SEAUPG 29 CONFERENCEHILTON HEAD ISLAND, SOUTH CAROLINA FullScale Accelerated Performance Testing for Superpave & Structural Validation Outcomes Ongoing Activities Future APT Experiments Background Superpave Performance Grade Purchase Specification PG 64 7 22 Hilton Head Island, South Carolina November 92, 29 FHWA TurnerFairbank Highway Research Center Office of Infrastructure R&D Nelson Gibson, nelson.gibson@dot.gov G* sinδ Energy In Energy Back Out Background How good is this? Value to the community? Energy Dissipated Fatigue Cracking G* sinδ Number of ALF Wheel Passes For First Crack to Appear At Surface 4, 35, 3, 25, 2, 5,, 5, PostSHRP 993 FHWA ALF Better fatigue performance at cooler and warmer 3, 2mm Thick Asphalt Pavements 35, "Softer" PG 58 9 34 "Stiffer" PG64 7 22 25, 5, 25, 22, 28C 9C C ALF Test Temperature Number of ALF Wheel Passes For First Crack to Appear At Surface 2,, 8, 6, 4, 2,, mm Thick Asphalt Pavements 5, "Softer" PG 58 9 34 "Stiffer" PG64 7 22,, 4, 28C 9C C ALF Test Temperature 2, Stiffer binder performed better at intermediate temperature Not intent of specification OBJECTIVES for FullScale Accelerated Performance Testing for Superpave & Structural Validation Recommendations that provide AASHTO with a binder purchase specification that is blind to the type of modification. Secondary Objectives Candidate Parameters Stiffness Reduction Time Sweep Stress Sweep Large Strain Time Sweep Surrogate Fracture CTOD Critical Tip Opening Displacement Strength Yield Energy
SEAUPG 29 CONFERENCEHILTON HEAD ISLAND, SOUTH CAROLINA Approach to Evaluation Laboratory Scale Tests Approach to Evaluation Laboratory Scale Tests Rheology FullScale ALF Performance Rheology FullScale ALF Performance AsBuilt AsDesigned Experimental Design (PAV) The Pavement Test Facility (PAV) 2
Lane 8, Control Lane 9, SBS 644 Lane, Air Blown Lane, SBSLG Lane 2, Terpolymer..... Reduced Frequency (Hz) SEAUPG 29 CONFERENCEHILTON HEAD ISLAND, SOUTH CAROLINA 64 o C (47 o F) psi Inflation, pounds No Wander Loading Conditions 9 o C (66 o F) 2 psi Inflation 6, pounds Wheel Wander Cumulative Crack Length (m) 2.. 8. 6. 4. 2.. 5,, 5, 2, 25, 3, 35, 4, Load Passes Lane, CRAZ / Control (No Cracking) Lane 2, Control Lane 3, Air Blown Lane 4, SBSLG Lane 5, CRTB Lane 6, Terpolymer Lane 7, Fiber Lane 7, Fiber (extrapolated) Ranked Cycles to 25 m Crack Length Cumulative Crack Length (m) 6. 5. 4. 3. 2. Lane 8, Control Lane 9, SBS 644 Lane, Air Blown Lane, SBSLG (No Cracking) 3, Passes Lane 2, Terpolymer (No Cracking) 4, Passes Interpolated Cycles Surface Initiation Confirmation of Bottom Up Cracking.. 5,, 5, 2, 25, 3, 35, 4, 45, Load Passes AsBuilt Consistency AsBuilt Consistency Crushed Aggregate Base Stiffness Falling Weight Deflectometer A4 Subgrade Stiffness HMA Stiffness E* Dynamic Modulus (MPa) 3
SEAUPG 29 CONFERENCEHILTON HEAD ISLAND, SOUTH CAROLINA Laboratory Scale Fatigue Performance Then Now Laboratory Scale Fatigue Performance Lab Fabricated 9 Field Compacted Nf (Number of cycles to C=.5) 8 7 6 5 4 f) T=25 C, f=hz TP Lab TP Top TP Bottom 75 75 35 5 Microstrain (Half Peak) Time Sweep and Stress Sweep Martono, W. and H. Bahia, Developing a Surrogate Test for Fatigue of Asphalt s, TRB 87 th Annual Meeting Compendium of Papers DVD (28) Time Sweep and Stress Sweep Martono, W. and H. Bahia, Developing a Surrogate Test for Fatigue of Asphalt s, TRB 87 th Annual Meeting Compendium of Papers DVD (28) Number of Cycles to Failure,,,,,, 3% 5% 7% 9% Stress Sweep, 722 Air Blown CRTB Terpolymer SBS LG Large Strain Time Sweep Surrogate Shenoy, A., (July 22) Fatigue Testing and Evaluation of Asphalt s Using the Dynamic Shear Rheometer, ASTM Journal of Testing and Evaluation, Vol. 3, No. 4, pp 3332 Large Strain Time Sweep Surrogate T=5 o C T=2 o C T=25 o C T=3 o C γ =.% γ = 8% γ =.% γ = 8% γ =.% γ = 8% γ =.% γ = 8% T o C where γ = 25% & G* = MPa X62 larger than PG 4
SEAUPG 29 CONFERENCEHILTON HEAD ISLAND, SOUTH CAROLINA Critical Tip Opening Displacement Critical Tip Opening Displacement CTOD is a measure of strain tolerance in the presence of a crack MTO Test Method LS299 4 mm W = 3 mm Force (N) 6 5 4 3 2 2 4 6 8 Displacement (mm) 5mmA 5mmB mma mmb 5mmA 5mmB Wt(j/m2) 6 4 2 8 6 4 y = 524263x + 6492.6 2 R 2 =.9675.2.4.6.8..2.4.6 L(m) Yield Energy Test Yield Energy Test Y Value How will Parameter Candidates be Compared to Each Other?.9.8.7.6.5.4.3.2. Proportional Trend Expected..2.3.4.5.6 X Value Rank Order Quality Indicator.73 +... Very Best Poor Completely Incorrect Parameter Superpave Time Sweep Stress Sweep LargeScale Time Sweep Surrogate CTOD BYET Rank Order Quality Indicator LabScale Fatigue FullScale ALF Ranking Cracking.8.6 +.6 +.8.6.4.6 +.8 +..4 +. +.8 5
SEAUPG 29 CONFERENCEHILTON HEAD ISLAND, SOUTH CAROLINA Ongoing: Accelerated Aging Objective: Double (X2) the number of data points Full Scale Fatigue Cracking UNAGED CRTB Air Blown Control Terpolymer SBS 644 SBSLG AGED CRTB Air Blown Control Terpolymer SBS 644 SBSLG Parameters Cumulative Crack Length (m) 8 6 4 2 Ongoing: Accelerated Aging Crack Length vs ALF Wheel Load Passes (Testing Condition: 9 o C & 6,6 lbs) L2S4 (Aged) L2S3 L5S4 (Aged) L5S3 LS4 (Aged) LS2 L4S4 (Aged) L4S3 L3S4 (Aged) L3S3 L6S4 (Aged) L6S3 5 5 2 25 Number of Load Passes Ongoing: Accelerated Aging PAV Aged ~ Ductile Recovered ~ Brittle Force (N) 24 2 8 5 2 9 6 3 Research where fullscale accelerated performance testing provides a role Your Input RANK More structure oriented experiments and less binderoriented experiments Shorter turnaround time; less ambitious Fatigue performance of High RAP HMA and HMA Overlays Cracking and durability of Ultrathin HMA overlays as pavement preservation Thinner & Cheaper Perpetual Pavements with Premium HMA Cost effectiveness of highmodulus highbinder HMA base Conclusions Similarly PG graded materials exhibit different fatigue cracking performance MEPDG evaluated construction variation and it is small Composite structure with wet process crumb rubber has the ability to arrest or slow propagating cracks Fibers are very effective in reducing fatigue cracking Aging has significant influence on ranked order of performance Conclusions There are better properties than G* sinδ CTOD and BYET appear to be among the most promising ONGOING: 32 s From LTPP Test Sections Asphalt Research Consortium tests cannot easily provide insight for structural effects ( mm vs. 5 mm thick), fiber modified mixtures and composite pavements with CRAZ Stresses the importance of testing mix Simplified fatigue testing in AMPT is of high value 6
SEAUPG 29 CONFERENCEHILTON HEAD ISLAND, SOUTH CAROLINA Acknowledgements State DOT & Industry TPF 5(9) SPR 2(74) Citgo Dow Dupont Koch Paramount TexPar Trifinery, GCA Trumbull Wright Asphalt Martin Color Fi Bit Mat Mathy Construction Hot Mix Industries FNF Construction Consulpav Rubber Producers Assoc. ISS RTG NAPA Nelson.Gibson@dot.gov Database Application (portable) has been developed 25page Detailed Report 6.5 Measured vs. AsBuilt Measured vs. AsDesigned Measured vs. As Built + Average Unbound Line of Equality LOG of MEPDG Cycles to 7.75% Cracking 6. 5.5 5. 4.5 mm Thick Lanes 5mm Thick Lanes 4. 4. 4.5 5. 5.5 6. 6.5 7. LOG Cycles to 7.75% Cracking (Measured 25% ALF Cracking) 7
SEAUPG 29 CONFERENCEHILTON HEAD ISLAND, SOUTH CAROLINA..5 Witczak Model Hirsch Model Correction Factor,.56 slope Abraham Maslow s Law of the Instrument Dynamic Modulus Ratio E*.95.9.85.8.75 3 4 5 6 7 8 9 Air Void Content (%),6 4.5 kip Measured Microstrain,4,2, 8 6 4.5 kip Longitudinal 4.5 kip Transverse Line of Equality 4 2 2 25 3 35 4 45 5 55 6 65 7 6 kip MEPDG AsBuilt Predicted Microstrain Yield Energy Test Johnson, C.M., H. Bahia and H. Wen, Practical Application of Viscoelastic Continuum Damage Theory to Asphalt Fatigue Characterization, Asphalt Paving Technology, Journal of the Association of Asphalt Paving Technologists (forthcoming).,,,,, XHead Displacement Control VECD Strain Control based on the idea that [in HMA mixtures], there is an energy threshold where a material s resistance to deformation and, resistance to damage are both overcome, leading to the propagation of cracking. NF AB 722 Fiber CRTB SBS LG Terpolymer 8
SEAUPG 29 CONFERENCEHILTON HEAD ISLAND, SOUTH CAROLINA Summary Summary Parameter Superpave Regression Significance ( p value) LabScale Fatigue Ranking 67% FullScale ALF Cracking 79% Parameter Superpave Correlation Coefficient, R LabScale Fatigue Ranking.56 FullScale ALF Cracking.68 Time Sweep 37% 98% Time Sweep +.29 +.94 Stress Sweep LargeScale Time Sweep Surrogate CTOD 67% 66% 43% 86% 8% 99% Stress Sweep LargeScale Time Sweep Surrogate CTOD.75.55 +.34.55.69 +.95 BYET 72% 92% BYET +.68 +.83 Summary Scatter Plot Superpave Parameter Superpave Time Sweep Stress Sweep LargeScale Time Sweep Surrogate CTOD BYET Coefficient of Determination, R 2 LabScale Fatigue Ranking.3.9.3.3.2.46 FullScale ALF Cracking.46.88.57.48.92.69 Scatter Plot Time Sweep Scatter Plot Stress Sweep 9
SEAUPG 29 CONFERENCEHILTON HEAD ISLAND, SOUTH CAROLINA Scatter Plot Large Strain Time Sweep Surrogate Scatter Plot CTOD Scatter Plot BYET A4 Validate tests for modified binder A5 Develop & validate models Historical Perspectives: SHRP A2A/A3A Identify Composition Quantify Composition Develop Physical Tests Correlation Composition and Physical Properties A Develop binder specification and protocols A3A Validate A2A with simulative large scale lab tests (Phase ) A5 Validate A2A with field data (Phase 2) Researchers were well aware of complexities and advanced tests & theories Introduction of DSR radically changed direction of research Linked dissipated energy to binder beam fatigue and ZaccaWigmore test road Linked stress relaxation to viscous component of stiffness to wheel tracking Field validation from early LTPP test sections showed no binder specification by itself can explain field performance Historical Perspectives: PostSHRP FHWA ALF Wheel Passes to 2mm Rut Depth,,,,,,, Historical Perspectives: PostSHRP NCHRP 9 HMA fatigue did not correlate to G* sinδ fatigue timesweep identified as potential candidate HMA permanent deformation did not correlate to G* /sinδ creep and recovery identified as potential candidate Unmodified AC s Modified s 2, 4, 6, 8,, 2, 4, 6, G*/sin(δ) at 2.25 rad/sec (Pa)
SEAUPG 29 CONFERENCEHILTON HEAD ISLAND, SOUTH CAROLINA Rutting: G* /sinδ Rutting: Zero & LowShear Viscosity G*/sinδ @ 64 o C 4, 2,, 8, 6, 4, 2, 5 5 2 Rutting (mm) G*/sin @ 64 o C 4, 2,, 8, 6, 4, 2, 2, 4, 6, Microstrain @ 5, Cycles Calculated from complex viscosity ZSV G" sinδ sinδ η = lim = lim & ηl = LSV ω ϖ ω ω =.rad / sec ω ω Slope Regression Significance (pvalue) Kendall s tau Kendall s tau Significance R R 2 G* /sinδ vs. Rutting Correct 92%.3 98%.55.3 G* /sinδ vs. Microstrain at 5, Cycles + Incorrect.2% +.2 64% +.78.6 Rutting: Zero & LowShear Viscosity Complex Viscosity (Pas) Low S.V. Zero S.V. 9 8 7 6 5 4 3 2 Complex Viscosity (Pas) Low S.V. Zero S.V. 9 8 7 6 5 4 3 2 5 5 2 2 3 4 5 6 Rutting (mm) Microstrain @ 5, Cycles Regression Kendall s Kendall s tau Slope Significance R R tau 2 Significance (pvalue) LSV vs. + 26% +.9 56% +.. Rutting Wrong direction Wrong direct ZSV vs. 7% +.2 68%.7.5 Rutting LSV vs..4%.4 6%.3 με @ 5k Cycles ZSV vs..4%.4 6%.3 με @ 5k Cycles Rutting: Shenoy s Estimate for NonRecovered Strain Shenoy, Aroon (2). Refinement of the Superpave specification parameter for performance grading of asphalt. J. Transportation Engineering, 27(5), 357362. sinδ tanδ sinδ NonRecovered Strain Derived from Oscillatory Test σ γ NR = tanδ sinδ % Rutting: Shenoy s Estimate for NonRecovered Strain parameter : tanδ sinδ J NR estimate = tanδ sinδ Rutting: Shenoy s Estimate for NonRecovered Strain 25 25 2 2 G* 5 5 tan δ sinδ tanδ sinδ 5 5 5 2 Rutting (mm) 5 2 4 6 Microstrain @ 5, Cycles) Slope Regression Significance (pvalue) Kendall s tau Kendall s tau Significance R R 2 G* tanδ sin δ vs. Rutting + Wrong direction 48% +.23 Wrong direction 68% +.22 Wrong direction.5 tanδ sinδ tanδ sinδ vs. με @ 5k Cycles 6.3%.33 8%.37.
SEAUPG 29 CONFERENCEHILTON HEAD ISLAND, SOUTH CAROLINA NonRecovered Compliance J NR.4 σ E = ε τ G = γ.2. ε ( t) D( t) = σ γ ( t) J ( t) = τ MAMTL JNR.8.6 Stress (Pa) 3, 25, 2, 5,, 5, 4 8 2 Time (s) Strain (%) 8, 7, 6, 5, 4, 3, 2,, 4 8 2 Time (s) J (Pa ).35.3.25.2.5..5 4 8 2 Time (s).4.2.2.4.6.8..2.4 TFHRC JNR 2