Post on 13-Aug-2020
ContFri
tinuingction C
g InvesCharacAggre
stigatiocteristegate
on of Pics of in Ohio
Rober
Ohio DepOffice of R
U.S. DepFeder
S
PolishinLimeso
rt Y. Liang
partment ofesearch an
partment ofral Highway
State Job N
FHWS
ng andstone
g, Ph.D.,
fof Transporta
nd Developm
andf Transportay Administra
Number: 134
Final ReWA/OH-200September 2
d
P.E.
or the ation ment
d the ation ation
4219
eport 09/10 2009
ii
iii
Continuing Investigation of Polishing and Friction Characteristics of Limestone
Aggregate in Ohio
Robert Y. Liang, Ph. D., P.E.
431 ASEC, Civil Engineering Department University of AKRON
Akron, OH, 44325
Credit Reference: Prepared in cooperation with the Ohio Department of Transportation and the U.S. Department of Transportation, Federal Highway Administration. Disclaimer Statement: The contents of this report reflect the views of the authors who are responsible for the facts and accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the Ohio Department of Transportation or the Federal Highway Administration .This report does not constitute a standard, specification or regulation.
September 2009
iv
ABSTRACT
Due to increased focus on maintaining highway operating safety and reducing wet weather
accidents, highway agencies have adopted a routing practice to monitor pavement surface
friction values. Skid Number (SN) is a measure of pavement surface friction determined
by LWST (Locked Wheel Skid Trailer). Once the SN falls below certain threshold criteria,
highway agencies would start a pavement resurfacing program to ensure that adequate
friction is maintained. The practice of monitoring and remedying the low skid resistance
pavement sections is important; however, it is a passive approach. A more proactive
approach would be to test the hot mix asphalt in the laboratory during its initial mix design
stage to ensure that aggregate combinations used in the mix design would provide
adequate friction over the expected life span of the pavement.
To achieve the screening task for polishing and friction behavior of the HMA during its
mix design stage, a laboratory-scale accelerated polishing device that can mimic the actual
abrasion and polishing behavior between the vehicle rubber tire and the HMA surface has
been developed. The developed polishing device possesses some practical characteristics.
These include the versatility of testing HMA specimens that can be prepared with the
roller compaction equipment or the gyratory Superpave compaction equipment, the test
duration is reasonably short, the test procedure including test specimen preparation and
v
friction measurements is relatively simple and repetitive, the developed test procedure
requires minimum labor efforts, and finally, the test results can provide a realistic
indication of the polishing and friction behavior of the hot mix asphalt specimens.
The validation of the developed accelerated polishing device included favorable
comparisons between the HMA polishing behavior from the new test device with the
aggregate polishing behavior from the British Polishing Wheel tests. Furthermore, a total
of eight pavement sections in service were identified for long–term friction and texture
measurements, which provide important data correlations between field performance and
laboratory obtained behaviour. The laboratory tests on the HMA samples prepared with
the job mix formula for the eight monitored pavement sections were completed and
analysed to confirm test repeatability of the new laboratory accelerated polishing device.
The statistical correlation analyses on laboratory and field data have resulted in the
development of useful empirical predictive equations for the Skid Number (SN) from the
DFT (Dynamic Friction Tester) measured skid number. Finally, the British Pendulum
Number (BPN) from the British Pendulum Tester (BPT) was strongly correlated with the
Mean Texture Depth (MTD) measured by the sand patch method.
Based on the correlation studies conducted in the study, together with other highway
agencies’ acceptance criteria for aggregate friction values, a set of tentative polishing and
friction screening criteria for use by Ohio DOT is presented in the report.
vi
TABLE OF CONTENTS
Page LIST OF TABLES..........................................................................................................xiii
LIST OF FIGURES…………………………………………………...........……..…....xv
CHAPTER
1. INTRODUCTION…………………………..………………………………………..1
1.1 Statement of Problem .................................................................................... 1
1.2 Objectives of the Study ................................................................................. 3
1.3 Scope of the Work ........................................................................................ 4
1.4 Outlines of the Report ................................................................................... 5
2. EXISTING LITERATURE METHODS AND EQUIPMENTE REVIEW…..……...9
2.1 Introduction ................................................................................................... 9
2.2 Background and Significance of Work ....................................................... 11
2.2.1 Mechanism of Polishing, Wearing and Skid Resistance ......................... 11
2.2.2 Factors Affecting Skid Resistance ........................................................... 14
2.2.3 Roughness and Texture ............................................................................ 16
2.2.4 Importance of Aggregate Characteristics to Surface Performance .......... 20
2.2.5 Aggregate Factors Affecting Pavement Friction ..................................... 22
vii
2.2.5.1 Aggregate Shape ................................................................................... 22
2.2.5.2 Aggregate Size and Gradation .............................................................. 23
2.2.5.3 Resistance to Polish-Wear Action ........................................................ 23
2.2.6 Models for Wet Pavement Friction .......................................................... 24
2.2.6.1 The Penn State Model ........................................................................... 24
2.2.6.2 The Rado Model ................................................................................... 26
2.2.6.3 The PIARC Model and the International Friction Index ...................... 28
2.2.7 Frictional Needs of Traffic ...................................................................... 30
2.2.8 Factors Affecting Wet-Pavement Safety ................................................. 31
2.2.9 Skid Resistance Requirements and Practices by Different Agencies.. .... 33
2.2.10 Air Void and Temperature Effect on Frictional Properties of Asphalt
Pavement Surface ………………………………………………… …………34
2.2.11 Overview of Polishing, Friction, and Texture Measurements ............... 37
2.2.11.1 Existing Accelerated Polishing Machines .......................................... 37
2.2.11.1.1 Polishing Devices for…………...:………………………………..38
2.2.11.1.1.1 British Polishing Wheel…………………………………………38
2.2.11.1.1.2 Michigan Indoor Wear Track ...................................................... 39
2.2.11.1.1.3 Micro-Deval Device .................................................................... 39
2.2.11.1.2 Polishing Devices for HMA:……………………………………..39
2.2.11.1.2.1 NCAT Polishing Machine ........................................................... 40
2.2.11.1.3 Polishing Devices for Aggregates and HMA:……………………40
2.2.11.1.3.1 North Carolina State University Wear and Polishing Machine .. 41
viii
2.2.11.1.3.2 Wehner/Schulze Polishing Machine ........................................... 41
2.2.11.1.3.3 Penn State Reciprocating Polishing Machine ............................. 42
2.2.11.2 Friction Measurement Methods .......................................................... 42
2.2.11.2.1 Locked Wheel Friction Devices…………………………………..44
2.2.11.2.2 Side Force Coefficient Devices…………………………………..46
2.2.11.2.3 Fixed Slip Devices………………………………………………..48
2.2.11.2.4 Variable Slip Devices……………………………………… ……53
2.2.11.2.5 Other Friction Measurement Methods……………………………55
2.2.11.2.5.1 Dynamic Friction Tester .............................................................. 55
2.2.11.2.5.2 Pendulum Devices ....................................................................... 56
2.2.11.2.5.3 Michigan Laboratory Friction Tester .......................................... 58
2.2.11.2.5.4 PTI Friction Tester ...................................................................... 59
2.2.11.2.5.5 Stopping Distance Method .......................................................... 60
2.2.11.2.5.6 Wehner/Schulze Friction Device ................................................. 61
2.2.11.3 Texture Measurement Methods .......................................................... 61
2.2.11.3.1 Microtexture Measurement……………………………………….62
2.2.11.3.2 Macrotexture Measurement………………………………………63
2.2.11.3.2.1 Volumetric Measurements .......................................................... 64
2.2.11.3.2.2 Outflow Meter ............................................................................. 65
2.2.11.3.2.3 Profile Tracers ............................................................................. 66
3. A NEW ACCELERATED POLISHING DEVICE FOR HMA SURFACES……...70
3.1 Introduction ................................................................................................. 70
ix
3.2 Existing Laboratory Scale Polishing Devices ............................................. 71
3.3 Equipment Development ............................................................................ 72
3.3.1 Equipment Description and Operational Procedure ................................ 73
3.3.2 Operation Conditions ............................................................................... 78
3.4 Equipment Characteristics and Validation ................................................. 79
3.4.1 Materials .................................................................................................. 80
3.4.2 Sample Preparation Procedure for HMA Specimens ............................... 80
3.4.3 Friction and Surface Texture Measurements ........................................... 81
3.4.4 Supplemental Image Analysis Techniques .............................................. 83
3.4.5 Repeatability of the Accelerated Polishing Equipment ........................... 83
3.4.6 Polishing Effect of the Accelerated Polishing Machine .......................... 84
3.4.7 Comparing HMA Surface and Aggregate Surface PolishingBehavior... 91
3.4.8 Comparing the Polishing Trend with the Aggregate Exposure Area ...... 93
3.4.9 Polishing Trend of HMA Samples Prepared by Two Compaction
Methods............................................................................................................. 96
3.4.10 Application of the Accelerated Polishing Device .................................. 97
3.4.10.1 Correlation with PV values ................................................................. 98
3.4.10.2 Correlation with SN values ............................................................... 100
3.5 Summary and Conclusions ....................................................................... 101
4. LABORATORY TEST RESULTS AND DATA ANALYSIS……………………104
4.1 Introduction ............................................................................................... 104
4.2 Pavement Sections and Material Properties .............................................. 105
x
4.3 Test Program ............................................................................................. 107
4.3.1 Sample Preparation Procedure for HMA Specimens ............................. 107
4.3.2 Friction and Texture Measurement Techniques .................................... 108
4.3.3 Accelerated Polishing Device ................................................................ 108
4.4 Laboratory Test Results ............................................................................ 109
4.5 Analysis of Test Results ........................................................................... 109
4.5.1 Analysis of Repeatability ....................................................................... 109
4.5.2 Analysis of Polishing Behavior (BPN) .................................................. 115
4.5.2.1 Rate of Friction Loss (Percent Hourly Drop in Polish Numbers) ....... 115
4.5.2.2 Absolute and Percent Value of Decrease (Initial Polish Number versus
Final Polish Number) ...................................................................................... 120
4.5.3 Surface Texture Behaviour .................................................................... 121
4.5.3.1 Rate of SurfaceTexture Loss(Percent Hourly Drop in TextureValues)121
4.5.3.2 Absolute and Percentage Value of Decrease (Initial Texture Value
versus Final Texture Value). ........................................................................... 125
4.5.4 Correlation Study between BPN and MTD ........................................... 125
4.6 Summary and Conclusions ....................................................................... 129
5. LABORATORY STUDY OF AIR VOID AND TEMPERATURE EFFECTS ON
HMA FRICTIONOPERTIES…………………………………………………….…131
5.1 Introduction ............................................................................................... 131
5.2 Background ............................................................................................... 132
5.3 Laboratory Testing Program ..................................................................... 136
xi
5.3.1 Materials ................................................................................................ 136
5.3.2 Test Program .......................................................................................... 136
5.3.3 Accelerated Polishing Machine ............................................................. 137
5.3.4 Friction Measurement Method ............................................................... 138
5.4 Test Results and Analysis ......................................................................... 138
5.4.1 Air Voids Effects ................................................................................... 138
5.4.2 Temperature Effects ............................................................................... 143
5.5 Summary and Conclusions ....................................................................... 147
6. CORRELATION STUDY BETWEEN FRICTION MEASUREMENTS BY LWST
AND DFT………………………..…………………………………………..……....149
6.1 Introduction ............................................................................................... 149
6.2 Friction and Texture Measurement Techniques in This Study ................. 151
6.3 Experimental Program .............................................................................. 152
6.4 Field Test Results and Data Analysis ....................................................... 155
6.4.1 Typical LWST, DFT, and CTM Test Results ........................................ 155
6.4.2 Analysis of Test Results......................................................................... 157
6.4.2.1 Simple Linear Regression ................................................................... 159
6.4.2.2 Multicollinearity ................................................................................. 162
6.4.2.3 Multiple Linear Regression................................................................. 162
6.5 Summary and Conclusions ....................................................................... 164
7. POLISHING MACHINE BASED ON HIGH-PRESSURE WATER JET………..166
xii
7.1 Introduction ............................................................................................... 166
7.2 Equipment Development .......................................................................... 170
7.2.1 Equipment Description and Operational Procedure .............................. 170
7.3 Equipment Characteristics and Validation ............................................... 175
7.3.1 Materials ................................................................................................ 175
7.3.2 Sample Preparation Procedure for HMA Specimens ............................. 176
7.3.3 Friction and Surface Texture Measurements ......................................... 176
7.3.4 Work Plan .............................................................................................. 177
7.4 Polishing Effect of the Accelerated Polishing Machine ........................... 177
7.5 Summary and Conclusions ....................................................................... 184
8. SUMMARY AND CONCLUSIONS……………………………………………..186
8.1 Summary of Work Done ........................................................................... 186
8.2 Observations and Conclusions .................................................................. 188
8.3 Recommendations for Implementation ..................................................... 194
8.4 Recommendations for Future Work ......................................................... 195
REFERENCES……………………………………………………………………….198
APPENDICES………………………………………………………………………..207
APPENDIX A. ................................................................................................ 207
APPENDIX B. ................................................................................................ 224
xiii
LIST OF TABLES
Table Page
3-1: A summary of the existing accelerated polishing machines .................................... 74
3-2: Range of and selected optimum operation parameters ............................................. 79
3-3: Repeatability tests for the limestone and gravel ........................................................ 86
3-4: Simple Linear Regression between Aggregate Friction Values (Liang and Chyi
2000) and H . MA Friction Values (This Study) for Columbus Limestone………...……92
3-5: Simple Linear Regression between Aggregate Friction Values (Liang and Chyi
2000) and HMA Friction Values (This Study) for Stocker Sand and Gravel .................. 92
3-6: Simple Linear Regression between Aggregate and HMA Friction Values and
Aggregate Exposure Area ................................................................................................. 96
3-7: Simple Linear Regression between Friction Values of Gyratory Compacted
Specimens and Friction Values of Roller Compacted Slab Specimens (Limestone
aggregate) ......................................................................................................................... 97
3-8: Simple Linear Regression between Friction Values of Gyratory Compacted
Specimens and Friction Values of Roller Compacted Slab Specimens (Sand and Gravel
aggregate) ......................................................................................................................... 97
3-9: TxDOT acceptance criterion of aggregates ............................................................... 99
3-10: Derived acceptance criteria of HMA based on BPN values ................................... 99
xiv
3-11: Derived acceptance criterion of HMA based on SN values ................................. 101
4-1: Asphalt concrete pavement sections and the associated JMFs ............................... 106
4-2: Repeatability tests for the eight different job mix formulas .................................... 111
4-3: Percent decrease in BPN and MTD between initial and final values ..................... 120
4-4: Percent decrease in BPN and MTD between initial and final values ..................... 126
4-5: Simple linear regression between BPN and MTD .................................................. 127
4-6: Simple linear regression between BPN and MTD.............................................. 128
5-1: Summary of the studies focused on the temperature effect on HMA frictional
properties ........................................................................................................................ 135
5-2: Test of Homogeneity of variances, 1-Way ANOVA Table, and Multiple
Comparisons for the Effect of HMA Air Void on BPN ................................................. 142
5-3: Test of Homogeneity of variances, 1-Way ANOVA Table, and Multiple
Comparisons for the Effect of Pavement, Rubber Slider, and Water Temperatures on
BPN ................................................................................................................................ 146
6-1: HMA Pavement sections identification .................................................................. 153
6-2: Sample of skid numbers measured using LWST for one pavement section ........... 156
6-3: Simple linear regression between SN(64)R and DFT64, DFT20, and MPD .......... 160
6-4: Multicollinearity check using Tolerance, VIF, and Condition Index on the
independent variables ..................................................................................................... 162
6-5: Multiple linear regression between MPD and DFT64, MPD and DFT20, and DFT20
and DFT64 ...................................................................................................................... 163
7-1: Work plan summary of the laboratory work .......................................................... 177
xv
LIST OF FIGURES
Figure Page
Figure 2-1: Schematic of adhesion and hysteresis of rubber-tire friction ......................... 13
Figure 2-2: The contribution of adhesion (microtexture) and hysteresis (macrotexture) to
the friction factor as a function of sliding speed (reproduced from Federal Aviation
Administration 1971) ........................................................................................................ 14
Figure 2-3: Texture wavelength influence on surface characteristics (reproduced from
PIARC 1987) .................................................................................................................... 18
Figure 2-4: Schematic representation of microtexture and macrotexture ......................... 19
Figure 2-5: Effect of microtexture and macrotexture on skid resistance as a function of
speed ................................................................................................................................. 19
Figure 2-6: Terms used to describe the texture of a road surface ..................................... 20
Figure 2-7: Weight phase diagram of Hot Mix Asphalt ................................................... 21
Figure 2-8: Penn State Model (NCHRP Synthesis 291, 2000) ......................................... 26
Figure 2-9: Rado Model .................................................................................................... 27
Figure 2-10: Friction-slip curve of a braking tire (reproduced from Federal Aviation
Administration 1971) ........................................................................................................ 43
Figure 2-11: Locked Wheel Skid Trailer .......................................................................... 45
Figure 2-12: Ribbed tire versus smooth tire ..................................................................... 45
Figure 2-13: Sideways Force Coefficient Routine Investigation Machine (SCRIM) ....... 47
Figure 2-14: Side force tester: the MuMeter (Tomita 1964) ............................................ 48
xvi
Figure 2-15: The Runway Friction Tester ........................................................................ 52
Figure 2-16: The Griptester device ................................................................................... 52
Figure 2-17: Saab Friction Tester ..................................................................................... 53
Figure 2-18: The portable friction tester for measuring the friction on small surfaces .... 53
Figure 2-19: Norsemeter road friction measurement trailer ............................................. 54
Figure 2-20: Dynamic Friction Tester: (a) bottom view, (b) general view, and (c)
controller ........................................................................................................................... 56
Figure 2-21: British Pendulum Tester ............................................................................... 58
Figure 2-22: North Carolina State University Variable Speed Friction Tester ................ 58
Figure 2-23: Michigan Laboratory Friction Tester ........................................................... 59
Figure 2-24: MTD determination using the sand patch method ....................................... 65
Figure 2-25: Outflow meter .............................................................................................. 66
Figure 2-26: Circular Texture Meter: (a) general view, and (b) bottom view .................. 68
Figure 2-27: Mean Segment Depth determination ............................................................ 68
Figure 3-1: Different views of the accelerated polishing machine using rubber shoes; all
units are in inches ............................................................................................................. 75
Figure 3-2: Overall view of the accelerated polishing machine using rubber shoes and
setups for testing slab specimen and gyratory compacted specimen ................................ 78
Figure 3-3: Gradation curves ............................................................................................ 80
Figure 3-4: Polishing, friction, and texture results of tests conducted on limestone slab
specimens .......................................................................................................................... 87
xvii
Figure 3-5: Polishing, friction, and texture results of tests conducted on limestone
gyratory compacted specimens ......................................................................................... 88
Figure 3-6: Polishing, friction, and texture results of tests conducted on Sand and Gravel
slab specimens .................................................................................................................. 89
Figure 3-7: Polishing, friction, and texture results of tests conducted on Sand and Gravel
gyratory compacted specimens ......................................................................................... 90
Figure 3-8: Image analysis results of tests conducted on Limestone gyratory compacted
specimens .......................................................................................................................... 94
Figure 3-9: Image analysis results of tests conducted on Sand and Gravel gyratory
compacted specimens ....................................................................................................... 95
Figure 4-1: Average percent hourly drop in BPN vs. polishing time for low polish
susceptibility aggregates ................................................................................................. 117
Figure 4-2: Average percent hourly drop in BPN vs. polishing time for medium polish
susceptibility aggregates ................................................................................................. 117
Figure 4-3: Average percent hourly drop in BPN vs. polishing time for high polish
susceptibility aggregates ................................................................................................. 118
Figure 4-4: Normalization of BPN wrt. the maximum difference in BPN for low polish
susceptibility aggregates ................................................................................................. 118
Figure 4-5: Normalization of BPN wrt. the maximum difference in BPN for medium
polish susceptibility aggregates ...................................................................................... 119
Figure 4-6: Normalization of BPN wrt. the maximum difference in BPN for high polish
susceptibility aggregates ................................................................................................. 119
xviii
Figure 4-7: Average percent hourly Drop in MTD vs. polishing time for low polish
susceptibility aggregates ................................................................................................. 122
Figure 4-8: Average percent hourly drop in MTD vs. polishing time for medium polish
susceptibility aggregates ................................................................................................. 122
Figure 4-9: Average percent hourly drop in MTD vs. polishing time for high polish
susceptibility aggregates ................................................................................................. 123
Figure 4-10: Normalization of MTD wrt. the maximum difference in MTD for low polish
susceptibility aggregates ................................................................................................. 123
Figure 4-11: Normalization of MTD wrt. the maximum difference in MTD for medium
polish susceptibility aggregates ...................................................................................... 124
Figure 4-12: Normalization of MTD wrt. the maximum difference in MTD for high
polish susceptibility aggregates ...................................................................................... 124
Figure 5-1: Gradation curve ............................................................................................ 137
Figure 5-2: BPN vs. polishing time at different air voids ............................................... 139
Figure 5-3: BPN vs. air voids ......................................................................................... 143
Figure 5-4: BPN vs. polishing time at different pavement, rubber slider, and water
temperatures .................................................................................................................... 144
Figure 5-5: BPN vs. temperature .................................................................................... 146
Figure 6-1: Simple linear regression ............................................................................... 160
Figure 7-1: GRAP polishing machine ............................................................................. 169
Figure 7-2: Schematic depiction of GRAP polishing machine concept ......................... 169
Figure 7-3: GRAP aggregate specimen .......................................................................... 170
xix
Figure 7-4: Schematic depiction of the concept of the high-pressure water jet polishing
machine using HMA specimens ..................................................................................... 171
Figure 7-5: Overall view of the accelerated polishing machine using high-pressure water
........................................................................................................................................ 173
Figure 7-6: Drum (chamber) for placing the test specimen ............................................ 174
Figure 7-7: Details on slab specimen mounting in the drum .......................................... 174
Figure 7-8: Details on gyratory compacted specimen mounting in the drum ................. 175
Figure 7-9: Aggregate gradation curve ........................................................................... 176
Figure 7-10: Friction values for trial number 1 (at 10 rpm and 1450 psi) ...................... 179
Figure 7-11: MPD trend for Trial number 1 (at 10 rpm and 1450 psi ) .......................... 180
Figure 7-12: Friction values for trial number 2 (at 10 rpm and 400 psi) ........................ 180
Figure 7-13: MPD trend for trial number 2 (at 10 rpm and 400 psi) .............................. 181
Figure 7-14: Friction values of trial number 3 (at 10 rpm 1000 psi) .............................. 181
Figure 7-15: MTD trend for trial number 3 (at 10 rpm 1000 psi) ................................... 182
Figure 7-16: Friction values for trial number 4 (at 10 rpm 500 psi) ............................... 182
Figure 7-17: MTD trend for trial number 3 (at 10 rpm 500 psi) ..................................... 183
Figure 7-18: Tested HMA roller-compacted slab specimen surface .............................. 183
Figure 7-19: Tested HMA roller-compacted slab specimen surface .............................. 184
1
CHAPTER I
1. INTRODUCTION
1.1 Statement of Problem
Asphalt concrete pavements, over passage of time and under traffic, gradually lose their
surface friction (skid resistance), creating a serious safety concern especially when
pavements are wet. Lack of adequate skid resistance of pavement surface can create a
serious safety concern to the travelling vehicles at high speed, especially when the
vehicle is braking suddenly on wet pavement surface where hydroplaning can occur.
Statistical data has shown that most of fatal accidents on highways are related to
hydroplaning when uncontrolled skidding and gliding of high speed vehicles occurs.
It has been known that skid resistance is affected by factors, such as bleeding of
asphalt, polished aggregate, smoothened macrostructure, rutting and inadequate cross
slope. However, aggregate and mixture characteristics remain the most dominant
controlling factor. The ability of different aggregates to resist polish and wear is related
to mineralogical and chemical compositions as well as physical properties, such as
texture, particle size, shape, and gradation.
2
In recognition of the importance of providing a high skid resistance pavement surface
to prevent hydroplaning, most highway agencies maintain a strong safety program in
enforcing regular measurement of skid resistance of asphalt pavement surface by the
use of the Locked Wheel Skid Trailer (LWST). Once the measured skid resistance is
below certain criteria specified by the highway agencies, a remedial action either in
terms of resurfacing of the pavement or rejuvenating surface texture would be taken to
restore the friction and skid resistance to the acceptable level. Although this practice is
commendable in that the highway pavement surface is maintained at adequate skid
resistance level; nevertheless, the cost associated with the regular
monitoring/measuring program and the remediation action could be extensively high.
An alternative approach would be to take a more proactive stand in screening the
polishing potential of aggregates and Hot Mix Asphalt (HMA) mix to ensure that the
final selected mix can provide sustained resistance to polishing while providing
adequate level of friction (or skid resistance) over the life span of the pavement. To
achieve this initial screening task during the mix design stage of the HMA, there is a
need to develop a laboratory-scale accelerated polishing device that can mimic the
actual abrasion and polishing behavior between the vehicle rubber tire and the HMA
surface. In developing this accelerated polishing device, there are some practical
considerations that ideally should be taken into account, including the versatility of
testing HMA specimens that can be prepared with the conventional compaction
methods (i.e., gyratory compaction or roller compaction), the specimens could be
prepared as part of mix design procedure (i.e., the gyratory compactor together with the
3
industry standard 6 inch diameter mold), test duration should be relatively short, test
procedure including test specimen preparation and friction measurement techniques
should be relatively simple and repetitive, the developed test procedure should require
minimum labor efforts, and finally, the test results should provide realistic indication
(screening outcome) of the polishing and friction behavior of the HMA specimens.
1.2 Objectives of the Study
The main objective of the research is to develop a practical, time efficient quality
control test procedure to screen potential polishing and friction performance of
aggregates during the mix design stage in the Superpave designed HMA. To this end,
the specific objectives are outlined below.
Develop new accelerated polishing equipment for Superpave HMA to facilitate rapid
simulation of polish and wear actions between vehicle tires and asphalt pavement
surface. The new accelerated polishing equipment should have the following attributes:
the capability to test small-size HMA specimens (e.g., gyratory compacted specimens
with 6 inch in diameter and 4 inch in height), the test can be completed in a reasonable
timeframe, simple procedure, and efficient test method with less labor effort.
Develop a complete test protocol to include sample preparation method, test sequence,
data precision and bias analysis, and acceptance criteria for adoption by the governing
state or federal highway agencies.
4
1.3 Scope of the Work
The research involved both laboratory experimental work and field measurements of
asphalt pavement friction/skid performance. In addition, significant development work
was involved in the design, fabrication, and trying out the new accelerated polishing
equipment for HMA specimens. The development work of appropriate test apparatus
has been subjected to several cycles of refinement and validation. Furthermore,
separate studies were carried out to investigate the air void and temperature effects on
Hot Mix Asphalt (HMA) frictional properties in laboratory settings.
For correlation study eight asphalt pavement sections at different locations in Ohio
were identified and the associated HMA mix design (Job Mix Formulas) were provided
by the Ohio Department of Transportation. The compacted Superpave HMA
specimens, using the provided mix formulas, were produced in large numbers for each
aggregate source for a series of laboratory tests using the developed accelerated
polishing equipment. The test sequence of each specimen was carried out as follows.
First, the initial value of British Pendulum Number (BPN) and surface texture
(roughness) of the newly compacted HMA specimens were measured. In addition,
image analysis was carried out to quantify the area of exposed aggregate surface. Next,
the specimen was subjected to accelerated polishing using the developed accelerated
polishing equipment. The entire test duration of each polishing action was optimized by
examining the evolution of BPN and texture values at various stages of test. The plan
was to polish until the specimen reached the residual state (in terms of friction and
5
texture values) is reached. The test results for all aggregate sources are plotted in terms
of BPN and texture values versus duration of polishing action.
The field work at the identified pavement sections involves annually measuring friction
and texturing using the Dynamic Friction Tester and the Circular Texture Meter,
respectively. The measured values are correlated with skid number measured by the
Locked Wheel Skid Trailer conducted by ODOT personnel at the same location.
Finally, the laboratory determined friction values at different test durations during
accelerated polishing are compared with history of skid number in the field. This
comparison serves as validation of the developed test procedure.
A different type of polishing equipment based on the concept of high-pressure water jet
has also been developed in this study to investigate the potential of using the high-
pressure water jet to accomplish the desired accelerated polishing on HMA surface.
The preliminary results of the initial trial of the equipment are reported in this report
and some surprising outcomes of using this equipment has pointed to the possible use
of the technique for rejuvenating the texture of the worn out existing pavement surface.
1.4 Outlines of the Report
Presented in chapter II is a review of the literature on the concepts and theories of
polishing of aggregate and HMA, the frictional characteristics of aggregate and HMA,
as well as the interrelationship between aggregate source, HMA friction properties, and
HMA surface texture properties. Chapter II also provides information on the different
6
equipment used for accelerated polishing of aggregates and asphalt mixtures, as well as
different friction and texture measurement devices. Relevant research and practice by
the state and federal highway agencies on the related topics was also covered in this
chapter.
Chapter III presents the development of a new laboratory-scale accelerated HMA
polishing device for the purpose of screening the polishing and friction performance of
the HMA mix, in terms of differentiating the contributing factors of the aggregate
source, aggregate gradation, and binder type and content that constitute the important
constituents of the HMA. The operation principles along with the operation conditions
of this equipment are described in detail. The performance of this accelerated polishing
test device is evaluated and described in detail as well. The tentative acceptance criteria
based on other agencies’ acceptance criteria and the established correlations in this
study are presented for screening the HMA specimens from the polishing and friction
standpoint. The advantages and limitations of the developed accelerated polishing
device are presented at the conclusion of this chapter.
Chapter IV presents the laboratory test results and the associated data analysis
conducted on laboratory-prepared gyratory-compacted HMA specimens that are made
of eight different job mix formulas using the developed laboratory-scale accelerated
HMA polishing device for the purpose of examining the repeatability of the newly
developed accelerated polishing device and to investigate the relationship between
friction and texture values.
7
Chapter V introduces the controlled laboratory test results for quantifying the effects of
air void and temperature on the measured friction properties of the gyratory compacted
HMA surfaces. The laboratory test procedures, including the materials used in
preparing the HMA specimens, the method used to polishing the HMA surface, and the
friction measurement techniques, are described in detail. Statistical analysis was carried
out to determine the significance of these two variables (air void and temperature) on
the measured friction values. Finally, the method for extrapolating the friction values
measured at any density and temperature to the friction values at other density and
temperature is proposed at the end of the chapter.
The main objective of Chapter VI is to present the measurements that consist of the
skid numbers, the friction numbers measured using the dynamic friction tester, and the
mean profile depth measured using the circular texture meter at the same time for the
same pavement surface location at the long-term monitoring pavement sections. From
these measured data, a statistical study was conducted to develop the relationship to
predict the skid number at 64 km/h (40 mph) using the ribbed tire locked wheel skid
trailer (SN(64)R) from one or the combination of the following three measurements: (a)
the friction number at 64 km/h (40 mph) using the dynamic friction tester (DFT64), (b)
the friction number at 20 km/h (12.5 mph) measured by the dynamic friction tester
(DFT20), and (c) the mean profile depth measured by the circular texture meter (MPD).
Either one of the three predictive equations can be used successfully to predict the
SN(64)R values.
8
Chapter VII presents the development of a different accelerated polishing equipment
that involves the use of high-pressure water jet. The operation principles along with the
operation conditions and the performance of the developed accelerated polishing test
device are evaluated and described. The ability of the developed device to simulate the
tire-pavement interaction is discussed and presented in this chapter as well.
Chapter VIII provides the summary of work done as well as conclusions and
recommendations for implementation.
9
CHAPTER II
2. EXISTING LITERATURE METHODS AND EQUIPMENTE REVIEW
2.1 Introduction
Asphalt concrete pavements under traffic loads and environmental weathering can
gradually lose surface friction (or skid resistance) due to tire and pavement interaction
through polishing and wearing as well as freezing and thawing induced degradation.
Although weather effect may be present, the primary cause of polishing and loss of
friction can be attributed to loss of microtexture and macrotexture of the pavement
surface through prolonged abrasion action between the vehicle tires and pavement
surface. Lack of adequate skid resistance of a pavement surface can create a serious
safety concern to the travelling vehicles at high speed, especially when the vehicle is
braking suddenly on wet pavement surface where hydroplaning can occur. Statistical
data has shown that most of fatal accidents on highways are related to hydroplaning
when uncontrolled skidding and gliding of high speed vehicles occurs. Hydroplaning
results when vehicle tires move fast relative to the wet pavement surface, such that
there is insufficient time to channel the moisture away from the center of the tire. The
result is that the tire is lifted by the water away from the road and all traction is lost.
10
In recognition of the importance of providing high skid resistance pavement surface to
prevent hydroplaning, most highway agencies maintain a strong safety program in
enforcing regular measurement of skid resistance of asphalt pavement surface by the
use of the Locked Wheel Skid Trailer (LWST). Once the measured skid resistance is
below certain criteria specified by the highway agencies, a remedial action either in
terms of resurfacing of the pavement or rejuvenating surface texture would be taken to
restore the friction and skid resistance to the acceptable level. Although this practice is
commendable in that the highway pavement surface is maintained at adequate skid
resistance level; nevertheless, the cost associated with the regular
monitoring/measuring program and the remediation action could be extensively high.
An alternative approach would be to take a more proactive stand in screening the
polishing potential of aggregates and Hot Mix Asphalt (HMA) mix to ensure that the
final selected mix can provide sustained resistance to polishing while providing
adequate level of friction over the life span of the pavement. To achieve this initial
screening task during the mix design stage of the HMA, there is a need to develop a
laboratory-scale accelerated polishing device that can mimic the actual abrasion and
polishing behaviour between the vehicle rubber tire and the HMA surface. In
developing this accelerated polishing device, there are some practical considerations
that ideally should be taken into account, including the versatility of testing HMA
specimens that can be prepared with the conventional compaction methods (i.e.,
gyratory compaction or roller compaction), the specimens could be prepared as part of
mix design procedure (i.e., the gyratory compactor together with the industry standard
11
6 inch diameter mold), test duration should be relatively short, test procedure including
test specimen preparation and friction measurement techniques should be relatively
simple and repetitive, the developed test procedure should require minimum labour
efforts, and finally, the test results should provide realistic indication (screening
outcome) of the polishing and friction behaviour of the HMA specimens.
2.2 Background and Significance of Work
The background and the pertinent literature review concerning the significance of work
is presented herein.
2.2.1 Mechanism of Polishing, Wearing and Skid Resistance
Polishing of aggregates is defined as the loss of small asperities of road surfaces. These
asperities are called microtexture. Wearing, on the other hand, is the loss of
macrotexture or surface irregularities. Most researchers agree that the principal
mechanism of polishing is the abrasion of the small aggregate asperities as a result of
the rubbing action under loaded tires with the fine road detritus as the abrasive agent.
The principal mechanism of wearing involves continuous abrasion resulting from loads
and environmental changes such as freezing/thawing, wetting/drying, and oxidation.
Polishing and wearing generally involve similar processes that vary only in the degree
and the rate of material loss. Friction, which is the force that resists the relative motion
between two bodies in contact, is an essential part of the tire-pavement interaction. Not
only does friction allow a vehicle to accelerate and maneuver, but also it exerts a major
12
determining factor in the ability to stop a vehicle. The factors influencing the
development of friction between rubber tires and a pavement surface include the
texture of the pavement surface, vehicle speed, and the presence of water. However,
pavement skid resistance (or pavement friction) is defined as the ability of a travelled
surface to prevent the loss of traction with the vehicle rubber tires . Skid resistance and
texture of pavement surface are two important parameters often measured during the
service life of the pavement to ensure that they meet the minimum required criteria for
safety reason. Theoretically, the friction that develops between a rubber tire and a
travelled pavement surface consists of two components; namely, adhesion and
hysteresis (Kummer, 1966). As depicted in Figure 2-1, adhesion is the shear force
between the tire and the pavement surface generated when the tire rubber slides over
the aggregate surface asperities due to microtexture and the aggregate particles indent
onto the rubber. In essence, adhesion can be viewed as the molecular bonds generated
when the tire rubber deforms under load. The second friction component, hysteresis, is
developed when the tire rubber deforms due to macrotexture (or irregularities) of the
pavement surface. In essence, it can be viewed as energy loss that occurs as the rubber
is alternately compressed and expanded as it slides over the irregular pavement surface
texture.
13
Figure 2-1: Schematic of adhesion and hysteresis of rubber-tire friction
Figure 2-2 is a schematic diagram of the contribution of adhesion and hysteresis to the
friction factor. At low speed, friction is due mainly to adhesion. On the other hand, at
high speed, the contribution of hysteresis becomes more significant. A pavement that is
covered with a thin film of lubricant would provide only hysteresis.
14
Figure 2-2: The contribution of adhesion (microtexture) and hysteresis
(macrotexture) to the friction factor as a function of sliding speed (reproduced from
Federal Aviation Administration 1971)
2.2.2 Factors Affecting Skid Resistance
The fiction between the rubber tire and the pavement surface is dependent on the two
materials in contact viz. the type of rubber used and the pavement surface. The type of
rubber used is important because it’s damping characteristics change with the type of
rubber and its chemical composition. The pavement surface, as determined by the
surface texture and the visco-elasticity of asphalt pavement, is also important in
determining the magnitude of both adhesion and hysteresis.
15
The adhesion component is dependent upon the following factors:
Interface lubrication: dry surfaces provide higher adhesion than wet surfaces. The
presence of lubrication tends to decrease the interface shear strength.
Sliding speed of rubber: adhesion increases with speed and reaches a maximum at a
“critical speed”. The critical speed ranges from 0.1 to 10 mph depending on the rubber
type and temperature. For speed above this “critical speed”, adhesion decreases.
Adhesion also decreases as the loading pressure increases. Higher loading pressure is
associated with an increase in the actual contact area; but that increase is not
proportional to the increase in the loading pressure.
The adhesion at a particular speed may increase, decrease, or remain unaffected by
temperature changes at the interface due to the visco-elastic nature of the rubber and
asphalt pavements.
On the other hand, the hysteresis component (Bazlamit, 1993) is dependent upon the
following factors:
Hysteresis will increase with the increase in the damping ability of rubber.
Unlike adhesion, hysteresis decreases as the temperature of the interface increases.
Hysteresis is virtually independent of the loading pressure and lubrication.
16
2.2.3 Roughness and Texture
Pavement texture is the feature of the road surface that ultimately determines most tire-
pavement interactions, including wet friction, noise, splash and spray, rolling
resistance, and tire wear (NCHRP Synthesis 291, 2000). Pavement texture has been
categorized into four ranges based on the wavelength of its components: microtexture,
macrotexture, megatexture, and roughness or evenness. At the 18th World Road
Congress, the Committee on Surface Characteristics of the World Road Association
(PIARC) proposed the definitions of the wavelength range for each of the categories as
shown in Figure 2-3 (PIARC 1987). The committee further proposed the range of the
texture wavelengths that are important for various tire-pavement interactions, which are
also shown in Figure 2-3. Wet pavement friction is primarily affected by the range
described by microtexture and macrotexture, as can be seen in a vast number of recent
studies, for example, by Davis (2001), Do and Marsac (2002), McDaniel and Coree
(2003), Luo (2003), Flintsch et al. (2003), Hanson and Prowell (2004), Kuttesch
(2004), Wilson and Dunn (2005), and Goodman et al. (2006). Because the range of
microtexture and macrotexture affects noise, splash and spray, and tire wear,
pavements designed with high friction values may have adverse effects on these
characteristics.
A detailed description of microtexture and macrotexture is presented herein. Tiny
grains of fine aggregate and features that make up the surface of coarse aggregate
provide what is known as the pavement microtexture. Thus, microtexture is a function
17
of aggregate gradation. In functional terms, microtexture is the most significant
contributor to low speed skid resistance and provides a gritty surface to penetrate thin
water films and produce good frictional resistance between the tire and the pavement.
Microtexture describes wavelength that ranges from 0.1mm to 0.5mm and it is
correlated to low speed friction. Features of the pavement surface that range from
approximately 0.5 mm to 50mm in length are classified as macrotexture. Macrotexture
was shown to be the primary determining factor of high speed wet skid resistance
(McGhee and Flintsch, 2003; Chelliah et al., 2003; NCHRP Synthesis 291, 2000;
Janoo, and Korhonen, 1999; Dewey, et al., 2001; Abe et al., 2002). Macrotexture is a
result of the large aggregate particles in the mixture and it is a function of aggregate
type. Macrotexture provides drainage channels for water expulsion between the tire and
the pavement thus allowing better frictional resistance and preventing hydroplaning.
Macrotexture can be estimated using volumetric or laser-based methods. Both the
microtexture and macrotexture of asphalt concrete pavements are influenced by the
properties of the coarse aggregates exposed at the wear surface since the coarse
aggregate in bituminous mixtures is more influential than other mix constituents in
determining skid resistance (Dewey et al., 2001; Crouch et al., 1996). A schematic
illustration of microtexture and macrotexture is shown in Figure 2-4.
Figure 2-5, on the other hand, presents the effect of microtexture and macrotexture
properties on skid resistance as a function of speed (Janoo and Korhonen, 1999).
Clearly, to maintain a constant high skid resistance value at various speed levels, the
pavement surface should have both good microtextures and macrotexture (Janoo and
18
Korhonen, 1999; NCHRP Synthesis 291, 2000). The change in the texture depends on
the aggregate resistance to fragmentation, wear, and polishing. Aggregate
fragmentation and wear depend on the toughness and hardness of the aggregate
minerals and the aggregate itself. Polishing depends on the difference in hardness of the
different minerals present in the aggregate (Janoo and Korhonen, 1999). Surface
texture can be defined in terms of microtexture and macrotexture; the terms used to
describe the texture of a road surface are shown in Figure 2-6.
Figure 2-3: Texture wavelength influence on surface characteristics (reproduced
from PIARC 1987)
19
Figure 2-4: Schematic representation of microtexture and macrotexture
Figure 2-5: Effect of microtexture and macrotexture on skid resistance as a function
of speed
20
Figure 2-6: Terms used to describe the texture of a road surface
2.2.4 Importance of Aggregate Characteristics to Surface Performance
Aggregates constitute more than 90% by weight of asphalt pavement materials as
shown in Figure 2-7. Strength and durability of aggregates often hold the primary
concern of the designer, especially in the bituminous construction (Smith and Fager,
1991; NCHRP Report 405, 1998). Consequently, aggregate plays a very significant
role in surface performance. The role of aggregate is to provide a macrotexture that will
induce tire hysteresis and facilitate water drainage in the tire-pavement contact area. It
is also to provide a microtexture that will maintain a level of friction.
21
Aggregate
Binder + Void
90%
10%
Figure 2-7: Weight phase diagram of Hot Mix Asphalt
Within the service life of an asphalt pavement, surface aggregates are subjected to
various types of stresses. These stresses could induce differential wear that may be
beneficial in restoring surface friction, or they could cause cracking, scaling, etc. In
order for the aggregates to withstand wear that causes differential changes in surface
macrotexture, the aggregate must be hard and tough and possess well-bonded grains so
that it will not be easily crushed or fractured under traffic loading stresses. The
aggregate must also be chemically stable. If the aggregate resists excessive wear but
undergoes slow differential wear, the macrotexture will be preserved and the
microtexture will be improved.
According to Gandhi et al. (1991), among the aggregate properties that affect friction,
polish values and acid solubility (carbonate content) were statistically significant.
However, the carbonate content gave a better correlation than polish values. Also,
when texture depth was included as an additional variable in the models with either
Binder+Void
22
solubility or polish values, the correlation coefficients increased slightly. American
Association of State Highway and Transportation Officials (AASHTO) guidelines
recommended the use of either the Acid Insoluble or Polish test for evaluating
aggregates. Thus, a requirement of a minimum polish value from 45 to 48 and
maximum carbonate content of aggregate from 10 to 25 percent will be in accordance
with the accepted national and international practice.
2.2.5 Aggregate Factors Affecting Pavement Friction
Excluding those asphalt pavements produced mainly from fine aggregates, the skid-
resistant properties of asphalt pavements depend primarily on the coarse aggregates.
According to a study (Beaton, 1976), four characteristics should be evaluated in the
selection of aggregates for skid-resistant asphalt pavements. These are: texture, shape,
size, and resistance to polish-wear action. Texture was discussed in previous section,
the other three characteristics (i.e., shape, size, and resistance to polish-wear action) are
explained below.
2.2.5.1 Aggregate Shape
Shape of an aggregate particle significantly affects its skid-resistant properties. Shape
of the aggregates also influences factors like hardness of grains, strength of the matrix,
and overall aggregate resistance to abrasion. Processing procedures also govern the
shape of both natural and synthetic aggregates. Angularity contributes to skid-resistant
qualities, but retention of angularity depends on characteristics like mineralogical
composition and amount of polish-wear produced by traffic.
23
2.2.5.2 Aggregate Size and Gradation
Aggregate size influences skid resistance qualities of the pavement. However, it must
be considered in relation to pavement type and mix design. Generally, larger-size
aggregates in asphalt pavement mixes have greater control over skid resistance than
smaller-size aggregates. As per Dahir (1979), open grading has been successfully used
to facilitate fast drainage of wet pavements in the tire-pavement contact area, by
reducing skid resistance-speed gradient.
2.2.5.3 Resistance to Polish-Wear Action
The ability of an aggregate to resist the polish-wear action of traffic has long been
recognized as the most important characteristic for use in pavement construction. When
an aggregate becomes smooth, it will have poor skid resistance. Also, if it polishes and
wears (abrades) too rapidly, the pavement will be slippery under wet conditions
(Hosking, 1976).
A study (Sherwood, 1970) showed that coarse grain sizes and differences in grain
hardness appear to combine to lead to differential wear and plucking out or shearing of
grains that result in a constantly renewed abrasive surface.
According to Shupe (1958), certain minerals are associated with good skid resistance
qualities. For example, the superior performance of dolomitic limestone over relatively
pure carbonate limestone.
24
2.2.6 Models for Wet Pavement Friction
Wet pavement friction is a measure of the force generated when a tire slides on a wet
pavement surface. Wet pavement friction is often referred to as “skid resistance”, and
the two terms are used interchangeably (NCHRP Synthesis 291, 2000). Wet pavement
friction decreases with increasing speed. This was first recognized by Moyer in 1934.
More specifically, skid resistance decreases as the velocity of the tire surface relative to
the pavement surface increases. This relative velocity is called the slip speed. There are
several models for determining pavement friction. A few of the most commonly used
models are described in this section.
2.2.6.1 The Penn State Model
The Penn State Model (Leu and Henry, 1983) describes the relationship of friction ()
to slip speed (S) by an exponential function:
SPNG
e 100
(2-1)
Where o is the intercept of the friction at zero speed, and PNG is the percent
normalized gradient (the speed gradient times 100 divided by the friction) defined by:
dS
dPNG
100 (2-2)
It was demonstrated that PNG is constant with speed and therefore Equation 2-1
follows by rearranging Equation 2-2 and integrating from S = 0 to S. Furthermore, it
25
was discovered that PNG is highly correlated with macrotexture and that o can be
predicted from microtexture.
Later version of the Penn State Model replaced the term (100/ PNG) by a speed
constant Sp:
pS
S
e
(2-3)
The PIARC Model (PIARC 1995) adopted the Penn State Model, but shifted the
intercept to 60 km/h:
pS
S
eFSF
60
60)( (2-4)
Where F(S) is the friction at slip speed S in km/h, and F60 is the friction at 60 km/h (36
mph).
Figure 2-8 shows the Penn State Model for two cases that have the same level of
friction at 60 km/h (36 mph), but behave very differently at other speeds, because of
differences in texture, resulting in different values for PNG and Sp. This example
demonstrates the need for specifying more than a single value, such as the friction at 60
km/h (36 mph), to describe the skid resistance of a pavement.
26
Figure 2-8: Penn State Model (NCHRP Synthesis 291, 2000)
2.2.6.2 The Rado Model
The Rado Model, on the other hand, depicts the entire braking maneuver using the
following equation (NCHRP Synthesis 291, 2000):
2)/ln(
)(
C
SS
peak
peak
eS (2-5)
Where peak= peak friction level,
Speak = slip speed at the peak (typically 15% of the vehicle speed), and
C = shape factor related to the harshness of the texture.
Figure 2-9 is a plot of Equation 2-5 with some typical values: peak = 0.6, Speak = 15
km/h (9 mph), and C = 0.5, with the forward speed of the test vehicle of 120 km/h (66
mph).
27
Figure 2-9: Rado Model
As a tire proceeds from the free rolling condition to the locked wheel condition under
braking, the friction increases from zero to a peak value and then decreases to the
locked wheel friction. Anti-lock brake systems release the brakes to attempt to operate
around the peak level of friction.
The Penn State and Rado Models together can be used to simulate the complete vehicle
braking in an emergency situation. The Rado Model is used at the beginning of the
braking manoeuvre until wheels are fully locked. If braking continues after the locked
wheel condition is reached, the vehicle speed (which then is equal to the slip speed)
decreases and the friction follows the Penn State Model until the vehicle stops (Luo
2003; NCHRP Synthesis 291, 2000).
28
2.2.6.3 The PIARC Model and the International Friction Index
The International PIARC Experiment to Compare and Harmonize Texture and Skid
Resistance Measurements (PIARC 1995) was conducted in Belgium and Spain in the
fall of 1992. Each friction tester was operated at three speeds: 30, 60, and 90 km/h (18,
36, and 54 mph), and each tester made two operated runs at each speed. All texture
measurements were made on dry surfaces before any water was applied to the roadway.
As a control, a microtexture measurement was made before and after the skid testers
made their tests. These data were used to show that there were no statistically
significant changes occurring during the testing.
The Rado Model at slip speeds above the peak and the Penn State Model are similar
and are dependant on the pavement characteristics. Because the Penn State Model is
less complex, it was chosen as the basis for the analysis of the data from the experiment
and the development of the International Friction Index (IFI). The IFI was developed as
a common scale for the reporting of pavement friction measurements. IFI is currently
being adopted worldwide as the standard skid resistance measure.
The IFI consists of two terms: (1) the speed constant, Sp, of wet pavement friction
which is a function of pavement macrotexture, and (2) the wet friction of a pavement,
F60, at 60 km/h, that depends on a measured friction value, the slip speed, and the
speed constant.
The wet pavement speed constant Sp in km/h is determined from macrotexture
measurement (Tx in mm) as follows:
29
TxbaS p (2-6)
where a and b are calibration constants that are specified in ASTM E 1960 for the
results of macrotexture testing in accordance with ASTM E 2157 and ASTM E 965.
The calibrated wet friction parameter (F60) can be estimated from the results of friction
testing. The relationship for calculating F60, per ASTM E 1960, is:
TxCeFRSBAF pS
S
60
60 (2-7)
where A, B, and C are calibration constants, and FRS is the measured friction at some
slip speed, S in km/h. The calibration constants are specified in ASTM E 1960 for the
results of friction testing in accordance with ASTM E 1911 and ASTM E 274.
The termpS
S
eFRS
60
in equation 2-7 is known as the adjusted value of friction from a
slip speed of S in km/h to 60 km/h for the equipment.
After the determination of F60 and Sp, the calibrated friction at any other slip speed,
F(S), can be calculated using equation 2-4.
The resulting parameters that make up the IFI (F60 and Sp) are sufficient to describe
the friction as a function of slip speed using equation 2-4. Note that a macrotexture
measurement is required to apply the IFI. The two parameters (F60 and Sp) distinguish
the difference between the two pavements shown in Figure 2-8.
30
Another advantage of the IFI is that the value of F60 for a pavement will be the same
regardless of the slip speed. That permits the test vehicle to operate at any safe speed.
Finally, the standard ASTM E1960 describes a procedure to calibrate devices that did
not participate in the experiment.
2.2.7 Frictional Needs of Traffic
The minimum skid-resistant pavement is generally required to satisfy the normal needs
of traffic without skid-related accidents. Normal needs of traffic encompass all the
driving, cornering, and braking manoeuvres by the majority of drivers under normal
traffic conditions. In providing skid resistance, the normal frictional needs of traffic
must be satisfied before steps can be taken to accommodate more severe demands.
Minimum frictional requirements of a pavement are those that satisfy the normal needs
of traffic. “Minimum” refers to the lowest acceptable friction level and specifically
implies that the level should be higher whenever possible. Minimum frictional
requirements are, therefore, defined if the normal needs of traffic can be described. As
outlined in NCHRP No. 37 three methods of determining the minimum frictional
requirements are as follows:
For any standard skid-resistance measurement method, a comparative study can be
made between the skid resistance requirements of different pavement sections. For
example, the skid resistance rate observed on a large sample of pavement surfaces,
representing the entire design speed range from 48 to 129 km/h (30 to 80 mph) can be
compared with the skid-resistance measured on other surfaces under clearly defined
31
pavement conditions. This method determines the friction level, which separates
pavements susceptible to skidding and skid-resistant pavement surfaces.
Driver behaviour pattern of a large driver population during acceleration, driving,
cornering, and deceleration can be investigated by concealed recorders carried on board
or located near the site being surveyed. This method yields an acceleration spectrum,
which defines normal, intermediate, and emergency needs according to magnitude and
their frequency of occurrence.
The frictional needs can be deduced from vehicle design and highway geometry, or the
superposition of the two, whenever the limiting needs are determined by these factors
and not by the driver, as for instance by the full-throttle acceleration of a particular type
of vehicle. The frictional needs for this manoeuvre are solely dictated by vehicle factors
such as weight-to-horsepower ratio, transmission ratios, and center-of-gravity location.
2.2.8 Factors Affecting Wet-Pavement Safety
Skid Number (SN) alone is not a good measure of wet pavement safety. Many other
factors affect safety under wet-pavement conditions, and it is only when these
conditions demand a particular level of traction that SN becomes important. Some of
these factors according to Wambold and Kulakowski (1991) are listed below.
Vehicle Speed: Friction demand increases with speed. The centrifugal forces generated
during the vehicle cornering, which have to be counteracted by tire-pavement friction
forces to prevent the vehicle from skidding off the road are proportional to the square
32
of vehicle speed. Also, pavement resistance decreases with increasing speed in an
approximately exponential manner.
Road Geometry: Friction demand on straight sections of roads is low, if road is level,
vehicles travel at low speeds, and if there are no intersections. The demand for friction
increases significantly if a grade or a curve is to be negotiated. Page and Butas (1986)
concluded that wet-pavement accident rates are significantly higher on curves than any
other type of geometric alignment. The effect of curvature on wet-pavement accident
rates was found to be particularly significant on pavements with SN values less than
25. Furthermore, for SN values less than 25, wet-pavement accident rates were
significantly greater for both uphill and downhill slopes steeper than 3 percent than for
flatter terrain.
Traffic Flow: Traffic volume does not have a significant influence on wet-pavement
accident rates. However, under special circumstances, like on undivided highways with
SN values less than 25, wet-pavement accident rates increase significantly when
average daily traffic exceeded 15,000. Traffic composition, particularly the percentage
of trucks has a significant effect on friction demand, since the stopping distances of
trucks are 1.3 to 2.8 times longer than those of passenger cars.
Vehicle Type: If equal stopping distance is required for all vehicles, then the friction
demand for buses and trucks is higher than that for passenger cars. The friction demand
is also higher for vehicles with lower degrees of understeer.
33
Driver Skills: Few drivers can operate their vehicles with 100 percent efficiency, i.e.,
using 100 percent of the available friction. Olson et al. (1984) found that truck driver
efficiencies ranged from 62 to 100 percent, but most of the drivers had little or no
practice in emergency braking situations. The concern over emergency braking skills
will be considerably alleviated when antilock brake systems (ABS) become a more
common feature.
2.2.9 Skid Resistance Requirements and Practices by Different Agencies
For the sake of uniformity, skid number at 40 mph using the locked wheel skid
trailer is normally used to compare between the minimum requirements set by
different state agencies. The Florida Department of Transportation (FDOT) Safety
Improvement Program Manual calls for desirable skid number values of 35 and
greater for facilities with posted speed limits greater than 45 mph. On roadways with
a posted speed limit less than or equal to 45 mph, the desirable skid number value is
30 or greater. In addition, the FDOT Friction Testing and Action Program calls for
skid number values of 35 and above, and pavements having mean skid number
values below 35 must be re-tested in one year. These friction requirements are
generally consistent with other state transportation departments (Jackson, 2003).
Oklahoma Department of Transportation (OKDOT) requires a minimum field skid
number of 35 which confirms the previous conclusion, while New York Department
of Transportation (NYDOT) uses a design target of minimum skid number of 32 at a
speed of 40 mph using ribbed tire. Indiana Department of Transportation (INDOT),
34
on the other hand, established a uniform minimum friction requirement of 20 for the
standard smooth tire at 40 mph. It was indicated that by the seven-year friction
measurements, this friction requirement can guarantee a reasonable and consistent
friction performance for INDOT network pavement (Li et al. 2004). One more
important thing to mention is that many state highway agencies have established
their minimum friction requirements based on the recommendation of the minimum
friction requirement by NCHRP-37 (Kummer and Meyer, 1967). Using standard
ribbed tire, NCHRP-37 recommended a minimum friction number of 37.
Texas Department of Transportation has done extensive research and summarized the
guidelines that different State Departments of Transportations follow for testing and
acceptance of aggregates for adequate provision of skid-resistant pavements
(Jayawickrama, et. al, 1998). The same has been reproduced and can be found in Liang
(2003).
2.2.10 Air Void and Temperature Effect on Frictional Properties of Asphalt Pavement
Surfaces
In recent years, a few notable research efforts have directed toward a better
understanding of the influencing factors on the HMA surface friction properties. In
2006, Goodman et al. reported that initial field BPN can be correlated with the
following variables: fineness modulus (FM), voids in mineral aggregate (VMA),
percent passing the 4.75mm sieve (P4.75) and bulk relative density (BRD).
Interestingly, the bulk relative density was found to significantly affect the measured
35
BPN values. Specifically, an increased BPN was observed through the use of more
densely graded aggregates (reduced FM), less void space between aggregate particles
(less VMA), use of finer gradation (more P4.75), and higher compactive effort
(increased BRD).
There have been some recent research efforts toward quantifying the temperature
effects on the measured pavement friction values. For example, Runkle and Mahone
(1980), Burchett and Rizenbergs (1980), and Bazlamit and Reza (2005) have found that
an increase in temperature can result in a corresponding decrease in skid resistance of a
pavement surface.
A very noteworthy study was conducted in connection with the Virginia Smart Road to
investigate the friction properties as affected by seasonal temperature differences.
Wang and Flintsch (2007) studied the surface friction and texture properties of 12
asphalt pavement sections placed at the Virginia Smart Road pavement facility over a
6-year time period. Both short-term (seasonal) and long-term (multi-year) variations of
the surface characteristics were investigated in terms of temperature and time effects.
Their investigation showed that pavement skid resistance decreases in summer (high
temperatures) and thus confirming that temperature can exert significant effects on the
seasonal and multi-year variations of pavement surface friction. Other studies related to
Virginia Smart road include Flintsch et al. (2005) and Luo (2003). In both reports, the
effect of pavement temperature on frictional properties of HMA pavement surfaces at
36
the seven HMA surfaces was confirmed. Their analysis showed that pavement friction
tends to decrease with an increase in pavement temperature.
An Indiana study conducted by Elkin et al. (1980) also confirmed that there was a
noticeable loss of skid resistance as pavement surface temperature increases beyond 32
°C ( F90 ) and especially above 38 °C ( F100 ). Hill and Henry (1978) conducted a
study on twenty one test surfaces in State College, Pennsylvania. The testing program
included daily skid measurements according to ASTM E 274 and the collection of daily
weather data. Pavement temperature was chosen as the temperature parameter. It was
found that an increase in pavement temperature of C10 (50 °F) can result in a decrease
in SN at 64 kmph (40 mph) of about 1.2 skid numbers. This decrease, however, is
outweighed by measurement error, particularly lateral placement of the test tire, which
accounts for as much as 4 skid numbers at a speed of 64 kmph (40 mph).
Despite the significant number of reports cited above to indicate the significant effects
of temperature on pavement surface friction values, there are contradictory findings
reported by Dahir et al. (1979) and Mitchell et al. (1986) as well. Dahir et al. (1979)
conducted a study at the Pennsylvania State University to investigate the short-term
(seasonal) variations of skid resistance. In this study, skid resistance measurements
according to ASTM E 274 were made on dry pavements. In addition, tire and pavement
temperatures were continuously monitored by using radiometers mounted on the tester.
Ambient and water temperatures were measured using appropriate thermometers. It
was concluded that temperature variations of the magnitude experienced in the study do
37
not seem to significantly affect the skid resistance measurements. Mitchell et al. (1986)
conducted a study on the pavement surfaces incorporated into Maryland’s highway
system. The primary objective of Mitchell’s study was to determine the parameters that
could be used to predict the influence of seasonal variations. In addition to friction
numbers measured according to ASTM E274, data relating to weather conditions and
pavement temperature were also recorded. Mitchell and his co-workers found that the
effect of pavement temperature on skid resistance appears to be of no significance
whatsoever.
2.2.11 Overview of Polishing, Friction, and Texture Measurements
Many equipment types and methods have been used over the years to achieve
accelerated polishing on aggregate and HMA specimens and to quantify friction and
texture properties (Wallman and Astrom, 2001). A review of these equipment and
methods is summarized in the following sections.
2.2.11.1 Existing Accelerated Polishing Machines
A review of the existing laboratory-scale accelerated polishing devices reveals that they
can be categorized into three groups: one is capable of polishing the aggregate samples,
the other one is capable of polishing the HMA samples, and the third is capable of
polishing both (aggregate and HMA specimens). A brief review of the existing devices
in each category follows.
38
2.2.11.1.1 Polishing Devices for Aggregates:
Within this category, there are three existing devices: British Polishing Wheel,
Michigan Indoor Wear Track, and Micro-Deval device.
2.2.11.1.1.1 British Polishing Wheel
Most polishing machines on aggregates specimens work on the principle of reducing
the microtexture of the aggregate. For example, the ASTM D3319 British Polishing
Wheel method allows the curved specimens (aggregate coupons) clamped around the
periphery of the wheel assembly to form a continuous strip of aggregate particles. The
wheel is then rotated against a rubber-tire wheel that provides the polishing action.
Silicon carbide grit No. 150, with a feeding rate of 6 ± 2 g/min along with distilled
water at a rate of 50 - 75 ml/min, is used to help accelerate the polishing. The aggregate
specimens are formed by mounting uniformly-sized coarse aggregate particles by hand
in a curved mold and holding them in place with a bonding agent (polyester or epoxy
resin). A catalyst could be used for faster curing of the resin. The companion British
Pendulum Tester (BPT) specified in ASTM E303-93 is used to measure specimen
friction values. The British polishing wheel is used for polishing microtexture of
aggregate coupons only; however, it does not have the ability to alter macrotexture of
aggregates or to test HMA specimens. In addition and as described above, the
procedure used to prepare the aggregate coupons for polishing test is tedious and time
consuming.
39
2.2.11.1.1.2 Michigan Indoor Wear Track
The Michigan Department of Transportation (MDOT) wear track device uses the full-
scale smooth tires to polish coarse aggregate specimens. After polishing, the specimens
are subsequently tested by a laboratory version of the ASTM towed friction tester.
According to Dewey et al. (2001), the circular wear track is very large, with a diameter
of 7 ft. It accommodates 16 trapezoidal specimens. The individual specimens have
parallel sides of 15.5 and 19.5 inch and non-parallel sides of 11 inch. Two wheels, with
normal forces of 800 lb, pivot around the center. This device is used for polishing
coarse aggregates only. It is by far the largest polishing device found. As can be
imagined, sample preparation procedure is not only cumbersome but also time-
consuming.
2.2.11.1.1.3 Micro-Deval Device
The Texas Transportation Institute (Luce et al., 2007) uses the Micro-Deval device as
the mechanism to polish aggregates. The results showed that the Micro-Deval test is an
effective method for polishing aggregates within a short time (180 minutes). The
Micro-Deval device can only polish aggregates and not HMA specimens.
2.2.11.1.2 Polishing Devices for HMA:
Within this category there is one device that is the National Center for Asphalt
Technology.
40
2.2.11.1.2.1 NCAT Polishing Machine
The National Center for Asphalt Technology (NCAT) laboratory-scale accelerated
polishing device is designed to polish the HMA surface. The NCAT (Voller and
Hanson, 2006) follows the same polishing principle as a Circular Track Polishing
Machine. The NCAT machine could polish the area sufficiently large to accommodate
the required measurements with the Dynamic Friction Tester (DFT) and Circular
Texture Meter (CTM) to measure friction and texture, respectively. The NCAT
polishing equipment uses three pneumatic tires made of resin or hard rubber, 8 inch in
diameter, to polish an annulus that occupies a nominal 24 inch square slab. With rubber
tires, water is used to wash the abraded rubber particles off the specimen surface during
polishing. Dead weights are used to produce a total vertical force of 150 lb through the
three wheels. Up to 100,000 revolutions at 40 rpm have been successfully applied to
reach the terminal friction values of the HMA surface. NCAT uses a modified linear
compactor to produce the slabs (24 inch square area) for polishing test. A somewhat
prolonged test time, up to 41.7 hours, has been recorded by NCAT in order to reach the
terminal friction values.
2.2.11.1.3 Polishing Devices for Aggregates and HMA:
Three devices exist within this category: NCSU Wear and Polishing Machine,
Wehner/Schulze Polishing Machine, and Penn State Reciprocating Polishing Machine.
41
2.2.11.1.3.1 North Carolina State University Wear and Polishing Machine
Circular Track Polishing Machines represent yet another type of polishing concept.
Some of these polishing machines can be used for polishing either aggregate specimens
or HMA specimens. The North Carolina State University (NCSU) Wear and Polishing
Machine, as specified in ASTM E660, utilizes four individually mounted, free rolling
wheel assemblies that pivot about a central shaft. The four wheels are loaded to 72 lb in
vertical force. The tires are 11 inch in diameter and made of smooth nylon. Twelve
specimens (aggregate or HMA mixes) are arranged around the perimeter of the track
for polishing. The overall diameter of the track, to the center of the polishing wheels, is
36 inch. After 8 hours of polishing action, the surface friction of each specimen is
measured using either the British Pendulum Tester (BPT) or the Variable Speed
Friction Tester (VST). The test does not use slurry or water. Although the device is
fairly large, it nevertheless polishes only a relatively small area of the specimen
surface.
2.2.11.1.3.2 Wehner/Schulze Polishing Machine
The Wehner/Schulze polishing machine was developed in Germany 30 years ago (Do
et al., 2007). It is comprised of two heads to facilitate polishing and friction
measurement, respectively. Specimens are cores with the diameter of 8.9 inch. They
can be taken from asphalt pavement or laboratory-prepared slabs (aggregate or asphalt
specimens). The polishing action is achieved by means of three rubber cones mounted
on a rotary disc, which rolls on the specimen surface. The rotation frequency is 500
rpm, giving a linear speed of (17 km/h) 10.6 mph. The contact pressure between the
42
cones and the specimen surface is 58.0 psi. The slip between the cone and the specimen
surface is between 0.5% and 1%, which is roughly the slip between rolling tires and
roads. A mix of water with quartz powder is sprayed on the specimen surface during
the polishing action. The surface is polished on a ring of roughly 6.3 inch in diameter
and 2.4 inch in width. At each stop, water is sprayed on the specimen surface and 500
rotations are performed using the cones to wash all debris. This machine is not
designed to handle typical specimen size compacted from the gyratory compactor.
2.2.11.1.3.3 Penn State Reciprocating Polishing Machine
The Penn State Reciprocating Polishing Machine (Nitta et al., 1990), ASTM E 1393,
represents a different style of polishing concept. It can be used in a laboratory or in the
field to polish aggregates or HMA. In essence, a 3.5 by 3.5 inch rubber pad is oscillated
back and forth on the specimen surface on which abrasive slurry is sprayed as well.
Some of the critiques about this device include the relatively small polishing area (4.5
inch by 6.5 inch), the polishing action can only affect the aggregate macrotexture, and
reciprocal movement.
2.2.11.2 Friction Measurement Methods
When a tire is braked from free rolling situation to locked wheel the friction force
experienced by the wheel hub changes depending on the slip (the ratio between slip
speed and operating speed). This is illustrated in the typical friction-slip curve shown in
Figure 2-10. The maximum friction is normally found at a slip rate of about 7-20% and
can be considerably higher than the locked wheel friction (100% slip).
43
Figure 2-10: Friction-slip curve of a braking tire (reproduced from Federal Aviation
Administration 1971)
Pavement surface friction can be measured using one out of four different principles:
locked wheel (100% slip), side force, fixed slip (normally between 10 and 20% slip),
and variable slip (0 to 100% slip). In addition, some of the systems detect the peak
friction and some vary the slip in an attempt to operate around the peak friction level.
Each method of measuring friction has advantages. Direct use of the values produced
by any one type of measurement relates to a different scenario. The locked wheel
method simulates emergency braking without anti-lock brakes, the side force method
measures the ability to maintain control in curves, and the fixed slip and variable slip
methods relate to braking with anti-lock brakes. These friction-measuring methods and
the respective devices are summarized and discussed herein.
44
2.2.11.2.1 Locked Wheel Friction Devices
The pavement surface skid resistance is most often measured by the skid trailer (Luo,
2003). The skid trailer consists of a truck containing a large water tank for wet testing
and a trailer with a locking mechanism on one wheel (see Figure 2-11). The locked
wheel trailer can be used to test the frictional properties of the pavement surface at any
speed up to 96.6 km/h (60 mph). The standard test procedure for the locked wheel skid
trailer is described in detail in the ASTM E274 specification. The test begins with the
attainment of the desired test speed, usually at 64 km/h (40 mph). An activator, located
inside the truck, is used to initiate the test sequence beginning with the application of a
thin layer of water to the pavement surface. Usually, a nominal water film of 0.5 mm is
used. After the correct amount of water has been applied, the test wheel is locked and
instrumentation on the trailer records the sliding force of the locked tire. This test
allows for the computation of the skid number by dividing the tractive force applied to
the tire to the vertical load applied to the tire.
Measuring skid resistance of wet pavement by the standard locked wheel skid trailer
device involves a complex tire-pavement interaction, affected by the variables such as
the tire used, pavement surface texture, age of pavement in service, and temperature of
the contacting surfaces (Davis, 2001). Among the testing variables of a locked wheel
device, one critical decision would be whether to use the ribbed (ASTM E501) or
smooth tires (ASTM E524). The smooth tire is sensitive to both the microtexture and
macrotexture of the pavement. On the other hand, the grooves in the ribbed tire provide
channels much larger than the pavement macrotexture for water flow. Consequently,
45
the friction measurement by the ribbed tire is insensitive to the macrotexture, but
sensitive to microtexture (Li et al., 2003). Figure 2-12 shows pictures of ribbed tire and
smooth tire.
Figure 2-11: Locked Wheel Skid Trailer
Figure 2-12: Ribbed tire versus smooth tire
46
2.2.11.2.2 Side Force Coefficient Devices
The side force coefficient is measured by a test that uses a freely rolling wheel to
determine the frictional properties of the pavement. This type of test uses a wheel that
is mounted at an angle to the direction of motion of the test vehicle. The force that is
produced on the sideways mounted wheel is used to calculate the friction coefficient of
the pavement surface. This method has been used for many years and can be performed
using a motorcycle and a sidecar (Alsopp, 1985).
Two examples of equipment that utilize the side force coefficient methodology to
measure surface friction are the Sideways Force Coefficient Routine Investigation
Machine (SCRIM) and the MuMeter. Both of these pieces of equipment were
developed in Britain (NCHRP Synthesis 291, 2000). The SCRIM (Figure 2-13) was
originally developed by Transport and Road Research Laboratory (TRRL) in United
Kingdom about 1953 and is a rebuilt truck with a measuring wheel placed between the
front and the rear axle. The measuring wheel is a special motorcycle wheel mounted
with a constant side slip angle of 20 degrees. During measurement the wheel is rotating
freely and the road surface friction is evaluated as the lateral force acting on the free
rolling wheel divided by the load on the wheel, the results is called the Sideway-Force
Coefficient (SFC). SCRIM uses the sideway-force method of measuring resistance to
skidding because it is more suitable for routine measurement (Hosking and Woodford,
1976b) than the locked wheel method. Tests are normally carried out at 50 km/h (31
mph).
47
Figure 2-13: Sideways Force Coefficient Routine Investigation Machine (SCRIM)
The MuMeter (shown in Figure 2-14) is used for airports and has been used since the
1970’s by the Department of Transportation in Arizona. The MuMeter is a lightweight
three-wheeled trailer that does not require a special towing vehicle. Two of the smooth-
treaded wheels on the trailer are used to measure the friction of the surface while the
third wheel is a stabilizing wheel. The two test tires are at an angle of 15 degrees to
allow for the sideways force coefficient to be measured. Attached to the trailer is a unit
that records the friction values as the test occurs. The MuMeter measures the dry
friction of the pavement surface, as well as the wet pavement friction if a wetting
system were included in the towing vehicle. The side force coefficient that is obtained
from this type of equipment is calculated by dividing the sideway force to the vertical
reaction between the test wheel and the road surface (Alsopp, 1985).
48
Figure 2-14: Side force tester: the MuMeter (Tomita 1964)
2.2.11.2.3 Fixed Slip Devices
Fixed slip friction measurement devices are primarily used in European countries.
Fixed slip testers operate with a constant rate of slip, typically around 10% to 20%.
This allows for the maximum friction value of the roadway surface to be measured.
The amount of slip is controlled by hydraulics or by allowing the chain drive of the
tester to be lower than that of the testing vehicle. Examples of fixed slip devices
include the Runway Friction Tester, the Griptester, the Saab Friction Tester, and the
Portable Friction tester.
The Runway Friction Tester (RFT) is an accurate and repeatable, self contained
continuous friction measuring equipment that provides continuous self-wetted
49
coefficients of friction on airport runways. The RFT, shown in Figure 2-15, is designed
for both maintenance and operational testing to evaluate surface friction changes due to
weather, high usage, aging, and contaminants. The RFT uses a two axis force
transducer mounted on a retractable fifth wheel assembly mounted under the rear truck
bed. The test assembly provides real time vertical load and horizontal tractive force
measurement. System electronics include a laptop computer and ink-jet printer. User-
friendly WindowsXP software allows the operator to control the entire test procedure
including test speed, self wetting or dry testing, test method, test type (manual or
automatic), and annotate airfield test conditions for later reference. All data is stored for
further processing in user-determined formats. Data stored includes the raw load and
traction data, speed, water flow, optionally GPS coordinates and texture data. Friction
numbers can be printed at any desired intervals. Friction data and speed are visible real
time during the test on both the laptop computer and the dash mounted information
display. The system includes 1000 litre (250 U.S. gallon) built in water tank, water
pump and laminar flow water nozzle for self wetting testing of up to 11,000 m (36,000
ft) of runway without refilling. The RFT uses the continuous peak friction test method.
First, the operator selects whether the test is to be dry or self-wetted. Next, the operator
should bring the vehicle up to a test speed of 65 km/h (40 mph) or 95 km/h (60 mph)
prior to reaching the test site. At 65 km/h (40 mph), the friction survey recording
should begin at 152 m (500 ft) from the threshold end of the runway, and terminated
approximately 152 m (500 ft) from the opposite end of the runway. At 95 km/h (60
mph), the friction survey recording should begin at 305 m (1000 ft) from the threshold
50
end of the runway, and terminated approximately 305 m (1000 ft) from the opposite
end of the runway. At the same distances for the respective speeds, the fifth wheel
should be lowered onto the runway pavement surface. While conducting the friction
survey, the vehicle must be held at a constant speed. Peak friction is continuously
calculated by the system’s on-board computer for each test run. Friction measurements
are displayed both graphically and numerically on the laptop’s 15-inch screen and/or
printer and stored on hard disk drive for later transfer to a memory stick.
The Griptester is an excellent small three-wheeled device with the "normal" axle being
connected to the recording wheel by a pair of gears that causes a braking effect on the
axle of the third wheel that can be measured as a "Grip Number". This equipment
(shown in Figure 2-16) can be pushed by hand on a pre-wetted road surface for small
area surveys. Test speeds can vary from 5 km/h to 130 km/h (3 mph to 81 mph)
depending upon application. The measured values can be affected by the test speed.
One can obtain more information on the use of the Griptester from (Lund, 1997 and
Wambold et al., 1995).
The Saab Friction Tester, used in Sweden, operates at a constant slip rate of 17% (see
Figure 2-17). A retractable test wheel subjected to a constant vertical load is located
behind the rear axle of the test car. A chain driven transmission, which allows for the
constant slip rate, is connected to the rear axle and the test wheel. The Saab Friction
Tester measures the wet friction of the pavement surface with the aid of a water pump
attached to the test vehicle. Computers inside the vehicle record the forces acting on the
51
wheel as a result of the slip. The frictional properties of the pavement are reported as
the Brake Force Coefficient (BFC) and are calculated by dividing the retarding or
braking force to the vertical reaction between the tire and the road (Alsopp, 1985).
The Portable Friction Tester (PFT) is a portable manually driven instrument for
measuring the friction on small surfaces (down to 1 m in length) and has been
developed at the Swedish National Road and Transport Research Institute (VTI). It
uses a fixed slip, usually between 17 and 21%. The PFT consists of a three-wheeled
pushcart with the measuring wheel mounted in front of the others (see Figure 2-18).
The friction between the measuring wheel and the surface of interest is presented by the
PFT as the friction coefficient, which is the frictional force on the measuring wheel
divided by the normal load on the same wheel. This friction coefficient is henceforth
referred to as the PFT friction value. Operating the PFT can sometimes be difficult,
especially when measuring on slopes and curves. The PFT is equipped with a
speedometer indicating the appropriate speed, as an aid to the operator. Before the
measurements are started, the cart should be pushed until a certain speed is reached that
should then be kept constant during the entire measurement. Pushing a button on the
handle of the cart starts the measurement; when the section has been measured, the
process is stopped by again pushing the button. The PFT has been compared with the
British pendulum tester (Lundkvist and Linden, 1994; Centrell, 1995; Astrom, 2000)
with the result that it is possible to convert friction values measured by the PFT into
British pendulum numbers, at least in the measurement conditions considered.
52
Furthermore, the repeatability of the PFT measurement has proved to be good and
measurement time is less than with the pendulum.
Figure 2-15: The Runway Friction Tester
(a) Underside of Griptester (b) Griptester attached to towing vehicle
Figure 2-16: The Griptester device
53
Figure 2-17: Saab Friction Tester
Figure 2-18: The portable friction tester for measuring the friction on small surfaces
2.2.11.2.4 Variable Slip Devices
Variable slip friction measurement devices measure the friction of the pavement
surface in a manner similar to the fixed slip devices. During testing, the slip rate of the
test wheel is varied to allow for a range of friction values to be recorded. Japan and
Norway are the primary users of variable slip friction devices to measure the condition
of pavement surfaces.
54
Some examples of variable slip friction measurement devices include the Norsemeter
ROAR used in Norway and the IMAG system used in Japan (NCHRP Synthesis 291,
2000). In Norway Norsemeter has developed a flexible friction measurement unit
called ROAR (Schmidt, 1999). In Figure 2-19, it can be seen in the form of a friction
measurement trailer, including the water supply for wet friction measurements. The
measuring wheel has a smooth tire with an outer diameter of 410 mm (16 inch), ASTM
E1551. It is placed on a separate unit, which include all mechanical parts necessary for
the measurement so that this small unit can act alone and for example measure dry
friction placed directly on a road maintenance truck. During one measurement cycle
(about 1 second), ROAR measures the complete friction-slip curve, from pure rolling to
locked wheel. The device can operate at speeds between 20 and 130 km/h (12.5 and 81
mph).
Figure 2-19: Norsemeter road friction measurement trailer
55
2.2.11.2.5 Other Friction Measurement Methods
In addition to what precedes, there are more available equipment and methods used for
friction measurements.
2.2.11.2.5.1 Dynamic Friction Tester
The measurement of surface friction can also be done by the dynamic friction tester
shown in Figure 2-20. Per ASTM E1911, the test equipment consists of a disk fitted
with three spring-loaded rubber sliders at a diameter of 284 mm (11.2 inch). The disk is
initially suspended above the pavement surface and is driven by a motor until the
desired tangential speed of the sliders is attained. Water then flows over the surface
being tested, so that wet friction is measured similar to the operation of the skid trailer.
The rotating disk is then dropped onto the wet surface with an applied vertical force of
3.6 kg (8 Ib) while the friction is continuously measured as the disk slows down to zero
speed. The friction force and the speed during the spin down are recorded in a file. The
DFT can be used to measure the friction as a function of speed over the range of zero to
90 km/h (0 to 55 mph). The DFT system can be used not only in the field but also on
the laboratory prepared HMA specimen that is at least 450 by 450 mm (17.75 by 17.75
inches) in surface area.
56
(a) Bottom view (b) General view (c) Controller
Figure 2-20: Dynamic Friction Tester: (a) bottom view, (b) general view, and (c)
controller
2.2.11.2.5.2 Pendulum Devices
The British Pendulum Tester per ASTM E303-93 is used widely to measure friction. It
can be used on the curved coupons from the polishing wheel, on flat specimens from a
circular polishing track or reciprocating polisher, or on actual roadway surfaces. The
British Pendulum tester consists of a rubber slider attached to the end of a pendulum
arm as shown in Figure 2-21. As the pendulum swings, it is propelled over the surface
of the specimen. As the rubber slider contacts the surface of the specimen, the kinetic
energy of the pendulum decreases due to friction. This energy loss is measured and
reported as the British pendulum number (BPN) on flat surfaces or the polished stone
value (PSV) for curved aggregate coupons from the polishing wheel. The slider travels
at roughly 10 km/h (6 mph), so is only capable of measuring low-speed friction. Due to
the small size of the rubber slider and its low speed, however, it is widely recognized as
a measure of the microtexture only. The benefits of higher macrotexture cannot be
evaluated. The test procedure is as follows. Typically, the specimen is first immersed
57
in water and then the test surface is cleaned. Next, the BPT device is levelled and
adjusted so that the contact path during the swing of the slider is within the range of
6.1125 mm ( 063.092.4 inch). The rubber slider is cleaned and wetted. Five
swings are made for each specimen, from which an average of the last four readings is
recorded as the BPN.
The North Carolina State University Variable Speed Friction Tester is another
pendulum-type tester; see Figure 2-22. Instead of a rubber slider there is a locked-
wheel smooth rubber tire at the end of the pendulum. A stream of water is sprayed at a
specific velocity in the path of contact of the wheel. By adjusting the velocity of the
water stream, different vehicle speeds can be simulated in the laboratory or in the field.
However, uneven pavement surfaces in the field may provide inaccurate
measurements.
The advantages of this device are that it is portable, has a low initial cost, and can test
in different orientations. Its disadvantages include that results from coarse macrotexture
are questionable, it can only simulate low-speed skidding, and it requires laborious
calibration.
58
Figure 2-21: British Pendulum Tester
Figure 2-22: North Carolina State University Variable Speed Friction Tester
2.2.11.2.5.3 Michigan Laboratory Friction Tester
As a companion to the wear track, the Michigan Department of Transportation
(MDOT) uses a laboratory scale version of a towed friction trailer to measure the
frictional resistance of specimens polished on the wear track. The device, shown in
Figure 2-23, consists of a tire in a stationary frame that is rotated at the equivalent of 64
km/h (40 mph) then is dropped onto a specimen clamped onto the frame under a spray
59
of water. The torque produced as the tire slows due to friction with the specimen is
measured, and the peak torque is used as an indicator of the surface friction of the
specimen (McDaniel and Coree, 2003). A major problem with this method for the
purposes of this study is that it is used on uniformly sized coarse aggregate only. It
provides no assessment of macrotexture effects, and there is no history of using it for
pavement specimens instead of aggregates only.
Figure 2-23: Michigan Laboratory Friction Tester
2.2.11.2.5.4 PTI Friction Tester
Along with the Penn State Reciprocating Polishing Machine, the Pennsylvania
Transportation Institute (PTI) also developed a companion friction tester. The PTI
friction tester used a freefalling weight to propel a rubber slider in a linear path along a
surface. The frictional resistance is determined from the speed of the slider across the
surface. A surface with higher friction would slow the slider more than a smooth
surface.
60
2.2.11.2.5.5 Stopping Distance Method
Stopping distance methods is a field technique in characterizing pavement surface skid
resistance (ASTM E445/E445M). In this method, a four-wheel passenger vehicle is
used in which all four wheels are equipped with a braking system. The pavement in the
test lane is wetted. The test vehicle is brought above the desired testing speed and is
permitted to coast onto the wetted section until the proper speed is attained. The brakes
are then promptly and forcefully applied to cause a quick lockup of the wheels and to
skid to a stop. The distance required to stop is recorded with the aid of instrumentation
and reported as the stopping distance (SD). Using the recorded stopping distance and
the velocity of the vehicle upon application of the brakes, the stopping distance number
(SDN) can be calculated:
100255
2
SD
VSDN (2-8)
Where:
SDN = Stopping Distance Number,
V = Speed of the vehicle at the moment of brake application in km/h, and
SD = Stopping distance in meters.
The SDN can be used to evaluate pavement friction but does not report a coefficient of
friction. It is helpful in determining the relative adequacy of friction of different
61
pavement surfaces but does not correlate to other skid resistance measurements (ASTM
E445).
2.2.11.2.5.6 Wehner/Schulze Friction Device
As mentioned earlier, this machine comprises two heads for the polishing and the
friction measurement. After the washing period, the specimen is moved manually to the
friction-measuring head. This head is composed of three small rubber pads (0.6 in2 area
for each pad) disposed at 120 on a rotary disc. The contact pressure between the
rubber pads and the specimen surface is approximately 29 psi. For the friction
measurement, the disc is launched until a speed of 100 km/h (62 mph) at its
circumference is reached. When the speed reaches 90 km/h (55.9 mph), water is
projected on the specimen surface. At 100 km/h (62 mph), the motor is stopped and the
disc is dropped until the rubber pads touch the specimen surface. The rotation is
stopped by friction between the rubber pads and the specimen surface and then the
friction–time curve is recorded.
2.2.11.3 Texture Measurement Methods
The levels of pavement texture that affect friction are microtexture and macrotexture. If
both microtexture and macrotexture are maintained at high levels, they can provide
resistance to skidding on wet pavements (NCHRP Synthesis 291, 2000). A study
conducted in Europe (Roe et al., 1998) reported that increased macrotexture reduces
total accidents, under both wet and dry conditions. There are a variety of ways to
measure pavement texture, ranging from simple, indirect estimates to extremely high-
62
tech direct measurements. Technological advances make direct measurement more
feasible now than in the recent past. Literature pertaining to these methods was
collected and is presented herein.
2.2.11.3.1 Microtexture Measurement
Currently there is no system capable of measuring microtexture profiles at highway
speeds. A profile of the microtexture of an in-service pavement surface also could be
misleading (NCHRP Synthesis 291, 2000). The portions of the pavement surface that
contact the tires are polished by traffic, and it is the microtexture of the surface of the
exposed aggregate that comes into contact with the tire that influences the friction. The
valleys are not subjected to polishing and their contribution to the overall microtexture
should not be included in prediction of friction.
Because of the difficulty in measuring microtexture profiles, a surrogate for
microtexture is generally preferred. In research at the Pennsylvania State University
(Henry and Leu, 1978), it was found that the British Pendulum Numbers (BPNs) were
highly correlated with the parameter o of the Penn State Model (Equation 2-1). The
parameter, o, is the zero speed intercept of the friction-speed curve and characterizes
the friction at low slip speeds. The slider of the BPT engages only the portion of the
asperities that are subject to polishing by traffic and therefore the BPN values could be
considered as the surrogate for microtexture.
The values of the friction measured by the dynamic friction tester when the slip speed
is 20 km/h (12.5 mph) are highly correlated with BPN values (NCHRP Synthesis 291,
63
2000). This indicates that DFT friction values at 20 km/h (12.5 mph) could also be used
as an indirect measurement of the microtexture or as the surrogate for microtexture.
In the United Kingdom, the SCRIM values are synonymous with microtexture. The
SCRIM is a side force coefficient measuring device and therefore the sliding speed of
the test tire is relatively low. The SCRIM operates at traffic speeds; however, because
the slip speed is low, it serves as a surrogate for a microtexture measurement.
The PIARC Model for the IFI avoids the need for measuring microtexture, if
macrotexture measures are available. A friction measurement at any slip speed,
together with the macrotexture parameter, determines the friction as function of slip
speed.
There is currently no practical procedure for the direct measurement of the
microtexture profile in traffic. Such a procedure would possibly enable testers to avoid
the friction measurement altogether by measuring microtexture and macrotexture in
order to predict the wet pavement friction as a function of speed. This would eliminate
the need to carry water and use a high-powered host vehicle (NCHRP Synthesis 291,
2000).
2.2.11.3.2 Macrotexture Measurement
The different macrotexture measurement methods are summarized below.
64
2.2.11.3.2.1 Volumetric Measurements
The Sand Patch Method (ASTM E965) is used to measure macrotexture of the
specimen surface. This method involves taking a known volume of a spreadable
material and spreading it out in a circle on the surface of the specimen. Measuring the
diameter gives the area of the circle. The Mean Texture Depth (MTD) is determined by
dividing the volume by the area. Figure 2-24 illustrates the procedure followed to
measure the MTD and the required tools for the sand patch method.
Another volumetric measure of macrotexture is the grease patch method developed by
the National Aeronautics and Space Administration (NASA). This method is similar in
concept to the sand patch, except that grease is spread over the surface in a rectangular
area between parallel strips of masking tape. Again, the average macrotexture depth is
determined by dividing the known volume of grease by the area covered.
A third volumetric approach uses the same concept of spreading a known volume of
material over a measured area is the silicone putty, or Silly Putty. In this method, a
fixed amount of putty is pressed onto the surface using a plastic disk with a round
recess whose volume equals that of the putty. When pressed onto a surface with
texture, the amount of macrotexture is indicated by the diameter of the putty.
T
su
cy
w
th
th
a
ou
an
ou
b
B
Fi
2.2.11.3.2.2
The outflow
urface macro
ylinder is sup
with water, a
hrough the v
hat water wil
smooth text
utflow mete
ny additiona
utflow mete
etween the b
Buhlmann, 19
gure 2-24: M
2 Outflow M
meter, show
otexture. A c
pposed to m
and the time
voids and ch
ll flow faster
ture that seal
er gives an a
al informati
r to truly giv
bottom of th
982 and NCH
MTD determ
Meter
wn in Figure
cylinder is p
meet the pave
e for a know
hannels on th
r if the pave
ls better to th
average value
on about th
ve an indicat
he rubber gas
HRP Synthe
65
mination usin
e 2-25, uses
placed on the
ement surface
wn volume o
he pavement
ment has a r
he rubber rin
e for the pav
he distributio
tion of surfa
sket and the
sis 291, 200
g the sand p
the rate of
e pavement.
e as would a
of water to
t surface is
rough texture
ng. As with t
vement mac
on or nature
ace macrotex
top of the p
0).
patch method
flow of wat
A rubber ri
a tire. The cy
escape from
measured. T
e than if the
the volumetr
crotexture, bu
e of the tex
xture, the wa
pavement lay
d
ter to estima
ing around th
ylinder is fill
m the cylind
The concept
pavement h
ric devices, th
ut cannot giv
xture. For th
ater must dra
yer (Yager an
ate
he
ed
der
is
has
he
ve
he
ain
nd
66
Figure 2-25: Outflow meter
2.2.11.3.2.3 Profile Tracers
The development of this method has been attempted by many investigators. In 1980,
Gillespie reported on a profile tracing device and showed some correlation between
profile parameters and skid resistance. In 1967, Hankins reported in the development
and testing of a highly sensitive profile tracing device to be used in the laboratory. By
analysing the plot of the profile parameters and skid resistance, it revealed a wide
scatter of points which limit the usefulness of this approach.
The pavement surface macrotexture can be accurately measured by an advanced laser
profiling technology. The circular texture meter (CTM) shown in Figure 2-26
represents one of such advanced laser based profiler or macrotexture measurement
device. The test procedure using the CTM has been standardized in ASTM E 2157. In
essence, the CTM uses a laser to measure the profile of a circle 284 mm (11.2 inch) in
diameter or 892 mm (35 inch) in circumference. The circumference is divided into
eight segments of 111.5 mm (4.4 inch) arc. The average MPD is determined for each of
67
the segments of the circle. The data process is carried out as follows: (a) the
circumferential annulus is divided into eight segments of 4.4 inch arc, (b) the Mean
Segment Depth is determined according to the schematic diagram in Figure 2-27 for
each segment in accordance to ASTM E1845, and (c) the final reported MPD of the
test specimen is the average of all eight Mean Segment Depths. In addition to the MPD
reported by the CTM, Root Mean Square (RMS) can also be reported. RMS is a
statistical value of how much the actual measured profile deviates from a best fit
modelled profile of the data. The capability of the CTM to quantify MPD and RMS
allows for the assessment of the “orientation” of texture, which in turn enables the
engineer to determine what types of features are supplying the macrotexture; i.e.,
negatively, positively, or neutrally textured (McGhee and Flintsch, 2003).
In 2004, Hanson et al. evaluated the circular texture meter for measuring surface
macrotexture of pavements. They reported that the CTM produces comparable results
to the ASTM E965 Sand Patch Test. When open-graded mixtures were excluded; this
study indicated that the offset was non-significant between CTM and Sand Patch test
results. The slope of the best fit line comparing the results was statistically significant,
and ranged from 0.93 (2003 data) to 1.01 (2000 data).
T
T
p
tr
(a
Figure 2-
The Kansas H
This method
otentiometer
ransversely u
a) Bottom vie
-26: Circular
Figure 2
Highway Co
uses a mo
r attached t
under the m
ew
r Texture M
2-27: Mean S
ommission d
otorized lath
to the micro
microscope.
68
eter: (a) gen
Segment De
developed a m
he and a ste
oscope focu
The operat
(b) G
neral view, an
epth determin
method calle
ereo microsc
using shaft.
or keeps th
General view
nd (b) bottom
nation
ed linear tra
cope with t
The specim
he surface u
w
m view
averse metho
the shaft of
men is mov
under consta
od.
f a
ed
ant
69
focus, thus changing the potentiometer voltage. This results in a tracing of the surface
macrotexture.
Texas Transportation Institute developed a device called texturemeter, consisting of a
series of evenly spaced vertical parallels in a frame. With the exception of two, the rods
can move vertically and are independent of one another. A string is attached to the
movable end and to the frame. The texturemeter produces a straight line on a smooth
surface and a dial indicator has been calibrated to zero. If there are any irregularities,
the string produces a zigzag line and results in a dial reading greater than zero. The
reading is proportional to the coarseness of the macrotexture: the coarser the surface,
the higher the reading.
70
CHAPTER III
3. A NEW ACCELERATED POLISHING DEVICE FOR HMA SURFACES
3.1 Introduction
With time and traffic, asphalt concrete pavements gradually lose their skid resistance,
creating a serious safety concern especially when pavements are wet. As the driving
speed and the Average Daily Traffic (ADT) increases, the chances of having skid-
related accidents also increase rapidly (Beaton, 1976 and Brilett, 1984). Thus, the
Federal Highway Administration has issued a Wet Skid Accident Reduction Program
to encourage each state highway agency to minimize wet weather skidding accidents
by identifying the sections of roadways with high occurrence of skid accidents and then
by either resurfacing of the pavement or rejuvenating surface texture to bring the
asphalt pavement surface to an adequate skid resistance properties. However, the cost
associated with the regular identifying and the remediation action could be extensively
high. An alternative approach would be to take a more proactive stand in screening the
polishing potential of aggregates and hot mix asphalt to ensure that the final selected
mix can provide sustained resistance to polishing while providing adequate level of
friction over the life span of the pavement. To achieve this initial screening task during
the mix design stage of the HMA, there is a need to develop a laboratory scale
71
accelerated polishing device that can mimic the actual abrasion and polishing behavior
between the vehicle rubber tire and the HMA surface.
The main objective of this chapter is to present a laboratory-scale accelerated HMA
polishing device for the purpose of screening the polishing and friction performance of
the HMA mix, in terms of aggregate source, aggregate gradation, and binder type and
content that constitute the important physical constituents of the HMA. This polishing
device should be capable of reproducing the tire-pavement interaction in a reasonable
timeframe, easy to operate with less labor effort, and has the ability to test small size
HMA specimens (e.g., 6 inch in diameter gyratory compacted specimens). The
operation principles along with the operation conditions of the polishing device are
described in detail. The performance of the accelerated polishing test device is
evaluated and described in detail as well. The acceptance criteria are presented for
screening the HMA specimens from the polishing and friction standpoint. The
advantages and limitations of the developed device are presented at the conclusion of
the chapter.
3.2 Existing Laboratory Scale Polishing Devices
In the past efforts in developing laboratory-scale accelerated polishing devices, several
key mechanisms involved in polishing either aggregate or HMA have been identified.
As reviewed in Ibrahim (2007), the skid resistance of asphalt concrete can be affected
by bleeding and flushing of bituminous binder to the surface, surface wear due to
studded tires, polishing of surface aggregate, rutting due to compaction, lateral
72
distortion, contamination (rubber, oil, water, etc.), smoothened macrostructure, and
inadequate cross slope. Among these factors, however, aggregate and mixture
characteristics remain the most dominant controlling factors. Research focused on
polishing and friction characteristics of aggregates (Colony, 1984; Colony, 1992; Liang
and Chyi, 2000; Liang, 2003; Dewey et al., 2001; and Do et al., 2003) have shown
strong correlations between the time history of skid number (SN) at the monitored
pavement sections and the factors such as traffic conditions, properties of asphalt
mixes, and geological features; i.e., predominant aggregate and physiographic area. In
the research on aggregate friction loss (Liang and Chyi, 2000), the aggregate polishing
propensity can be identified by means of a suite of test procedures, including the use of
mineralogical analysis using thin sections and Acid Insoluble Residue test (ASTM D
3042-03). Thus, the current understanding of aggregate polishing and friction behavior
is well established.
A review of the existing laboratory-scale accelerated polishing devices reveals that they
can be categorized into three groups: one is capable of polishing the aggregate samples,
one is capable of polishing the HMA samples, and the third is capable of polishing both
(aggregate and HMA specimens). A brief review of the existing devices is summarized
in Table 3-1.
3.3 Equipment Development
In the following section, the newly developed equipment and its operational procedure
along with the operation conditions are discussed in detail.
73
3.3.1 Equipment Description and Operational Procedure
The guiding principle of developing the laboratory-scale accelerated polishing
equipment is that the evolution history of friction loss of the asphalt pavement surface
can be accurately replicated and measured in realistic short test duration. In essence, the
abrasive action between the rubber tire of a vehicle and asphalt concrete pavement
surface will be enacted in the accelerated polishing device. The deign of the polishing
equipment allows for pressing polishing shoes (pads) made of Styrene-Butadiene-
Rubber (SBR) onto the surface of the HMA specimen at a constant vertical force while
rotating these rubber pads at a constant rotational speed. It should be noted that the
polishing device is designed to accommodate two specific specimen dimensions: an 18
inch by 18 inch by 2 inch high roller compacted slab specimen or a 6 inch diameter by
4 inch high Superpave gyratory compacted specimen. As a result of different specimen
sizes, the rubber pads are designed differently. For the gyratory compacted specimen, a
solid rubber disk of 6 inch in diameter and 1.5 inch thick is used. For the slab
specimen, a rubber ring of approximately 13 inch in outside diameter and 9 inch in
inside diameter is used to fit with the required polishing area for the DFT and CTM.
The schematic diagram of the equipment is presented in Figure 3-1, in which Figure
3-1(a) and Figure 3-1(b) show the top view and elevation view, respectively.
Details of the rubber pad dimension for the gyratory specimen and large slab
specimen are shown in Figure 3-1(c) and Figure 3-1(d), respectively. A photograph
of the completely fabricated accelerated polishing device is shown in
74
Figure 3-2(a), with the close-up view of two types of specimens mounted in
positions shown in Figure 3-2(b) and Figure 3-2(c),respectively.
Table 3-1: A summary of the existing accelerated polishing machines
75
1.5 HP MOTOR
ADD WEIGHTFOR ADDITIONALPRESSURE
A33.75
33.75
(a) Top view of the accelerated polishing machine using rubber shoes
34.25
20.00
14.25
42.625
26.375
9.437
ADJUST HEIGHT OFRUBBER PADS WITH CRANK
1/2 NPT NIPPLEFOR FLUID
HMA SAMPLE
SHROUD
BUILT FORM
DOLLY28.00
ELECTRIC BOX
B BB1B1
(b) Section A-A for the machine details
Figure 3-1: Different views of the accelerated polishing machine using rubber
shoes; all units are in inches
76
( 13.75 DRIVE PLATE)
( 1.500 OD X 0.075 ID SHAFT)
( 0.25 HOLES FOR WATER, 2 PLACES)
( 13.50 OD OF RUBBERBLOCK MOUNT RING)
( 8.75 ID OF RUBBERBLOCK MOUNT RING)
(13 OD X 9 ID RUBBER SHOE)
(c) Slab specimen rubber shoe
(13.75 DRIVE PLATE)
( 13.50 OD X 8.75 IDSPACER RINGS)
( 0.25 HOLES FOR WATER, 2 PLACES)
( 6.00 SBR RUBBER SHOE)
( 1.500 OD X 0.075 ID SHAFT)
(GROOVE FOR WATER)
(d) Gyratory specimen rubber shoe
Figure 3-1: Different views of the accelerated polishing machine using rubber
shoes; all units are in inches
77
(a) Overall view of the accelerated polishing machine using rubber shoes
(b) Details on slab specimen mounting
78
(c) Details on gyratory compacted specimen mounting
Figure 3-2: Overall view of the accelerated polishing machine using rubber shoes
and setups for testing slab specimen and gyratory compacted specimen
3.3.2 Operation Conditions
The design of the device includes individual control for the vertical force on the
specimen, the rotational speed of the rubber pad, and the rate of water spray on the
specimen surface during polishing. The range of these individual controls is presented
in Table 3-2.
In an effort to determine an optimum operation condition, the authors have conducted a
series of tests on the HMA specimens using different combinations of operation
79
conditions, including the type of rubber to be used for the rubber pads, the rotational
speed of the rubber pad, the vertical force applied by the rubber pad to the HMA
specimens, and the rate of water spray. The final selected operation conditions are
summarized in Table 3-2 as well. These optimum operation conditions would ensure
that the rubber pad would not experience rocking motion, and that more or less flat
contact surface between the rubber pad and the specimen is maintained. Furthermore,
the water spray is to ensure that the rubber debris was washed off and that rubber-
specimen surface was not overheated.
Table 3-2: Range of and selected optimum operation parameters
3.4 Equipment Characteristics and Validation
In this section, the repeatability of the test results from the developed accelerated
polishing device is discussed. Furthermore, the ability of the polishing device to
replicate the polishing trend of the aggregate is demonstrated.
80
3.4.1 Materials
In the evaluation study of the developed device, two aggregate sources (limestone and
gravel) and two asphalt binder grades (PG 70-22 and PG 64-22) were used to compact
the HMA specimens. The gradation curve for the aggregate is shown in Figure 3-3. The
optimum binder content is determined by using Marshall Design method. The optimum
binder content is 5.9% and 6.3% for the mix consisting of limestone and PG 70-22
binder and the mix consisting of gravel and PG 64-22 binder, respectively.
Figure 3-3: Gradation curves
3.4.2 Sample Preparation Procedure for HMA Specimens
The mixing procedure of the loose mix is as follows. First, the aggregates are separated
by dry sieving into the desired sizes using the mechanical shaker. The aggregates are
then washed and heated to about 330˚F. Aggregates are weighed and blended according
0
20
40
60
80
100
120
0.001 0.01 0.1 1
Grain size (in)
Pe
rce
nt
fine
r
Limestone
Sand & Gravel
81
to the gradation curve shown in Figure 3-3. The weighed aggregate mix is then put in
the oven at 330˚F for 3 hours for achieving a uniform aggregate temperature. The
mixing bowl and the mixing paddle are also heated to 300˚F. The asphalt binder is
heated in the oven at a temperature of 350˚F for 2-3 hours. At this point, the aggregate
is placed in the mixing bowl and blended quickly with the asphalt binder until a
uniform blend is obtained.
The gyratory compactor is used to compact the loose mix into a 6-inch cylindrical
specimen, while a roller compactor is used to compact the loose mix into a 18 inch by
18 inch by 2 inch slab specimen.
3.4.3 Friction and Surface Texture Measurements
Different types of measuring techniques are used to measure friction and texture of the
HMA surface for the two types of specimen sizes due to different polishing rubber pads
used (see the description of the rubber polishing pads). For the 6 inch cylindrical
gyratory compacted HMA specimens, the British Pendulum Tester and the sand patch
method are used for measuring friction and surface texture. For the 18 inch by 18 inch
by 2 inch roller compacted specimens, the DFT and CTM are used for measuring
friction and texture, respectively. A brief description of each measurement technique is
presented herein.
The BPT (ASTM E303-93) test procedure is as follows. Typically, the specimen is
first immersed in water and then the test surface is cleaned. Next, the BPT device is
82
levelled and adjusted so that the contact path during the swing of the slider is within the
range of 063.092.4 inch. The rubber slider is cleaned and wetted. Five swings are
made for each specimen, from which an average of the last four readings is recorded as
the British Pendulum Number (BPN).
The sand patch method (ASTM E965) is a technique to measure macrotexture of the
HMA surface. This method involves taking a known volume of a spreadable material
and spreading it out in a circle on the surface of the specimen. The Mean Texture
Depth (MTD) is determined by dividing the volume of the spread material by the
surface area covered by the spread material. This technique is used for the 6 inch
gyratory compacted specimen due to the ease associated with a small area to cover.
The test procedure of DFT is given in ASTM E1911. The DFT device consists of a disk
fitted with three spring-loaded rubber sliders at a diameter of 11.2 inch. The disk is
initially suspended above the pavement surface and is driven by a motor until the
desired tangential speed of the sliders (about 55 mph) is attained. Water is then flowed
over the test surface. The rotating disk is then dropped onto the wet surface with the
applied vertical force of 8 lb while the friction force and speed of the rotating disk are
continuously measured and recorded as the disk slows down to stop (zero speed).
The CTM is described in ASTM E2157. In essence, it uses laser techniques to measure
the surface texture profile of an annulus surface area (i.e., 11.2 inch in outside diameter
and 10 inch in inside diameter). The data process is carried out as follows: (a) the
circumferential annulus is divided into eight segments of 4.4 inch arc, (b) the Mean
83
Segment Depth is determined for each segment in accordance to ASTM E1845, and (c)
the final reported MPD of the test specimen is the average of all eight Mean Segment
Depths.
3.4.4 Supplemental Image Analysis Techniques
Digital image analysis techniques are used to quantify the percentage of the exposed
aggregate area of the specimen surface after being subjected to polishing by the
accelerated polishing device. The percent of exposed aggregate area is defined as the
area of the exposed aggregate surface divided by the total area of the HMA specimen
surface that is being polished by the device. The typical image analysis procedure
involves first taking the digital images of the specimen surface using the Olympus C-
5060 Wide Zoom high-performance 5.1-megapixel digital camera. The digital images
are then opened in the software (Scion image provided by National Institutes of Health)
and converted into binary images for the subsequent calculation of exposed aggregate
area.
3.4.5 Repeatability of the Accelerated Polishing Equipment
The repeatability of the polishing results using the developed accelerated polishing
device was examined. For each set of specimens made of the same mix formula
(aggregate source, aggregate gradation, optimum binder content, binder type, and
compaction method and effort), three replicate specimens were tested. The friction
values obtained from the BPT, the MTD measured by the sand patch method and the
84
image analysis results (Agg. %) from the three replicates are statistically analyzed
using Homogeneity of Variance (Levene statistic), one-way Analysis of Variance
(ANOVA), and Multiple Comparisons to check for the repeatability of test results.
Homogeneity of Variance and one-way ANOVA are used to check if there is any
significant difference between the variances and the means of at least two specimens
for each set of specimens (three specimens) made of the same JMF. Multiple
Comparisons, on the other hand, is used to check if there is any significant difference
between the means of different two-specimen combinations of the three specimens
made of the same JMF. The software Statistical Package for the Social Sciences
(SPSS) was employed for obtaining the statistical analysis results. Table 3-3
summarizes the statistical analysis results. It can be seen that the difference between the
variances and the means of the results (in terms of BPT, MTD, and IA) for the three
replicate specimens is insignificant for all cases when considering the friction values
(BPN) and aggregate exposure area (IA) and insignificant for the vast majority of the
cases when considering the macrotexture values (MTD), thus supporting the
repeatability of the polishing action provided by the accelerated polishing device.
3.4.6 Polishing Effect of the Accelerated Polishing Machine
The polishing effect of the accelerated polishing machine is examined in this section.
For the HMA slab specimens made with limestone aggregates, the friction values (FN)
obtained from the DFT at different measuring speeds versus the polishing duration is
shown in Figure 3-4(a). The MPD measured by the CTM is plotted versus duration of
85
polishing in Figure 3-4(b). It can be seen from Figure 3-4(a) that friction decreases with
polishing duration until it reaches the residual value. Corresponding to the friction
decrease, there is a similar trend of MPD decrease as well.
For the 6-inch HMA gyratory compacted specimens made with limestone aggregates,
the friction values (BPN) obtained from the BPT and the MTD measured by the sand
patch method are plotted against the polishing duration in Figure 3-5(a) and Figure
3-5(b), respectively. It can be seen that both the friction values and the MTD decrease
as the polishing duration increases.
The test results of HMA specimens made of the sand and gravel aggregates are shown
in Figure 3-6(a) and (b) and Figure 3-7(a) and (b) for the slab specimens and the
gyratory compacted specimens, respectively. The general trend of the polishing
behavior for the specimens made of the limestone and the sand and gravel is similar.
Furthermore, it can be seen that the polishing behavior for large slab specimens and
gyratory compacted specimens exhibit the similar decreasing pattern.
86
Table 3-3: Repeatability tests for the limestone and gravel
a. significant at the p-value smaller than 0.05
87
(a) FN by DFT vs. polishing time at different speeds
(b) MPD by CTM vs. polishing time
Figure 3-4: Polishing, friction, and texture results of tests conducted on limestone
slab specimens
30
40
50
60
70
80
90
100
0 100 200 300 400 500 600 700 800 900 1000
Polishing Time (min.)
FN
me
as
ure
d u
sin
g D
FT
0 km/hr
10 km/hr
20 km/hr
30 km/hr
40 km/hr
50 km/hr
64 km/hr
0.85
0.90
0.95
1.00
1.05
0 200 400 600 800 1000
Polishing Time (min.)
MP
D (
mm
)
Friction Measurement
Speed
88
(a) BPN by BPT vs. polishing time
(b) MTD by sand patch technique vs. polishing time
Figure 3-5: Polishing, friction, and texture results of tests conducted on limestone
gyratory compacted specimens
45
50
55
60
65
70
75
80
0 100 200 300 400 500
Polishing Time (min.)
BP
N
0.1
0.12
0.14
0.16
0 100 200 300 400 500
Polishing Time (min.)
MT
D (cm
)
89
(a) FN by DFT vs. polishing time at different speeds
(b) MPD by CTM vs. polishing time
Figure 3-6: Polishing, friction, and texture results of tests conducted on Sand and
Gravel slab specimens
0.60
0.65
0.70
0.75
0 200 400 600 800 1000
Polishing Time (min.)
MP
D (m
m)
Friction Measurement
Speed
90
(a) BPN by BPT vs. polishing time
(b) MTD by sand patch technique vs. polishing time
Figure 3-7: Polishing, friction, and texture results of tests conducted on Sand and
Gravel gyratory compacted specimens
55
60
65
70
75
0 100 200 300 400 500
Polishing Time (min.)
BP
N
0.075
0.077
0.079
0.081
0.083
0.085
0 100 200 300 400 500
Polishing Time (min.)
MT
D (
cm)
91
3.4.7 Comparing HMA Surface and Aggregate Surface Polishing Behavior
In a previous study (Liang and Chyi, 2000), different aggregates sources were tested
for polishing and friction behavior using the accelerated British Polishing Wheel
(ASTM E3319). The results of polishing behavior of two aggregates from (Liang and
Chyi, 2000) and the current study of the HMA specimens made with the same two
aggregate sources are statistically compared in Table 3-4 and Table 3-5 for Limestone
and Sand and Gravel aggregates, respectively. It can be seen that the polish (friction)
values of the aggregates, denoted by PV, are highly correlated to the friction values of
the HMA made with the same aggregates, denoted by either BPN for the gyratory
compacted specimens or FN_SPEED (where SPEED refers to the friction at that
measuring speed) for the roller compacted slab specimens with friction measured at
different speeds. The fact that aggregates constitute more than 90% by weight of the
HMA leads us to believe that aggregate would be a dominant controlling factor on
friction of HMA surface. The high correlations presented in Table 3-4 and Table 3-5
support this observation. Based on the ANOVA analysis results presented in Table 3-4
and Table 3-5 for the Limestone and Sand and Gravel aggregates, respectively, the
overall significance of the models as indicated by the F-value (i.e., as F goes up, P goes
down, thus indicating more regression confidence in that there is a difference between
the two means) and the P-value (the probability of getting a value of the test statistic as
extreme as or more extreme than that observed by chance alone, if the null hypothesis
Ho, is true) is found to be significant at the 0.05 significance level. Based on the
comparisons presented in this section, the developed laboratory-scale accelerated
92
polishing device is shown to be able to polish the HMA surface and provide similar test
results as if the polishing tests were performed on the aggregates only.
Table 3-4: Simple Linear Regression between Aggregate Friction Values (Liang
and Chyi 2000) and HMA Friction Values (This Study) for Columbus Limestone
Table 3-5: Simple Linear Regression between Aggregate Friction Values (Liang
and Chyi 2000) and HMA Friction Values (This Study) for Stocker Sand and
Gravel
93
3.4.8 Comparing the Polishing Trend with the Aggregate Exposure Area
Image analysis is carried out to quantify the area of exposed aggregate (Agg. %) of
the gyratory compacted HMA specimen surface during different stages of the
polishing test. The percent of aggregate exposure area of Limestone specimens is
measured from the digitized images shown in Figure 3-8(a) and then plotted against
the polishing duration in Figure 3-8(b). Similarly, the digitized images of the Sand
and Gravel specimens shown in Figure 3-9(a) are used to plot the percent of
aggregate exposure area versus the polishing duration in Figure 3-9(b). From
Figures 3-8(b) and 3-9(b), one can see that the more polishing duration the
specimen is subjected to, the more aggregates area is exposed until reaching the
maximum percentage.
A statistical analysis is conducted to exam the correlations between the friction values
of aggregate (PV) and HMA (BPN) and the percent of exposed aggregate area (Agg.
%) at different stages of polishing. The regression models and the coefficient of
determination (R2) for each model are presented in Table 3-6 together with the
ANOVA analysis results. It can be seen that PV and BPN correlates well with the
percent of exposed aggregate area at different stages of polishing.
94
(a) The captured images shown on the left and the digitized images shown on the
right at different polishing times
(b) Aggregate exposure area vs. polishing time
Figure 3-8: Image analysis results of tests conducted on Limestone gyratory
compacted specimens
0
5
10
15
20
25
0 100 200 300 400 500 600
Polishing Time (min.)
Ag
gre
ga
te e
xp
os
ure
(%
)
0 min
240 min
480 min
95
(a) The captured images shown on the left and the digitized images shown on the
right at different polishing times
(b) Aggregate exposure area versus polishing time
Figure 3-9: Image analysis results of tests conducted on Sand and Gravel gyratory
compacted specimens
0
5
10
15
20
25
0 100 200 300 400 500 600
Polishing Time (min.)
Ag
gre
gate
exp
osu
r (%
)
0 min
240 min
480 min
96
3.4.9 Polishing Trend of HMA Samples Prepared by Two Compaction Methods
The friction values of the two types of specimen sizes (slab specimens and 6-inch
specimens) each with different type of compaction method (i.e., roller compaction vs.
gyratory compaction) have been found to be correlated and the coefficients of
determination have been found to be significant as can be seen from Table 3-7 and
Table 3-8 for Limestone and Sand and Gravel aggregates, respectively. It is very
interesting to note that the correlativity is more significant between BPN and
FN_SPEED at low speeds; for example, at 0, 6, and 12.5 mph. This high correlation is
reasonable considering that BPT actually measures the friction values at low speed; i.e.,
at 6 mph. Based on the ANOVA analysis shown in Tables 6 and 7, the overall
significance of the models as presented by the F-value and P-value was found to be
significant at the 0.05 significance level.
Table 3-6: Simple Linear Regression between Aggregate and HMA Friction Values
and Aggregate Exposure Area
97
Table 3-7: Simple Linear Regression between Friction Values of Gyratory
Compacted Specimens and Friction Values of Roller Compacted Slab Specimens
(Limestone aggregate)
Table 3-8: Simple Linear Regression between Friction Values of Gyratory
Compacted Specimens and Friction Values of Roller Compacted Slab Specimens
(Sand and Gravel aggregate)
3.4.10 Application of the Accelerated Polishing Device
In order to use the test results from the developed accelerated polishing device in
screening the aggregate source and mix design, there is a need to correlate the friction
values measured on the HMA surface in the laboratory to the friction values measured
either on the aggregate samples in the laboratory or the skid number measured on
98
pavement surface. This is because there are more experiences in using either the PV
values determined from BPT or SN determined from LWST. In this section, different
paths will be developed to allow the use of the developed accelerated polishing device
for qualifying the aggregate source and mix design from the polishing and friction
point of view.
3.4.10.1 Correlation with PV values
Texas Department of Transportation (TxDOT) has adopted Table 3-9 for acceptance of
aggregate. These standards were based on the polish values (PV) and the Average
Daily Traffic (ADT).
Through the developed correlation between aggregate polish values (PV) from a
previous study (Liang and Chyi, 2000) and gyratory compacted HMA friction numbers
(BPN, shown in Tables 3-4 and 3-5), the TxDOT criteria based on PV values can be
transformed to represent acceptance criteria of HMA based on BPN values. The
derived acceptance criteria, based on BPN values of HMA surfaces, are shown in Table
3-10. Accordingly, four categories of acceptance criteria can be formulated.
Highly acceptable (BPN greater than 51 for HMA prepared using limestone aggregate
and 62 for HMA prepared using gravel aggregate).
Acceptable (BPN greater than 47 for HMA prepared using limestone aggregate and 60
for HMA prepared using gravel aggregate).
99
Marginally acceptable (BPN greater than 43 for HMA prepared using limestone
aggregate and 57 for HMA prepared using gravel aggregate).
Unacceptable (BPN less than 43 for HMA prepared using limestone aggregate and 57
for HMA prepared using gravel aggregate).
Table 3-9: TxDOT acceptance criterion of aggregates
Table 3-10: Derived acceptance criteria of HMA based on BPN values
100
3.4.10.2 Correlation with SN values
Friction values measured by the British Pendulum Tester (BPN) and skid numbers
measured by the skid trailer (SN) do not correspond exactly; nevertheless, Kissoff
(1988) has developed an approximate relationship (see Equation 3-1) that could be used
to relate BPN to SN. Therefore, the above acceptability criteria based on BPN values
can be altered according to Kissoff’s relationship to develop SN-based acceptability
criteria as summarized below. The derived acceptance criteria, based on SN values of
HMA surfaces, are shown in Table 3-11. Accordingly, four categories of acceptance
criteria can be formulated.
690.9)(862.0 BPNSN (3-1)
Highly acceptable (BPN greater than 34 for HMA prepared using limestone aggregate
and 44 for HMA prepared using gravel aggregate).
Acceptable (BPN greater than 31 for HMA prepared using limestone aggregate and 42
for HMA prepared using gravel aggregate).
Marginally acceptable (BPN greater than 27 for HMA prepared using limestone
aggregate and 39 for HMA prepared using gravel aggregate).
Unacceptable (BPN less than 27 for HMA prepared using limestone aggregate and 39
for HMA prepared using gravel aggregate).
101
Table 3-11: Derived acceptance criterion of HMA based on SN values
3.5 Summary and Conclusions
Presented in this chapter is the development of an accelerated laboratory-scale
polishing device that is capable of mimicking the polishing action of the HMA surface
by a vehicle tire in a short duration, thus allowing for screening the aggregate source
and mix design formula to ensure adequate friction (or skid resistance) of the HMA
over the expected life span of the pavement surface. The accelerated polishing machine
is capable of testing two different sizes of HMA specimens: 18 inch by 18 inch by 2
inch high slab specimens compacted using the roller compactor and 6 inch diameter
and 4 inch high gyratory compacted HMA specimens. Although the device can handle
two different specimen dimensions, each with different type of compaction method, the
intended routine test sequence is geared toward the testing of 6 inch gyratory
compacted HMA specimens due to its ease of sample preparation. The design
principles of the testing device, together with the optimized operation conditions, are
outlined in detail in this chapter. Evaluation of the capability of the developed
102
accelerated polishing device has been conducted through a series of designed testing
and comparisons, as summarized below.
Repeatability of the machine was checked and affirmed using one-way ANOVA test.
The polishing effect of the machine was confirmed through examination of the test
results conducted on Limestone and Sand and Gravel aggregates.
Good correlation of the polishing and friction behavior was found between aggregate
specimens and the HMA specimens made with the same aggregates. Therefore, it
maybe reasonable to conclude that the new accelerated polishing machine can
accomplish the intended tire/pavement wearing and polishing mechanisms.
Image analysis validated the polishing action.
Good correlation was found between the two specimen sizes using different
compaction methods.
The developed accelerated polishing device can be used effectively for screening
polishing and friction properties of HMA mix (i.e., aggregate source, binder type and
content, etc.) during the HMA mix design stage. Furthermore, the device possesses the
following advantages:
●Tests are repeatable.
●The device can test small-size HMA specimens (i.e., 6 inch diameter and 4 inch high
gyratory compacted specimens).
103
●The test can be completed in a reasonable timeframe.
●The test procedure is simple.
●The test method is efficient (i.e., less labor effort).
●The device can simulate the tire/pavement interaction.
The correlation study between friction values measured by DFT at low measuring
speed and the BPT suggests that the BPT can be used for measuring friction of HMA at
low speeds.
A set of tentative acceptance criteria of gyratory compacted HMA specimens was
developed through two different correlations (i.e., by correlating BPN with PV or
relating BPN with SN). These acceptance criteria are divided into two parts based on
the aggregate type used; i.e., limestone or gravel. The acceptance criteria consist of four
categories: 1. highly acceptable, 2. acceptable, 3. marginally acceptable, and 4.
unacceptable, which can be used to screen and select pertinent aggregate and HMA
mix design for adequate polishing resistance and friction values.
104
CHAPTER IV
4. LABORATORY TEST RESULTS AND DATA ANALYSIS
4.1 Introduction
In adequate friction on the pavement surface is a major cause for wet weather related
accidents on highways. To minimize wet weather related accidents, most highway
agencies have developed a strategic goal of maintaining high skid resistance on the
pavement surface. In general, the state DOTS have maintained a friction measurement
program, through the use of the locked wheel skid trailer (LWST), to measure the skid
number (SN). Once the measured SN is below a threshold value, then state DOTs
would take remedial actions by resurfacing the pavement surface with high skid
resistance HMA (hot mix asphalt) to ensure driving safety.
Although the practice of monitoring pavement surface friction by means of SN and
taking remedial actions by pavement resurfacing is important; nevertheless, it is a
passive solution toward the problem. A more proactive approach to solving the
problem would be to use high polish-resistant and high friction aggregate and the
accompanied HMA (hot mix asphalt) mix design during the initial stage of the material
acceptance process. A significant number of research efforts have been devoted toward
the development of appropriate mix design protocols, such as SuperPave mix design
105
but there has been a lack of more intensive research efforts to develop appropriate test
procedures to screen polishing and friction behavior of HMA. As described in chapter
3, the authors developed a laboratory scale accelerated polishing device for HMA and
for simulating the actual tire-asphalt polishing action. The friction measurement of the
laboratory prepared HMA specimens, however, can only be carried out by laboratory
scale portable devices such as British Pendulum Tester (BPT) or the Dynamic Friction
Tester (DFT). Since the BPT is widely used by highway agencies in developing
acceptance criteria for friction, there is a need to gain a better understanding of the
numerous influencing factors on the measured BPN (British Pendulum Number)
thorough the BPT. The main objective of this chapter is to examine the relationship
between the BPN, measured by the BPT, and the texture properties (i.e., MTD, Mean
Texture Depth) of the HMA surface, measured by the sand patch method.
4.2 Pavement Sections and Material Properties
To facilitate the selection of the aggregates and the accompanied JMF for mix design of
HMA, the authors have consulted with the Ohio Department of Transportation’s
pavement construction records to identify a total of eight pavement sections for long-
term monitoring of friction behavior. The selection of these pavement sections is based
on the criteria that each of the pavement sections provides adequate documentation of
traffic counts as well as the construction materials used and the mix design. Table 4-1
provides information on these identified pavement sections. Based on field friction
data, these aggregate sources can be roughly classified into three categories: low (L),
106
medium (M), and high (H). Details of the JMF of each mix design can be found in
Appendix A.
Table 4-1: Asphalt concrete pavement sections and the associated JMFs
107
4.3 Test Program
The eight job mix formulas obtained from these eight pavement sections are used to
prepare the gyratory compacted HMA specimens that are 6 inch in diameter and 4 inch
in height for the accelerated laboratory polishing and the accompanied measurement of
BPN and MTD. Once the gyratory compacted specimens are prepared, the initial
values of BPN and MTD are measured. Each specimen is then subjected to polishing in
the accelerated polishing machine. After each hour of polishing, the BPN and MTD are
taken again. This polishing and measuring process continues for 8 hours for each
specimen tested.
4.3.1 Sample Preparation Procedure for HMA Specimens
The mixing procedure of the loose mix is as follows. First, the aggregates are separated
by dry sieving into the desired sizes using the mechanical shaker. The aggregates are
then washed and heated to about 330˚F. Aggregates are weighed and blended according
to the specified gradation curve. The properly proportioned aggregate mix is then put in
the oven at 330˚F for 3 hours for achieving a uniform aggregate temperature. The
mixing bowl and the mixing paddle are also heated to 300˚F. The asphalt binder is
heated in the oven at a temperature of 350˚F for 2 to 3 hours. At this point, the
aggregate is placed in the mixing bowl and blended quickly with the asphalt binder
until a uniform blending is obtained. The gyratory compactor is used to compact the
108
loose mix into a 6 inch diameter and 4 inch high cylindrical specimen according to the
ODOT compaction specifications.
4.3.2 Friction and Texture Measurement Techniques
The BPT (ASTM E303-93) can be described as follows. Typically, the test specimen is
first immersed in water and then the test surface was cleaned. Next, the BPT device is
levelled and adjusted so that the contact path between the rubber slider and specimen
surface is within the range of 063.092.4 inch. The rubber slider is cleaned and wetted.
Five swings are made for each specimen being tested, from which an average of the last
four readings is recorded as the BPN.
The sand patch method (ASTM E965) is a technique to measure macrotexture of the
HMA surface. This method involves taking a known volume of a spreadable material
and spreading it out in a circle on the surface of the specimen. The Mean Texture
Depth (MTD) is determined by dividing the volume of the spread material by the
surface area covered by the spread material.
4.3.3 Accelerated Polishing Device
The accelerated polishing device uses the rubber pad to brush against the HMA
specimen surface at constant rotational speed and constant normal pressure. Different
surface friction and texture properties can be produced by subjecting the HMA
specimens to different duration of polishing action that represents the entire lifespan of
109
the pavement surface in the field. Details of the accelerated polishing device can be
found in Chapter III.
4.4 Laboratory Test Results
The laboratory-prepared, gyratory-compacted HMA specimens are subjected to
accelerated polishing using the developed accelerated polishing machine. Typically,
after each one hour of polishing, the specimens are taken out from the polishing
machine for measuring the friction value (i.e., BPN) and texture properties (i.e., MTD)
It should be noted that for each mix type (JMF) studied, a total of three replicate
specimens are prepared and tested to ascertain the repeatability of the test results as
well as to obtain quantitative data for correlation analysis between BPN and MTD.
Appendix B provides the numerical values of the BPN and MTD for each hour of
polishing for all eight hours using the eight different JMFs labelled according to their
polish susceptibility.
4.5 Analysis of Test Results
The conducted analysis of the obtained results is shown below.
4.5.1 Analysis of Repeatability
The repeatability of the test results was examined in this section. For each set of
specimens made of the same JMF, three replicate specimens were tested. The friction
(BPN) values obtained from the BPT and the MTD measured by the sand patch method
from the three replicates are statistically analysed using Homogeneity of Variance
110
(Levene statistic), one-way Analysis of Variance (ANOVA), and Multiple
Comparisons to check for the repeatability of test results. Homogeneity of Variance
and one-way ANOVA are used to check if there is any significant difference between
the variances and the means of at least two specimens for each set of specimens (three
specimens per set) made of the same JMF. Multiple Comparisons, on the other hand, is
used to check if there is any significant difference between the means of different two-
specimen combinations of the three specimens set made of the same JMF. The software
Statistical Package for the Social Sciences (SPSS) was employed for obtaining the
statistical analysis results. Table 4-2 summarizes the statistical analysis results. It can
be seen that the difference between the variances and the means of the results (in terms
of BPT and MTD) for the three replicate specimens is insignificant for all cases when
the friction values (BPN) is considered, and insignificant for the vast majority of the
cases when the macrotexture values (MTD) is considered. Therefore, the repeatability
of the test results is confirmed.
111
Table 4-2: Repeatability tests for the eight different job mix formulas
a. significant at the p-value smaller than 0.05
112
Table 4-2: Repeatability tests for the eight different job mix formulas (continued)
a . significant at the p-value smaller than 0.05
113
Table 4-2: Repeatability tests for the eight different job mix formulas (continued)
a. significant at the p-value smaller than 0.05
114
Table 4-2: Repeatability tests for the eight different job mix formulas (continued)
a. significant at the p-value smaller than 0.05
115
4.5.2 Analysis of Polishing Behavior (BPN)
The polishing behavior in terms of the friction values (BPNs) was analyzed and
presented herein.
4.5.2.1 Rate of Friction Loss (Percent Hourly Drop in Polish Numbers)
The percent hourly drop in BPN is calculated by Equation (4-1).
hournBPNat
hournBPNathournBPNatPNrlyDropinBPercentHou
th
thth
)(
)1()( (4-1)
The percent hourly drop in BPN was calculated for each of the eight mixes. The
average percent hourly drop for the aggregates in each of the three categories (i.e., L,
M, and H) of polish susceptibility is calculated for further analysis.
The calculated average percent hourly drop in BPN was plotted against the polishing
duration in minutes as shown in Figure 4-1, Figure 4-2, and Figure 4-3 for low (L),
medium (M), and high (H) polish susceptibility aggregates, respectively. It can be seen
that the percent decrease in polish number is the highest during the first hour of
polishing. With the passage of time, the drop in BPN decreases up to the sixth hour of
polishing. After 6 hours of polishing, the BPN appears to have reached a constant value
even with additional polishing action (duration). The pattern of the BPN loss over time
may be explained in the following manner. The exceedingly high rate of drop of BPN
values in the first hour of polishing could be attributed to the presence of surface
impurities on the HMA specimen surface. With the passage of time during which
116
further sacrificial polishing occurs, these aggregates are cleaned off. Also certain
angular protrusions on the surface of the aggregates wear off or break off during this
time to reveal a much smoother surface. As a result, the rate of drop of BPN values
decreases to eventually reaching a negligible rate.
The normalized results of measured BPN values are prepared based on the loss of BPN
values at any given time of polishing divided by the maximum loss of BPN values at
the end of eight hours of polishing. Figure 4-4, Figure 4-5, and Figure 4-6 show the
plots of each polish susceptibility level (low, medium, and high). It can be seen from
these curves that they follow a similar trend of polishing with time. Roughly, those
HMA classified as low susceptibility (L) lost 51 to 62% of the total BPN loss in the
first hour of polishing. Those medium polish susceptibility aggregate (M) lost in the
first hour of polishing the 21 to 41% of the total BPN loss. The high polish
susceptibility aggregate (H) lost in the first hour of polishing about 44% of the total
BPN loss.
117
Figure 4-1: Average percent hourly drop in BPN vs. polishing time for low polish
susceptibility aggregates
Figure 4-2: Average percent hourly drop in BPN vs. polishing time for medium
polish susceptibility aggregates
0
2
4
6
8
10
12
60 120 180 240 300 360 420 480
Polishing Time (min.)
Pe
rce
nt
Dro
p in
BP
N (L
ow
Po
lish
ing
)
0
2
4
6
8
10
60 120 180 240 300 360 420 480
Polishing Time (min.)
Pe
rce
nt
Dro
p in
BP
N (
Me
diu
m P
olis
hin
g)
118
Figure 4-3: Average percent hourly drop in BPN vs. polishing time for high polish
susceptibility aggregates
Figure 4-4: Normalization of BPN wrt. the maximum difference in BPN for low
polish susceptibility aggregates
0
2
4
6
8
10
12
14
16
18
60 120 180 240 300 360 420 480
Polishing Time (min.)
Per
cen
t Dro
p in
BP
N (
Hig
h P
olis
hin
g)
(BPNi-BPNt)/(BPNi-BPNf)
0
0.25
0.5
0.75
1
1.25
0 50 100 150 200 250 300 350 400 450 500
Polishing Time (min.)
No
rmal
ized
BP
N L1
L2
L3
119
Figure 4-5: Normalization of BPN wrt. the maximum difference in BPN for
medium polish susceptibility aggregates
Figure 4-6: Normalization of BPN wrt. the maximum difference in BPN for high
polish susceptibility aggregates
(BPNi-BPNt)/(BPNi-BPNf)
0
0.25
0.5
0.75
1
1.25
0 50 100 150 200 250 300 350 400 450 500
Polishing Time (min.)
No
rma
lize
d B
PN
M1
M2
M3
M4
(BPNi-BPNt)/(BPNi-BPNf)
0.00
0.25
0.50
0.75
1.00
1.25
0 50 100 150 200 250 300 350 400 450 500
Polishing Time (min.)
No
rma
lize
d B
PN
H1
120
4.5.2.2 Absolute and Percent Value of Decrease (Initial Polish Number versus Final
Polish Number)
Table 4-3 presents the absolute decrease and the percent decrease in BPN between
initial and final values for different HMA mixes. As expected, for low polish
susceptibility aggregates the percent decrease between initial and final BPN is less than
medium polish susceptibility aggregates which, in turn, is less that high polish
susceptibility aggregates. The same conclusion can be drawn when the average
absolute decrease and the average percent decrease of each polish susceptibility
aggregate (i.e., L1, L2, and L3, and M1, M2, M3, and M4, and H1) is calculated.
Table 4-3: Percent decrease in BPN and MTD between initial and final values
121
4.5.3 Surface Texture Behaviour
The polishing behaviour in terms of the texture values (BPNs) was analysed and
presented herein.
4.5.3.1 Rate of Surface Texture Loss (Percent Hourly Drop in Texture Values)
For each polish susceptibility aggregate, the average percent hourly drop in
macrotexture (MTD) is plotted against the polishing duration in Figure 4-7, Figure 4-8,
and Figure 4-9, respectively. It is observed that the percent decrease in macrotexture is
the maximum during the first hour of polishing. With the passage of time, the drop in
MTD decreases and stabilizes after the fifth hour of polishing. After which, the HMA
surface is said to have reached the residual state that further polishing action will not
significantly reduce the macrotexture. The normalized plot in terms of the loss of
macrotexture at a polishing duration divided by the maximum loss of macrotexture at
the end of eight hours of polishing is presented in Figure 4-10, Figure 4-11, and Figure
4-12 for three polish susceptibility categories. It can be seen from these curves that for
all polish susceptibility categories that they follow a similar trend of MDT versus
polishing duration. In the first hour of polishing for the low polish susceptibility
aggregate, the loss of MTD values is anywhere between 38 to 58% of the total loss. For
the medium polish susceptibility aggregate, the loss of MTD values is anywhere
between 40 to 59% of the total loss. For the high polish susceptibility aggregate, the
loss of MTD values in the first hour is 69% of the total loss.
122
Figure 4-7: Average percent hourly Drop in MTD vs. polishing time for low polish
susceptibility aggregates
Figure 4-8: Average percent hourly drop in MTD vs. polishing time for medium
polish susceptibility aggregates
0
2
4
6
8
10
12
60 120 180 240 300 360 420 480
Polishing Time (min.)
Pe
rce
nt
Dro
p in
MT
D (
Lo
w P
olis
hin
g)
0
2
4
6
8
10
12
14
60 120 180 240 300 360 420 480
Polishing Time (min.)
Pe
rce
nt
Dro
p in
MT
D (
Me
diu
m P
olis
hin
g)
123
Figure 4-9: Average percent hourly drop in MTD vs. polishing time for high polish
susceptibility aggregates
Figure 4-10: Normalization of MTD wrt. the maximum difference in MTD for low
polish susceptibility aggregates
0
5
10
15
20
25
60 120 180 240 300 360 420 480
Polishing Time (min.)
Pe
rce
nt
Dro
p in
MT
D (
Hig
h P
olis
hin
g)
(MTDi-MTDt)/(MTDi-MTDf)
0
0.25
0.5
0.75
1
1.25
0 50 100 150 200 250 300 350 400 450 500
Polishing Time (min.)
No
rmal
ize
d M
TD L1
L2
L3
124
Figure 4-11: Normalization of MTD wrt. the maximum difference in MTD for
medium polish susceptibility aggregates
Figure 4-12: Normalization of MTD wrt. the maximum difference in MTD for high
polish susceptibility aggregates
(MTDi-MTDt)/(MTDi-MTDf)
0
0.25
0.5
0.75
1
1.25
0 50 100 150 200 250 300 350 400 450 500
Polishing Time (min.)
No
rma
lize
d M
TD
M1
M2
M3
M4
(MTDi-MTDt)/(MTDi-MTDf)
0
0.25
0.5
0.75
1
1.25
0 50 100 150 200 250 300 350 400 450 500
Polishing Time (min.)
No
rmal
ized
MT
D
H1
125
4.5.3.2 Absolute and Percentage Value of Decrease (Initial Texture Value versus
Final Texture Value)Table4-4 presents the absolute decrease and the percent
decrease in MTD between initial and final values for different HMA mixes
studied. As expected, the percent decrease between initial and final MTD for
low polish susceptibility aggregates is less than that for medium polish
susceptibility aggregates which, in turn, is less that for high polish
susceptibility aggregates. The same conclusion can be drawn when the
average absolute decrease and the average percent decrease for each polish
susceptibility aggregate is calculated.
4.5.4 Correlation Study between BPN and MTD
A simple linear regression analysis was performed between the corresponding BPN and
MTD values. The regression analysis results shown in Table 4-5 confirm that BPN and
MTD are highly correlated with high coefficient of determination (R2), In addition,
good correlation was observed in Table 4-6 between the change in BPN (BPN) and
the change in MTD (MTD).
126
Table 4-4: Percent decrease in BPN and MTD between initial and final values
127
Table 4-5: Simple linear regression between BPN and MTD
128
Table 4-6: Simple linear regression between BPN and MTD
129
4.6 Summary and Conclusions
In this chapter, the polishing behavior of laboratory-prepared, gyratory-compacted
HMA specimens made of eight different job mix formulas has been studied in terms of
friction values (BPN) and macrotexture data (MTD). In addition, the potential
relationship between BPN and MTD has also been investigated. The conclusions that
can be drawn from this study are summarized below.
It has been observed that the decrease in friction (BPN values) and surface
macrotexture (MTD values) is the maximum during the first hour of polishing in the
accelerated polishing equipment. With the passage of time, the drop in BPN and MTD
decreases and eventually becomes negligible at the 6th hour of polishing. This behavior
is attributed to the presence of surface impurities on the HMA specimen surface. With
the passage of polishing time during which further sacrificial polishing occurs, these
aggregates are cleaned off. Also certain angular protrusions on the surface of the
aggregates wear off during this time to reveal a much smoother surface. As a result, the
rate of decrease in both BPN and MTD becomes smaller with polishing time and
eventually becomes negligible at the 6th hour of polishing action.
The macrotexture (MTD values) of HMA surface was found to be strongly correlated
with the surface friction (BPN values). In addition, good correlation was also found
between change in BPN values (BPN) and change in surface texture (MTD) for each
130
HMA mix as well as for each polish susceptibility category. Therefore, the results
presented in this chapter provide strong quantitative evidence in supporting the strong
interrelationship between the friction and texture properties of the HMA surfaces.
131
CHAPTER V
5. LABORATORY STUDY OF AIR VOID AND TEMPERATURE EFFECTS ON
HMA FRICTION PROPERTIES
5.1 Introduction
Friction is often considered to be governed by the surface characteristics of Hot Mix
Asphalt (HMA) surface; however, some recent studies by Goodman et al. (2006),
Wang and Flintsch (2007), Flintsch et al. (2005), and Luo (2003) have shown that other
factors, such as density (air void) and temperature, may also affect surface friction
properties of HMA surfaces. Pavement skid resistance is generally defined as the
ability to prevent the loss of traction between the tire and pavement surface. Pavement
friction is the result of a complex interplay between two principal frictional force
components: adhesion and hysteresis (Figure 2-1). Although there are other
components of pavement friction (e.g., tire rubber shear), they are insignificant when
compared to the adhesion and hysteresis force components. Thus, friction can be
viewed as the sum of the adhesion and hysteresis frictional forces.
Air voids of the newly laid asphalt pavement layers generally decreases from original
7-8% to 3-4% due to compaction by traffic. Therefore, the possible magnitude of
friction change as the air void decreases by 50% over the life span of a pavement needs
132
to be investigated. Furthermore, surface temperature may vary greatly (e.g., from -11.1
°C to 50 °C or 12 °F to 122 °F) during a life span of the pavement, there is a need to
quantify the temperature effect on the measured friction values as well.
This chapter presents the controlled laboratory test results for quantifying the effects of
air void and temperature on the measured friction properties of the gyratory compacted
HMA surfaces. The laboratory test procedures, including the materials used in
preparing the HMA specimens, the method used to polishing the HMA surface, and the
friction measurement techniques, are described in detail. Statistical analysis was used
to determine the significance of these two variables (air void and temperature) on the
measured friction values. Finally, the method for extrapolating the friction values
measured at any density and temperature to the friction values at other density and
temperature is proposed at the end of the chapter.
5.2 Background
In the United States, the most common method of measuring the skid resistance of a
pavement surface is by means of the skid trailer (Luo 2003). The standard test
procedure for the Locked Wheel Skid Trailer (LWST) is described in detail in the
ASTM E 274 specification. The test begins with the skid trailer reaching the desired
test speed; usually 64 km/h (40 mph). An activator, located inside the truck, is used to
initiate the test sequence by starting with spraying a thin layer of water to the pavement
surface. After a correct amount of water has been applied, the test wheel is locked
(usually the left tire) and instrumentation in the trailer records the sliding force of the
133
locked tire. This test allows for the computation of the skid number (SN) by dividing
the tractive force applied to the tire by the vertical load applied to the tire.
A different method for friction measurement that can be conducted on pavement
surface or on laboratory prepared HMA surface is the British Pendulum Tester (BPT)
specified in ASTM E 303. Recent studies by Goodman et al. (2006) and Bazlamit and
Reza (2005) have utilized BPT for their respective friction studies. Although British
Pendulum Numbers (BPNs) and SNs do not correspond exactly; nevertheless, Kissoff
(1988) has developed an approximated relationship that could be used to relate BPN to
SN. Therefore, the BPN values studied in this chapter could be extrapolated to the SN
values based on Kissoff correlations.
In recent years, a few notable research efforts have been directed toward a better
understanding of the influencing factors on the HMA surface friction properties. In
2006, Goodman et al. reported that initial field BPN can be correlated with the
following variables: fineness modulus (FM), voids in mineral aggregate (VMA),
percent passing the 4.75mm sieve (P4.75) and bulk relative density (BRD).
Interestingly, the bulk relative density was found to significantly affect the measured
BPN values.
There have been some recent research efforts toward quantifying the temperature
effects on the measured pavement friction values. For example, Runkle and Mahone
(1980), Burchett and Rizenbergs (1980), and Bazlamit and Reza (2005) have found that
an increase in temperature can result in a corresponding decrease in skid resistance of a
134
pavement surface. Despite the significant number of reports cited above to indicate the
significant effects of temperature on pavement surface friction values, there are
contradictory findings reported by Dahir et al. (1979) and Mitchell et al. (1986) as well.
It was concluded that temperature variations do not seem to significantly affect the skid
resistance measurements. A comprehensive summary of the studies focused on the
temperature effect on HMA frictional properties is presented in Table5-1 It should be
noted that Ta, Tp, Tt, and Tw stand for air temperature, pavement temperature, tire
temperature, and water temperature, respectively.
135
Table 5-1: Summary of the studies focused on the temperature effect on HMA
frictional properties
Based on the review presented, it seems that the effect of air void on surface friction is
qualitatively understood. However, a quantitative assessment has not been reported in
the literature. Regarding the temperature effects on surface friction, the contradictory
conclusions found in the reports may need further clarification. As the effects of
temperature on the measured pavement skid resistance can be broken down into four
components: air temperature, water temperature, tire temperature, and pavement
136
temperature, it could be argued that a controlled laboratory test program may provide
quantitative temperature effects that could not be readily obtained from field study. The
contribution of this chapter lies in providing the controlled laboratory study results to
augment existing knowledge on the effects of air void and temperature on the friction
properties of the compacted HMA surface.
5.3 Laboratory Testing Program
The testing program of the laboratory experiments is shown herein.
5.3.1 Materials
The aggregate used in this study was a limestone with a gradation curve shown in
Figure 5-1. The asphalt binder used was PG 64-22. Based on the Superpave
Specifications adopted by Ohio Department of Transportation, an optimum binder
content of 6.1% was used to compact the HMA specimens. To achieve the desirable air
void, the number of gyrations was varied in preparing the test specimens for friction
measurement.
5.3.2 Test Program
The test program consists of varying the air voids and temperature as test variables.
Each test variable condition (either air void or temperature) is tested in triplicate
specimens to ensure data repeatability. The effect of air void was assessed by preparing
gyratory compacted specimens (6 inches in diameter) at different densities (air voids)
by compacting the specimens to different numbers of gyrations. The range of air voids
137
studied in the test program is 0.8% to 5.4%. The effect of temperature was studied by
varying the temperature of the test specimen, the rubber pad of the friction measuring
device (BPT), and the spraying water. The temperature range studied in this test
program is 4.4 °C (40 °F) to 48.9 °C (120 °F).
Figure 5-1: Gradation curve
5.3.3 Accelerated Polishing Machine
The present study utilizes a laboratory-scale accelerated polishing machine to polish
the HMA specimens to mimic different stages of the actual pavement surface under
traffic induced polishing and wearing. The accelerated polishing machine uses the
rubber pad to brush against the HMA specimen surface at constant rotational speed and
constant normal pressure (refer to Chapter III). Different surface friction and texture
properties can be produced by polishing the initially compacted HMA specimens to
different durations of polishing as was demonstrated in Chapter III. The accelerated
0
20
40
60
80
100
120
0.001 0.01 0.1 1
Grain size (in)
Pe
rce
nt
pa
ss
ing
138
polishing machine allowed the polishing of the laboratory prepared HMA specimens to
different stages of polishing action, thus mimicking the actual pavement surface under
traffic polishing over the entire life span of the pavement surface.
5.3.4 Friction Measurement Method
The measurement of surface friction values of HMA specimens follows the procedures
outlined in ASTM E 303 for the BPT. It should be noted that five measurements were
made for each specimen, from which an average of the last four readings was recorded
as the BPN.
5.4 Test Results and Analysis
The test results and data analysis of the air void and temperature effects on HMA
frictional properties are discussed in the following sections.
5.4.1 Air Voids Effects
For the effects of air voids, the friction measurement was made for the HMA
specimens compacted to three different air voids (0.8%, 2.8%, and 5.4%) and polished
to three different polishing stages: 0 minutes (initial unpolished stage), 240 minutes
(partially polished stage), and 480 minutes (completely polished stage). These air voids
were chosen to cover a wide range of realistic densities in pavement surface during the
life span of the pavement.
139
The variation of friction with air void in the unpolished, partially polished, and
completely polished conditions is plotted in Figure 5-2. At each air void and polishing
stage, the average from 12 readings (i.e., 3 specimens x 4 repeated readings) is plotted.
It can be seen that the friction (BPN) value increases with an increase in air void at all
polishing stages.
Figure 5-2: BPN vs. polishing time at different air voids
A set of statistical analysis was conducted using the Statistical Package for the Social
Sciences (SPSS) windows based program, including homogeneity of variances test,
One-Way Analysis of Variance (ANOVA), and Post Hoc tests. Levene's test is used to
test if n samples would have equal variances (homogeneity of variance). The analysis
of variance assumes that variances are equal across groups or levels. The Levene test is
intended to verify the validity of that assumption. One-Way ANOVA is used to
compare the means of several populations. Post Hoc test is used to evaluate whether the
Pavement Density Effect on BPN
45
50
55
60
65
70
75
80
0 100 200 300 400 500
Polishing Time (min.)
BP
N
at 5.4% air void
at 2.8% air void
at 0.8% air void
140
levels or groups within the factor are significantly different or not. It can be performed
for factors with three or more levels or groups (Kutner et al. 2004). Table 5-2 provides
a summary of statistical analyses performed to support and validate that the effect of
changing HMA air voids on the measured friction values at three polishing stages is
significant. From Table 5-2 under the column of homogeneity of variances, it can be
seen that variances are not significantly different (i.e., equal variances). It is also
evident from the One-Way ANOVA table that the difference between means is
significant. Finally, it can be seen that the mean difference between any two groups is
significant for most of the cases, as seen from the column with the heading of multiple
comparisons. It is noted that the column “Group” under “Multiple Comparisons” in
Table 5-2 refers to the different air voids (AV) used; in other words, I denotes AV of
0.8%, II denotes AV of 2.8%, and III denotes AV of 5.4%. All observations are made
at the 0.05 significance level.
As reviewed in the background section of this chapter, the LWST test is conducted on
the actual pavement surface to monitor the friction values in terms of SN. Therefore, a
useful equation is developed herein to enable extrapolating SN obtained at a given air
void to the SN at other air voids Toward this goal, a linear curve fit is developed and
shown in Figure 5-3, where BPN is taken from the intermediate (partially) polished
state (after 240 minutes of polishing) that is typical of a pavement that has been in
service. The value of the coefficient of determination, R2, is 0.9967. The slope of the
fitting line is 1.4192. Thus, one can deduce the following general equation that relates
the BPN at any other air void (BPNAV) to the BPN at measured air void (BPNMAV)
141
)5(419.1 AVBPNBPN MAVAV (5-1)
The relationship between SN and BPN, such as the one proposed by Kissoff (1988)
given in Equation 5-2, can be used to convert Equation 5-1 into Equation 5-3.
690.9)(862.0 BPNSN (5-2)
)5(223.1 AVSNSN MAVAV (5-3)
Equation 5-3 can be used for obtaining the SN at any air void (SNAV) given the
SN at the measured air void (SNMAV) and vice versa.
142
Table 5-2: Test of Homogeneity of variances, 1-Way ANOVA Table, and Multiple
Comparisons for the Effect of HMA Air Void on BPN
143
Figure 5-3: BPN vs. air voids
5.4.2 Temperature Effects
For the temperature effect, the test results obtained from the laboratory work include
friction (BPN) at three different temperatures: 40, 75, and 140 F , and three stages of
polishing. The variation of temperature was achieved by placing the HMA specimen,
the rubber slider of the BPT, and water in the oven.
The variation of friction (BPN) values with temperatures in the unpolished, partially
polished, and completely polished conditions is plotted in Figure 5-4, in which HMA
specimen, sliding rubber pad, and water temperatures are controlled. It can be seen that
the BPN values decrease with an increase in temperature.
y = 1.4192x + 50.742
R2 = 0.9967
51
52
53
54
55
56
57
58
59
0 1 2 3 4 5 6
Air Voids (%)
BP
N
144
Figure 5-4: BPN vs. polishing time at different pavement, rubber slider, and water
temperatures
A set of statistical analysis was conducted similar to the analysis carried out for the air
void effects. The statistical analysis results of temperature effects on BPN values are
summarized in Table 5-3. It can be seen that variances are not significantly different
(i.e., equal variances). It can also be seen from the One-Way ANOVA analysis that the
difference between means is significant. Finally, the mean difference between any two
groups is significant, as can be seen from the multiple comparisons column in the table.
A linear curve fit between the BPN and test temperature is shown in Figure 5-5, where
BPN is taken at an intermediate (partially) polished state. The value of the coefficient
of determination, R2, is 0.9951. The straight line of the best fit has a slope of -0.042.
45
50
55
60
65
70
75
80
0 100 200 300 400 500
Polishing Time (min.)
BP
Nat 40 Fat 75 F
at 140 F
145
Thus, a linear relationship as given in Equation 5-4 can be used to reflect the
temperature effect on the BPN values
)68(042.0 TBPNBPN MTT (5-4)
where BPNT is the BPN at any temperature and BPNMT is the BPN at the measured
temperature. Using the same correlation equation (Equation 5-2), Equation 5-4 can be
converted into Equation 5-5 for modifying the SN values for the temperature effect
)68(036.0 TSNSN MTT (5-5)
where SNT is the value of skid number at any temperature and SNMT is the value of
skid number at the measured temperature.
146
Table 5-3: Test of Homogeneity of variances, 1-Way ANOVA Table, and Multiple
Comparisons for the Effect of Pavement, Rubber Slider, and Water Temperatures on
BPN
Figure 5-5: BPN vs. temperature
y = -0.042x + 62.069
R2 = 0.9951
55
56
57
58
59
60
61
0 20 40 60 80 100 120 140 160
Temperature (F)
BP
N
147
5.5 Summary and Conclusions
This chapter presented the test results of a laboratory test program to investigate the
effects of the air void and temperature on the HMA surface friction properties. The
surface friction values were determined using the British Pendulum Tester, while an in-
house accelerated polishing machine was used to polish the laboratory prepared Hot
Mix Asphalt specimen surfaces to mimic different stages of actual pavement surface
during the life span of the pavement. The laboratory HMA specimens were prepared
using the gyratory compactor with different number of gyrations to achieve different air
void of the test specimens. The temperature of the test specimens and the rubber slider
of the BPT, as well as the spraying water temperatures were carefully controlled in the
laboratory test program. The laboratory test results were analyzed statistically to
ascertain the significance of each test variable (air void and temperature) on the
measured friction values. Finally, a linear regression analysis of test data has yielded
useful equations for extrapolating the measured BPN or SN values at a given air void
or temperature to other different air void and/or temperature. Specific conclusions from
this chapter are enumerated below.
The effect of air voids on the measured HMA frictional properties (BPN values) was
found to be statistically significant at the 0.05 significance level. Basically, there was
an increase in friction (BPN values) corresponding to an increase in air void.
For the air void difference between 3.5% and 7.5%, the corresponding difference in
BPN is about 5.7 and the corresponding difference in SN is 4.9. Therefore, the
148
extrapolation relationships given in Equations 5-1 and 5-3 are recommended to correct
the measured BPN or SN for the desired air void other than the one during the
measurement.
The effect of temperature of the HMA surface, the rubber slider of BPT, and the
spraying water on the measured HMA friction (BPN values) was found to be
statistically significant. Essentially, there was a decrease in friction corresponding to an
increase in temperature.
For temperature difference between -11.1 °C to 50 °C (12 °F to 122 °F), the
corresponding difference in BPN is about 4.6 and the corresponding difference in SN is
4.0. Therefore, the extrapolation relationships given in Equations 5-4 and 5-5 are
recommended to correct the measured BPN or SN for the desired temperature other
than the one during the measurement.
149
CHAPTER VI
6. CORRELATION STUDY BETWEEN FRICTION MEASUREMENTS BY
LWST AND DFT
6.1 Introduction
As a part of a safety monitoring program, pavement surfaces require frequent
measurement of friction by means of some types of devices and standard procedures.
Among the widely used friction measuring devices and the corresponding standardized
procedures are the Locked Wheel Skid Trailer (LWST) and Dynamic Friction Tester
(DFT). The LWST (ASTM E 274) can only be used on actual pavement surface due to
the nature of operation. The DFT (ASTM E 1911), on the other hand, is a portable
device and therefore can be used on both field pavement surface as well as in the
laboratory prepared Hot Mix Asphalt (HMA) surface that is 0.6 m by 0.6 m (2 ft by 2
ft) surface area. The LWST is carried with either ribbed or smooth tire with skid
number measured at the speed of 64 kmph (40 mph). The DFT utilizes the standardized
rubber pads to measure the friction values at the speed ranging from 0 to 90 kmph (0 to
55 mph). There is a growing trend among the highway engineers to seek possible
correlations between the friction values measured by the LWST and the DFT so that
150
the findings obtained in the laboratory study through DFT can be related to actual
pavement performance measured through both LWST and DFT.
If the correlation can be developed accurately between the SN measured by LWST and
the DFT measured friction values, some advantages may be realized. For example, we
could use the portable Dynamic Friction Tester for pavement friction monitoring while
enjoying the advantages of DFT (i.e., easy to use, quick, accurate, and portable).
Additionally, the correlation between laboratory measured DFT and field measured SN
would allow for better interpretation of laboratory test results by relating them to field
observed performance in terms of SN values. Finally, the DFT provides friction values
at both high speed and low speed range, i.e., the friction at high speeds (DFT64, where
64 means 64 km per hour) is known to reflect macrotexture effect, while DFT friction
values at low speeds (DFT20) is known to reflect microtexture effect. One additional
advantage of the DFT device is that a companion texture measurement device (Circular
Texture Meter) can be used to measure the Mean Profile Depth (MPD) of the same
surface area of the DFT measurement.
Researchers have developed correlations between the friction properties measured by
different friction measuring devices. For instance, Wambold et al. (1998) developed a
regression equation between Friction Number (FN) measured by the DFT when the slip
speed is 20 kmph (12.5 mph) and British Pendulum Numbers (BPNs); the coefficient
of determination was found to be 86%. Henry et al. (2000) determined the regression
equation of the friction number F60 calculated from MPD and DFT measurement
151
values (i.e., DFT20, at a slip speed of 20 km/hr and DFT60, at a slip speed of 60
kmph). The resulting regression equation is found accurate provided that pavements
have Mean Profile Depth (MPD) in the range of 0.2 - 2.2 mm. Also, there has been
work to correlate the skid numbers (SN) measured by the LWST with the friction
numbers measured using the DFT (Wambold et al., 1995).
The main objective of this chapter is to present the measurements, compiled by the
authors that consist of the SN, DFT friction numbers and MPD measured by the CTM
at the same time for the same surface. From these measured data, a statistical study
was conducted to develop the relationship to predict the skid number at 64 kmph (40
mph) using the ribbed tire Locked Wheel Skid Trailer (SN(64)R) from one or the
combination of the following three measurements: the friction number at 64 kmph (40
mph) using the Dynamic Friction Tester (DFT64), the friction number at 20 kmph
(12.5 mph) measured by the Dynamic Friction Tester (DFT20), and the mean profile
depth measured by the Circular Texture Meter (MPD).
6.2 Friction and Texture Measurement Techniques in This Study
In this chapter, the techniques used for measuring friction are the locked wheel skid
trailer (ASTM E 274) using the ribbed tire and the dynamic friction tester according to
ASTM E 1911. On the other hand, the circular texture meter (ASTM E 2157) was used
to measure the macrotetexture (MPD) of the pavement surfaces.
152
6.3 Experimental Program
The experimental program is designed to conduct field work on several
selected pavement sections throughout the state of Ohio. The pertinent
information of the selected pavement sections for this study is summarized
in Table 6-1, which include information such as aggregate polish
susceptibility (extracted from previous study by Liang and Chyi, 2000) and
aggregate source used in each pavement section, the location of the tested
pavement section such as route and section milemarker, and the number of
data points obtained so far from each pavement section, among other
information. Each data point represents one set of measurement, which
consists of the skid number measured using the LWST, the friction number
measured using the DFT, and the mean profile depth measured using the
CTM. In general, all field measurements are taken on the left wheel path.
The DFT and CTM measurements are the average of two runs on the left
wheel path.
153
Table 6-1: HMA Pavement sections identification
Polish Susceptibility
Aggregate Source DistrictLocation:
Route (Section)
No. of Measurements
Data Collected in 2006
Existing Pavement Sections
Possible High Polish (Gravel)
Chesterville @ Stockport
10 007 (37.3-
39.0) 8
Possible Medium Polish (Limestone)
Sandusky Crushed @ Parkertown
3 250 (3.55-
5.11) 6
Possible Medium Polish (Dolomite)
Stoneco @ Maumee 2 025(15.68-22) 40
Possible Low Polish (Gravel)
Martin Marietta @ Apple Grove
11 250 (22.5-
25.5) 10
Data Collected in 2007
Existing Pavement Sections
Possible High Polish (Gravel)
Chesterville @ Stockport
10 007 (37.3-
39.0) 8
Possible Medium Polish (Limestone)
Sandusky Crushed @ Parkertown
3 250 (3.55-
5.11) 6
Possible Medium Polish (Dolomite)
Stoneco @ Maumee 2 025(15.68-22) 36
Possible Low Polish (Gravel)
Martin Marietta @ Apple Grove
11 250 (22.5-
25.5) 10
New Pavement Sections
Possible Medium Polish (Limestone)
Sandusky Crushed @ Parkertown
3 162 (14.00-
19.00) 18
Possible Low Polish (Gravel)
Stocker Sand & Gravel @
Gnadenhutten 11 022(5.00-8.00) 12
Possible Medium Polish (Dolomite)
Stoneco @ Maumee 2 064 (8.90-
12.40) 14
Low Polish (Trap Rock)
Ontario Trap Rock @ London
12 090 (28.25-
29.21) 6
154
Data Collected in 2008
Existing Pavement Sections
Possible High Polish (Gravel)
Chesterville @ Stockport
10 007 (37.3-
39.0) 8
Possible Medium Polish (Limestone)
Sandusky Crushed @ Parkertown
3 250 (3.55-
5.11) 6
Possible Medium Polish (Dolomite)
Stoneco @ Maumee 2 025(15.68-22) 36
Possible Low Polish (Gravel)
Martin Marietta @ Apple Grove
11 250 (22.5-
25.5) 10
New Pavement Sections
Possible Medium Polish (Limestone)
Sandusky Crushed @ Parkertown
3 162 (14.00-
19.00) 18
Possible Low Polish (Gravel)
Stocker Sand & Gravel @
Gnadenhutten 11 022(5.00-8.00) 12
Possible Medium Polish (Dolomite)
Stoneco @ Maumee 2 064 (8.90-
12.40) 14
Low Polish (Trap Rock)
Ontario Trap Rock @ London
12 090 (28.25-
29.21) 6
The sequence of field work at each selected pavement section generally involves three
tasks. The first task involves the use of locked wheel skid trailer to measure the skid
number at 64 kmph (40 mph) using the ribbed tire. The second task involves the
measurement of the MPD, at the same spot of skid number measurement, using the
CTM. The third task involves the measurement of the friction number using the DFT
over the speed range of 0 to 90 kmph (0 to 55 mph). In carrying out the field work at
155
each selected pavement section, the authors were careful to ensure that the same spots
(as close as possible) on each pavement section were used for all three measurements.
Further, to minimize the short-term (weather-related) and long-term (traffic-related)
effects on the three measurements, they were all conducted at about the same time.
6.4 Field Test Results and Data Analysis
A summary of the field-test results and data analysis is shown herein.
6.4.1 Typical LWST, DFT, and CTM Test Results
A sample of SN measured by the LWST for one pavement section is shown in Table
6-2 . It can be seen that general information about the location, milemarker, speed at
which the test was conducted, test date, pavement condition at the time of the test,
temperature, and the direction are provided. A sample output of a DFT run is shown in
Table 6-2(a) . The coefficient of friction can be obtained from the DFT software at the
desired speed. The sample output of one CTM run is shown in 6-2(b) with MPD in mm
plotted on the y-axis. The SN, DFT friction values at different speed and the MPD
obtained at the selected pavement sections are used in the subsequent data analysis to
establish appropriate predictive equations for SN.
156
Table 6-2: Sample of skid numbers measured using LWST for one pavement
section
a)Dynamic Friction Tester output
157
(b) Circular Texture Meter output
Figure 6-2: Dynamic Friction Tester and Circular Texture Meter outputs
(continued)
6.4.2 Analysis of Test Results
In order to develop correlations between the variables studied it is reasonable to start
with simple linear regression analysis to assess any noticeable relationship between the
individual variables and to get rid of any possible multicollinearity (linear dependency
between individual variables). Subsequently, multiple linear regression analysis was
158
performed along with some diagnostics for capturing outliers and potential influential
data points. In summary, the statistical data analyses performed include the following
five sets of analysis.
Simple linear regression analysis was performed among the following variables: SN
measured using LWST at 64 kmph (SN(64)R), FN measured using the DFT at 64
kmph (DFT64), FN measured using the DFT at 20 kmph (DFT20), and MPD measured
using the CTM (MPD). DFT64 was thought to reflect the macrotexture effect while
DFT20 was to account for microtexture effect (Wambold, 1995).
Multicollinearity analysis was carried out to detect any variable interdependency.
Multiple linear regression analysis was carried out in order to predict SN(64)R from
the following combinations of variables: (a) MPD and DFT20, (b) MPD and DFT64,
(c) DFT20 and DFT64, and (d) MPD, DFT20 and DFT64.
Problem points diagnostic for detecting any outliers and potential influential points was
also conducted.
Normality check and constant variance check on the residuals (i.e., measured -
predicted) of the dependent variable were performed to ensure that assumptions of a
normal distribution and constant variance were valid for regression analyses.
The software Statistical Package for the Social Sciences (SPSS) was used to carry out
the statistical analyses.
159
It should be noted that in presenting the statistical analysis results, not only the
coefficient of determination (R2) is reported, but also the adjusted coefficient of
determination (R2a) value. Readers can consult a typical statistical textbook for the
definition of R2 and R2a.
6.4.2.1 Simple Linear Regression
The obtained simple linear regression models for the identified variables, along
with the corresponding R2 and the R2a , are summarized in Table 6-3 and shown in
Fifure6-2(a),Figure 6-2(b),and Figure 6-2(c). It can be seen that there is a linear
relationship between the dependant variable and the independent variables. This
observation is further validated by the Analysis of Variance (ANOVA) that is used
to assess the usefulness of a model presented in Table 3, specifically by the F-value
(i.e., as F goes up, P goes down, thus indicating more regression confidence in that
there is a difference between the two means) and the P-value (the probability of
getting a value of the test statistic as extreme as or more extreme than that observed
by chance alone, if the null hypothesis Ho, is true) based on the 0.05 significance
level.
The SN(64)R is found to be significantly correlated to DFT64 (R2 = 63%). This could
be explained by the elimination of the speed effect when using the same speed in both
friction measurements, i.e. at 64 kmph (40 mph). The observed low R2 between
SN(64)R and MPD could be attributed to the fact that friction measurement by the
ribbed tire is insensitive to the macrotexture represented by MPD. The relatively higher
160
R2 between SN(64)R and DFT20, which can be viewed as an indirect measurement of
the microtexture, can be attributed to the fact that the ribbed tire is more sensitive to
microtexture than to macrotexture.
Table 6-3: Simple linear regression between SN(64)R and DFT64, DFT20, and
MPD
Correlation Model R2 (%)
R2a
(%) ANOVA Table
F P-valuea SN(64)R vs.
MPD SN(64)R=46.334+7.8
56 MPD 7.8 7.4 23.239 0
SN(64)R vs. DFT20
SN(64)R=27.174+0.441 DFT20 50.9 50.8 286.472 0
SN(64)R vs. DFT64
SN(64)R=15.460+0.726 DFT64 63 62.9 470.126 0
Figure 6-1: Simple linear regression(continued)
161
Figure 6-1: Simple linear regression
162
6.4.2.2 Multicollinearity
Prior to any multiple linear regression analysis, multicollinearity needs to be checked.
Multicollinearity can be a problem when test of significance for regression coefficient
is run. Three multicollinearity diagnostics that can be run using the SPSS are:
tolerance, Variation Inflation Factor (VIF), and condition index. Multicollinearity is
considered a problem when tolerance is less than 0.1, VIF is greater than 10, and/or
condition index is greater than 30. From analysis presented in Table 6-4 and the above
mentioned criteria, one can conclude that multicollinearity is not a problem for the data
set examined herein.
Table 6-4: Multicollinearity check using Tolerance, VIF, and Condition Index on
the independent variables
Model Collinearity Statistics
Tolerance VIF
MPD 0.900 1.11
DFT20 0.146 6.867
DFT64 0.150 6.687
6.4.2.3 Multiple Linear Regression
Several multiple linear regression analyses were performed. The first multiple linear
regression was carried out in order to predict SN(64)R from MPD, DFT20, and DFT64
using the entire data set. The outcome of the multiple linear regression is shown in the
following regression equation:
163
SN(64)R=13.748+2.216(MPD)-0.95(DFT20)+0.836(DFT64) (6-1) Where
SN(64)R= skid number measured using LWST at 64 kmph (40 mph) with ribbed tire
DFT64= friction number measured using the dynamic friction tester at 64 kmph (40
mph)
DFT20= friction number measured using the dynamic friction tester at 20 kmph (12.5
mph); and
MPD= mean profile depth measured using the circular texture meter (mm)
R2 and R2a were found to be 63.8% and 63.4%, respectively.
Table 6-5: Multiple linear regression between MPD and DFT64, MPD and DFT20,
and DFT20 and DFT64
Correlation Model R2 (%)
R2a
(%)
ANOVA Table
F P-valuea
SN(64)R vs. MPD and DFT20
SN(64)R=26.672 +1.726 MPD+0.429 DFT20
51.3 50.9 144.661 0
SN(64)R vs. MPD and DFT64
SN(64)R= 15.014+ 1.921MPD+0.709
DFT64 63.4 63.2 238.583 0
SN(64)R vs. DFT20 DFT64
SN(64)R=14.517 -0.075 DFT20+0.828 DFT64
63.2 63.4 236.453 0
a. Dependant Variable :SNLWST
164
6.5 Summary and Conclusions
The major objective of this chapter was to present a comprehensive set of data
collected by the authors and the accompanied statistical analyses of the data set to
develop predictive equations for SN from other variables, such as DFT and MTD. The
specific conclusions that can be made from the statistical analysis are summarized as
follows:
Skid numbers (SN(64)R) measured using LWST were correlated to friction numbers
measured using the DFT at different speeds. It was observed that SN(64)R could be
significantly correlated to DFT64. This could be explained by the fact that both
frictions were measured at the same measurement speed, thus eliminating the potential
of speed effect. In the same token, the low coefficient of determination observed for the
relationship between SN(64)R and DFT20 can be due to the difference in the
measuring speed.
The coefficient of determination between SN(64)R and MPD was found to be low. It
could be due to the fact that friction measurement by the ribbed tire of the LWST is
insensitive to the macrotexture. On the other hand, the coefficient of determination
between SN(64)R and DFT20 was high, suggesting that the ribbed tire is more
sensitive to microtexture than macrotexture, since DFT(20) reflects the effect of
microtexture.
165
Several regression equations were examined for predicting SN(64)R from (a) MPD and
DFT20, (b) MPD and DFT64, (c) DFT20 and DFT64, and (d) MPD, DFT20, and
DFT64. However, it was found that MPD did not add much to the regression model,
neither did the DFT20 (see conclusions 1 and 2). Therefore, SN(64)R can be directly
predicted from the DFT64 as shown in Table 6-3. For slightly higher accuracy
SN(64)R can be predicted from MPD, DFT20, and DFT64 as shown in equation 6-1.
166
CHAPTER VII
7. POLISHING MACHINE BASED ON HIGH-PRESSURE WATER JET
7.1 Introduction
The design concept of the first accelerated polishing device using rubber shoes
(discussed in details in Chapter III) is based on the Ohio Department of Transportation
requirements; that is, time efficient (test duration should be no longer than one day for
test to be completed), testing Hot Mix Asphalt specimens rather than just aggregate
samples (thus, providing more authentic indicator of the performance of HMA in the
field), ease of operation, and repeatability of test results. Based on the above set of
criteria, one version of accelerated polishing equipment (i.e., using the rubber pads as
polishing devices) was developed and verified in the previous chapters. The equipment
is based on a simple concept to duplicate the tire-pavement interaction in a laboratory
setting. The deign of the polishing equipment allows for pressing polishing shoes made
of Styrene-Butadiene-Rubber (SBR) onto the surface of the HMA specimen at a
constant vertical force while rotating the rubber shoes at a constant rotational speed. It
should be noted that the polishing device is designed to accommodate two specific
specimen dimensions: an 18 inch by 18 inch by 2 inch high roller-compacted slab
specimen or a 6 inch diameter by 4 inch high Superpave gyratory-compacted
167
specimen. As a result of different specimen sizes, the rubber shoes are designed
differently. For the slab specimen, a rubber ring of approximately 13 inch in outside
diameter and 9 inch in inside diameter is used to match with the area needed for the
dynamic friction tester (DFT) and circular texture meter (CTM). For the gyratory
compacted specimen, a solid rubber disk of 6 inch in diameter and 1.5 inch thick is
used. A means was provided to allow for controlled feeding of water onto the surface
of the specimen. A timing device was used to time the duration of the
polishing/abrasion action. Details of the design of the equipment and validation of the
equipment operation and test results on different HMA mixes are fully documented in
previous chapters.
In addition to the mechanical polishing device developed and studied in this
dissertation, a completely new polishing concept using the high pressure water jet is
also explored in this study. The use of a high pressure water jet in polishing HMA can
be traced to the work conducted jointly by the French LRPC (Laboratoir Ponts et
Chaussees) and Quebec MTQ (Ministere des Transport du Quebec). In their work, the
machine referred to as “GRAP” was shown to achieve the polishing action by
projecting a stream of water and very fine abrasive agent under pressure (around 1450
psi) with a given angle of incidence. Figure 7-1 provides a photograph of the prototype
machine in the laboratory. The test method is well described in Delalande (1992). The
polishing concept of the GRAP polishing machine is illustrated schematically in Figure
7-2. The water supply should be around 11.7 litters per minute (3 gallons per minute) .
The surface is swept by displacement of the projection nozzle due to a cross motion
168
table. Polishing is obtained after twenty sweep cycles (2 hours and 45 minutes
including preparation of specimens and friction measurements; however, the actual
polishing time is 45 minutes) of rectangular shape. The machine is composed of the
projection housing, volumetric powder measurer, generator for water under pressure,
water-abrasive projection nozzle, cross motion table, retrieval tray, and electric control
panel with programmable controller. Results from Delalande (1992) showed that the
limit polishing states achieved by the GRAP is comparable to that achieved by the
Accelerated British Polishing Equipment. Do et al (2001) also reported similar
reasonable test results form GRAP. The success reported by Delalande (1992) and Do,
et al (2001) provide inventive for an independent investigation of the water jet based
approach to polish HMA surface. The design of high pressure water jet based polishing
equipment and its fabrication conducted in this study is reported in this chapter,
together with some preliminary test results for assessing its applicability for HMA
surface polishing. This chapter also provides preliminary findings concerning the test
variables, such as the rotational speed, the water jet pressure, the abrasive agent used,
and the impact angel that were experimentally investigated in this study.
It is very important to note that the GRAP polishing equipment was designed to polish
aggregate specimens. Specimens are 100 mm by150 mm (4 inch by 6 inch) rectangular
plates. These specimens are made ofcoarse aggregates of similar sizes and fixed in a
resinmatrix, as shown in Figure 7-3. The new device developed in this study, however,
aims at polishing HMA surface.
169
Figure 7-1: GRAP polishing machine
Figure 7-2: Schematic depiction of GRAP polishing machine concept
170
Figure 7-3: GRAP aggregate specimen
7.2 Equipment Development
In the following section, equipment description and operational procedure are
discussed in detail.
7.2.1 Equipment Description and Operational Procedure
The guiding principle of the laboratory-scale accelerated polishing equipment using a
high-pressure water jet is that the evolution history of friction loss of the HMA surface
can be simulated and measured in a realistic short test duration. The deign of the
polishing equipment allows for projecting high-pressure water (around 1450 psi) and
fine abrasive agent onto the specimen surface at the desired angle (usually 40 degree).
A picture illustrating the proposed concept of the high-pressure water jet accelerated
polishing machine for HMA testing is shown in Figure 7-4. It should be noted that the
polishing device is designed to accommodate two specific specimen dimensions: an 18
171
inch by 18 inch by 2 inch high roller-compacted slab specimen or a 6 inch diameter by
4 inch high Superpave gyratory-compacted specimen. As a result of different specimen
sizes, the spray nozzle can polish different patterns (suitable for the different friction
and texture measuring devices) as the rotary platform rotates different specimens with
different dimensions. Therefore, the friction and texture properties of the HMA
specimen surface are determined by two approaches for the two specimen sizes: (a) the
dynamic friction tester and circular texture meter for 18 inch by 18 inch slab
specimens, and (b) British pendulum tester and sand patch method for 6 inch in
diameter gyratory compacted specimens.
Figure 7-4: Schematic depiction of the concept of the high-pressure water jet
polishing machine using HMA specimens
The high pressure asphalt polisher includes the following equipment components. The
tester includes a tubular steel frame, approximately 77 inch by 52 inch by 32 inch high,
172
to support a drum deck and a top deck. A rotary deck is inside the drum. The top deck
is double hinged which allows the deck to be lifted for access to the rotary deck. The
test specimens are placed and registered on the rotary deck for testing. The rotary deck
is inside a 45.5 inch diameter cylindrical drum that is used to help contain the water
and debris spray. The water spray nozzle is hosed inside the protection drum. Floor
levelers with antiskid pads are included for positioning the machine. A winch type
system is used to raise and lower the samples onto the rotary deck. The polisher is
equipped with a high pressure pump/5HP motor assembly, rated for 2000 PSI max, and
set at approximately 1450 PSI with 11.7 litters per minute (3.0 gallons per minute) of
water flow. The water pressure can be adjusted through a pressure gage. A nozzle
which sprays a 2 inch minimum fan type pattern and draws grit into the spray is used.
The nozzle is supported in such a way that it can be adjusted for use for either the 18
inch square specimen or the 6 inch in diameter specimen. A 56.7 litter (15 gallon) grit
tank and grit suction system is provided as well. The amount of grit mixed into the flow
can be adjusted as needed. An auxiliary spray nozzle is provided to washout residual
grit from the tank. A 2 inch NPT drain with fittings and a 2 inch hose are designed and
fabricated to drain water from the drum. A 1/2 HP variable speed drive motor with a
gear reducer is used to rotate the rotary deck at a speed of approximately 10 rpm. The
speed can be varied between approximately 6 to 16 rpm. Electrical circuit box housing:
Emergency-Stop, On Push Button for the sample rotation and the pump, and Time
Meter are provided. A motor drive to control the motor speed is also included.
173
A photograph of the completely fabricated accelerated polishing device using high-
pressure water jet with all components labeled is shown in Figure 7-5. The photograph
of the drum (chamber) used to hold either specimen size (18 inch by 18 inch by 2 inch
high roller-compacted slab specimen or 6 inch diameter by 4 inch high Superpave
gyratory-compacted specimen) is shown in Figure 7-6. The close up view of the
specimens of the two specimen sizes being subjected to high pressure water jet
polishing is shown in Figure 7-7 and Figure 7-8, respectively.
Figure 7-5: Overall view of the accelerated polishing machine using high-pressure
water
F
Figure 7-6: D
Figure 7-7: D
Drum (cham
Details on sla
174
mber) for pla
ab specimen
acing the test
n mounting i
t specimen
n the drum
In
su
In
an
sl
cu
co
th
Figure 7-
7.3 Equipm
n this section
urface are pr
7.3.1 Mater
n the evaluat
nd the aspha
lab specimen
urve of the
ontent of 5.9
he Marshall D
-8: Details o
ent Characte
n, the investi
resented.
ials
tion study of
alt binder gr
ns or the Sup
aggregate u
9% used for
Design meth
on gyratory c
eristics and V
igation resul
f the develop
rade (PG 70
perpave gyra
used for HM
r compacting
hod.
175
compacted sp
Validation
lts of using t
ped polishing
0-22) were u
atory-compa
MA is shown
g the HMA
pecimen mo
the develope
g device, a li
used to com
acted 6 inch
n in Figure
specimens w
ounting in the
ed polisher to
imestone agg
mpact the rol
specimens.
7-9. The op
were determ
e drum
o polish HM
gregate sour
ller-compact
The gradatio
ptimum bind
mined by usin
MA
rce
ed
on
der
ng
176
Figure 7-9: Aggregate gradation curve
7.3.2 Sample Preparation Procedure for HMA Specimens
The mixing procedure of the loose mix is identical to the mixing procedure method
presented in Chapter III.
7.3.3 Friction and Surface Texture Measurements
Different types of measuring techniques are used to measure friction and texture of the
HMA surface for the two types of specimen sizes. For the 18 inch by 18 inch by 2 inch
roller-compacted specimens, the dynamic friction tester and the circular texture meter
are used for measuring friction and texture, respectively. For the 6 inch cylindrical
gyratory-compacted HMA specimens, the British Pendulum Tester and the sand patch
method are used for measuring friction and surface texture, respectively.
0
20
40
60
80
100
120
0.001 0.01 0.1 1
Grain size (in)
Pe
rce
nt
pa
ss
ing
177
7.3.4 Work Plan
The work conducted in the development stage of the high pressure water jet based
accelerated polishing machine included the polishing of laboratory prepared HMA
specimens (i.e., roller-compacted HMA specimens and gyratory-compacted HMA
specimens) at variable rotational speed of the rotary deck housed inside the protection
drum. Also, the water pressure varied in the work plan. The water jet that impacts onto
the specimen surface is set at an angle of 40 degree to the horizontal specimen surface.
The fine grit was used. A summary of the trial test program designed to investigate the
feasibility of the developed machine is shown in Table 7-1.
Table 7-1: Work plan summary of the laboratory work
7.4 Polishing Effect of the Accelerated Polishing Machine
The polishing effect of the accelerated polishing machine is examined in this section.
For the HMA slab specimens made with limestone aggregate, the friction values (FN)
obtained from the DFT at different measuring speeds versus the polishing duration for
the trial number one (10 rpm rotation speed and 1450 psi of water pressure) are shown
178
in Figure 7-10. For the same test (Trial No. 1), the MPD measured by the CTM is
plotted versus duration of polishing in Figure 7-11. It can be seen from Figure 7-10 that
friction increases with polishing duration. Corresponding to the friction increase, there
is a similar trend of MPD increase with polishing duration as well. This polishing
behavior is opposite to what is expected. A similar trend is observed for trial number
two (10 rpm rotation speed and 400 psi of water pressure) when the slab specimen was
tested. Figure 7-12 and Figure 7-13 show the increasing trend of both friction and
texture values with the polishing duration for trial number 2 test results.
For the 6 inch HMA gyratory-compacted specimens made with limestone aggregate,
the friction values (BPN) obtained from the BPT and the MTD measured by the sand
patch method when the rotation speed was 10 rpm and the water pressure was 1000 psi
(trial number 3) are plotted against the polishing duration in Figure 7-14 and Figure
7-15, respectively. It can be seen that both the friction values and the MTD increase as
the polishing duration increases.
The test results of HMA specimens at 10-rpm rotation speed and the water pressure of
500 psi (trial number 4) are shown in Figure 7-16 and Figure 7-17. The similar trend as
trial number 3 is observed. Furthermore, one can see that the polishing behavior for
large slab specimens and gyratory-compacted specimens is similar.
The photographs of the tested HMA specimens are presented in Figure 7-18 and Figure
7-19 for the roller-compacted and gyratory-compacted specimens, respectively. It is
179
strikingly clear that high pressure water jet appears to have the effect of rejuvenate and
renew the surface such that there is the accompanied improvement of friction values.
Figure 7-10: Friction values for trial number 1 (at 10 rpm and 1450 psi)
40
60
80
100
120
0 20 40 60 80 100 120
Polishing Time (min.)
FN
me
as
ure
d u
sin
g D
FT
0 km/hr
10 km/hr
20 km/h
30 km/hr
40 km/hr
50 km/hr
64 km/hr
Friction Measurement
Speed
180
Figure 7-11: MPD trend for Trial number 1 (at 10 rpm and 1450 psi )
Figure 7-12: Friction values for trial number 2 (at 10 rpm and 400 psi)
1.00
1.10
1.20
1.30
1.40
1.50
0 20 40 60 80 100 120
Polishing Time (min.)
MP
D m
easu
red
usin
g C
TM
(m
m)
0
20
40
60
80
100
0 40 80 120 160 200 240
Polishing Time (min.)
FN
me
asu
red
by
DF
T 0 km/hr
10 km/hr
20 km/h
30 km/hr
40 km/hr
50 km/hr
64 km/hr
Friction Measurement
Speed
181
Figure 7-13: MPD trend for trial number 2 (at 10 rpm and 400 psi)
Figure 7-14: Friction values of trial number 3 (at 10 rpm 1000 psi)
0.90
0.95
1.00
1.05
1.10
1.15
0 40 80 120 160 200 240
Polishing Time (min.)
MP
D m
ea
su
red
by
CT
M (m
m)
65
70
75
80
85
0 50 100 150 200
Polishing Time (min.)
BP
N m
easu
red
by B
PT
182
Figure 7-15: MTD trend for trial number 3 (at 10 rpm 1000 psi)
Figure 7-16: Friction values for trial number 4 (at 10 rpm 500 psi)
1.7
1.8
1.9
0 100 200 300 400 500
Polishing Time (min.)
MT
D m
ea
sure
d b
y s
an
d p
atc
h m
eth
od
(m
m)
60
65
70
75
0 50 100 150 200
Polishing Time (min.)
BP
N m
easu
red
by B
PT
Fig
Figu
1.
1.
1.
1.
MT
D m
ea
su
red
by
sa
nd
pa
tch
me
tho
d
(mm
)
gure 7-17: M
ure 7-18: Tes
.5
.6
.7
.8
0
MTD trend fo
sted HMA ro
50
183
or trial numb
oller-compac
100
Polishing Ti
ber 3 (at 10 r
cted slab spe
0
ime (min.)
rpm 500 psi)
ecimen surfa
150
)
ace
200
184
Figure 7-19: Tested HMA roller-compacted slab specimen surface
7.5 Summary and Conclusions
Presented in this chapter is the development of an accelerated laboratory-scale
polishing machine using the concept of high-pressure water jet to polish HMA surface
in a short duration, with the intention to use this equipment for screening the aggregate
source and mix design formula to ensure adequate friction (or skid resistance) of the
HMA over the expected life span of the pavement surface. The accelerated polishing
machine is designed such that it is capable of testing two different sizes of HMA
specimens: 18 inch by 18 inch by 2 inch high slab specimens compacted using the
roller compactor and 6 inch diameter and 4 inch high gyratory-compacted HMA
specimens. The design principles of the testing device, together with the evaluation
185
results of the capability of the developed polishing machine to simulate the real
polishing action, are described in detail in this chapter.
The preliminary findings based on four trial tests (two on large slab specimens and two
on small size gyratory compacted specimens), however, indicate that both friction and
texture values tend to increase with the polishing durations for two combinations of
pressure and rotation speed of the rotary deck. It seems that the more polishing action
the specimen is subjected to, the aggregate edges are created such that the surface
texture values are increased with the accompanied increase in friction values. It may be
of interest (but which is outside the scope of this study) to investigate if the similar
trend exists for HMA prepared with aggregate source that is sand and gravel.
One interesting side effect from the finding observed in this study is the concept of
using controlled high pressure water jet to rejuvenate (create) the desirable rough
surface texture for restoring the surface friction of those worn surface course made of
limestone aggregates. If this found to be technically feasible to carry out in the field
with confirmed friction restoration benefit, then it can be used to improve the friction
of the existing asphalt pavement surface, rather than the conventional approach of
resurfacing or reconstructing the pavement surface course. A substantial cost saving
can be realized in this approach of using high pressure water jet in maintaining or
improving surface friction.
186
CHAPTER VIII
8. SUMMARY AND CONCLUSIONS
Preserving adequate friction and texture values during the entire life expectancy of
asphalt pavements is of key importance in reducing skid-related accidents. The
important tasks accomplished in the course of this research are briefly summarized
below. Furthermore, conclusions and recommendations for practical implementation as
well as future research are presented at the end of this chapter.
8.1 Summary of Work Done
Presented in chapter II is a review of the literature on the concepts and theories of
polishing of aggregate and HMA, the frictional properties of aggregate and HMA, as
well as the interrelationship between aggregate source, HMA friction properties, and
HMA surface texture properties. Chapter II also provided information on the different
equipment used for accelerated polishing of aggregates and asphalt mixtures, as well as
different friction and texture measurement devices. Relevant research and practice by
the state federal highway agencies on the related topics was also covered in this
chapter.
187
Presented in Chapter III is the development of an accelerated laboratory-scale polishing
device that is capable of mimicking the polishing action of the HMA surface by a
vehicle tire in a short duration, thus allowing for screening the aggregate source and
mix design formula to ensure adequate friction (or skid resistance) of the HMA over
the expected life span of the pavement surface.
In Chapter IV, the polishing behavior of laboratory-prepared, gyratory-compacted
HMA specimens made of eight different job mix formulas was studied in terms of
friction values (BPN) and macrotexture data (MTD). In addition, the potential
relationship between BPN and MTD was also investigated.
Chapter V presented the test results of a laboratory test program to investigate the
effects of the air void and temperature on the HMA surface friction properties. The
surface friction values were determined using the British Pendulum Tester, while an in-
house accelerated polishing machine was used to polish the laboratory prepared Hot
Mix Asphalt specimen surfaces to mimic different stages of actual pavement surface
during the life span of the pavement. The laboratory HMA specimens were prepared
using the gyratory compactor with different number of gyrations to achieve different air
void of the test specimens. The temperature of the test specimens and the rubber slider
of the BPT, as well as the spraying water temperatures were carefully controlled in the
laboratory test program. The laboratory test results were analyzed statistically to
ascertain the significance of each test variable (air void and temperature) on the
measured friction values. Finally, a linear regression analysis of test data has yielded
188
useful equations for extrapolating the measured BPN or SN values at a given air void
or temperature to other different air void and/or temperature.
The major objective of Chapter VI was to present a comprehensive set of data
conducted by the authors and the accompanied statistical analyses of the data set to
develop predictive equations for SN from other variables, such as DFT and MTD.
Presented in Chapter VII is the development of an accelerated laboratory-scale
polishing machine using the concept of high-pressure water jet to polish HMA surface
in a short duration, with the intention to use this equipment for screening the aggregate
source and mix design formula to ensure adequate friction (or skid resistance) of the
HMA over the expected life span of the pavement surface. The accelerated polishing
machine is designed such that it is capable of testing two different sizes of HMA
specimens: 18 inch by 18 inch by 2 inch high slab specimens compacted using the
roller compactor and 6 inch diameter and 4 inch high gyratory-compacted HMA
specimens. The design principles of the testing device, together with the evaluation
results of the capability of the developed polishing machine to simulate the real
polishing action, are described in detail in this chapter.
8.2 Observations and Conclusions
The specific remarks or statements from the work conducted in this research can be
mentioned succinctly as follows:
189
The accelerated polishing machine is capable of testing two different sizes of
HMA specimens: 18 inch by 18 inch by 2 inch high slab specimens compacted
using the roller compactor and 6 inch diameter and 4 inch high gyratory
compacted HMA specimens. Although the device can handle two different
specimen dimensions, each with different type of compaction method, the
intended routine test sequence is geared toward the testing of 6 inch gyratory
compacted HMA specimens due to its ease of sample preparation.
Repeatability of the machine was checked and affirmed using one-way ANOVA
test.
The polishing effect of the machine was confirmed through examination of the
test results conducted on Limestone and Sand and Gravel aggregates.
Good correlation of the polishing and friction behavior was found between
aggregate specimens and the HMA specimens made with the same aggregates.
Therefore, it may be reasonable to conclude that the new accelerated polishing
machine can accomplish the intended tire/pavement wearing and polishing
mechanisms.
Image analysis validated the polishing action.
Good correlation was found between the two specimen sizes using different
compaction methods.
The developed accelerated polishing device can be used effectively for screening
polishing and friction properties of HMA mix (i.e., aggregate source, binder type
and content, etc.) during the HMA mix design stage.
190
The test procedure can be completed in a reasonable timeframe, and is simple
and efficient (i.e., less labor effort).
The correlation study between friction values measured by DFT at low
measuring speed and the BPT suggests that the BPT can be used for measuring
friction of HMA at low speeds.
A set of tentative acceptance criteria of gyratory compacted HMA specimens was
developed through two different correlations (i.e., by correlating BPN with PV or
relating BPN with SN). These acceptance criteria are divided into two parts
based on the aggregate type used; i.e., limestone or gravel. The acceptance
criteria consist of four categories: 1. highly acceptable, 2. acceptable, 3.
marginally acceptable, and 4. unacceptable, which can be used to screen and
select pertinent aggregate and HMA mix design for adequate polishing resistance
and friction values.
It has been observed that the decrease in friction (BPN values) and surface
macrotexture (MTD values) is the maximum during the first hour of polishing in
the accelerated polishing equipment. With the passage of time, the drop in BPN
and MTD decreases and eventually becomes negligible at the 6th hour of
polishing. This behavior is attributed to the presence of surface impurities on the
HMA specimen surface. With the passage of polishing time during which further
sacrificial polishing occurs, these aggregates are cleaned off. Also certain angular
protrusions on the surface of the aggregates wear off during this time to reveal a
much smoother surface. As a result, the rate of decrease in both BPN and MTD
191
becomes smaller with polishing time and eventually becomes negligible at 6th
hour of polishing action.
The macrotexture (MTD values) of HMA surface was found to be strongly
correlated with the surface friction (BPN values). In addition, good correlation
was also found between change in BPN values (BPN) and change in surface
texture (MTD) for each HMA mix as well as for each polish susceptibility
category. Therefore, the results presented in this report provide strong
quantitative evidence in supporting the strong interrelationship between the
friction and texture properties of the HMA surfaces.
The effect of air voids on the measured HMA frictional properties (BPN values)
was found to be statistically significant at the 0.05 significance level. Basically,
there was an increase in friction (BPN values) corresponding to an increase in air
void.
For the air void difference between 3.5% and 7.5%, the corresponding difference
in BPN is about 5.7 and the corresponding difference in SN is 4.9. Therefore, the
extrapolation relationships given in Equations 5-1 and 5-3 are recommended to
correct the measured BPN or SN for the desired air void other than the one
during the measurement.
The effect of temperature of the HMA surface, the rubber slider of BPT, and the
spraying water on the measured HMA friction (BPN values) was found to be
statistically significant. Essentially, there was a decrease in friction
corresponding to an increase in temperature.
192
For temperature difference between -11.1 °C to 50 °C (12 °F to 122 °F), the
corresponding difference in BPN is about 4.6 and the corresponding difference in
SN is 4.0. Therefore, the extrapolation relationships given in Equations 5-4 and
5-5 are recommended to correct the measured BPN or SN for the desired
temperature other than the one during the measurement.
Skid numbers (SN(64)R) measured using LWST were correlated to friction
numbers measured using the DFT at different speeds. It was observed that
SN(64)R could be significantly correlated to DFT64. This could be explained by
the fact that both frictions were measured at the same measurement speed, thus
eliminating the potential of speed effect. In the same token, the low coefficient of
determination observed for the relationship between SN(64)R and DFT20 can be
due to the difference in the measuring speed.
The coefficient of determination between SN(64)R and MPD was found to be
low. It could be due to the fact that friction measurement by the ribbed tire of the
SWLT is insensitive to the macrotexture. On the other hand, the coefficient of
determination between SN(64)R and DFT20 was high, suggesting that the ribbed
tire is more sensitive to microtexture than macrotexture, since DFT(20) reflects
the effect of microtexture.
Several regression equations were examined for predicting SN(64)R from (a)
MPD and DFT20, (b) MPD and DFT64, (c) DFT20 and DFT64, and (d) MPD,
DFT20, and DFT64. However, it was found that MPD did not add much to the
regression model, neither did the DFT20 (see conclusions 1 and 2). Therefore,
193
SN(64)R can be directly predicted from the DFT64 as shown in Table 6-3. For
slightly higher accuracy SN(64)R can be predicted from MPD, DFT20, and
DFT64 as shown in equation 6-2.
The preliminary findings based on four trial tests (two on large slab specimens
and two on small size gyratory compacted specimens), however, indicate that
both friction and texture values tend to increase with the polishing durations for
two combinations of pressure and rotation speed of the rotary deck. It seems that
the more polishing action the specimen is subjected to, the aggregate edges are
created such that the surface texture values are increased with the accompanied
increase in friction values. It may be of interest (but which is outside the scope of
this study) to investigate if the similar trend exist for HMA prepared with
aggregate source that is sand and gravel.
One interesting side benefit from the findings observed in this study is the
concept of using controlled high pressure water jet to rejuvenate (create) the
desirable rough surface texture for restoring the surface friction of those worn
surface course made of limestone aggregates. If this is found to be technically
feasible to carry out in the field with confirmed friction restoration benefit, then
it can be used to improve the friction of the existing asphalt pavement surface,
rather than the conventional approach of resurfacing or reconstructing the
pavement surface course. A substantial cost saving can be realized in this
approach of using high pressure water jet in maintaining or improving surface
friction.
194
8.3 Recommendations for Implementation
Using the current research results and with the eventual implementation of the friction
related Supplemental Specifications by ODOT, the contractors can develop their HMA
mix that is capable of providing satisfactory friction performance over the expected life
cycle of a pavement surface course. This in turn can eliminate any need for early
pavement resurfacing, even though the structural capacity of pavement is adequate, due
to premature loss of skid resistance. There should be tremendous cost saving as a result
of elimination of premature pavement resurfacing and prevention of unnecessary lane
closure due to resurfacing. More importantly, the wet weather related accidents can be
reduced because of maintaining high skid resistance pavement surface throughout
Ohio’s highways. Safety in the form of consistent and acceptable friction of pavement
over the life expectancy of pavement can be achieved.
The developed friction/polishing procedure can be easily incorporated as a part of
gyratory mix design procedure, as the polishing and friction test procedure only needs
the six inch diameter gyratory compacted HMA specimens. Therefore, there should be
no extra effort in preparing test specimens. Furthermore, the test duration is relatively
short, so that the final test results can be obtained on the same day of the test. Thus, the
test procedure minimizes the needs for additional labor or equipment for test specimen
preparation.
195
The developed test procedure and equipment can also be adopted by other DOTs in
their friction polishing program to ensure safety in the form of consistent and adequate
friction of pavement using their respective available aggregate sources.
To reach full adoption and implementation of the current UA accelerated polishing
equipment by ODOT and paving contractors, the following areas need to be addressed.
The correlations need to be strengthened by including long-term data from
actual pavement performance in the field.
The time-scale difference between the laboratory friction-time curve and actual
pavement friction-time curve (as affected by traffic count and environment
condition) needs to be fully understood and incorporated in the acceptance
criteria.
The current accelerated test equipment (research grade) should be further
improved, ideally through partnership with commercial equipment
manufacturers, to develop the second generation for routine testing by the
general users such as contractors, testing agencies, and highway agencies.
The developed extrapolation relationships given in Chapter V are
recommended to correct the measured BPN or SN for the desired air void or
temperature other than the one during the measurement.
8.4 Recommendations for Future Work
196
Currently, the University of Akron research team, under the sponsorship of ODOT, has
developed an efficient accelerated laboratory polishing equipment with the
accompanied friction measurement method for determining the polishing and friction
behavior of HMA with Ohio typical aggregate used in the pavement surface course.
The test protocols of the accelerated laboratory polishing equipment with the associated
acceptance criteria were developed in the current project. However, it is imperative that
a long-term validation effort be established to continue to collect field performance
data (i.e., SN from Locked Wheel Skid Trailer, Friction number measured by the
Dynamic Friction Tester and British Pendulum Tester, and surface texture measured by
Circular Texture Meter) on existing test sections (a total of eight pavement sections
throughout Ohio have already been identified and monitored for two years under the
current project). The corresponding laboratory test data on the HMA from these eight
test sections have been compiled as well. However, since polish rates are non-linear
and mix gradation, aggregate source, and binder type and content are varied, a long-
term validation of the new method through correlation with field data is needed to
confirm whether or not the expected results match actual field performance. The focus
of this research would be on validation of the long-term applicability of the laboratory
accelerated polishing test method developed by the UA Research Team.
The main objectives of future research are to validate the applicability of the developed
laboratory test protocol and acceptance criteria associated with the newly developed
accelerated polishing equipment through a correlation and comparison study with field
performance data. If necessary, improved test protocol and the acceptance criteria need
197
to be developed based on the long-term laboratory vs. field correlation study results.
The specific objectives of the future effort can be enumerated as follows.
Continue to improve and refine the laboratory test protocols to ensure ease of
implementation by the potential users, such as the contractors, the aggregate
producers, and DOT material engineers.
Validate the acceptance criteria by relating laboratory measured time-series
friction loss behavior to the time history of field performance data in the
previously selected pavement test sections throughout Ohio.
Develop ODOT Supplemental Specifications incorporating the developed
equipment and test procedures for friction/polishing criteria during the mix
design of the hot mix asphalt for a surface course.
Finally, one interesting side effect from the findings observed in Chapter VII is the
concept of using controlled high-pressure water jet to rejuvenate (create) the desirable
rough surface texture for restoring the surface friction of those worn surface course
made of limestone aggregates. If this found to be technically feasible to carry out in the
field with confirmed friction restoration benefit, then it can be used to improve the
friction of the existing asphalt pavement surface, rather than the conventional approach
of resurfacing or reconstructing the pavement surface course. A substantial cost saving
can be realized in this approach of using high-pressure water jet in maintaining or
improving surface friction. In order to come to this worthy finding, further research is
needed in this area.
198
REFERENCES
1. Abe, H., Tamai, A., Henry, J. J., and Wambold, J. (2002) “Measurement of Pavement Macrotexture with Circular Texture Meter” In Transportation Research Record: Journal of the Transportation Research Board, No. 1764, TRB, National Research Council, Washington, D.C., pp. 201-209.
2. Alsopp, H. C. (1985) “Road Surface Friction Measurements, Methods, and Machines” Anti-Lock Braking Systems for Road Vehicles, Institution of Mechanical Engineers, London: Mechanical Engineering Publications Limited, pp. 61-72.
3. American Society for Testing and Materials “Measuring Pavement Macrotexture Depth Using a Volumetric Technique” ASTM Standard Test Method E-965-96 (2001), Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2005.
4. American Society for Testing and Materials “Measuring Surface Friction Properties Using the British Pendulum Tester” ASTM Standard Test Method E-303-93 (2003), Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2005.
5. American Society for Testing and Materials “Standard Practice for Calculating International Friction Index of a Pavement Surface” ASTM Standard Test Method E-1960-03, Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2005.
6. American Society for Testing and Materials, “Insoluble Residue in Carbonate Aggregates,” ASTM Standard Test Method D-3042-03, Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2003.
7. American Society for Testing and Materials, “Accelerated Polishing of Aggregates Using the British Wheel,” ASTM Standard Test Method D-3319, Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 1999.
8. American Society for Testing and Materials, “Accelerated Polishing of Aggregates or Pavement Surfaces Using a Small-Wheel, Circular Track Polishing Machine,” ASTM Standard Test Method E-660, Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2005.
9. American Society for Testing and Materials, “Determining the Polishability of Bituminous Pavement Surfaces and Specimens by Means of the Penn State Reciprocating Polishing Machine,” ASTM Standard Test Method E-1393, Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2005.
199
10. American Society for Testing and Materials, “Standard Practice for Calculating Pavement Macrotexture Mean Profile Depth,” ASTM Standard Test Method E 1845-01, Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2005.
11. American Society for Testing and Materials, “Standard Specification for Special Purpose, Smooth-Tread Tire, Operated on Fixed Braking Slip Continuous Friction Measuring Equipment,” ASTM Standard Test Method E1551-06, Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2006.
12. American Society for Testing and Materials, “Standard Test Method for Measuring Pavement Macrotexture Properties Using the Circular Track Meter,” ASTM Standard Test Method E-2157-01, Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2005.
13. American Society for Testing and Materials, “Stopping Distance on Paved Surfaces Using a Passenger Vehicle Equipped With Full-Scale Tires,” ASTM Standard Test Method E445/E445M-88(2001), Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2001.
14. American Society for Testing and Materials, Measuring Pavement Surface Frictional Properties Using the Dynamic Friction Tester, ASMT Standard Test Method E 1911-98 (2002), Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2005.
15. American Society for Testing and Materials, Skid Resistant of Paved Surfaces Using a Full-Scale Tire, ASMT Standard Test Method E 274-97, Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2005.
16. American Society for Testing and Materials, Standard Specification for Standard Rib Tire for Pavement Skid-Resistance Tests, ASTM Standard Test Method E 501-94, Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2005.
17. American Society for Testing and Materials, Standard Specification for Standard Smooth Tire for Pavement Skid-Resistance Tests, ASTM Standard Test Method E 524-88, Book of ASTM Standards, Volume 04.03, Philadelphia, PA, 2005.
18. Astrom, H. (2000) “Evaluating the VTI portable friction tester for road marking friction measurement” [In Swedish], VTI notat 45-2000, Swedish National Road and Transport Research Institute, Linkoping, Sweden.
19. Bazlamit, S. M. (1993) “Normalization of Skid Number Observations for Improved Pavement-Management Decisions”.
200
20. Bazlamit, S. M., and Reza, F. (2005) “Changes in Asphalt Pavement Friction Components and Adjustment of Skid Number for Temperature” Journal of Transportation Engineering, Vol. 131, No. 6, PP. 470-476, ASCE.
21. Beaton, J. L. (1976) “Providing Skid Resistance Pavement” Transportation Research Record 622, TRB, National Research Council, Washington, D. C., pp. 39-50.
22. Brilett, F. (1984) “Skid Resistance of Highway Pavements: Assessment of Fourteen Years of National skid-Resistance surveying” Bulletin des Liaison des Ponts, Dec.
23. Burchett, J. L., and Rizenbergs, R. L. (1980) “Seasonal Variations in the Skid Resistance of Pavements in Kentucky” In Transportation Research Record: Journal of the Transportation Research Board, No. 788, TRB, National Research Council, Washington, D.C., 12p.
24. Centrell, P. (1995) “Road marking friction—A comparison between the SRT pendulum and the VTI portable friction tester” [In Swedish], VTI notat 57, Swedish National Road and Transport Research Institute, Linkoping, Sweden.
25. Chelliah, T., Stephanos, P., Shah, M. G., and Smith T. (2003) “Developing a Design Policy to Improve Pavement Surface Characteristics” CD-ROM, Transportation Research Board, National Research Council, Washington, D.C., 24p.
26. Colony, D. C. (1984) “Skid Resistance of Bituminous surfaces in Ohio” Final Report of ODOT Project No. 3775, Ohio Department of Transportation.
27. Colony, D. C. (1992) “Influence of Traffic, Surface Age and Environment on skid Number” Final Report of ODOT Project No. 14460(O), Ohio Department of Transportation.
28. Crouch, L. K., Shirley, G., Head, G., and Goodwin, W. A. (1996) “Aggregate Polishing Resistance Pre-Evaluation”, Transportation Research Record, TRB, 1530.
29. Dahir, S. H., Henry, J. J., and Meyer, W. E. (1979) “Final Report: Seasonal Skid Resistance Variations” Pennsylvania Department of Transportation (PENNDOT), Harrisburg.
30. David, A., Kuemmel, R. C., Sontag, J. A., Crovetti, Y. B., John R J., and Allex S. (2000) “Noise and Texture on PCC Pavements” Final report multi state study, Wisconsin Department of Transportation Kinsman Blvd., Madison, Report Number WI/SPR-08-99.
201
31. Davis, R. M. (2001) “Comparison of Surface Characteristics of Hot-Mix Asphalt Pavement Surfaces at the Virginia Smart Road” Virginia Polytechnic Institute and State University, MSc Thesis, 235p.
32. Delalande, G. (1992) “Résistance des granulats au polissage – Méthode d’essai par projection” Bulletin de Liaison des LPC, 177, 73-80.
33. Dewey, G. R., Robords, A. C., Armour, B. T., and Muethel, R. (2001) “Aggregate Wear and Pavement Friction” Transportation Research Board, Annual Meeting CD-ROM, 17p.
34. Do, M. T., and Marsac, P. (2002) “Assessment of the Polishing of the Aggregate Microtexture by Means of Geometric Parameters” CD-ROM, Transportation Research Board, National Research Council, Washington, D.C., 19p.
35. Do, M. T., Kerzreho, J. P., Balay, J. M., and Gothie, M. (2003) “Full Scale Tests for the Assessment of Wear of Pavement Surface” Transportation Research Board, Annual Meeting CD-ROM.
36. Do, M. T., Ledee V., and Hardy, D. (2001) “Tire/Road Friction – Assessment of the Polishing of Road Aggregates” LCPC, Route de Bouaye, BP 4129, 44341 Bouguenais, FRANCE; e-mail: minh-tan.do@lcpc.fr.
37. Do, M., Tang, Z., Kane, M., and Larrad, F. (2007) “Pavement Polishing-Development of a Dedicated Laboratory Test and its Correlation with Road Results” Wear 263, pp. 36-42.
38. Elkin, B. L., Kercher, K. J., and Gulen, S. (1980) “Seasonal Variation in Skid Resistance of Bituminous Surfaces in Indiana” In Transportation Research Record: Journal of the Transportation Research Board, No. 777, TRB, National Research Council, Washington, D.C., pp. 50-58.
39. Federal Aviation Administration (1971) “Measurement of Runway Friction Characteristics on Wet. Icy, or Snow-Covered Runways” U.S. Department of Transportation. Report No. FS- 160-65-68-1.
40. Flintsch, G. W., Leon, E. D., McGhee, K. K., and Al-Qadi, L. L. (2003) “Pavement Surface Macrotexture Measurement and Application” CD-ROM, Transportation Research Board, National Research Council, Washington, D.C., 24p.
41. Flintsch, G. W., Luo, Y., and Al-Qadi, I. L. (2005) “Analysis of the Effect of Pavement Temperature on the Frictional Properties of Flexible Pavement
202
Surfaces” CD-ROM, Transportation Research Board, National Research Council, Washington, D.C., 15p.
42. Gandhi, P. M., Colucci, B., and Gandhi, S. P. (1991) “Polishing of Aggregates and Wet-Weather Accident Rates for Flexible Pavments”, Transportation Research Record, TRB, 1300.
43. Gillespie, T. D., et al. (1980) “Calibration of Response-Type Road Roughness Measuring Systems” National Cooperative Highway Research Program Report 228.
44. Goodman, S. N., Hassan, Y., and Abd El Halim, A. O. (2006) “Preliminary Estimation of Asphalt Pavement Frictional Properties from Superpave Gyratory Specimens and Mix Parameters” CD-ROM, Transportation Research Board, National Research Council, Washington, D.C., 14p.
45. Hanson, D. I., and Prowell, B. D. (2004) “Evaluation of circular texture meter for measuring surface texture of pavements” NCAT Report 04-05.
46. Henry, J. J. (2000) “Evaluation of Pavement Friction Characteristics” NCHRP Synthesis of Highway Practice 291, Transportation Research Board, National Research Council.
47. Henry, J. J., Abe, H., Kameyama, S., Tamai, A., Kasahara, A., and Saito, K. (2000) “Determination of the International Friction Index (IFI) Using the Circular Testure Meter (CTM) and the Dynamic Friction Tester (DFT)” Permanent International Association of Road Congresses (PIARC) IVth Symposium on Surface Characteristics, Nantes, France, Publication No. 109.01.06.B- 2000.
48. Henry, J. J., and Leu, M. C. (1978) “Prediction of Skid Resistance as a Function of Speed from Pavement Texture” Transportation Research Record 666, Transportation Research Board, National Research Council, Washington D. C., pp. 7-10.
49. Hill, B. J., and Henry, J. J. (1978) “Short Term, Weather-Related Skid Resistance Variation” In Transportation Research Record: Journal of the Transportation Research Board, No. 836, TRB, National Research Council, Washington, D.C., pp. 76-81.
50. Hosking, J. R. (1976) “Aggregates for Skid Resistant Roads” Report 693, Transport and Road Research Laboratory, Crowthorne, Berkshire, United Kingdom.
203
51. Hosking, J. R., and Woodford, G. C. (1976b) “Measurement of skidding resistance: Part I. Guide to the use of SCRIM” No. TRRL Report LR 737, Crowthorne, UK: Department of the Environment, Transport and Road Research Laboratory.
52. Ibrahim, M. A. (2007) “Evaluating Skid Resistance of Different Asphalt Concrete Mixes,” Building and Environment 42, pp. 325-329.
53. Jackson, N. M. (2003) “Evaluation of Variable Friction Properties of FDOT Friction Courses: Phase I” FDOT Contract # BD-273, University of North Florida, 23p.
54. Janoo, V. C., and Korhonen, C. (1999) “Performance Testing of Hot-Mix Asphalt Aggregates” Special Report 99-20, US Army Corps of Engineers Cold Regions Research & Engineering Laboratory.
55. Jayawickrama, P.W., and Thomas, B. (1998) “Correction of Field Skid Measurements for Seasonal Variations in Texas” Transportation Research Record 1639, Transportation Research Board, National Research Council, Washington, D.C., pp. 147-154.
56. Kandhal, P. S., and Parker, F. (1998) “Aggregate Tests Related to Asphalt Concrete Performance in Pavements” NCHRP Report No. 405, TRB, National Research Council, Washington, D.C.
57. Khasawneh, M. A., and Liang, R. (2008) “A New Accelerated Polishing Equipment for HMA Surface” under review to be submitted to Journal of Materials, ASTM, 2008
58. Khasawneh, M. A., and Liang, R. (2008) “Friction Measurements by Locked Wheel Skid Trailer and Dynamic Friction Tester: A Correlation Study” under review to be submitted to Journal of Transportation Engineering, ASCE, 2008.
59. Kissoff, N. V. (1988) “Investigation of regional differences in Ohio pavement skid resistance through simulation modeling” PhD dissertation, University of Toledo, Toledo, Ohio.
60. Kummer, H. W. (1966) “Unified Theory of Rubber and Tire Friction” Pennsylvania State University College of Engineering, University Park.
61. Kummer, H. W., and Meyer, W. E. (1967) “Tentative Skid-Resistance Requirements for Main Rural Highways” NCHRP-37, Highway Research Board, National Research Council.
204
62. Kutner, M. H., Nachtsheim, C. J., and Neter, J. (2004) “Applied Linear Regression Models” McGraw-Hill Irwin, Boston, MA.
63. Kuttesch, J. S. (2004) “Quantifying the Relationship between Skid Resistance and Wet Weather Accidents for Virginia Data” Virginia Polytechnic Institute and State University, MSc Thesis, 140p.
64. Leu, M. C., and Henry, J. J. (1983) “Prediction of Skid Resistance as a Function of Speed from Pavement Texture” Transportation Research Record 946, Transportation Research Board, National Research Council, Washington D. C.
65. Li, S., Zhu, K., Noureldin, S., and Kim, D. (2003) “Pavement Surface Friction Test Using Standard Smooth Tire: The Indiana Experience” CD-ROM, Transportation Research Board, National Research Council, Washington, D.C., 2003, 24p.
66. Liang, R. Y. (2003) “Blending Proportions of High skid and Low skid Aggregate” Final Report of ODOT Project No. FHWA/OH-2003/014, Ohio Department of Transportation.
67. Liang, R.Y., and Chyi, L. L. (2000) “Polishing and Friction Characteristics of aggregates Produced in Ohio” Final Report of ODOT Project No. FHWA/OH-2000/001, Ohio Department of Transportation.
68. Luce, A., Mahmoud, E., Masad, E., Chowdhury, A., and Little, D. (2007) “Relationship of Aggregate Texture to Asphalt Pavement Skid Resistance” Transportation Research Board, Annual Meeting CD-ROM.
69. Lund, B. (1997) “Friction Test. Comparative Testing with 3 Different Equipment Carried Out During the summer 1996” report 82, Road Directorate, Danish Road Institute.
70. Lundkvist, S. O., and Linden, S. A. (1994) “Road marking Friction – A Comparison between the SRT Pendulum and the VTI Portable Friction Tester” VTI Notat Nr 65A-1994, Vag- och transportforskningsinstitute, Linkoping.
71. Luo, Y. (2003) “Effect of Pavement Temperature on Frictional Properties of Hot-Mix-Asphalt Pavement Surfaces at the Virginia Smart Road” Virginia Polytechnic Institute and State University, MSc Thesis, 183p.
72. McDaniel, R. S., and Coree, B. J. (2003) “Identification of Laboratory Techniques to Optimize Superpave HMA Surface Friction Characteristics” Final Report (SQDH 2003 – 6 HL 2003 – 19)
205
73. McGhee, K. K., and Gerardo, W. F. (2003) “High-Speed Texture Measurement of Pavements” Final Report No VTRC 03-R9, Virginia Transportation Research Council, Charlottesville, VA.
74. Mitchell, J. C., Phillips, M. I., and Shah, G. N. (1986) “Seasonal Variation of Friction Numbers” Report No. FHWA/MD-86/02, Maryland Department of Transportation (MDOT), Baltimore.
75. Moyer, R. A. (1934) “Skidding Characteristics of Automobile Tires on Roadway Surfaces and Their Relation to Highway Safety” Bulletin 120, Iowa Engineering Experiment Station, Ames.
76. Nitta, N. Saito, K. and Isozaki, S. (1990) “Evaluating the Polishing Properties of Aggregates and Bituminous Pavement Surfaces by Means of the Penn State Reciprocating Polishing Machine” Surface Characteristics of Roadways, International Research and Technologies, ASTM SPT 1031, W. E. Meyer, and J. Reichert, Eds., American Society for Testing and Materials, Philadelphia, PA, pp. 113-126.
77. Olson, P. L., Cleveland, D. E., Fancher, P. S., Kostynick, L. P., and Schneiter, L. W. (1984) “NCHRP Report No. 270: Parameters Affecting Stopping Sight Distance” TRB, National Research Council, Washington, D.C.
78. Page, B. G., and Butas, L. F. (1986) “Evaluation of Friction Requirements for California State 25 Highways in Terms of Highway Geometrics” Report No. FHWA/CA/TL 86/01, Federal 26 Highway Administration, Washington, D.C.
79. Runkle, S. N., and Mahone, D. C. (1980) “Variation in Skid Resistance over Time” Virginia Highway & Transportation Research Council, pp. 10-13.
80. Schmidt, B. (1999) “Friktionsmalinger. Sammenlingende Malinger Mellem ROAR och Stradograf” Rapport 90, Vejdirektoratet, Vejteknisk Institut.
81. Sherwood, W. C., and Mahone, D. C. (1970) “Predetermining the Polish Resistance of Limestone Aggregates” Highway Research Board 341, HRB, National Research Council, Washington, D. C.
82. Shupe, J. W. (1958) “A Laboratory Investigation of factors Affecting the Slipperiness of Bituminous Paving Mixtures” JHRP report, Project No. C-36-53D. Purdue University.
83. Smith, B. J., and Fager, G. A. (1991) “Physical Characteristics of Polish Resistance of Selected Aggregates” Transportation Research Board, TRB, 1301.
206
84. Tomita, M. (1964) “Friction Coefficients between Tires and Pavement Surfaces” U.S. Navy Civil Engineering Laboratory, Technical Report R303.
85. Voller, T. W., and Hanson, D. I. (2006) “Development of Laboratory Procedure for Measuring Friction of HMA Mixtures – Phase I,” Final Report of NCAT No. 06-06, National Center for Asphalt Technology.
86. Wallman, C., and Astrom, H. (2001) “Friction Measurements Methods and the Correlation between Road Friction and Traffic Safety” Swedish National Road and Transport Research Institute, 47p.
87. Wambold, J. C., and Kulakowski, B. T. (1991) “Limitations of Using Skid Number in Accident Analysis and Pavement Management” Transportation Research Record 1311, , TRB, National Research Council, Washington, D.C. pp. 43-49.
88. Wambold, J. C., Antle, C. E., Henry, J. J., and Rado, Z. (1995) “International PIARC Experiment to Compare and Harmonize Texture and Skid Resistance Measurements” Final Report, PIARC, Paris, France.
89. Wambold, J. C., Henry, J. J., and Anderson, A. (1998) “Third Year Joint Winter Runway Friction Program” National Aeronautics and Space Administration (NASA), Washington, D.C.
90. Wang, H., and Flintsch, G. W. (2007) “Investigation of Short and Long-Term Variations of Pavement Surface Characteristics at the Virginia Smart Road” CD-ROM, Transportation Research Board, National Research Council, Washington, D.C., 17p.
91. Wilson, D. J., and Dunn, R. C. M. (2005) “Polishing Aggregate to Equilibrium Skid Resistance” Road and Transport Research, Vol. 14 No. 2, pp. 55-71.
92. World Road Association, PIARC (1987) “Report of the Committee on Surface Characteristics” XVIII World Road Congress, Brussels, Belgium.
93. Yager, T. J., and Bühlmann, F. (1982) “Macrotexture and Drainage Measurements on a Variety of Concrete and Asphalt Surfaces” Pavement Surface Characteristics and Materials, ASTM STP 763. C. M. Hayden, Ed., American Society for Testing and Materials, Philadelphia, PA, pp. 16-30.
207
APPENDICES
APPENDIX A.
JOB MIX FORMULAS
Eight pavement sections were identified in different Districts in Ohio during this
research project. The selection of these pavement sections is based on the criteria that
each of the pavement sections has adequate documentation of traffic counts as well as
the construction materials (i.e., Job Mix Formulas) used.
The eight JMFs for the current study were selected to have a wide range of polish
susceptibility: for example; three aggregate sources have possible low polish
susceptibility denoted by L1, L2, and L3, four aggregate sources have possible medium
polish susceptibility denoted by M1, M2, M3, and M4, and one aggregate source has
possible high polish susceptibility denoted by H1.
Appendix A summarizes all eight job mix formulas (aggregate gradation, optimum
binder content, and other volumetric properties of HMA) used in this research that were
provided by Ohio Department of Transportation.
208
Table A-1: Percent passing, optimum binder content and volumetric properties for
low polish susceptibility aggregate
209
Figure A-1: Gradation curve for low polish susceptibility aggregate (L1)
0
20
40
60
80
100
120
0.001 0.01 0.1 1
Grain size (in)
Pe
rce
nt
pa
ss
ing
210
Table A-2: Percent passing, optimum binder content and volumetric properties for
low polish susceptibility aggregate (L2)
211
Figure A-2: Gradation curve for low polish susceptibility aggregate (L2)
0
20
40
60
80
100
120
0.001 0.01 0.1 1
Grain size (in)
Pe
rce
nt
pa
ss
ing
212
Table A-3: Percent passing, optimum binder content and volumetric properties for
low polish susceptibility aggregate (L3)
213
Figure A-3: Gradation curve for low polish susceptibility aggregate (L3)
0
20
40
60
80
100
120
0.001 0.01 0.1 1
Grain size (in)
Pe
rce
nt
pa
ss
ing
214
Table A-4: Percent passing, optimum binder content and volumetric properties for
medium polish susceptibility aggregate (M1)
215
Figure A-4: Gradation curve for medium polish susceptibility aggregate (M1)
0
20
40
60
80
100
120
0.001 0.01 0.1 1
Grain size (in)
Perc
ent passin
g
216
Table A-5: Percent passing, optimum binder content and volumetric properties for
medium polish susceptibility aggregate (M2)
217
Figure A-5: Gradation curve for medium polish susceptibility aggregate (M2)
0
20
40
60
80
100
120
0.001 0.01 0.1 1
Grain size (in)
Pe
rce
nt
pa
ss
ing
218
Table A-6: Percent passing, optimum binder content and volumetric properties for
medium polish susceptibility aggregate (M3)
219
Figure A-6: Gradation curve for medium polish susceptibility aggregate (M3)
0
20
40
60
80
100
120
0.001 0.01 0.1 1
Grain size (in)
Pe
rce
nt
pa
ss
ing
220
Table A-7: Percent passing, optimum binder content and volumetric properties for
medium polish susceptibility aggregate (M4)
221
Figure A-7: Gradation curve for medium polish susceptibility aggregate (M4)
0
20
40
60
80
100
120
0.001 0.01 0.1 1
Grain size (in)
Pe
rce
nt
pa
ss
ing
222
Table A-8: Percent passing, optimum binder content and volumetric properties for
high polish susceptibility aggregate (H1)
223
Figure A-8: Gradation curve for high polish susceptibility aggregate (H1)
0
20
40
60
80
100
120
0.001 0.01 0.1 1
Grain size (in)
Pe
rce
nt
pa
ss
ing
224
APPENDIX B.
LABORATORY TEST RESULTS
The laboratory-prepared gyratory-compacted HMA specimens are polished for eight
hours using the developed accelerated polishing machine. Specimens are then tested
after each hour of polishing by the British pendulum tester and sand patch method.
Three specimens are tested for each JMF and their average is reported as the BPN and
MTD, which is a measure of the polish value and macrotexture, respectively.
Appendix B provides information on numerical values of the BPN and MTD for each
hour of polishing for all eight hours using the eight different JMFs labelled according
to their polish susceptibility. For all the JMFs studied, a residual friction (BPN) and
macrotexture (MTD) values are found to be reached at the end of eight hours of
polishing.
225
Table B-1: BPN for 8-Hour Polishing for Job mix formula # 1 (L1)
226
Table B-2: MTD for 8-Hour Polishing for Job mix formula # 1 (L1)
227
Table B-3: BPN for 8-Hour Polishing for Job mix formula # 2 (L2)
228
Table B-4: MTD for 8-Hour Polishing for Job mix formula # 2 (L2)
229
Table B-5: BPN for 8-Hour Polishing for Job mix formula # 3 (L3)
230
Table B-6: MTD for 8-Hour Polishing for Job mix formula # 3 (L3)
231
Table B-7: BPN for 8-Hour Polishing for Job mix formula # 4 (M1)
232
Table B-8: MTD for 8-Hour Polishing for Job mix formula # 4 (M1)
233
Table B-9: BPN for 8-Hour Polishing for Job mix formula # 5 (M2)
234
Table B-10: MTD for 8-Hour Polishing for Job mix formula # 5 (M2)
235
Table B-11: BPN for 8-Hour Polishing for Job mix formula # 6 (M3)
236
Table B-12: MTD for 8-Hour Polishing for Job mix formula # 6 (M3)
237
Table B-13: BPN for 8-Hour Polishing for Job mix formula # 7 (M4)
238
Table B-14: MTD for 8-Hour Polishing for Job mix formula # 7 (M4)
239
Table B-15: BPN for 8-Hour Polishing for Job mix formula # 8 (H1)
240
Table B-16: MTD for 8-Hour Polishing for Job mix formula # 8 (H1)