side variation in profile) to the overall roughness. When HRI is low compared to MRI, roll is significant. This is common among asphalt pavements (Karamihas et al. 1995). Certain types of pavement distress, such as longitudinal cracking, may also cause significant differences between HRI and MRI. Figure 26 compares the HRI to MRI for all of the profile measurements that are covered in this appendix (290 pairs of roughness values). The figure shows a best fit line with a zero intercept and a line of equality. The slope of the line is 0.845, which is typical for asphalt pavement. RN has shown a closer relationship to road user opinion than the other indexes (Sayers and Karamihas 1996a). As such, it may help distinguish the segments from each other by ride quality. Further, the effect on RN may help quantify the impact of that distress on ride when a particular type of distress dominates the roughness of a section. In particular, a very low RN value coupled with moderate IRI values indicates a high level of short wavelength roughness and potential sensitivity to narrow dips and measurement errors caused by coarse surface texture. Table 21 provides the roughness values. The table also lists the date of each measurement and the time in years since the site was opened to traffic. Negative values indicate measurements that were made before rehabilitation. 73 HRI (inches/mile) (in/mi) 200 Line of Equality 150 HRI = 0.845•MRI 100 50 0 0 50 100 150 MRI (inches/mile) (in/mi) 200 Figure 26. Comparison of HRI to MRI Table 21. Roughness Values Section Date Years 04B901 04B901 04B901 04B901 04B901 04B901 04B901 04B902 04B902 04B902 04B902 04B902 04B902 04B902 04B903 04B903 04B903 04B903 04B903 04B903 04B903 29‐Jan‐97 5‐Dec‐97 7‐Dec‐98 10‐Nov‐99 12‐Nov‐01 9‐Feb‐04 17‐Dec‐04 29‐Jan‐97 5‐Dec‐97 7‐Dec‐98 10‐Nov‐99 12‐Nov‐01 9‐Feb‐04 17‐Dec‐04 29‐Jan‐97 5‐Dec‐97 7‐Dec‐98 10‐Nov‐99 12‐Nov‐01 9‐Feb‐04 17‐Dec‐04 1.84 2.68 3.69 4.61 6.62 8.86 9.71 1.84 2.68 3.69 4.61 6.62 8.86 9.71 1.84 2.68 3.69 4.61 6.62 8.86 9.71 Left IRI (inch/mi) Ave St Dev 66 1.0 68 1.5 66 1.1 68 1.0 77 2.1 128 4.9 90 4.2 58 1.3 56 1.6 54 0.8 63 3.1 73 10.3 143 28.1 79 2.3 59 0.6 59 1.8 60 0.5 60 0.9 66 3.2 119 1.8 143 15.4 Right IRI (inch/mi) Ave St Dev 65 1.2 63 1.5 68 1.2 66 1.9 70 2.1 88 4.2 103 4.1 61 1.4 59 1.1 61 0.5 61 0.7 63 0.6 69 2.9 89 0.6 69 1.1 67 0.5 68 1.0 68 1.0 67 1.3 68 1.9 71 2.9 74 MRI HRI RN (inch/mi) 65 65 67 67 74 108 97 60 58 58 62 68 106 84 64 63 64 64 66 93 107 (inch/mi) 58 57 59 58 63 88 77 50 49 49 53 57 86 66 54 53 54 53 54 75 87 3.79 3.67 3.68 3.72 3.51 2.41 2.82 3.87 3.82 3.82 3.68 3.46 2.56 2.94 3.77 3.71 3.73 3.80 3.69 2.37 2.42 Table 21. Roughness Values (Continued) Section 04B959 04B959 04B959 04B959 04B959 04B959 04B959 04B960 04B960 04B960 04B960 04B960 04B960 04B960 04B961 04B961 04B961 04B961 04B961 04B961 04B961 04B962 04B962 04B962 04B962 04B962 04B962 04B962 04B962 04B964 04B964 04B964 04B964 04B964 04B964 04B964 04B964 Date 29‐Jan‐97 5‐Dec‐97 7‐Dec‐98 10‐Nov‐99 12‐Nov‐01 9‐Feb‐04 17‐Dec‐04 29‐Jan‐97 5‐Dec‐97 7‐Dec‐98 10‐Nov‐99 12‐Nov‐01 9‐Feb‐04 17‐Dec‐04 29‐Jan‐97 5‐Dec‐97 7‐Dec‐98 10‐Nov‐99 12‐Nov‐01 9‐Feb‐04 17‐Dec‐04 29‐Jan‐97 2‐Feb‐97 5‐Dec‐97 7‐Dec‐98 10‐Nov‐99 12‐Nov‐01 9‐Feb‐04 17‐Dec‐04 29‐Jan‐97 2‐Feb‐97 5‐Dec‐97 7‐Dec‐98 10‐Nov‐99 12‐Nov‐01 9‐Feb‐04 17‐Dec‐04 Years 1.84 2.68 3.69 4.61 6.62 8.86 9.71 1.84 2.68 3.69 4.61 6.62 8.86 9.71 1.84 2.68 3.69 4.61 6.62 8.86 9.71 1.83 1.84 2.68 3.69 4.61 6.62 8.86 9.71 1.83 1.84 2.68 3.69 4.61 6.62 8.86 9.71 Left IRI (inch/mi) Ave St Dev 43 0.6 45 0.6 46 0.5 47 0.5 47 0.6 50 0.9 47 1.1 47 1.1 47 0.6 45 1.0 47 0.7 46 0.6 52 0.7 54 1.3 52 0.3 56 0.8 55 1.5 52 1.0 54 1.1 56 1.1 59 2.8 52 1.9 51 1.0 54 2.0 53 1.4 54 1.7 81 5.5 149 12.3 281 49.8 41 0.3 41 0.8 45 1.2 44 0.4 46 0.8 47 0.8 49 1.4 52 1.8 Right IRI (inch/mi) Ave St Dev 51 0.7 53 0.7 55 0.8 55 1.1 61 1.1 60 1.4 73 4.2 45 0.9 46 0.7 46 0.8 46 0.6 45 0.4 48 0.7 47 1.1 58 0.6 56 0.6 59 1.7 60 0.9 57 0.4 58 0.4 59 1.3 59 2.0 58 0.7 60 0.9 61 1.6 58 0.9 59 1.1 72 1.8 76 3.0 42 0.9 40 0.5 41 0.5 47 1.5 46 1.4 46 0.6 47 1.1 49 2.4 75 MRI HRI RN (inch/mi) 47 49 50 51 54 55 60 46 46 46 47 46 50 51 55 56 57 56 55 57 59 55 54 57 57 56 70 111 178 41 41 43 45 46 47 48 50 (inch/mi) 41 43 44 44 46 46 51 39 39 39 40 40 41 42 49 50 51 50 50 51 52 48 47 50 49 49 58 92 152 37 36 37 38 39 40 41 41 4.22 4.08 4.09 4.13 4.10 3.95 3.81 4.07 3.95 3.99 4.05 4.06 3.75 3.66 4.00 3.86 3.88 3.99 3.97 3.82 3.77 3.84 3.86 3.71 3.73 3.79 3.06 2.26 1.77 4.24 4.23 4.02 4.03 4.02 4.01 3.79 3.76