Heavy Duty Pavement Design Using CalME

John Harvey Master Class

New Directions Heavy Duty Pavement Design Wednesday 1 pm

Caltrans context for heavy duty pavement design

• Old pavement – 1/3 concrete pavement, urban

high-volume freeways, 30-50 years old

– 2/3 asphalt surfaced: composite, semi-rigid, full-depth and conventional flexible structures, original structures 20-90 years old

• 90 % of work is overlays, crack PCC/seat/overlay, mill & overlay, recycling, reconstruction

• Cost and Constructability – Urban reconstruction, rehabilitation, preservation work

done at night, 55 hour weekends, or 24/7 closures – Rural projects have more time – Need thinner structures to carry same load, to get faster

construction, reduced cost

• Materials – PMB and RHMA surfaces – High RAP contents – Varying bitumen supplies – Performance related specs

• Consider construction compaction and variability

Drivers for Heavy Duty Design Method

CalME Overview • Introduced in 2006 • Focus on heavy duty rehabilitation and preservation • Incremental – Recursive approach

– Simulation runs one increment of axle loads/temperatures at a time (4 hour period of one day per month)

– Calculates damage and permanent deformation

– Adjusted stiffnesses are used as input for the next increment

– Calculations continued through simulation of entire life of pavement

– Can include simulation of pavement preservation treatments

CalME Overview • Response calculation engines

– Layer elastic calculations of critical stresses, strains – Regression models for reflection cracking strains

• Damage models – Asphalt fatigue damage – Asphalt permanent shear strain – Cemented soil fatigue, crushing – Full-depth reclamation damage

(foamed asphalt done, cemented, engineered emulsion in progress)

– Cold In-Place Recycling damage • Distress models

– Asphalt fatigue cracking – Asphalt rutting – Unbound layers rutting

CalME Overview • Aging and rest period models for asphalt • Transfer functions, shift factors calibration:

– 27 original Heavy Vehicle Simulator (HVS) test sections

– 26 Westrack sections; NCAT, MnROAD, CEDEX track validation

– 4 California field sections • Monte Carlo simulation

for reliability – Within project variability – Uses FWD back-calculated

stiffnesses for variability of existing layers

• Sensitivity analysis for between project/contractor

• Mechanistic models – Linear elastic solid mechanics:

Openpave layer elastic, Odemark-Boussinesq, – For calculating the primary response (stress,

strain, displacement)

• Empirical models – Relating the calculated response to pavement

performance (fatigue damage, cracking, rutting)

• Both components must be verified against reality in two step process using APT data

CalME Model Framework

Extrapolation of WIM spectra to 80,000 lane-km network

• Excellent WIM data since 1995

• 115 stations • Assignment tree

uses truck classification data to estimate axle load spectra for every km of entire state network

Obtain the Annual Average Daily Truck Traffic on California State Highways Report from

http://www.dot.ca.gov/hq/traffops/saferesr/trafdata/

AADT > 70000 or

(4-8)/(9-15) Ratio >1or

Truck Percentage < 10%

Calculate (4-8)/(9-15) Ratio (=number of trucks with 2,3,4 axles divided by number of trucks with 5 or more

axles) and Truck Percentage (=AADTT/AADT*100) based on truck data in the year of 2000

Group 3

No

Yes

Truck Percentage > 25%

On I-5, US-97 in District 2 or I-40, HWY-58 in District 8?

Yes

Yes

On Highway 86 in District 8? Yes

No

On I-5 in District 3 (Sacramento county), 6, or 10; or on US-99 in District 3, 6, or 10; or on I-505, or on I-80 in the Placer county (District 3), or on HWY-46 in District 5, or on I-580 in District

10?

Group 2a Group 2b

No

Yes

Group 2

No

Group 1 or 2No

In the coastal regions or in urban areas? Specifically, in Districts 1, 4, 5,

7, 11, or 12 , or on highways in the Sacramento County of District 3, or on highways in the San Joaquin County,

Stanislaus County, or Merced County in District 10?

No

Group 1

Yes

Truck Percentage < 10% and (4-8)/(9-15) Ratio > 1.2?

Group 1b

No

Group 1a

Yes

On I-5, US-97, or I-505?

Group 2aa

Group 2ab

Yes

No

On I-5, US-97, or I-505?

Group 2ba

Group 2bb

Yes

No

On I-5, US-97, or I-505?

Group 3a

Group 3b

Yes

No

On I-10 (postmile>30 in Riverside County), I-15 (postmile >44 in San Bernardino County), HWY-18 (postmile < 48 in San Bernardino County), I-40, Highways 58, 74 (postmile>40 in Riverside County), 78, 86, 111, 127, 177, 178, 195, 247,

US-95, or US-395 in District 8?

Yes

No

Tandem Axle Load, kN

Nor

mal

ized

Fre

quen

cy

0 20 60 100 140 180 220 260 300

0.0

0.06

0.12

0.18

0.24

Group1

Tandem Axle Load, kN

Nor

mal

ized

Fre

quen

cy

0 20 60 100 140 180 220 260 300

0.0

0.06

0.12

0.18

0.24

Group2

Tandem Axle Load, kN

Nor

mal

ized

Fre

quen

cy

0 20 60 100 140 180 220 260 300

0.0

0.06

0.12

0.18

0.24

Group3

Pavement Temperature • California is divided into six climate zones • 30-year hourly surface temperature database

developed based on EICM climate model runs • In-depth temperatures calculated with 1-D FEM on

the fly for every hour and different years

30 Years of pre-calculated Surface temperature

1-D FEM to calculate in-depth temperature

• Using: Incremental-recursive – Characterize material in terms of entire damage process

from start to end for different strain/stress/temperature levels

– Simulate damage process each step of entire life – Correlation of damage to distress – Calibrate using data from entire damage process

• Not Using: Linear sum of cycles (Miner’s Law) – Characterize material in terms of repetitions to failure

(distress) for different strain/stress/temp levels for first year – Sum up damage as nactual/Nfailure – Calibrate initial state and final distress state

Incremental-Recursive Process vs Linear Sum of Cycles

Incremental-Recursive: Use Entire Damage Process from Lab Characterization Test

• Example for asphalt fatigue shown • Considers damage and crack propagation in test

-10

-8

-6

-4

-2

0

2

0 2 4 6 8 10 12 14 16 18

Repetitions [Ln(n)]

Stiff

ness

Rat

io [L

n(-ln

(SR

))]

AR4000-D (AV = 5.61%, 699 microstrain)RAC-G (AV = 6.41%, 698 microstrain)MAC15-G (AV = 5.70%, 696 microstrain)MB15-G (AV = 6.45%, 702 microstrain)MB4-G (AV = 6.01%, 740 microstrain)Three-Stage Weibull Fitted Lines

SR = 0.1000

SR = 0.5000

SR = 0.9999

AR-4000-D RAC-G Modified binders

Time hardening process

0

0.05

0.1

0.15

0.2

0.25

0.3

0 2000 4000 6000 8000 10000 12000 14000 16000

Number of load applications

"Dam

age"

12

3

Stepping through time: Time hardening process

Calibration using entire damage process not just end state of percent of wheelpath cracked

Includes pavement preservation simulation for life cycle costing with Monte Carlo reliability calcs

Use to Date and Next Steps • Use to Date:

– California • Official state design method for flexible surfaces as of 2013, training

and implementation underway • Major issue: lab testing capability for performance related specs • Used on 3 long life rehab/reconstruct Interstate projects in last 4

years, 40 year design lives, approximately $100M total work

– Approx. 20 evaluation, teaching and APT use licenses; Skanska use in design/build projects in Sweden and eastern Europe

• Improvements for late 2016 – Web based version – Fatigue, stiffness, rutting data for high RAP Superpave mixes – Improved models for aging, consideration of thixotropy (rest

periods), stabilized full-depth reclamation – Initial models for consideration of interlayers, raveling

References: technical reports

• Calibration of Incremental-Recursive Flexible Damage Models in CalME Using HVS Experiments http://www.ucprc.ucdavis.edu/PDF/CalME%20Calib%20Rpt_RR-2005-06.pdf

• Calibration of CalME Models Using WesTrack Performance Data http://www.ucprc.ucdavis.edu/PDF/WesTrack%20CalME%20Calib_UCPRC-RR-2006-14_final.pdf

• Mn/ROAD Case Study Using CalBack and CalME http://www.ucprc.ucdavis.edu/PDF/UCPRC-TM-2008-16.pdf

• Calibration of the CalME Rutting Model Using 2000 NCAT Data http://www.ucprc.ucdavis.edu/PDF/UCPRC-TM-2008-04.pdf

• Calibration of CalME Models Using Field Data Collected from US 101 near Redwood National Park, Humboldt County http://www.ucprc.ucdavis.edu/PDF/UCPRC-TM-2008-14.pdf

References: some articles about CalME and its use

• Ullidtz, P., J. Harvey, I. Basheer, B.W. Tsai, C. Monismith, “Simulation of the WesTrack Experiment Using CalME,” Proceedings of the 3rd International Conference on Accelerated Pavement Testing, Madrid, Spain, October, 2008

• Ullidtz, P., J. Harvey, I. Basheer, D. Jones, R. Z. Wu, J. Lea, Q. Lu, “CalME: “A New Mechanistic-Empirical Design Program for Flexible Pavement Rehabilitation,” Transportation Research Record 2153, December, 2010, pp 143-152.

• Ullidtz, P., Mateos, A., Ayuso, J., Harvey, J., Basheer, I. Simulation of Full Scale Accelerated Pavement Test from CEDEX Using the Californian Predictive Pavement Design System (CalME). 11th International Conference on Asphalt Pavements, 284-294.

• Wu, R. and Harvey, J. Calibration of Asphalt Concrete Cracking Models for California Mechanistic-Empirical Design (CalME). 7th International Conference on Cracking in Pavements, RILEM. pp 537-547.

• Lu, Qing, Yu Zhang, and John Harvey. "Estimation of truck traffic inputs for mechanistic-empirical pavement design in California." Transportation Research Record: Journal of the Transportation Research Board 2095 (2009): 62-72.

• Lea, J. and J. Harvey. 2012. Simplified Thermal Modeling Approach Used in CalME. Paper No. 12-2938, Annual Meeting of the Transportation Research Board, Washington DC.

• Basheer, I., Harvey, J., Hogan, R., Signore, J., Jones, D. and Holland, J. Transition to Mechanistic Empirical Pavement Design in California. 11th International Conference on Asphalt Pavements, Nagoya, Japan, 81-90.

• Harvey, J., Lu, Q., Lea, J. D., Ullidtz, P., Wu, R. Z., Basheer, I. Features of Mechanistic Empirical Asphalt Pavement Models for New Design and Rehabilitation in California. 11th International Conference on Asphalt Pavements, Nagoya, Japan, 91-100.

• Ullidtz, P., Harvey, J., Basheer, I., Lea, J. D., Wu, R. Z. Process of Developing a Mechanistic-Empirical Asphalt Pavement Design System for California. 11th International Conference on Asphalt Pavements, Nagoya, Japan, 101-111.

• Coleri, E, RZ Wu, JT Harvey, J Signore, Calibration of incremental-recursive rutting prediction models in CalME using Heavy Vehicle Simulator experiments, Advances in Pavement Design through Full-scale Accelerated Pavement Testing, 4th International Conference on Accelerated Pavement Testing, October, 2012, Davis, CA

• Khazanovich, L, D Tompkins, P Saxena, R Wu, E Coleri, JT Harvey, 2013, Use of the Mechanistic-Empirical Pavement Design Guide and CalME to Mitigate Rutting in Asphalt Overlays of Concrete Pavements,, Transportation Research Record: Journal of the Transportation Research Board 2368, pp 36-44.

• Coleri, E, J Harvey, J Signore, V Mandapaka, K Sahasi, Comparison of rutting and cracking performance estimates for Hveem and Superpave mix designs utilizing CalME mechanistic analysis, Tranportation Research Board, January 2015 (presented, publication pending).

• Mandapaka, V., Basheer, I., Sahasi, K., Ullidtz, P., Harvey, J. and Sivaneswaran, N. Mechanistic-Empirical and Life-Cycle Cost Analysis for Optimizing Flexible Pavement Maintenance and Rehabilitation. ASCE Journal of Transportation Engineering, Volume 138(5).

• Mandapaka, V, I Basheer, K Sahasi, P Vacura, BW Tsai, CL Monismith, J Harvey, P Ullidtz, Application of four-point bending beam fatigue test for the design and construction of a long-life asphalt concrete rehabilitation project in Northern California, 3rd International Workshop on Four Point Bending, Davis, CA, October, 2012.

Questions?

Reports at www.ucprc.ucdavis.edu