Are AUSTROADS Pavement Design Performance Models Adequately Calibrated

Are AUSTROADS Pavement Design Performance Models Adequately Calibrated for New Zealand? Dr Bryan Pidwerbesky General Manager - Technical Fulton Hogan Ltd AUSTROADS PTF Workshop, Wellington 04 December 2014 www. fultonhogan. com

Outline • • • Asphalt fatigue strain criterion Subgrade strain criterion Terminal rut depth Back-calculation from FWD deflection bowls Conclusion/summary www. fultonhogan. com

AUSTROADS Pavement Design Guide Asphalt 1 Cemented Material 2 Unbound subbase Subgrade 3 1 Horizontal tensile strain in bottom of asphalt – fatigue cracking 2 Horizontal tensile strain in bottom of cemented material - cracking 3 Vertical compressive strain in top of subgrade - rutting & shape loss www. fultonhogan. com

Causes of Cracking in Asphalt • Inadequate load supporting capacity: – Loss of base, subbase or subgrade support (eg water ingress) → high deflection and/or deformation – Inadequate thickness of the pavement to take the loads – Increase in loading – Poor construction www. fultonhogan. com

Causes of Cracking in Asphalt • Reflective cracking (from underlying asphalt, stabilised base or subgrade) www. fultonhogan. com

Causes of Cracking in Asphalt • Brittle Failures – Old oxidized asphalt – Asphalt too stiff for environmental conditions Outside Wheel Path Little Shape loss www. fultonhogan. com

Causes of Cracking in Asphalt • Thermal-induced cracking www. fultonhogan. com

Causes of Cracking in Asphalt • Classic fatigue-induced cracking is rare in New Zealand • Cracks normally start at top of asphalt – Start as very fine cracks created during roller compaction – Largest tensile strain is at top of asphalt – rarely bottom up www. fultonhogan. com

Causes of Cracking in Asphalt • Fatigue strain Very small strains (~100 με) per loading • Flexure strain Larger strains exceed maximum tensile strain capacity • Thermal-induced strain Environmental factors www. fultonhogan. com

Asphalt fatigue criterion The History of Asphalt Fatigue Criterion www. fultonhogan. com

Asphalt fatigue criterion The History of Asphalt Fatigue Criterion Fatigue lives for different mixes at 0°C showing derived bitumen strain Pell, P. S. (1962) Fatigue Characteristics of Bitumen and Bituminous Mixes. Int’l Conference on Structural Design of Asphalt Pavements, Ann Arbor, USA. www. fultonhogan. com

Asphalt Fatigue Relationship • 1960’s - Laboratory-derived fatigue relationship • 1970’s - Adjusted to predict fatigue life in pavements using a shift factor F • • • N = allowable number of load repetitions µε = tensile microstrain produced by the load VB = % by volume of binder in asphalt Smix =mix stiffness modulus (MPa) F = range of values www. fultonhogan. com

Shift Factors • Shell Pavement Design Manual (1978) : F=10 • Saunders, L. R. A Modern Basis for Pavement Design (1982) : F = 10 • AUSTROADS Pavement Design Guide (1992) Ignored shift factor (F = 10 was considered) • Baburamani, ARR 334 Asphalt Fatigue Life Predictions Models (1999) F = 10 to 20 • AUSTROADS Pavement Design Guide (2001 draft) : F = 5 • Saleh (2012) : F = 5. 7 www. fultonhogan. com

AUSTROADS Guide (2014) Table 6. 15 Suggested Reliability Factors for Asphalt Fatigue RF 80% 2. 5 Desired project reliability 85% 90% 95% 2. 0 1. 5 1. 0 97. 5% 0. 67 Desired project reliability has two components: • a shift factor relating mean laboratory fatigue life to a mean in-service fatigue life, taking account of differences between laboratory test conditions and conditions applying to in-service pavement; • a reliability factor relating mean in-service fatigue life to in-service predicted life at a desired project reliability, taking into account factors such as construction variability, environment and traffic loading • “for lightly-trafficked roads load-induced fatigue cracking is uncommon. ” www. fultonhogan. com

Reliability factor/shift factor is too low • Confusion about what constitutes fatigue cracking • Fatigue cracking is result of millions of very small resilient strains under wheel loadings, at significantly less than horizontal strain capacity of bound material • In majority of cases, crack-induced failures are actually due to excess deflection/flexure of the underlying pavement &/or subgrade, causing significant tensile strain in asphalt that exceeds its tensile strain capacity • Fatigue criterion not applicable to thin asphalt surfacings www. fultonhogan. com

Appropriate shift/reliability factor • Saunders (1982) • Saleh (2012) • Experience 10 5. 7 5 -10 Recommended Reliability Factors RF 80% 10 Desired project reliability 85% 90% 95% 5 4 3 www. fultonhogan. com 97. 5% 2. 5

Subgrade strain criterion “. . . the primary function of a road structure is to protect the underlying soil from excessive stresses produced by traffic loads. . ” “It is therefore necessary to limit the deformation in the soil and this may be done by limiting the value of the vertical compressive stress reaching the top of the subgrade. . ” “… the value of the vertical stress in the subgrade is one of the critical quantities determining the performance of a flexible pavement. ” Peattie, K. R. (1962) A Fundamental Approach to the Design of Flexible Pavements. Proc. Int’l Conference on the Structural Design of Asphalt Pavements, Ann Arbor www. fultonhogan. com

Subgrade strain criterion “Deformations of the surface under the action of repeated loadings by traffic is controlled by limiting the vertical compressive stress or strain in the subgrade, and if necessary on the other granular layers in the structure. ” “…irrespective of the construction, the maximum vertical compressive strain in the top of the subgrade is 9 x 10 -4, and for roads carrying greater traffic volumes, a permissible compressive strain should be 6. 5 x 10 -4. ” Dormon, G. M. (1962) The Extension to Practice of Fundamental Procedure for the Design of Flexible Pavements. Proc. Int’l Conference on the Structural Design of Asphalt Pavements, Univ. of Michigan, Ann Arbor. www. fultonhogan. com

160 Granular Overlay (mm) 140 120 100 80 60 40 20 0 11. 4 11. 9 12. 4 12. 9 Chainage (km) Granular Overlay (mm) - Austroads (GMP-Rigorous) 13. 4 Granular Overlay (mm) - TNZ Precedent Method www. fultonhogan. com

Subgrade strain criterion For unbound or stabilised granular pavements, subgrade strain criteria is conservative Actual measured strains are greater than permissible strains calculated according to the criteria. Vertical compressive strains in the basecourse can be as large (in magnitude) as vertical compressive strains in the subgrade Recommendation Strains in the basecourse should be explicitly considered in the AUSTROADS pavement design procedure www. fultonhogan. com

Terminal rut depth Permanent subgrade strain/load is too small to measure Subgrade strain criterion based on resilient subgrade strain because that is a much larger magnitude & can be measured Assumed relationship between resilient & permanent subgrade strain Accumulation of permanent subgrade strain manifests itself as pavement rutting Thickness designs assume terminal rut depth is 20 -25 mm www. fultonhogan. com

Terminal rut depth “Implicit in the design procedure for these pavements (Section 8. 3 and, specifically, Figure 8. 4 of the Guide) is a terminal condition which is considered to be unacceptable and, hence, signifies the end of life for the pavement. ” “The view of the MEC Review Committee at the time was that, in terms of rutting, it represented an average rut depth of about 20 mm. ” AUSTROADS (2004) Technical Basis of AUSTROADS Pavement Design Guide. APT 33/04 Severity Level Low Moderate High Rut (mm) 6 – 12. 5 - 25 >25 mm Typical Definitions of Rutting (FHWA, 2011) www. fultonhogan. com

Terminal rut depth 30 Actual 1 25 Extrapolate 1 20 Actual 2 15 Extrapolate 2 10 Terminal Rut 5 0 Maintenance Intervention 0 2 4 6 8 10 12 14 16 18 20 22 24 www. fultonhogan. com

Deflection & Back-calculation techniques based on FWD deflection bowls inaccurate for estimating pavement & subgrade properties: • Transfer functions are based on regression analyses & are never calibrated for specific projects • Transposition of independent & dependent variables • CBR’s derived from back calculation only intended to be relative & approximate, & used only in the context of pavement design overlays • Derived CBR value is only for modeling requirements & cannot accurately reflect actual subgrade CBR - it has to be measured in lab or inferred from in situ tests www. fultonhogan. com

Deflection & Back-calculation Example data from actual projects shows variability in subgrade CBR values derived from different techniques Subgrade Bearing Capacity Parameter CBR inferred from in-situ Scala Penetrometer Rehabilitation Project A B C 4% 4 -5% 4% Isotropic Modulus Backcalculated 69 MPa 35 MPa 86 MPa Anisotropic Modulus Equivalent(1) Laboratory soaked Subgrade CBR assumed for design 100 MPa 52 MPa 15% 25% 5 4 113 MPa 5 (1) Modulus back-calculated from FWD deflection bowl: 10 th percentile isotropic subgrade stiffness converted to practical equivalent anisotropic stiffness (EISO=0. 67 x. EANISO(vert)) (Tonkin & Taylor, 1998) www. fultonhogan. com

Deflection & Back-calculation Subgrade strain criterion was only ever intended to be used for design purposes & provides reasonable values, given cumulative effect of assumptions made during design process Predictions of material properties and remaining life from backcalculation procedures (based on FWD deflection bowls) poorly correlated with actual performance Recommendation To use back-calculation procedures based on FWD deflections for estimating remaining life of a specific pavement contractually, models & algorithms used in procedure must be robustly validated for specific conditions of each site www. fultonhogan. com

Conclusion/Summary For fatigue cracking in bitumen-bound layers, project reliability factors should be in range of 2. 5 to 5 (at least) for New Zealand Asphalt fatigue criterion is not applicable to thin surfacings Vertical compressive & shear strains within unbound & modified pavement layers should be explicitly considered as a critical parameter in flexible pavement design Terminal rut depth for unbound granular/ stabilised flexible pavements is 20 mm Back-calculation procedures based on FWD deflection data may be used to estimate remaining life of a pavement ONLY after models & algorithms have been robustly validated for specific conditions of each site www. fultonhogan. com