Afzal Waseem, Xian-Xun (Arnold) Yuan and Medhat Shehata Department of Civil Engineering, Ryerson University

EVALUATION OF PRE-OVERLAY RUT DEPTH FOR LOCAL CALIBRATION OF THE MEPDG RUTTING MODEL
Presented by Dr. Arnold Yuan Assistant Professor, Ryerson University E-mail: arnold.yuan@ryerson.ca

Canadian Society of Civil Engineering 2013 Annual Conference, Montreal, QB, Canada, May 29 June 1, 2013

Overview
What have we achieved ?
Evaluated level-3 pre-overlay rut value for rehabilitated flexible pavement sections of Ontario.

What do we present ?
This presents the result from this investigation based on the MTOs PMS-2 data.

The goal of the investigation was to minimize the impact of an uncertain pre-overlay rut value on the accuracy of level-1 local calibration.

Canadian Society of Civil Engineering 2013 Annual Conference, Montreal, QB, Canada, May 29 June 1, 2013

What is MEPDG Method ?

MECHANISTIC EMPIRICAL RESULT

Pre-overlay rut Pre-overlay rut refers to the terminal rut depth at the end of previous life cycle of a rehabilitated flexible pavement section.

1. 2. 3. 4.

AC bottom-up cracking AC top-down cracking AC thermal facture Total Cracking

5. Permanent deformation
6. International Roughness Index

Need and Significance

Research Need:
The rut values were not measured in Ontario until 2002. Majority of Ontarios PMS-2 flexible pavement sections starts their life cycles before 2002. Pre-overlay rut value overlooked in previous local calibration studies.

Significance
Pre-overlay rut is very sensitive input parameter. Pre-overlay rut is a very important parameter in local calibration

Effects of Pre-overlay Rut

Sensitivity Analysis showing the effect of Pre-overlay Rut on Permanent deformation
9 8 Permanent Deformation (mm) 7 6 5 4 3 2 1 0 1997 Rut = 4 mm Rut = 5 mm Rut = 6 mm Rut = 7 mm Rut = 8 mm Rut = 9 mm Rut = 10 mm Rut = 11 mm

1999

2000

2001 2003 2004 2005 Pavement Life (in calendar year)

2007

2008

2010

Sensitivity of permanent deformation to pre-overlay rut value (ranging from 4 mm to 12 mm) for a typical rehabilitated flexible pavement

IRI, transverse cracking and fatigue cracking are also sensitive to Pre-overlay Rut value.

Inputs and Database

Location: Latitude, Longitude, Elevation

Climate Information Air Temperature Precipitation Moisture content Sunshine Wind speed Freezing Index

Wind speed

Sunshine

Whether Stations in DARWin-METM

Air Temperature

Precipitation Moisture content

Vehicle Inputs Axle spacing Axle configuration Traffic Wander Axle width Dual tire spacing Tire Pressure Operational speed Vehicle class

MTO Guide for Ontario`s Default MEPDG Inputs

MTO PMS-2 Database

Traffic Information AADTT Growth rate Truck distribution Truck traffic classification No. of Lanes Monthly and Hourly Adjustment factor

W/T depth

AC Layer MTO PMS-2 Database Thickness Unit weight Air void Effective binder content Poissons Ratio Tensile Strength Thermal conductivity Heat capacity Sieve Passing Base, Sub-base and MTO PMS-2 subgrade Layers Database Thickness Resilient Modulus

Inputs and Database Location of Pavement Sections

51 flexible pavement sections on Google Maps

Research Methodology
Research Tool:
MEPDG analysis were run in DARWin-METM (the official AASHTOWare for MEPDG)

Methodology:
Two approaches are taken in this study: 1) Frequency distribution analysis 2) Regression analysis Finally, local calibration of the rutting model is performed by using different pre-overlay rut values.

Criteria for Identification of Terminal Rut

Identification of terminal rut was not clearly evident in data due to inconsistencies in PMS-2 Database; therefore identification of terminal rut was done: o The identification was done on the basis of DMI and RCI. o Abrupt increase of DMI is an indication of an overlay and rutting value corresponding to the previous year was identified as the pre-overly rut of this section

Statistical Estimation
Number of Pavement Sections

Frequency Distribution:
A total of 1175 flexible pavement sections identified from the identified pre-overlay rut.
Functions Minimum Maximum Mean Mode Standard Deviation Total Rut (mm) 2.360 15.333 7.053 6.261 1.835

300 250 200

262 244 227

150
100 50 2 3 0 18

135 112 80 51 26 11 1 1

2 3 4 5 6 7 8 9 10 11 12 13 14 15 Pre-Overlay Rut (mm)

Histogram of the life-cycle maximum rut depth from PMS-2 database

On the basis of frequency distribution, a mean value of 7 mm can be used as a Level 3 input for the pre-overlay rut value for pavement sections in Ontario

Statistical Estimation
Regression Analysis:
A total of 864 flexible pavement sections from the PMS-2 database. The regression analysis of the rut depth expressed by the DMI and RCI were conducted. Regression resulted in poor correlation with a R2 value of only 0.22.

This poor fitness of this data can be explained by many factors as the both DMI and RCI (before 1997) are subjective values.

Effects of Pre-overlay Rut on Local Calibration

MEPDG Scenario Analysis:
DARWin-METM default, globally calibrated models were used in this study.

Preliminary 51 rehabilitated flexible pavement sections were analyzed with a preliminary total rut value of 4 mm
This pre-overlay rut value was selected because it was recommended by the MTO staff DARWin-METM analyses were repeated for different preoverlay rut depths ranging from 4 to 12 mm at an incremental step of 1 mm.

Effects of Pre-overlay Rut on Local Calibration

MEPDG Scenario Analysis:
Parameter Mean BIAS Root Mean Square 4.709 3.772 Error (RMSE) Std. Deviation (SD) 3.323 2.914
20

4 5 3.206 2.095

6 1.281 3.222 2.663

20

Pre-overlay Rut (mm) 7 8 9 0.477 0.067 -0.403 3.004 2.494 2.672 2.355 2.594 2.269
20

10 -0.801 2.598 2.213

11 -1.235 2.638 2.198

12 -0.605 3.701 2.912

Observed rutting (mm)

Observed rutting (mm)

15

15

Observed rutting (mm)

y = 0.2322x + 4.769 R = 0.1501

y = 0.207x + 4.6359 R = 0.1538

y = 0.207x + 4.6359 R = 0.1538

15

10

10

10

0 0 5 10 15 20

0
0 5 10 15 20

Predicted Rutting (mm)

Predicted rutting (mm)

5 10 15 Predicted rutting (mm)

20

(a) Pre-overlay rut = 7 mm

(b) Pre-overlay rut = 8 mm

(c) Pre-overlay rut = 9 mm

A pre-overlay rut of 7 to 9 mm will improve accuracy and precision of rut prediction of globally calibrated model.

CONCLUSIONS
DARWin-METM, the study has shown that both the bias and RMSE reduced drastically when the average terminal rut depth of 7 mm was used for the preoverlay value. The level-3 pre-overlay rut value of 7 mm is recommended for rehabilitated flexible pavement sections of Ontario.

REFERENCE
Waseem, A., Yuan, X.-X., and Shehata, M. (2013). "Evaluation of Pre-overlay Rut with MEPDG Rutting Model for Flexible Pavements using Ontario's Long Term PMS Data." CSCE 3rd Specialty Conference on Engineering Mechanics and Materials, Montreal, Quebec
http://www.scribd.com/doc/149970728/Evaluation-of-Pre-overlay-Rut-Depth-for-LocalCalibration-of-the-MEPDG-Rutting-Model-in-Ontario-Pavements