1. February 5, 20161 Fatigue of Asphalt Mixtures, Endurance Limit, Polymer Modifications, Healing 1.E+02 1.E+03 1.E+04 1.E+05 1.E-07 1.E-05 1.E-03 1.E-01 1.E+011.E+031.E+05 Reduced Frequency [ rad /s ] |G*| [ MPa ]. 0 15 30 45  Phase  º ]. |G*| Lab|G*| GB Phase_LabPhase GB

2. February 5, 20162 Fatigue of asphalt mixtures 4 p bending 2 p bending 4 p bending 2 p bending direct tension indirecttension

3. February 5, 20163 Fatigue resistance is a specimen, not a material property p bending strain

4. February 5, 20164 Area in 2 and 4 p test subjected to fatigue 2 point bending test 4 point bending test Normalized height over sample Relative strain amplitude

5. February 5, 20165 indirect tension load contr. 4 p bending load contr. 4 p bending displ contr. N 10 6 10 4 10 2 80201  10 1 10 2 10 3 Fatigue resistance is a specimen, not a material property

6. February 5, 20166 Interpretation of fatigue tests Input: half sine or haversine displacement signal Output: residual stress develops; after some time peak load is 50% of initial peak load and load signal becomes sinusoidal

7. February 5, 20167 Test set-up

8. February 5, 20168 T = 30 o C Measured load Modelled load Displacement

9. February 5, 20169 T = 0 o C Measured load Modelled load Displacement

10. February 5, 201610 Input: half sine/haversine load Output: creep displacement; Creep will most probably be failure mechanism Interpretation of fatigue tests

11. February 5, 201611 Conclusion Reported nr of load repetitions to failure is depending on type of test. Reported nr of load repetitions to failure is depending on type of test. Reported nr of load repetitions to failure is depending on whether test is load or displacement controlled Reported nr of load repetitions to failure is depending on whether test is load or displacement controlled Reported applied strains are depending on how test is performed (full sine vs half sine) Reported applied strains are depending on how test is performed (full sine vs half sine) Reported endurance limits are in fact specimen properties Reported endurance limits are in fact specimen properties

12. February 5, 201612 What is reality? LINTRACK Linear tracking device 35 – 100 kN wheel load 20 km/h Uni and bi directional Temperature controlled Built at end/beginning 1980’s/1990’s

13. February 5, 201613 Transversal strain vs nr of load repetitions; residual strains are not taken into account in common design procedures

14. February 5, 201614 Residual strains do not seem to develop in longitudinal direction

15. February 5, 201615 Substantial corrections needed to match lab result with practice log  log N Field fatigue Lab fatigue Shift factor (healing, lateral wander, damage propagation, stress redistribution etc. 2.5 - 40)

16. February 5, 201616 From lab to in situ fatigue relations the big unknown, advanced theories required n can be estimated with confidence log N log 

17. February 5, 201617 Theory on crack growth in visco- elastic media is of help Crack growth law: dc/dN = AK n Crack growth law: dc/dN = AK n A = f(S mix, m,  t,  ) A = f(S mix, m,  t,  ) n = f(2/m and void content) n = f(2/m and void content) Fatigue law: N = k 1 (  ) -n Fatigue law: N = k 1 (  ) -n k 1 = f(A, m, specimen geometry) it is a specimen dependent parameter ! k 1 = f(A, m, specimen geometry) it is a specimen dependent parameter ! m is slope of master curve for stiffness! m is slope of master curve for stiffness! =

18. February 5, 201618 Geometry dependency of lab fatigue relationship N = h (1-n/2) F c  -n / A Smix n F c = c0/h  cf/h d(c/h) / [1.99(c/h) 0.5 – 2.47(c/h) 1.5 - 12.97(c/h) 2.5 – 23.17(c/h) 3.5 - 24.8(c/h) 4.5 ] n The thicker the beam the lower the fatigue life at same strain level

19. February 5, 201619 Size effect on 4p beam bending tests

20. February 5, 201620 Endurance limits Size 0.5 Size 1 Size 1.5 58  m/m 44  m/m 37  m/m

21. February 5, 201621 Effect of beam thickness on flexural strength

22. February 5, 201622 SBS

23. February 5, 201623 Simple tests to determine effect of SBS

24. February 5, 201624 Simple experiment to show strength of SBS modified asphalt mixtures

25. February 5, 201625 0.8 1.6 0.4 0.8

26. February 5, 201626

27. February 5, 201627 Triaxial Test, Cohesion “C” and Angle of Internal Friction “  ” Failure Envelope

28. February 5, 201628 General Case Failure envelope can also be generated by means of tension and compression tests. Failure envelope can also be generated by means of tension and compression tests. In generalized case,  is replaced by bulk stress I 1 and  is replaced by deviator stress parameter J 2 In generalized case,  is replaced by bulk stress I 1 and  is replaced by deviator stress parameter J 2

29. February 5, 201629 Parameters used in the Failure Envelope Graphs I 1 =  1 +  2 +  3 I1I1I1I1  J2 J2 J2 J2

30. February 5, 201630 Importance of Failure Envelope I1I1I1I1  J2 J2 J2 J2 Whenmat A and When mat A and mat B have about same stiffness, then an almost similar point in I 1 – J 2 space will be obtained but mat B will perform much better Mixtures did not differ too much in terms of stiffness

31. February 5, 201631 Failure Envelopes at 5 o C and Strain Rate of 0.01 %/s

32. February 5, 201632 Failure Envelopes at 40 o C and Strain Rate 0.01 %/s

33. February 5, 201633 R initial Loading cycles N f,50 Fatigue Life in terms Stress Ratio R limit N

34. February 5, 201634 Fatigue Test Results at 20 o C and 8 Hz

35. February 5, 201635 Endurance Limits at 8 Hz and 20 o C Mixture S m,initial (GPa) ε limit (10 -6 m/m) 599-40 8.9 8.9 50 50 602-42 10.8 10.8 80 80 604-41 10.1 10.1 75 75

36. February 5, 201636 Analyzed Pavement Structures Variable thickness; Stiffness of mixtures 40, 41 and 42;  = 0.35 E = 300 MPa; h = 300 mm;  = 0.35 E = 100 MPa;  = 0.35 F = 50 kN; r = 150 mm

37. February 5, 201637 Required Asphalt Thickness 42/194 41/211 40/294

38. February 5, 201638 Conclusion Modifying asphalt mixtures with specially designed polymers can result in a significant reduction of the asphalt layer thickness or a significant increase in pavement life when thickness is kept the same Modifying asphalt mixtures with specially designed polymers can result in a significant reduction of the asphalt layer thickness or a significant increase in pavement life when thickness is kept the same

39. February 5, 201639 Healing of bituminous mastic (bitumen + fine aggregate < 63  m Healing of bituminous mastic (bitumen + fine aggregate < 63  m) O hr1 hr 3 hr18 hr Source: PhD Qiu

40. February 5, 201640 Healing of mastics

41. February 5, 201641 Fatigue and healing Mixture Rest period S retest / S 1st test N retest / N 1st test pen 40/60 void content 20 o C / 10 Hz end of test: S n /S i = 0.3 3.5 months at 15 o C 0.830.03 pen 40/60 void content 5 o C / 8 Hz end of test: S n / S N = 0.5 18 months at 15 o C 1 0.6 – 0.9 Source: PhD Pramesti and Li

42. February 5, 201642 Conclusions Healing is recovery of strength and should not be confused with recovery of stiffness Healing is recovery of strength and should not be confused with recovery of stiffness Most probably, stiffness recovery is due to thixotropy Most probably, stiffness recovery is due to thixotropy Healing of asphalt mixtures is mainly a flow driven process Healing of asphalt mixtures is mainly a flow driven process Long rest periods are beneficial but only at elevated temperatures Long rest periods are beneficial but only at elevated temperatures Temperature is more important than time Temperature is more important than time Healing (strength recovery) of asphalt mixtures is limited Healing (strength recovery) of asphalt mixtures is limited

43. February 5, 201643 Thank you for your attention