1. Ames, IA
August 15-16, 2013
Establishment of Relation between Pavement Surface Friction and Mixture Design Properties
Mozhdeh Rajaei
Nima Roohi Sefidmazgi
Hussain Bahia, Ph.D.

2. Outline Background and Introduction Materials and Methods
Results and Conclusions

3. Background and Introduction

4. Friction in Pavement
The cost for highway accidents in 2000 exceeded $230 billion. Many of these crashes are tied to wet road conditions, and inadequate friction characteristics (Noyce et al., 2007).
Factors affecting friction:
pavement surface characteristics,
vehicle characteristics,
tire characteristics, and
environmental conditions

5. Surface Texture
Surface texture refers to the combination of different aggregate shapes and sizes used in asphalt mixtures. Surface texture is defined in terms of wavelength (λ, distance along the surface) and amplitude (a, height above the surface).
(Henry 2000)

6. Factors Affecting Texture
Property
Texture Range
Nominal Maximum Aggregate Size (NMAS)
Macro-Texture
Mixture Coarse Aggregate Type
Micro-Texture
Mixture Fine Aggregate Type
Asphalt Binder Content
Aggregate Gradation
Mixture Air Voids
(Sandberg 2000, Hall 2009, Ahamed 2009)

7. Objectives of Study
Relate lab/field friction measures to mixture properties
Relating lab texture measurements to field friction measures

8. Materials and Methods
Stationary Laser Profilometer (SLP) Circular Track Meter (CTM) Friction Number (FN)

9. Methods Field Friction Measurement: Field Texture Measurement:
Friction Number (FN)
Field Texture Measurement:
Circular Track Meter (CTM)
Laboratory Texture Measurement:
Stationary Laser Profilometer (SLP)

10. Pavement Friction Measures
Friction number (FN):
The average coefficient of friction measured by a locked-wheel test device as specified in ASTM E274.
This device was developed to use in situations with no anti-lock brakes.
Skid Trailer was K.J. Law Profiler.
FN (V) = 100 * μ = 100 * F/W
V is velocity of the test tire, km/hr (65 km/hr).
μ is the coefficient of friction.
F is the tractive horizontal force applied to the tire, kg.
W is the vertical load applied to the tire, kg.

11. Stationary Laser Profilometer (SLP)
Used both in the laboratory and in the field.
Texture measurements described in ISO

12. Lab Texture Measure Mean Profile Depth (MPD):
MPD values yield a two-dimensional representation of the surface texture (ISO ).
MPD texture parameters provide averaged values for surface texture but do not quantify the distribution of asperities at the pavement surface.

13. Circular Track Meter (CTM)
Measures pavement macro-texture in circular area
SLP make’s linear measure.
Standardized under ASTM E

14. Materials Field sections and corresponding cores from across WI.
Field Sections and corresponding cores from MnROAD, MN.
Both dense graded and porous/gap graded.
Different mixture properties:
Varying NMAS, Pb, gradation and density.

15. Weibull Distribution
In order to describe gradation with the minimum number of variables, a Weibull distribution is fitted to gradation curve.
x is the aggregate size (mm)
κ is the shape factor
λ is the scale factor
Increasing κ

16. Analysis and Discussion

17. Regression Analysis
Regression analysis has been performed using Minitab16.
The magnitude of the statistical parameter, p-value, for each variable is an indicator of the significance of that variable
p-value closer to zero indicates high significance.
Significance of values approaching 1.0 is negligible.

18. Statistical Model Lab Friction vs. Mix Design Properties
Laser MPD = Gmb Pb NMAS κ λ
Where MPD is the mean profile depth in millimeter,
Gmb is bulk asphalt mixture density (g/cm3),
Pb is the binder percent,
NMAS is the nominal maximum aggregate size in millimeter,
κ and λ are the Weibull distribution parameters.
Predictor
Coefficient
Standard Error Coefficient
T value
P value
Constant
9.467
4.040
2.340
0.036
Gmb
-3.197
1.115
-2.870
0.013 Pb
-0.356
0.315
-1.130
0.278
NMAS
0.085
0.043
1.940
0.074 κ
1.354
0.234
5.800
0.000 λ
-1.476
0.360
-4.100
0.001

19. Model Quality of Fit

20. Discussion of Model NMAS: Positive coefficient
Gradations with higher NMAS, (generally coarser gradations) result in higher MPD
Bulk specific gravity (Gmb): Negative Coefficient
Lower Gmb, (thus higher air voids and porosity), leads to higher MPD.
Binder Content (Pb): Negative Coefficient
Lower Pb reduces binder film thickness around aggregates as well as reducing the aggregate packing level, thus increasing MPD.
Weibull Shape Factor (κ): Positive coefficient
Higher κ values will result in gradations closer to a one-sized gradation and further from the maximum density line, thus resulting in higher MPD values.
Weibull Scale Factor (λ): Negative coefficient
Decreasing λ will generally result in finer gradation. If all curves above max density line, lower λ will being further from line.

21. Discussion of Model: Gradation
It hypothesized that what is important for gradation is being further from the maximum density line, and not only the overall coarseness or fineness of the gradation.

22. Filed Friction vs. Lab Friction
CTM MPD is the mean profile depth (mm) measured using the CTM in the field,
SLP MPD is the mean profile depth (mm) measured using the laser profilometer in the laboratory.
Almost equivalent for dataset

23. Relating Laboratory Texture to Field Friction
FN is the smooth-tire friction number from field measurements,
CTM MPD is the mean profile depth (mm) measured using the CTM in the field.
SLP MPD is the mean profile depth (mm) measured using the laser profilometer in the laboratory
Using SLP-CTM
Model

24. Conclusions
Using statistical analysis mixture design parameters (i.e. volumetric and aggregate gradation properties) could be related to laboratory texture measurements (MPD).
Knowing mixture design properties can lead to the estimation of road texture parameters.
It was shown that increasing the distance of the gradation curve from the maximum density line (on both the coarse or fine side) is more important than the overall coarseness or fineness of the gradation in terms of increasing the expected texture.

25. Conclusions
Laboratory measured friction parameters (MPD) can be related to field friction values (FN) using regression analysis.
Utilizing the models developed in this study, by further investigation, mixture designers can have a guideline to estimate friction.
Models developed in this study showed that the measurements for field and laboratory compacted samples from SLP device can be used to estimate friction parameters.
A limited data set were used to develop models in this study, therefore more tests and analysis are needed to verify the results.

26. Acknowledgements
This research was sponsored by CFIRE under project I.D and the Western Research Institute "Asphalt Research Consortium".
Authors would also like to acknowledge the contributions of Mr. Timothy Miller, formerly of UW-Madison; as well as MnROAD for use of their test track database friction measurements.

27. Thank You! Questions? www.uwmarc.org Mozhdeh Rajaei rajaei@wisc.edu
Nima Roohi Sefidmazgi
Hussain Bahia 27