Phenomenological models for binder rutting and fatigue. University of Wisconsin Research Team

Transcription

1 Phenomenological models for binder rutting and fatigue University of Wisconsin Research Team

2 Outline of Talk The need for measuring damage behavior Current efforts, and evolution of ideas Gaps and what to do about them

3 Performance Grading and Modified (Engineered) Bitumen Cold Temp. Cracking Fatigue Cracking Rutting Mixing & Paving Typical Asphalt Stiffness, G* Ideal (modified) Asphalt PMB Temperature PMB w WMA

4 Review of Superpave Binder Tests DT Direct Tension Test DSR Dynamic Shear Rheometer BBR RV Rotational Viscometer Bending Beam Rheometer Related to Performance! Climate -- PG HT-LT Traffic Speed DSR Traffic Volume PG shift Traffic loading NA Pavement Structure NA Assumption: Bitumen in Linear VE range

5 Current focus is on Linear Visco-elasticity: G*, delta Complex Modulus (Pa) 1E+06 1E+05 1E+04 1E+03 1E+02 1E+01 Low SBS % No polymer network Viscous above 60 C Reasonably workable High SBS % Polymer network Elastic during compaction Workability problem Courtesy of Nynas 1E << Viscous Phase Angle delta ( ) Elastic >> 45

6 Performance Grading Redefined (2001) Linear VE is not sufficient (NCHRP 9-10) Bitumen damage resistance is very important Modified bitumen are best in damage resistance f Stress LVE Non-Linear -- Damage Strain f

7 RUTTING 7

8 Current Test: Cyclic Loading Torque (stress) Can only give total energy: Total W= i2 sin G* W delayed elastic+ W viscous Strain 2% Strain 25% Strain 50% Stress/ Strain 2 2

9 The new tests : Creep and Recovery NCHRP 9-10 >> (MSCR) Shear Stress, kpa 68 DSR Dynamic Shear Rheometer 0.1 Shear Stain, mm/mm Accumulated Strain

10 Binders with same G*/sin but different recovery >> % Recovery in MSCR 14% 12% PG82- Oxidized (1 cycle) 10% PG82 - Oxdized (100 cycle) Strain 8% 6% 4% PG82- PEs (1 cycle) PG82-PEs (100 cycle) PG82-SBSr (1 cycle) PG82-SBr (100 cycle) 2% 0% Time (s)

11 Binder Rutting Parameter The creep compliance, J(t), in terms of its elastic component (J e ), the delayed-elastic (J de ), and the viscous component (Jnr): J ( t) J J J e de v nr Calculate the viscous, non-recoverable, compliance (Jnr).

12 New AASHTO Standard for MSCR 100Pa, 3200 Pa Data Model Prediction time (s)

13 Effect of Stress and Total Strain on Mixture Rutting - R. Delgadillo, Ph.D Thesis 2008 Normalized Mixture Permanent Strain (1/MPa) B MPa (100psi) B MPa (150psi) B MPa (200psi) time (s) Polymer Modified Binder, Fine Mixtures

14 Knowledge Gaps: Effect of Loading Time and Stress hz r J (1/kPa) Linear even at 30 KPa! 0.1 J 0.1 kpa J 1 kpa J 10 kpa J 20 kpa J 30 kpa Non-linear Zone at 10 Kpa time (s)

15 Current Modeling of Creep and Recovery, R. Delgadillo, ), ( 2 1 p m p m t k t k t Nonlinear Power Model: = shear strain (dimensionless) t = time (s) = shear stress (kpa) ki, mi, pi = model parameters ) ( ), ( ), ( t J t J t r p

16 Failure of model for some binders We have to separate recoverable from permanent kpa 10 kpa Fit1 Fit J p time (s) J r ( t) t t ( t, ) t t

17 Application of the models Data Model 1 Prediction time (s) Data Model Prediction time (s)

18 Gaps in the Knowledge Nonlinearity effects Stress level Total strain / loading time? Why loading behavior different than unloading behavior? How can we model this? Is elastic recovery important? Can binder elasticity improve resistance to rutting of mixture?

19 Longitudinal Cracking In the Wheel Path Alligator cracking Fatigue 19

20 Evolution of binder fatigue testing Pre- SHRP, none --- or Penetration SHRP : Linear Visco-elastic G* sin, Rheological index, m, ect. NCHRP 9-10 Binder fatigue time sweep ETG FHWA : Stress Sweep DENT Canada Ontario - FHWA Asphalt Research Consortium Binder Yield Energy Test ( BYET) Linear Amplitude Sweep + VECD >>ETG >> AASHTO

21 Binder only Fatigue Test Results (DSR) 1.6E E+07 G*, Pa 1.2E E+07 G* (Pa) 8.0E E E E E Cycles Number of Cycles in the DSR

22 Binder Yield Energy Test (BYET)

23 Yield Energy Evaluation of different binders

24 Effect of Rate on Monotonic Test- Feb-March / s 0.01/ s Slippage began to occur at shear rates of 8% per second Limited by equipment capabilities

25 Linear Amplitude Sweep (LAS) Need 2 tests for VECD analysis + Relaxation /freq Sweep Amplitude Sweep

26 VECD Analysis of LAS Damage Curve Damage equation following Yong-Rak Kim & D. Little work (2006) N ( ) D 0 i 1 sin i i i 1 i 1 D t 1 2 I G * sin G * 1 t t 1

27 VECD Fatigue Prediction Model Work with E. Kutay With the VECD curve fit to a simple numeric equation: Fatigue life can be predicted using: k = 1 + (1 C 2 )

28 Calculating A and B

29 Simulated Binder Fatigue

30 Normal Force Issues Stress Control

31 Normal Force Issues Strain Control

32 Gaps in the Knowledge What is causing the double peak in the BYET? Is this relevant to pavement conditions? What causes this? VECD appears to be a good path forward, however: B = - 2 What strain/stress should we use? Do we need to separate non-linearity from damage?

33 33

34 AASHTO Draft Procedure

35 Other Damage Chracterizations Fracture: Single Edge Notch Beam ( SENB) What is the role of fillers? Adhesion / Cohesion PATTI >>> Bitumen Bond Strength Test

36 SENB Example Results Load (mn) Sample 1 Sample 2 Sample 3 Sample PPS = 252 pulses/step = 2 Binder Mastic Limestone Displacement (mm) Load (mn) Granite Displacement (mm)

37 Bitumen Bond Strength Testing Apparatus Graded Scale for Air Flow Control

38 Example of Test Results Cohesive Failure Adhesive Failure January 05,

39 Bitumen Adhesion Constructed 27/04 and trafficked for 1 warm day Constructed 29/04 Temperature drop 30/04 Source: Gerrie Van Zyl RSA

40 Evolution of PG Bitumen Specification Performance Indicators Temperature S, m, Tg, Fracture Energy Adhesion Nf Jnr PAV-aged RTFO Jnr Unaged pa-s pa-s Tmin Tavg Tmax 135 C 160 C

41 Thank You! Acknowledgments ARC- WRI- FHWA Carl Johnson Northwest Asphalt Haifang Wen Washington State Univ. Cassie Hintz Cristian Clotopel Raul Velasquez Aaron Coenen