CA2330431A1 - In situ shear strength test facility - Google Patents

Abstract

The past decade has seen many significant improvements in asphalt pavement technology, particularly through large-scale research efforts such as the Strategic Highway Research Programs in the United States and Canada. However, there remains much room for improvement, particularly in the use of shear properties for the design, construction, monitoring and performance prediction of asphalt concrete pavements.
The primary objective of this investigation was the design, fabrication and validation of an advanced field test facility known as the In-Situ Shear Stiffness Test (InSiSST TM) for asphalt concrete pavements. The development process took a stepwise approach including the analysis of current testing devices and their related deficiencies.
The resulting test facility is portable, stable, and rugged - requiring only a single operator with no heavy lifting or complex set-up. Test results are instantly available and initial validation testing has indicated excellent accuracy and repeatability. Future testing with the InSiSST TM facility will provide invaluable input to pavement performance models and may allow the development of a strength or stiffness-based quality control/assurance (QC/QA) specification.

Description

The Ottawa-Carleton dnstitute for Civil Engineering Design, Development and Validation of the Ia-Situ Shear stif, fness Test (InSiSST'~"') Facility for Asphalt Concrete Pavements Researched and Written by:
Stephen Norman Goodman, B.A.Sc., E.LT.
A thesis submitted to the Faculty of Graduate Studies and Research in partial fulfillment of the requirements for the degree of Master of Engineering Department of Civil and Environmental Engineering Carleton University Ottawa, Canada September 2000 ~ 2000, Stephen Norman Goodman TABLE OF CONTENTS - BRIEF
Abstract.......................................................................
........................................................
iii Acknowledgements...............................................................
..............................................
iv Table of Contents - Brief ...............................................................................
......................
v Table of Contents -Detailed.......................................................................
........................
vi List of Figures........................................................................
.............................................
xi List of Tables ...............................................................................
.....................................xiii Notations and Abbreviations ...............................................................................
.............xiv List of Appendices.....................................................................
.......................................xvi CHAPTER 1:
INTRODUCTION...................................................................
.....................

CHAPTER 2: LITERATURE REVIEW
...........................................................................

CHAPTER 3: REVIEW OF PREVIOUS WORK AND ANALYTICAL MODELLING

CHAPTER 4: DEVELOPMENT OF THE IN-SITU SHEAR STIFFNESS
TEST

(InSISSTT'~s) ...............................................................................
......................................1O8 CHAPTER 5: PRELIMINARY TESTING AND
VALIDATION..................................142 CHAPTER 6: CONCLUSIONS AND
RECOMMENDATIONS...................................155 References.....................................................................
...................................................165 Appendices.....................................................................
..................................................175 v TABLE OF CONTENTS - DETAILED
Abstract.......................................................................
........................................................
iii Acknowled~ements...............................................................
..............................................
a iv Table of Contents - Brief ...............................................................................
......................
v Table of Contents -Detailed.......................................................................
........................
vi Ltst of Fyures.........................................................................
............................................
~a xt List of Tables ...............................................................................
.....................................
xiii Notations and Abbreviations ...............................................................................
.............
xiv List of Appendices.....................................................................
.......................................
xvi CHAPTER l:
INTRODUCTION...................................................................
.....................

1.1 Asphalt Concrete Pavements in Canada ..........................................................."."..."wwww 1.1.1 Asphalt Concrete Pavements Defined ...............................................................................
............

1.1.2 Climatic Conditions ...............................................................................
........................................

1.1.3 Transportation and the Canadian Economy..........................................................".,.www-wwww6 1.2 Pavement Structural Design and Loading Conditions..............................8 ..............................

1.2.1 Pavement Structural Desi~n.........................................................................
..................................8 1.2.2 Pavement Loading Conditions.....................................................................
..................................9 1.3 United States Strategic Highway Research ............................12 Program (US-SHRP)............

1.3.1 Background and Reason for Implementation ........................_......12 ...............................................

1.3.2 The SUPERPAVET"' Mix Design System.........................................................................
..........

1.3.3 Long Term Pavement Performance (US-LTPP) ...............................13 Project...............................

1.3.4 Introduction to the AASHTO 2002 Pavement ................................
Design Guide...................... 1~

1.4 Canadian Strategic Highway Research Program .............................19 (C-SHRP)....................

1.4.1 Background and Reason for Implementation ...........................,~,..
.............................................. 19 1.4.2 Canadian Long Term Pavement Performance ................................20 (C-LTPP) Project................

1.5 Specific Problem Definitions and Need for .............................21 New Test Facitity................

1.5.1 Improved Characterization of Pavement Structure................................
and Design Inputs ........ 22 1.x.2 Simple Performance Test for Superpave Verification................................23 and QC/QA Testing 1.5.3 The Need to Measure Field Properties ...............................................................................
.........23 1.6 The Innovations Deserving Exploratory Analysis.............................24 (IDEA) Program .......

1.7 Organization and Scope of Thesis ...............................................................................
.........25 1.7.1 Chapter 1:
Introduction...................................................................
.............................................25 1.7.2 Chapter 2: Literature Review.........................................................................
..............................26 1.7.3 Chapter 3: Review of Previous Work and Analytical Modellinc ................................................26 1.7.4 Chapter 4: Development of the In-Situ Shear Stiffness Test (InSiSSTTM) ........................"".,....27 1.7.5 Chapter 5: Preliminary Testing and Validation ...........................................................................27 1.7.6 Chapter 6: Conclusions and Recommendations................................................................
...........27 CHAPTER 2: LITERATURE REVIEW
........................................................................... 28 vt 2.1 Permanent Deformation of Asphalt Concrete Pavements.....................................................28 2.2 Manifestations of Rutting ...............................................................................
......................28 2.3 Asphalt Surface and Overlay Rutting ...............................................................................
....30 2.3.1 Location of Pavement Rutting ...............................................................................
...................... 30 2.3.2 Surface/Overlay Rutting Mechanism # 1 - Traffic Induced Densification ..................................31 2.3.3 Surface/Overlay Rutting Mechanism # 2 - Shear (Plastic) Flow ..............................................w 32 2.3.4 The Rutting Cycle..........................................................................
.............................................. 33 2.4 Quantification (Measurement) of Rutting....................................................._....wwww.~~~~~~34 2.5 Categories for Rutting Variable Classification......................................................_....ww...~

2.6 Category A-Bituminous Materials and Additives......................................................_..ww36 2. .1 Effect of Chemistry......................................................................
................................................36 2.6.2 Effect of Penetration/Viscosity..........................................................
..........................................

2. .3 Effect of Modifiers ...............................................................................
.......................................

2.6.4 Effect of Other Additives......................................................................
.......................................

2.7 Category B - Mineral Aggregates ...............................................................................
..........39 2.7.1 Effect of Source Properties.....................................................................
.....................................39 2.7.2 Effect of Consensus Properties ......................................................................, ..............................

2.8 Category C - Mix Design (Volumetric) Parameters........................................................~....41 2.8.1 Introduction to Volumetric Parameters.....................................................................
...................41 2.8.2 Effect of Air Voids ...............................................................................
.......................................45 2.8.3 Effect of Asphalt Cement Content........................................................................
.......................

2.8.4 Effect of Gradation ...............................................................................
.......................................

2. .~ Effect of VMA and VFA............................................................................
.................................

2.8.6 Effect of Dust Content........................................................................
.........................................

2.8.7 Effect of Laboratory Density and Compaction..............................
..............................................

2.9 Category D - Strength/Resistance Properties............................51 of Mix ...............................

2.9.1 Effect of Marshall Testing........................................................................
...................................51 2.9.2 Effect of Shear Strength and Stiffness ...............................
..........................................................~ 1 x.9.3 Effect of Resilient Modulus and Indirect ...............................53 Tensile Strength...........................

2.9.4 Effect of Creep..........................................................................
...................................................
~4 x.10 Category E - Pavement Structural and Geometric............................54 Design......................

x.10.1 Effect of Order of Rigidity of Pavement ......................,........54 Layers.........................................

2.10.2 Effect of Pavement Layer Thickness ...............................................................................
..........56 2.10.3 Effect of Surface (Wearing) Course vs.
...............................56 Base Coarse .................................

2.10.4 Effect of Pavement Alignment ...............................................................................
...................57 2.11 Category F - Construction-Related Factors .............................57 ...........................................

2.11.1 Effect of Compaction and other Construction................................57 Practices............................

?.11.2 Effect of Quality Control/Quality Assurance................................59 (QC/QA).............................

2.12 Category G - Environmental Factors ...............................................................................
...60 ''.12.1 Effect of Temperature....................................................................
............................................60 o ' ...............................................................................
..............................61 _.1_._ Effect of Atemg ...........

2.12.3 Effect of Moisture Damage (Stripping) ................................62 .....................................................

vii 2.13 Category H - Traffic (Load) Related Factors......................................................................63 2.13.1 Effect of Tire Contact Pressure (Load Magnitude)....................................................................6 2.13.2 Effect of Tire Material ...............................................................................
................................64 2.13.3 Effect of Number of Load Applications (ESAL's)....................................................................64 2.13.4 Effect of Rate of Loading ...............................................................................
........................... 65 2.14 Category X - Combinations of the Other Categories .........................................................65 2.15 Summary of Rutting Variable Relationships ......................................................................66 2.16 State-of-the-Practice: Asphalt Rutting Testers ...................................................................68 2.16.1 LCPC (French) Rut Tester.........................................................................
................................68 2.16.2 Hamburg Wheel Track Tester and Couch ..........................................70 Wheel Track Tester.......

2.16.3 Georgia Loaded Wheel Tester and Asphalt..........................................
Pavement Analyzer..... 72 2.16.4 Accelerated Load Facility.......................................................................
...................................74 2.16.5 Superpave Shear Tester ...............................................................................
..............................75 2.17 Deficiencies with Current Testing/Modelling......................................76 Practices ...............

2.17.1 Discussion of Empirical Rut Testers ..........................................76 ...............................................

2.17.2 Discussion of Existing Shear Tests..........................................................................
..................78 CHAPTER 3: REVIEW OF PREVIOUS WORK AND ANALYTICAL MODELLING 80 3.1 Introduction and Chapter Overview.......................................................................
...............80 3.2 Review of Previous Work - Laboratory Torsion Testing of Asphalt Concrete ....................81 3.2.1 Introduction ...............................................................................
..................................................81 3.2.2 Deriving Shear Properties from Laboratory Torsion ..82 Tests........................................................

3.2.3 Major Findings of Laboratory Torsion Testing ...........................................................................

3.3 Analysis of Laboratory Mix, Shear and Rutting Database..85 ........................:........................

3.3.1 Relation of Mix Characteristics to Shear Properties....................................................................

3.3.2 Relation of Shear Properties to Ruttin~
...............................................................................
........94 3.4 Review of Previous Work - The Carleton In-Situ Shear ..99 Strenth Test (CiSSST)..............

3.4.1 Introduction ...............................................................................
..................................................99 3.4.2 Deriving Shear Properties from Field Torsion Tests 100 .................................................................

3.4.3 Main Results of Previous Experiment ...............................................................................
........102 3.4.4 Advantages of CiSSST
Prototype......................................................................
........................102 3.5 Improved Analytical Framework to Determine Asphalt Shear Properties from the Surface Plate Loading Method.........................................................................
......................................103 3.5.1 Introduction ...............................................................................
................................................103 3.5.2 Reissner-Sagoci Problem........................................................................
...................................103 3.5.3 Finite Element Modellin; and Verification ...............................................................................

CHAPTER 4: DEVELOPMENT OF THE IN-SITU SHEAR STIFFNESS TEST

(InSiSSTT"') ...............................................................................
......................................108 4.1 Introduction and Chapter Overview.......................................................................
.............108 4.2 Critical Analysis of CiSSST Prototype Deficiencies..........................................................108 4.2.1 Chassis Design and Weight ...............................................................................
........................109 viii 4.2.2 Stabilization of Test Device.........................................................................
..............................109 4.2.3 Epoxy System Used for Loading Plate Attachment...................................................................11 4.2.4 Data Collection, Control System and Available .......110 Test Program...........................................

4.2.5 Overall Test Device Performance....................................................................
..........................111 4.3 Design Objectives for InSiSSTTM Test Facility..................................................................112 4.3.1 Mitigation of CiSSST Deficiencies ...............................................................................
............112 4.3.2 Reasonable Cost ...............................................................................
.........................................1 i2 4.3.3 Portability and Safety ...............................................................................
.................................112 4.3.4 Number of Operators and Ease of Use ...............................................................................
.......

4.3.5 Minimal Test Time and Damage to Pavement Surface........113 .....................................................

4.3.6 Correlate Results to Pavement Performance Indicators.............................................................

4.4 Design of InSiSSTTM Facility ...............................................................................
..............114 4.4.1 Introduction and Overall Design.........................................................................
.......................

4.4.2 The Primary Force Generation System (Powertrain).................................................................

4.4.3 The Transportation System.........................................................................
...............................

4.4.4 The Test Frame and Positioning System ...............................................................................
....120 4.4.5 The Stabilization System ...............................................................................
............................

4.4.6 Epoxy System.........................................................................
...................................................126 4.4.7 The Test Control/Data Collection System ...............................................................................
..127 4.4.8 Overall System Integration....................................................................
....................................129 4.4.9 Cost...........................................................................
.................................................................129 4.5 Fabrication, Debugging and "Shakedown"
Testing............................................................129 4.5.1 Positioning System Debugging......................................................................
............................

4.5.2 Jacking System Debugging......................................................................
..................................

4.5.3 Test System Debugging......................................................................
.......................................

4.5.4 Shakedown Testino........................................................................
............................................

4.6 Field Test Procedure ...............................................................................
............................132 4.6.1 Equipment Checklist......................................................................
............................................132 4.6.2 Transportation Safety.........................................................................
........................................ 132 4.6.3 Securing the Test Site ...............................................................................
.................................133 4.6.4 Preparation of Pavement Surface and Bonding the Loading Plates........................................... 133 4.6.5 Rutting/Density Surveys (Optional) ...............................................................................
........... 134 4.6.6 InSiSSTTM Test Procedure......................................................................
................................... 136 4.6.7 Leaving the Test Site ...............................................................................
.................................. 141 CHAPTER 5: Preliminary Testing and Validation ......................................................... 142 5.1 Introduction and Overview ...............................................................................
..................142 5.2 Analytical Models vs. Field Test Results........................................................................
....142 5.2.1 Verification of Linear Elastic Assumption................................
................................................142 5.2.2 InSiSSTTM vs. CiSSST
...............................................................................
...............................143 5.2.3 Practical Calculation of Shear Modulus ................................146 Using Equation 8........................

5.2.4 Asphalt Modulus vs. Torque Per Unit Twist................................
............................................. 147 5.2.5 Effect of Loading Plate Diameter ...............................................................................
...............

5.2.6 Discussion of Field Test Results and Analytical................................
Modelling...................... 153 5.2.7 Comparison of Field and Laboratory Results................................
............................................ 154 CHAPTER 6: Conclusions and Recommendations......................................................... 155 6.1 Review of Project Objectives.....................................................................
.........................155 tx 6.2 Review of Permanent Deformation and Previous Investigations .......................................156 6.3 Asphalt Mix Properties and Shear Characteristics..............................................................15 6.4 Asphalt Shear Characteristics and Rutting........................................................................
..159 6.5 Modelling In-Situ Shear Properties.....................................................................
................159 6.6 Design, Development and Verification of the InSiSSTTM..................................................160 6.7 Recommendations for Future Modifications......................................161 to INSISSTTM.........

6.7.1 Environmental Chamber........................................................................
....................................

6.7.2 Hydraulics.....................................................................
.............................................................162 6.7.3 Shear Vane...........................................................................
......................................................

6.8 Recommendations for Further Testing ......................................163 ..........,..............................

6.8.1 Additional Verification Testing........................................................................
.........................

6.8.2 Long Term Performance....................................................................
........................................163 6.8.3 Additional Testing for QC/QA
Specification..........................................
Development............ 164 References.....................................................................
...................................................

Appendices.....................................................................
.................................................. 175 x LIST OF FIGURES
Figure 1: Typical Cross Section for Asphalt Concrete Pavement2 .................................-.-...

Figure 2: Canadian Soil Temperature Zones ..........................--...--.---.--.--wwwwwwwww-4 Figure 3: Canadian Soil Moisture Zones .............................................................................

Figure 4: Canada's National Highway System and Major US .
Border Crossings .............. 7 Figure 5: Common Loading Conditions of Asphalt Pavements 10 ...............................---.---..

Figure 6: Transverse Profiles of Various Rutting Manifestations30 ..............................--.--..

Figure 7: Illustration of Surface Rutting........................----.-.--.-----wwwwwwwwwwww-~31 Figure 8: The Progression of Rutting with Traffic Loading 33 (Rutting Cycle)....................

Figure 9: Phase Diagram of Mix Constituents in Compacted 42 Specimen ..........................

Figure 10: Bitumen Stiffness vs. Mix Temperature for Three 60 Compaction Devices ........

Figure 11: LCPC Rutting Tester.........................................................................
...............69 Figure 12: Hamburg Wheel Tracking Tester.....................................................................71 Figure 13: Asphalt Pavement Analyzer.......................................................................
......73 Figure 14: Accelerated Load Facility (ALF) ...........................----.-----.--.-.wwwwwwwww~4 Figure 15: Superpave Shear Tester.........................................................................
...........
7~

Figure 16: Torsion Test Equipment at Carleton University .
.................................-----.---.~- 81 Figure 17: Typical Failure of Asphalt Specimen in Torsion .
Test Device........................ 82 Figure 18: Determination of Shear Properties from Different.
Test Methods................... 84 Figure 19: Laboratory Rutting vs. Asphalt Mix Shear Modulu......................................--.

Figure 20: Laboratory Rutting vs. Asphalt Mix Shear Strength.....................................---Figure 21: The Carleton In-Situ Shear Strength Test (CiSSST)..
Facility ........................ 99 Figure 22: Loading and Boundary Conditions of CiSSST
.....................................--.----..101 Figure 23: Load Plate Attached to Asphalt Concrete Pavement104 (ACP) Surface.............

xi Figure 24: Induced Failure in Asphalt Concrete Pavement (ACP) Surface .................... 104 Figure 25: Differential Element Shear Stresses from ............
Reissner-Sagoci Problem. 105 Figure 26: Initial Finite Element Model Verification..........................................--.-....-..-Figure 27: Side View of InSiSSTTM
...............................................................................
.

Figure 28: Top View of InSiSSTTM......................................................................
...........

Figure 29: InSiSSTT~f positioning System ......................................................................

Figure 30: Plan View of the Lower Positioning System.....-.-.....
.................................... 121 Figure 31: Plan View of InSiSSTTM Test Frame ......................................---.~.wwwwww Figure 32: Outline of Rutting and Density Survey...................................----~~..wwww~~-.

Figure 33: InSiSSTTM Trailer over Test Plates..............................-.......---.--wwwww-~~~.

Figure 34: Chocking the Trailer Tire...........................................................................
....

Figure 35: Attach Torque Cell Cable to Torque Cell ......--......
........................................ 139 Figure 36: InSiSSTTM Controls (Computer not shown) .......--~...~
.................................... 139 Figure 37: Connecting Collar from Torque Cell to ..............
Test Plate ......................... 140 Figure 38: Taking Pavement Temperature with IR Thermometer..............
.................... 140 Figure 39: Comparison of Field Results with Reissner-Sagoci Model ........................... 143 Figure 40: Typical Torque vs. Twist Angle Graph from...............
InSiSSTTM ............... 147 Figure 41: Determining the Tangent of the Torque-Twist...............
"S" Curve.............. 147 Figure 42: Torque vs. Angular Displacement for InSiSSTTM...............
Tests ................. 151 xn LIST OF TABLES
Table l: Classification Criteria for Transverse Profiles ........................-.....-..........--..-.---.- 29 Table 2: Categories for Rutting Variable Classification...........................................-----.-.. 35 Table 3: Various Asphalt Cement Modifiers.......................-.................----......---.-...--........ 39 Table 4: Summary Table of Rutting Variables and Qualitative Relationships.................. 67 Table 5: Characteristics of Rut Testers.......................................-.------........--.--.wwwwww 77 Table 6: Mix Properties Available from Zahw (1995) Database....................................... 86 Table 7: Engineering Properties Available from Zahw (1995) Database.......................... 86 Table 8: Mix Properties Yielding Greatest Correlation to Shear Properties ..................... 87 Table 9: Regression Statistics for Shear Modulus (Equation 4)........................................ 93 Table 10: Regression Statistics for Shear Strength (Equation 5) ...................................... 93 Table 11: Contribution of Individual Variables Toward Shear Properties........................ 94 Table 12: Rutting Models for Shear Strength and Modulus..........................................-... 95 Table 13: Target Test Strain Rates and Associated Motor Speeds.................................. 128 Table 14: Equipment Checklist......................................................................
.................. 132 Table 15: Comparison of CiSSST and InSiSSTTM Results ............................................. 144 Table 16: Shear Modulus vs. Torque Per Unit Twist ............................................-....-..-. 148 Table 17: Comparison of Torque Per Unit Twist for 92mm and 125mm Plates............. 152 Table 18: Results of InSiSST Testing with 125 mm Plates............................................. 153 xiii NOTATIONS AND ABBREVIATIONS
Absorbed Binder Volume, Vba............................................................................
...............

Accelerated Load Facility (ALF)..........................................................................
.............

Aggregate Volume, VS.............................................................................
..........................

Air Voids, V~
...............................................................................
......................................

American Association of State Highway and Transportation(AASHTO)...........
Officials 9 Apparent Aggregate Volume, VS~............................................................................
..........

Asphalt concrete pavement (ACP) ...............................................................................
.......

Asphalt Pavement Analyzer (APA)..........................................................................
.........

Average Rate of Densification (ARD)..........................................................................
.....

Binder Volume, Vb ...............................................................................
.............................

Bulk Aggregate Volume, Vsb............................................................................
.................

Canadian Long Term Pavement Performance (C-LTPP)..................................................

Canadian Strategic Highway Research Program .............................
(C-SHRP) ................ 19 Carleton In-Situ Shear Strength Test (CISSST) .............................
................................... 99 Coefficient of variation (COV)..........................................................................
..............

Effective Aggregate Volume, Vse ...............................................................................
.......

Effective Binder Volume, Vbe............................................................................
................

Federal Highway Administration (FHWA) .......................................................................

Field Shear Test (FST)..........................................................................
.............................

Fine aggregate angularity (FAA)..........................................................................
.............

Georgia Loaded Wheel Test (GLWT) ...............................................................................

Gross Domestic Product (GDP)..........................................................................
.................

Gyratory shear index (GSI)..........................................................................
......................

Gyratory testing machine (GTM) ...............................................................................
.......

Hot-mix asphalt concrete (HMAC) ...............................................................................
......

Innovations Deserving Exploratory Analysis ...............................
(IDEA) Program......... ?4 In-Situ Shear Strength/Stiffness Test (InSiSSTT"'')~",~"""~"""_~",.......
................-....~.~~~." ?7 Just-In-Time Delivery (JIT)..........................................................................
.......................

Laboratoire Central des Ponts et Chaussees ...............................
(LCPC).......................... 68 xiv Linear variable differential transducer (LVDT) .-.-...wwww.
............................................ 72 National Aggregate Association (NAA)..........................................................................
..

National Center for Asphalt Technology (NCAT) ......................................-..---...~ww.-~.~.

National Cooperative Highway Research Program (NCHRP).........................................-National Highway System (NHS)..........................................................................
..............

Penetration-Viscosity Ratio (PVR)..........................................................................
..........

Percent Air Voids, Va ...............................................................................
.........................

Present serviceability index or rating (PSI or .-.--~....~www.
PSR)...................................... 17 Quality control and quality assurance (QC/QA)......................................--.~._~..ww.-~
~3, 114 Resilient modulus Mj.............................................................................
............................
~3 Superior Performing Asphalt Pavements (SuperpaveTM~""~~~"".........
) ---_~..,~..~...,~,~~.""~~, 12 Superpave Shear Tester (SST)..........................................................................
.................

Transportation Association of Canada (TAC) .....................................................................

Transportation Research Board (TRB)..........................................................................
....
?4 United States Long Term Pavement Performance (US-LTPP)......................
................... 13 United States Strategic Highway Research Program ......................
(US-SHRP) .............. 12 Voids Filled with Asphalt (VFA) ...............................................................................
.......

Voids in the Mineral Aggregate (VMA)..........................................................................
..

xv LIST OF APPENDICES
Appendix A: Correlation Matrix for Zahw Database ..............................................~....... 164 Appendix B l: Selected Variables for Mixes 1 through 5................................................ 175 Appendix B?: Selected Variables for Mixes 6 through 12.............................................. 176 Appendix C 1: Shear and Rutting Properties for Mixes 1 through 5................................ 177 Appendix C?: Shear and Rutting Properties for Mixes 6 through 12.............................. 178 Appendix D: CiSSST Test Results ...............................................................................
... 179 Appendix E: InSiSSTTM Test Results........................................................................
...... 180 xm CHAPTER 1: INTRODUCTION
1.1 Asphalt Concrete Pavements in Canada 1.1.1 Asphalt Concrete Pavements Defined The term "pavement," although seemingly obvious in its usage, may have different meaning to different people or agencies. For example, pavement may simply refer to the surface layer of a road system, or may encompass additional underlying layers. The Transportation Association of Canada (TAC) has defined the term "pavement" as consisting of all structural elements or layers, including the shoulders, above the subgrade. While the subgrade is not part of the pavement structure by this definition, its characteristics such as strength or load can-ying capacity, drainage, etc. are implied (TAC 1997). This thesis has adopted the TAC
pavement definition wherever pertinent.
The term "asphalt concrete" refers to a conglomeration of asphalt cement (binder), aggregate and air (void space). Unless otherwise stated, the term asphalt concrete refers to hot-mix asphalt concrete (HMAC), the most common type of asphalt concrete used in transportation systems, which is mixed and placed at elevated temperatures.
Therefore, for the purposes of this thesis, an "asphalt concrete pavement (ACP)" is a pavement structure whose upper layers are constructed with hot-mix asphalt concrete unless otherwise stated. Figure 1 displays a typical cross section of an ACP, also commonly referred to as a flexible pavement.

_' .

r ' , .
~

, , ~
.

y ~ n1 t r --~ t ( ' r.~~ '-1 -.
.
~, r ~ J
K7 .~
, .

9. ~
a1 . ., 1 ..
' .

' o - 1 .1 .. ' a . , .'-_ '.

1 ~_ ' r.a ..
' ' . v W s a v ~ ' .

8 ~ --;
a -;-fl ~ :
" ' -' Q
-~ .
, .

p b ~ , 4 ~ -. ' _ y s~ a ._ , -._:, .

! a ,_. y ~ y L
! C~ ~

G

_ 0~ .~ O

Cn p ~ ~ ~ U
..
, ~

a , .
. d . o . , ~ .

~r . . . ...
a , '- ~ ~ ~ a ~ . .- ., o U
.0 ~
. Q
' .
.
.
.

_ c :.
.
_ ' _ ~

.~., ' ' , _ .

I . _ y U O
' ~ .. y . .y V V~ 4.L.
~ ~ _ v' v .
.

C .
3 0 ..
~_ _ v O
. L
_ _.
--. ~1 , _ W , e3 ~ W ~ ~a .

n ~ t t '' ..
''1 ' 1 , . '.
' ~

C~
' -y~ L

y ..a O
. GO

a r C ~ 'y ' cti J

1.1.2 Climatic Conditions Pavements in Canada are subjected to particularly harsh climatic conditions.
Furthermore, these harsh conditions are not consistent throughout Canada due to its enormous size. For example, the Canadian Meteorological Centre (?000) reports that northern cities such as Yellowknife in the Northwest Terntories consistently experience temperatures of -30 degrees Celsius (°C) for 3 months of the year with extreme temperatures of -51°C not uncommon. Central cities such as Regina, Saskatchewan may experience annual pavement temperature ranges of up to 80°C.
Finally, coastal cities such as St. John's, Newfoundland and Vancouver, British Columbia are subjected to 1.6 and 1.2 metres of rain respectively per year.
Figure 2 displays soil temperature zones across Canada. As shown, no less than 7 individual temperature zones are present, ranging from Arctic (extreme cold) in the north to Mild along the Canadian-US border. With the exception of some of the Atlantic Provinces, each province or terntory contains at least 2 of these zones with many of the provinces containing 5 zones.
Figure 3 displays the distribution of soil moisture across Canada. As with temperature, the distribution of soil moisture is extreme ranging from aquic-perhumid areas where the soil is fully saturated for long periods of the year to subaquic-arid regions with severe groundwater deficits.
The large variation in climatic conditions across Canada presents pavement designers and contractors with unique regional challenges.

n J
,; f., ~ i l :'_ m a ~
C ~, w 'r j > ',.. _ '. t ;~.'>. .-. .
L'L n ' U r ~ ' .
.... ~ U ~~ t Q J , _.. ~ :x r C r ~ 1 ~' .
w r-'~ C~
U Ci ,has (' o p "..t " '',.
a.' G ~ eu -i c -r w t7C w a. w " ~ j 1 ~.i..
j U Q: ~'' i~ cCy L a Q
w W -r U fd J J U G ~ 1~ , Q: - G f} C7 CJ -~
~Y c7p ~ ~).'..~ ~P , I f ._.......: I o.:~ . f ~ 2,".~
.C t i'~' '".x,~ , C/~
__ ._ ~ 2 .", tr' r~~ ' f 1 I 1 -i _ '4 ~ 1 ( _. S' ~1 :..~
.'f a. 1 i. j_..~...J ~. .;~ I .. L C/~
.. _ ... ~ ,. ~ ..3~ 'r G"~.,. n-~ L
z" ~ t ;
_ if ,~'~-.rt ~ t E .,~ ~ r~ '~ '~
f ~ i..~ r ~ ~ ri ~ f ~'~~ < ~ s~ . ~ ~ s. '-, v ~"'~ C~
Y' A ' i ~i Jv-t ~ t,, u, - a r .. ,.. '.;, ~ r -~<J >Y s ,..i ~ r .-..
~ ,_S ' f I~ b t ~"h x"r ~ 'tFv ~' ~"., CC
~...~~ L ~:~ ~ i ' '' r r ~ 'w/a~ ~'" ~W'' a CC Vj Z n ~, r, ~r.~ ~'f ~p,~~, ~,s C~
_ it , ~'t k~~q "'~ ~"r _ ~ ~- c q .. ..> > . ~ I ~~. t 1, ~' i ~~ ''1 I Y T z~" ~"~
U
6 1 ' . 51 ,:'x ~ ~'~f ~' r ;_ ' '._-; '~, ~1 r~~ ~~,d .t J N ~1 S ~.' v ! r. , ~ IJ~ ~ '3'~ G
' f .~ J~ G ~ ~ _ ~ ~ < ~ '..p4F y j ~~ ~ . ~~
// K Y 'r~~ L, k,.
~S>:*~~~~ 1~ ~
b p rlu(~ ~ ~pC~:;a"4~t~~' ~,~'~~t~ ~
~~i",,Ls~~ k~u,~ -r ~
1 ~~ ~r xp y. .
" \. "~ i ~' ,d T r',~'S~~us~ ~F~~ "k. ~, ' _= S, ~,, r a ~"y~d4 ~'""~ ~'/~r~" ~ er ' :s< ,~.n »~
~~1. ~.\f f J, t'1 i ! "~ ' ~..,z,c "'~' ~rt ~y6 ~rpv~ ><...
7 > S
x rf f ~ , i , a ~r ' 'w1~ t ,. !. ~
,,- ~ . ...
~~'r." '.- . t _ ; ', lr ~
,';. f ' .'i ~'~ '.: , . _. _ ~
w, a' ~
J
I U ~ 'j U .S'- ~ ' 6. N .~'-, 1;.. C n ~D °~' '~' 'm C is m J y 6 t1'C; .. , _ N ~ ~ =. . U
C,St =_ .. C ~ ~ ~ ..
C j.~"~L, a )r ...
w .U ~ ' C ~_ ' _.
(f] °v0 7 .~ ~ ' i1 4~ . ~ ..
111 b 3 . ii ._.
m 6 ~ 'n ~ _ i g ~ ~ ~ . ~
I =~ 'v U y 'n c a' ,' ' V G W
f Cl. r n ~ Gr c 'it 1 ~ z ' ' a' 4 ~ A ~ c ~ ,s > n ~ ' .i~ , c Z ~ ;; ~.c°' c_ ~ ~ ~ . r,, r~ '( _ -- ' t'1 -~ rn lL K 4! ~ ~ C L ? 4 VI j a ~ ;.
r a ~ 1 ~ ~. t ~ , ~~,, ~ . ~ '~ 4 ..
'T

_ .
.~ '" f ,.. Cv ''~
Q, ._.., y r ~ -= 7'r 'Q=I G~ C~
r.w r ~ " j ' ~ , ~"r "C7 __ ~ w O C~
r 5 ~ ~ f~ ~~yr ~ CSC
f) ~' yx7 r r. f ~ ~ ~" . ~ z ~ r . it V
. (' .... ~ ' _ ~it ; . _ r '.. n ~
r .r ~.rr.> ~.... ' ;~ i.~~ _ ~ V
t _ ~ . .~; r v ~ O it n ~~ a . ~ ~ a ~ ~'Y O
t i yr ~ ' 47 , y .T ~ . ~ (( : _ 4 ',. : j r s/: "~
~~ ~ I!' ~ J' ~ L r' 'i ~. S wI
t S Y( . c ~C
J
~-~' 3, tS ~ ~;~ ~ '' ., ~r<,~ ~ y. 4,.~-.. .Yr"iJ: CCS G~
f ~ ~i Vii' i ~ Y _ ~
r 1;f J ~ t_-"'s''~ ~~' .~ ~a: r~
.., r r's ( , '~ , ~.( b~ .~ ~,. y C~
',b "~~ ~ ,~ , . f ~t ~ : i y ~, ,1 ~ s1 ~ ~ ~ :., ~O ~
., ( ~'~ 4..J t y -i ' f I
i ! Y 1 J y5 ~ ~ Y ~ d...p~:''. ~ ~ ~~ . ~t ,' ,. 'a ~ iS J
' ,-v Y'~,r ,''~".~~, ~- ~ T'. ' . ~L,.. ~
~ . A , .,, r _r-~..._ Cf.i - t ' ~ r Jx'. y r y"~ t.., r,.
n 1 f S y ~ ' ~~Y ~ J~ ~~ ~yr.' .. ~
n.. (~ ~ _ ~ ~ l '~1 ~'~ ~..I
. ,,~ t ~~ x: ~ w y~ . :
,. ' ' ' ....., '.Z,. J ~' i '~,::.~5 ;.~.r ,'; , ~r.~
.;
~~
,..y,=, :r.
. ,, 1.1.3 Transportation and the Canadian Economy In 1999, transportation industries accounted for 30.6 billion Canadian dollars (4.1 %) of Canada's Gross Domestic Product (Transport Canada 1999). The trucking sector accounted for the largest proportion of the transportation industries at 1.7% of the GDP ($12.5 billion). The average annual growth of the trucking sector between 1994 and 1999 was 7.7%, more than double of any other transportation sector including rail, marine and air. In 1998/1999 alone, the trucking sector annual growth was 8.2%. These statistics clearly indicate the immense importance of trucking to the Canadian economy and that this importance is Growing at a high rate.
The National Highway System (NHS) is a network of roads identified by the Council of Ministers Responsible for Transportation and Highway Safety during a multi-stage policy study initiated in September 1987. The network consists of existing primary routes that provide interprovincial and international trade and travel by connecting a capital city or major provincial population or commercial centre in Canada with another capital city or major population centre, a major point of entry or exit to the United States highway network or another transportation mode served directly by the highway mode (Transport Canada 1999). The NHS is illustrated in Figure 4 Although the NHS accounts for only 24500 kilometres of the entire Canadian road network of over 900000 kilometres (less than 3%), the NHS experiences nearly one quarter of the total vehicle-kilometres driven (Transport Canada 1999).
Ontario and Quebec alone account for 60% of NHS traffic and these traffic levels are increasing every year.

The large and continuing increase in truck traffic on Canadian roads may be largely attributed to a new revolution in the way business is done in North America that started in the early 1990's. The main thrust of this new revolution was the implementation of a new manufacturing process referred to as "lean production"
- a process that has been shown to improve productivity, efficiency and profits (Earns 199?). While lean production involves numerous new procedures, an essential component is a process called "Just-In-Time Delivery," or JIT. In essence, JIT
delivery systems require delivery of inventory only when needed, permitting smaller storage space and faster model change in response to consumer demands.
Therefore, the nations highways become linear warehouses for manufacturing companies. This trend is not expected to reverse in the near future.
Unfortunately, although local government spending on transportation has increased over the past five years, spending at the federal and provincial/territorial levels has declined (Transport Canada 1999), leaving an overall decrease in funds available for highway maintenance. Therefore, as truck traffic increases and overall government spending decreases, the pavement industry will face increasingly difficult challenges to provide an adequate highway network for the public.
1.2 Pavement Structural Design and Loading Conditions 1.2.1 Pavement Structural Design Methods of pavement structural design may be classified into three categories as follows (TAC 1997).
i) Experience-based methods using standard sections;

ii) Empirical methods in which relationships between some measured pavement response, usually deflection, or field observations of performance, and structural thickness are utilized;
iii) Theory-based methods, using calculated stresses, strains, or deflections.
These are also known as mechanistic-empirical methods.
Currently, most flexible pavement structural design methods are mostly empirical methods that have been improved over the past 40 years to include deflection measurements, subgrade compressive strains and asphalt layer tensile strains. Material properties are typically characterized using elastic or resilient modulus. The most common methods are the Asphalt Institute Thickness Design Method (Asphalt Institute 1991) and the American Association of State Highway and Transportation Officials (AASHTO) Flexible Pavement Design Method (AASHTO 1993). Traffic loading in both methods consists of uniform vertical pressures applied to a multi-layered elastic system.
t.2.2 Pavement Loading Conditions Although traditional asphalt pavement analysis and design methods focus on uniform vertical stresses applied by traffic loading, there are actually 10 loading conditions commonly applied to pavements in service. These conditions are illustrated in Figure ~ (Gerrard and Harnson 1970).

1~

iii .~..-\~~.~~r/
I(a) UNIFORM YEATICAL PRESSURE P Ilbl UNIFORM VERTICAL DISPLACEMENT
N
/I \\\
MAX. STRESS PZ
t i ~ , ,_ 2la) LINEAR VERTICAL PRESSURE RZ 2(b) LINEAR VERTICAL DISPUCEM~NT
S
' HAX. STRESS p1 -,~~ r_ 3(n) LINEAR RADIAL SHEAR STRESS p1 3(b) LINEAR RADUIL SHEAR DISPLACEMENT
MAX. STRESS pz t..~-v vi ~J
4(a) LINEAR TORSIDNAL SHEAR STRESS p2 ~(b) LINEAR TORSIONAL SHEAR DISPLACEMENT
s w -r.
-w. -i 5fa) UNIFORM UNIDIRECTIONAL SHEAR STRESS p3 Slb) UNIFORM UNIDIRECTIONAL SHEAR
DISPLACEMENT
Figure 5: Common Loading Conditions of Asphalt Pavements (from Gerrard and Harrison 1970) According to Gerrard and Harnson (1970), loading by uniform vertical pressure (sub figure la) is typical of pneumatic tires and flexible foundations, while loading by uniform vertical displacement (sub figure 1b) corresponds to relatively rigid foundations. In addition to uniform vertical pressures/displacements, linear vertical pressures and displacements are also applied to pavement structures (sub figures 2a and ?b, respectively). Loading by linear vertical pressure represents moments about the horizontal axis applied to flexible pavements, while linear vertical displacements represent moments about the horizontal axis applied to rigid pavements.
In addition to vertical loading, pavements are also subjected to numerous loading conditions in shear. Linear radial shear stresses (sub figure 3a) are developed at the surface of pavements due to the grip of pneumatic tires.
Measurements by Bonse and Kuhn (1959), as well as Marwick and Starks (1941), indicate that the magnitude of the maximum stress is of the order of the tire inflation pressure. Linear radial shear stresses are present both at rest and during constant linear velocity. Linear radial shear displacement (sub figure 3b), when coupled with the uniform vertical pressure loading, gives the exact solution to the problem of a flexible foundation with a rough base (Gerrard and Harrison 1970).
The state of stress defined by linear torsional shear stress (sub figure 4a) is imposed as an automobile turns or enters a curved section of road. Linear torsional shear displacement on the other hand, may be applied to the analysis of vane shear tests at subgrade failure loads (subfigure 4b).
The final set of loading conditions consist of uniform unidirectional shear stress and displacement (subfigures Sa and Sb). Unidirectional shear stress is applied during braking, acceleration and traction of pneumatic tires, while unidirectional shear displacement represents lateral loads applied to foundations.
According to Figure 5, six of the ten possible pavement loading conditions involve shear, however, current practices only consider uniform vertical pressure.
1.3 United States Strategic Highway Research Program (US-SHRP) 1.3.1 Background and Reason for Implementation The United States Strategic Highway Research Program (US-SHRP) was a ~
year, $150 million dollar research program designed to improve the performance and durability of highways and make them safer for motorists and highway workers.
US-SHRP was initiated in response to the continuing deterioration of highway infrastructure in the United States and was intended to make significant advances in traditional highway engineering and technology through the concentration of new research funds in four key technical areas - Asphalt, Pavement Performance, Concrete and Structures, and Highway Operations (C-SHRP 1998). A total of 130 new products emerged from the US-SHRP research in the form of new equipment, processes, test methods, manuals and specifications for the design, maintenance and operations of highways (US-SHRP 1992).
1.3.2 The SUPERPAVET~' Mix Design System SuperpaveTM (Superior Performing Asphalt Pavements) was one of the major products of the SHRP asphalt research program. Unveiled in 1992, the Superpave system represented a fundamentally new system for designing asphalt concrete mixes. The performance-based nature of the system not only promoted improved pavement life, but also the potential ability to predict pavement performance based on accelerated testing (C-SHRP 1999).
Briefly, the Superpave system incorporates performance-based asphalt materials characterization with the design environmental conditions to improve performance by controlling rutting, low temperature cracking and fatigue cracking (Asphalt Institute 1997a). The Superpave system consists of three main components - the performance graded (PG) asphalt binder specification, the mixture design and analysis system, and a computer software system (Asphalt Institute 1997b).
Detailed discussion of the Superpave system is beyond the scope of this thesis and has been documented in countless other reports. It should be mentioned however, that by 2001, the AASHTO Task Force on SHRP Implementation predicts that over 80% of the hot mix asphalt produced and constructed in the United States will be designed with the Superpave system (AASHTO 1999). It is therefore clear that Superpave will be the asphalt mix design system in the United States for the foreseeable future. In Canada, Superpave implementation has progressed at a slower rate; however, it appears that Canadian agencies will also adopt Superpave as the new mix design system in the coming years.
1.3.3 Long Term Pavement Performance (US-LTPP) Project As pan of the US-SHRP, a comprehensive 20-year study of in-service pavements was initiated in 1987 to understand why some pavements perform better than others, with the ultimate Goal of building and maintaining a cost-effective highway system. This field experiment, known as the Long Term Pavement Performance (US-LTPPj project, is unprecedented in scope, consisting of over asphalt and Portland cement concrete pavement test sections across the United States and Canada (FHWA 1998a).
The original US-LTPP research plan set forth six objectives for the program (FHWA 1999a):
i) Evaluate existing pavement design methods.
ii) Develop improved design methodologies and strategies for the rehabilitation of existing pavements.
iii) Develop improved design equations for new and reconstructed pavements.
iv) Determine the effects of loading, environment, material properties and variability, construction quality, and maintenance levels on pavement distress and performance.
v) Determine the effects of specific design features on pavement performance.
vi) Establish a national long-term database to support LTS-SHRP's objectives and to meet the future needs of the highway industry.
To support these objectives, three types of studies were established: General Pavement Studies (GPS), Specific Pavement Studies (SPS) and the Seasonal Monitoring Program (SMP). The GPS experiments focus of the most commonly used structural designs for pavement. Eight types of existing in-service pavements are currently being monitored throughout North America. The performance of these structural designs is tested against an array of climatic, geologic;
maintenance, rehabilitation, traffic and other service conditions (FHWA 1999a).

In contrast, the SPS test sections were specially constructed to investigate certain pavement engineering factors. These sections allow critical design factors to be controlled and performance to be monitored for the initial date of construction. It is anticipated that the results from the SPS experiment will provide a better understanding of how selected maintenance, rehabilitation, and design factors affect pavement performance.
The SMP experiment sections were also specially constructed to provide data needed to determine the impacts of temperature and moisture variations on pavement response.
The primary product of the LTPP experiment is the Information Management System (IMS) database that contains the data collected from each of the three LTPP
studies. Administered by the Federal Highway Administration (FHWA), the IMS
database is available to anyone at no cost. To make the data more accessible and user friendly, portions of the IMS database meeting all quality control levels are released on CD-ROM under the name "DataPave." The latest version of DataPave - version 2.0 released in September of 1999, contains twice as much IMS data as its predecessor (FHWA 1999b).
1.3.4 Introduction to the AASHTO 2002 Pavement Design Guide As mentioned, most pavement design procedures are based on the AASHTO
Guide for the Design of Pavement Structures (TAC 1997). All previous and current versions of this guide have been based on performance equations developed at the AASHO Road Test in the 1950's. While previous versions of the guide have served well for almost four decades, there are a number of serious limitations to their continued use as the nation's primary pavement design procedures as outlined by McGee (1999):
~ Pavement rehabilitation design procedures were not considered at the AASHO
Road Test. Full consideration of rehabilitation design is required to meet today's needs.
~ Since the road test was conducted at one specific geographic location, it is difficult to address the effects of differences in climatic conditions on pavement performance. For example, at the road test a significant amount of distress occurred in the pavements during the spring thaw, a condition that does not exist in a significant portion of the country.
~ One type of subgrade was used for all of the test sections at the road test.
Many types exist nationally.
~ Only unstabilized, dense granular bases were included in the main pavement sections (limited use of treated bases was included for flexible pavements).
Various stabilized types now are used routinely.
~ Vehicle, suspension, axle configurations, and tire types were representative of the types used in the late 1950's. Many of these are outmoded in the 1990s.
~ Pavement designs, materials, and construction were representative of those used at the time of the road test. No subdrainage was included in the road test sections.

An additional problem with earlier AASHTO procedures is the order-of-magnitude difference between AASHO Road Test traffic loads and the loads carned by modern new and rehabilitated pavements. Road test pavements sustained at most some 10 million-axle load applications; less than cart-ied by some modern pavements in their first year of use due to the explosive growth of truck traffic over the last 40 years. Equations forming the basis of the earlier procedures were based on regression analyses of the road test data. Thus, application of the procedure to modern traffic streams meant the designer often was projecting the design methodology far beyond the data and experience providing the basis for the procedure. Clearly, the result was that the designer may have been working "in the dark" for highly trafficked projects. Such projects may well have been either "under designed" or "over designed" with the result of significant economic loss (McGee 1999).
Another major extrapolation is design life. Because of the short duration of the road test, the long-term effects of climate and aging of materials were not addressed. The AASHO Road Test was conducted over ? years, while the design lives for many of today's pavements are 20 to 50 years.
Finally, earlier AASHTO procedures relate the thickness of the pavement surface layers (asphalt layers or concrete slab) to performance. However, the observed performance of pavements reveals that many pavements need rehabilitation for reasons that are not directly related to pavement thickness (i.e.
rutting, thermal cracking, faulting etc.). Further, the primary measure of pavement performance in the earlier procedures is present serviceability (PSI or PSR) and the dominant factor effecting serviceability is pavement ride. Yet, in many cases pavement managers find that distress factors other than ride, such as cracking and rutting, control when pavement rehabilitation is required. To improve the reliability of design and to meet the needs of asset management, the management criteria and the pavement design procedure must relate to the same performance factors. To help alleviate these problems, the 2002 Guide will use the international roughness index (IRI) as a major pavement performance measure (McGee 1999).
The AASHTO Joint Task Force on Pavements (JTFP) has responsibility for the development and implementation of pavement design technologies. In recognition of the limitations of earlier Guides, the JTFP initiated an effort to develop an improved Guide by the year 2002. At the time of this writing, the National Cooperative Highway Research Program (NCHRP) is developing a major revision and update to the current AASHTO pavement design guide, under NCHRP
project 1-37A, due for release in 2002. A draft version of the new guide was completed in April 1999 although it has not been formally published (McGee 1999).
Unlike previous design guides, the 2002 guide will incorporate mechanistic-empirical concepts to better characterize the pavement structure and its constituent materi als.
Although this move represents a major step forward toward a more accurate pavement design and analysis system, the 2002 guide will only focus on vertical loading conditions on a multi-layered elastic system. Researchers concede that shear loading is important to pavement performance, however, Witczak (2000) explained that the 2002 guide will not include shear properties or loading conditions as the guide is being developed from already existing databases and test procedures.

1.4 Canadian Strategic Highway Research Program (C-SHRP) 1.4.1 Background and Reason for Implementation In 1987, The Canadian Strategic Highway Research Program (C-SHRP) was created in response to the commencement of SHRP in the United States. The objective of C-SHRP is to improve the performance and durability of highways and to make them safer to motorists and highway workers by extracting the benefits of the United States Strategic Highway Research Program (US-SHRP) and by solving highway problems having a high priority in Canada that were related to, but not duplicates of, US-SHRP projects (RTAC 1986).
C-SHRP is a dedicated program of the Council of Deputy Ministers Responsible for Transportation and Highway Safety and is managed by the C-SHRP
Executive Committee. Unlike US-SHRP, C-SHRP was always envisioned as a 15 year program with three 5-year program phases (C-SHRP 1998). Due to delays with the US-SHRP, the C-SHRP Executive Committee extended the first program phase by two years. C-SHRP Phase 1 ran from April 1987 until March 1994 and involved coordinating Canadian involvement with the US-SHRP research as well as conducting independent Canadian research related to US-SHRP. The complimentary C-SHRP research produced an additional 8 research products.
Phase 1 also saw the initiation of the Canadian Long Term Pavement Performance (C-LTPP) project, an independent experiment designed with Canadian pavement design and climatic conditions in mind.
The second phase of C-SHRP was completed between April 1994 and March of 1999. The focus of Phase 2 was technology transfer in the form of evaluating SHRP/C-SHRP research results and applying the findings to mainstream practice.
The C-LTPP project continued with a focus on data collection and management, with initial analysis of performance through Bayesian modelling procedures (Kaweski and Nickeson 1997).
The third and final phase of C-SHRP is currently underway and will conclude in April of 2004. As with Phase 2, technology transfer of SHRP products will continue as a primary focus, however, the range of products evaluated and promoted will be expanded to include products of the FHWA and AASHTO. The C-LTPP
experiment will conclude in 2004 and the resulting database will be completed for use by pavement designers and researchers to provide more cost effective pavement designs.
1.4.2 Canadian Long Term Pavement Performance (C-LTPP) Project The Canadian Long Term Pavement Performance (C-LTPP) project was initiated in 1987 as an independent Canadian experiment to investigate pavement performance. However, whereas the US-LTPP project covered all pavement types, the overall goal of the C-LTPP project is to increase pavement life through the development of cost-effective pavement rehabilitation procedures, based upon systematic observation of in-service pavement performance. As the majority of Canadian roads are asphalt concrete pavements, only new asphalt concrete overlays were selected for investigation under C-LTPP. Therefore, if comparing to the US-LTPP project, the C-LTPP test sites could be considered SPS-5 sections, although the C-LTPP sites are independent of US-LTPP.
In formulating the overall goal of C-LTPP, four distinct objectives were identified (C-SHRP 1997):

i) to evaluate Canadian practice in the rehabilitation of flexible pavements, and to subsequently develop improved methodologies and strategies;
ii) to develop pavement performance prediction models and validate other models or calibrate then to suit Canadian conditions;
iii) to establish common methodologies for long-term pavement evaluation, and to provide a national framework for continued pavement research initiatives;
iv) to establish a national pavement database to support the preceding C-LTPP
objectives as well as future needs.
A total of 24 test sites were selected for C-LTPP each with 2 to 4 adjacent test sections for a total of 65 test sections within the experiment. Each test section received an asphalt concrete overlay for the experiment. The use of adjacent sites allows for the comparison of different rehabilitation methods under identical traffic loading, climate and soil conditions. The alternative rehabilitation strategies employed on the C-LTPP test sections included variable overlay thickness, hot and cold-mix recycling, milling, inclusion of performance enhancing additives, or a combination thereof.
As with the US-LTPP, the primary product of C-LTPP is the C-LTPP
pavement performance database.
1.5 Specific Problem Definitions and Need for New Test Facility The preceding sections were included to provide some context of current research activities in the area of pavement performance, materials and design. While the past decade has seen many significant improvements in asphalt pavement technology, there remains much room for improvement, particularly in the use of shear properties to design, construct, monitor and predict the performance of ACP's. The following sections provide rationale for the current investigation; that being to design, develop and verify an advanced in-situ shear strength/stiffness test for asphalt concrete pavements.
1.5.1 Improved Characterization of Pavement Structure and Design Inputs As outlined in Section 1.2, pavement design procedures in use today only consider loading in the vertical direction, despite the fact that six of the ten common pavement loading conditions described involve shear stresses or displacements.
As indicated, the NCHRP is developing a major revision and update to the current AASHTO Guide for the Design of Pavement Structures (AASHTO 1993) under NCHRP project 1-37A due for release in 2002. Unlike previous design guides, the 2002 guide will incorporate mechanistic-empirical concepts to better characterize the pavement structure and its constituent materials. Although this move represents a major step forward toward a more accurate pavement design and analysis system, the 2002 guide will only focus on vertical loading conditions on a multi-layered elastic system. Although the vertical loading condition represents a large portion of the applied stress, the measurement of shear parameters of asphalt pavements and incorporation into analysis and design should significantly improve the reliability of pavement design, providing more cost effective pavements. Clearly, the inclusion of shear loading into the design process will be a complex undertaking that will require extensive research prior to implementation. Development of an in-situ test device will provide an important step toward this goal.

1.5.2 Simple Performance Test for Superpave Verification and QC/QA Testing One major issue not addressed in the original Superpave system was the adoption of a strength or durability test. Unlike the Marshall or Hveem methods that utilize stability or flow tests, Superpave was originally based solely on volumetric properties with no strength test for performance verification. To address this deficiency, the NCHRP initiated project 9-19 "Superpave Support and Performance Models Management" to recommend a laboratory "simple performance test" suitable for evaluating the rutting and fatigue performance of asphalt mixes and to ultimately provide input to performance models. In addition to performance testing, the simple performance test should also be capable for use in quality control and quality assurance (QC/QA) testing.
However, there is currently no in-situ performance test for Superpave or any other asphalt mix design system. The development of an in-situ shear strength/stiffness test would provide an excellent complimentary field test device allowing both faster and more cost effective performance testing and QC/QA
testing due to its portable nature.
1.5.3 The Need to Measure Field Properties Peck and Lowe ( 1960) perhaps gave the best reasoning for the introduction of field shear testing (of soils) as opposed to laboratory testing during the Research Conference on Shear Strength of Cohesive Soils in Colorado:
"It seems apparent that there are numerous unanswered questions with regard to the shear strength of undisturbed soils. Many of these arise because of doubts regarding the applicability of laboratory f ridings to field conditions. It is recognized that the mere act of obtaining a sample front a natural deposit radically alters the state of stress and induces strains, and that natural deposits are rarely homogeneous. Yet there seems to be an inclination to feel that the really fundamental research on shear strength of undisturbed soils must be done irz the laboratory, and tlzat the results of the laboratory studies nzay be applied to field conditions with a minimum of evidence to support the extrapolation. The panel discussions have indicated that there may be dangerous pitfalls in tlais path. "
1.6 The Innovations Deserving Exploratory Analysis (IDEA) Program The Innovations Deserving Exploratory Analysis (IDEA) programs, managed by the Transportation Research Board (TRB), provide start-up funding for promising but unproven concepts in surface transportation systems (TRB ?000).
The goal of the >DEA programs is to seek out and support new transportation solutions unlikely to be funded through traditional programs. IDEA programs differ from the more traditional research programs in the following ways:
1. They offer an arena for innovation. Topics are not restricted; good ideas that support the general goals of safe and efficient surface transportation are eligible.
?. Their impact is timely. Fledgling ideas take flight only when their development is nurtured. The IDEA programs foster good ideas at a critical early stage in the hope that they soon will take off on their own.
3. The proposal process is simple and accessible. Proposals are accepted at any time and awards are made twice each year. There are no prerequisites for submitting proposals; good ideas are welcome from anyone.

There are IDEA programs covering four major transportation areas - Highways, High Speed Rail, Intelligent Transportation Systems and Transit. As part of the NCHRP, the Highway-IDEA program is managed by the TRB and jointly supported by the FHWA and the member states of AASHTO. The program seeks advances in the construction, safety, maintenance, and management of highway systems (TRB
?000).
In September of 1997, a proposal entitled "Design, Development and Verification of an Advanced In-Situ Shear Strength Test for Asphalt Concrete Pavements" was submitted to the TRB for a NCHRP Highway IDEA Concept Exploration project (Goodman and Abd El Halim 1997). In February of 1999, Carleton University was awarded NCHRP IDEA Project #55 to develop the in-situ shear strength/stiffness test facility. The work completed for that project constituted a substantial portion of this thesis.
1.7 Organization and Scope of Thesis 1.7.1 Chapter 1: Introduction Chapter 1 commences with a definition of asphalt concrete pavements and their importance to the Canadian economy, followed by a description of the various pavement loading conditions. A brief introduction of the United States and Canadian Strategic Highway Research Programs (US-SHRP and C-SHRP) is then provided to introduce recent and ongoing research efforts, as well as give context to the specific problems associated with current practice and the need for a new field test device for asphalt concrete pavements. Chapter 1 finishes with an introduction to the Innovations Deserving Exploratory Analysis (IDEA) Program, which provided substantial funding toward the development of the test facility, as well as the organization and scope of this thesis.
1.7.2 Chapter 2: Literature Review Chapter 2 consists of an extensive literature review completed to examine all aspects of the asphalt pavement rutting phenomenon including its manifestations, mechanisms, procedures for quantification, and a detailed review of the contribution of numerous variables toward rutting resistance. Furthermore, a state-of-the-practice review of current laboratory and field rutting test methods is provided with a discussion of the various limitations associated with these devices.
1.7.3 Chapter 3: Review of Previous Work and Analytical Modelling The foundation for the current investigation was laid by two previous investigations completed at Carleton University. The first was a comprehensive laboratory investigation of asphalt shear properties and pavement rutting completed by Zahw (1995). Chapter 3 begins with a review of that investigation, followed by the results of a new study to investigate the relationships between asphalt mix characteristics, shear properties and pavement rutting, using data collected during his research. The second underlying research effort involved the construction of a first generation in-situ shear strength test device by Abdel Naby (1995). A
review of the device, known as the Carleton In-Situ Shear Strength Test (CiSSST), is provided including its main benefits and the results of his research concerning in-situ shear strength and its relation to pavement performance. The chapter concludes by introducing an improved analytical approach to derive asphalt pavement shear properties from the surface plate loading condition using closed form equations and the finite element method.

1.7.4 Chapter 4: Development of the In-Situ Shear Stiffness Test (InSiSSTTM) The primary basis for this thesis was the design and development of an advanced in-situ shear stiffness test for asphalt pavements. The design of the new device, entitled the In-Situ Shear Stiffness Test (InSiSSTTM) was conceived through an analysis of the deficiencies observed with the CiSSST prototype built in 1995.
The advantages and benefits of the InSiSSTTM are described, as well as some of the challenges experienced during its fabrication. Chapter 4 concludes with an initial set of instructions and procedures to carry out field testing with the InSiSSTT''' 1.7.5 Chapter 5: Preliminary Testing and Validation Once fabrication of the InSiSSTTM was complete, preliminary field tests were completed for validation purposes. Chapter ~ first presents the results of an exercise to validate the linear elastic assumption made by the analytical models presented in Chapter 3. Next, the results of comparison testing with the CiSSST and InSiSSTT"' devices are presented, as well as an interesting observation concerning the test plate diameter. Chapter 5 concludes with a comparison of field shear properties to those observed in the laboratory by Zahw (1995).
1.7.6 Chapter 6: Conclusions and Recommendations The final chapter summarizes the project objectives and the rutting phenomenon, as well as the conclusions observed during the investigation.
Recommendations for additional modifications to the InSiSSTTM and future testing are also presented.

CHAPTER 2: LITERATURE REVIEW
2.1 Permanent Deformation of Asphalt Concrete Pavements Although the term "rutting" is often used interchangeably with permanent deformation, it is only one of four observed manifestations as described below:
~ Rutting is characterized by channelized depressions (troughs) that run longitudinally in the wheelpaths. Rutting may or may not be accompanied by shoving of the pavement adjacent to the wheelpaths.
~ Correcgations and Slcovircg is characterized by ripples along the pavement surface formed by alternating areas of settlement and/or heave.
~ Grade Depressions and Settlement are manifested as irregular or localized areas of settlement (not specifically in the wheelpaths).
~ Upheaval or Swell consists of localized upward expansion of pavement due to swelling of underlying soils (base or subgrade) through moisture infiltration or frost heave.
However, rutting is the most common form of permanent deformation analysed.
Therefore, the terms rutting and permanent deformation will be considered the same for the purposes of this thesis.
2.2 Manifestations of Rutting Rutting itself is manifested in a number of forms depending on which of the pavement layers was responsible for the deformation. A novel and accurate method of determining the layer at which rutting occurred was investigated by Simpson et. al.
(1995). During their investigation, it was observed that the shape of the pavement transverse profile was theoretically indicative of where the rutting originated within the pavement structure. Using data from the US-LTPP Program, it was shown that the transverse profiles generally fit into one of four categories representing (a) subgrade rutting, (b) base rutting, (c) surface rutting, or (d) heave. Illustrations of these categories are shown in Figure 6.
Classification of transverse profiles into the four above categories was completed using the algebraic area between the transverse profile and the straight line connecting its end points as illustrated in Figure 6. Profile areas that were entirely negative proved to be the result of subgrade settlement while areas that were entirely positive were the result of heave. Furthermore, marginally positive profile areas were the result of surface rutting, whereas marginally negative areas were the result of base layer rutting.
The criteria used to classify transverse profiles from the US-LTPP database are shown in Table 1. Of the 134 US-LTPP sections analyzed, only six transverse profiles did not agree with the classification system. Therefore, over 95% of transverse profiles were correctly classified using Table 1.
Table 1: Classification Criteria for Transverse Profiles (from Simpson et. al. 1995) Transverse Profile Layer Responsible for Rutting Area of DistortionRatio of Positive to (mm2) Negative Area of Distortion Subgrade < -4500 < 0.4 Granular Base -4500 to 700 0.4 to 1.25 Surface (Asphalt)700 to 5000 1.25 to 3.0 Heave > 5000 > 3.0 a. SUBGRADE
b. BASE
c.SURFACE
d, HEAVE
Figure 6: Transverse Profiles of Various Rutting Manifestations (from Simpson et. al. 1995) 2.3 Asphalt Surface and Overlay Rutting 2.3.1 Location of Pavement Rutting Although rutting due to heave, subgrade or granular base failure can and does occur, rutting in properly constructed pavements is usually observed due to deformation of the asphalt layers. This statement was confirmed through a comprehensive National Rutting Study that was completed in United States by the National Center for Asphalt Technology (NCAT) in 1987. As part of the investigation, trench cuts were made in selected roads to observe where in the pavement structure that rutting was occurring. Reports by Cross and Brown (199?, 1991) revealed that most of the rutting occurred in the top 75 to 100 mm (3 to 4 in.) of the pavement, therefore, almost exclusively in the asphalt concrete liver.
The amount of rutting in the base coarse was insufficient to measure. The results of the MCAT study have been confirmed in other studies including Gervais and Abd El 1-lalim (1990). Figure 7 illustrates a cross section of an asphalt pavement section displaying surface rutting. Note that the bottom edge of the asphalt layer is almost completely flat while the surface is severely rutted.
Figure 7: Illustration of Surface Rutting (from Gervais and Abd EI Halim 1990) 2.3.2 Surface/Overlay Rutting Mechanism # 1 - Traffic Induced Densitication With the exception of wear rutting caused by studded tires, surface or overlay rutting is caused by two main mechanisms. The first is called "traffic-induced," or ".......... ,_...."~"~;,.,~~,~,~r.~4:::.:-,_~:.~w,.:..

"post" densification. Initial densification of the various pavement layers, including the asphalt concrete, occurs during the construction phase in the form of compaction. During compaction, the layers are compacted to form a dense, consistent structure to resist traffic loading. However, layers that are not compacted to a high degree retain significant void space. Under subsequent traffic loading, the layers continue to consolidate as the excess voids are removed. This process is known as traffic induced densification.
For properly constructed pavements, L1S-SHRP (1994) reports that traffic induced densification is generally not considered to be the cause of substantial rutti n a.
2.3.3 Surface/Overlay Rutting Mechanism # 2 - Shear (Plastic) Flow The second and more critical rutting mechanism is shear or "plastic°' flow of the asphalt concrete mix under traffic loading. Shear flow involves lateral movement of the asphalt cement and reorientation of aggregate particles under traffic loading to a new or more stable equilibrium. The movement (strain) incuwed by the asphalt concrete is not recovered once unloaded, resulting in permanent deformation. For properly constructed pavements, shear deformations caused primarily by large shear stresses in the upper portions of the asphalt/aggregate layers are dominant (US-SHRP 1994j.
The shear flow phenomenon was the underlying reason for the development of the InSiSSTT'''. It is hypothesized that increasing the resistance of the asphalt mix to shear deformation will reduce rutting and improve long-term pavement performance significantly.

2.3.4 The Rutting Cycle The two rutting mechanisms mentioned above are not mutually exclusive.
Indeed, a rutting cycle is observed in most instances during the design life of a properly constructed asphalt pavement. The three stages of permanent deformation with traffic loading are listed below (Carpenter 1993):
1. Primary - initial compaction and traffic induced densification ?. Secondary - stable shear period 3. Tertiary - rapid unstable shear failure Rutting is initiated with continued densification of the asphalt layer under traffic loading. Kandhal et. al. (1993) explain that during this stage, rutting is directly proportional to traffic.

0 ~6 _ 0 ~2 d 0' iv c 0.08 a 0 06 O.Oc .
0 02 ~
Initial l)cn.iCir.aion 0 05 1 i5 2 25 3 35 4 45 5 Load Repetitions (Millions Figure 8: The Progression of Rutting with Traffic Loading (Rutting Cycle) (from Carpenter 1993) s4 The second phase involves a stable shear period during which the rate of rutting decreases with increasing traffic until a condition of plastic flow occurs and the rate of rutting main increases (rapid unstable shear failure). Therefore, prevention of the onset of the tertiary flow stage should increase pavement life considerably with respect to permanent deformation.
2.4 Quantification (Measurement) of Rutting Rutting is quantified using various methods and measures for evaluation and modelling purposes. Perhaps the most common measures are average rcet depth or naccxintccm race depth in millimetres or inches as measured transversely across the pavement width using straightedges, profilometers, or other depth measuring devices.
A second commonly used measure is average raetting rcate; the rut depth divided by the amount of traffic to which a particular pavement has been subjected.
This measure allows different pavements to be compared directly. Previous work by >=3rown and Cross (1998), and Parker and Brown (1990) has shown that expressing the rate of rutting as a function of the square root of total traffic yields higher correlation with observed pavement behaviour when compared to other expressions (such as the log of traffic).
As previously presented, Simpson et. al. (1995) quantified rutting by deteumining the total area of deformation from transverse profiles obtained for the US-LTPP
Project.

2.5 Categories for Rutting Variable Classification The rutting phenomenon is very complex. At present, there is no single independent variable that completely captures or predicts rutting. In addition, Kandhal et. al. (1993) emphasized that a single "bad" property, such as excessive asphalt content, can nullify other good properties, such as coarse aggregate with 100 r'c fractured faced count. Furthermore, there are also numerous interactions between the independent variables, making analysis or modelling more difficult.
For this investigation, the factors affecting permanent deformation were classified into eight unique cate;ories and a ninth category encompassing combinations of the first eight. The categories are listed in 'fable 2 in no particular order of importance.
Table 2: Categories for Rutting Variable Classification Category ~ Variable Grouping i A Bituminous Materials and Additives I

B Mineral Aggregates ~ Mix Design Parameters o C

D Strength/Resistance Properties of Mix E Pavement Structural Design I

F J Construction-Related G Environmental-H Traffic-Related X Combinations of Above Categories A breakdown of the individual categories and their variables is presented in the following sections.

2.6 Category A - Bituminous Materials and Additives As defined by the TAC (I997j, bituminous materials are petroleum-based products of oil refining or naturally occurring asphalts. Asphalt is a dark brown to black solid or semisolid cementitious material that softens and liquefies when heated.
Asphalt cement (ACj receives further refinement to meet specifications for paving and related purposes.
Although liquid asphalts (made from asphalt cement] are available, asphalt cement is the only material used for the production of hot-mix asphalt concrete (HMAC). Liquid asphalts are used primarily for cold-mix asphalt concrete, a less durable material not regularly used for road construction in Canada.
Therefore, only asphalt cement and HMAC-related variables were considered during this investigation.
2.6.1 Effect of Chemistry Asphalt cement consists of asphaltenes, oily constituents and asphaltic resins (TAC 1997). At present, data that conclusively relates the chemical properties of asphalt cement to permanent deformation is relatively limited.
A 1983 study performed on Interstate 90 in Montana compared the performance of asphalt cements from each of that state's oil refineries.
Reports by Bruce (1987) and Jennings et. al. (1988] showed that asphalt cements containinj lower levels of asphaltenes and saturates were more susceptible to rutting. No conclusive relationships were observed for other constituents such as polar aromatics or naphthene aromatics. However, the authors cautioned that mix design asphalt content could have been an overriding factor for that investigation.

., -, W
2.6.2 Effect of Penetration/Viscosity Penetration is an empirical measure of asphalt cement hardness. In theory, the greater the penetration, the more susceptible the asphalt concrete pavement is to permanent deformation. However, the penetration test is performed at a standard temperature of 2~°C whereas rutting typically occurs at more elevated temperatures between 40°C and 60°C. Therefore, while various asphalt cements may yield similar penetration values, their high temperature performance may be markedly different. This has been the case to date since little data conclusively relates penetration to permanent deformation.
Viscosity is a measure of the resistance of the asphalt binder to floe (shear) at a specified temperature. Unlike penetration, viscosity measurements are often made at high temperatures to better capture the asphalt cement properties at temperatures more indicative of rutting (see Category G for the effect of temperature on pavement rutting).
Results of Bruce (1987) and Jennings et. al. (1988] did indicate that higher penetration and viscosity numbers seemed to result in 'renter rutting.
However. this trend may have been influenced by the aspl;alt content in the mixes.
The NCAT National Rutting Study completed by Cross and Brown (1992, 1991) indicated that hi'Ther penetration values were cowelated with increased rutting for mixes with more than '?.~~i'o air voids in-place. However, the degree to which penetration correlated with rutting was much lower than that observed for ag're~~ate properties. Conversely, a rutting study connpleted in Saskatchewan by Huber and Heiman (1987) concluded that penetration and viscosity did not demonstrate a significant effect on rutting performance.
Nievelt and Thamfld (1988) concluded that asphalt cements with higher viscosity values produced asphalt mixes with greater rutting resistance as tested using wheel-tracking tests on samples at multiple test temperatures.
It should be mentioned that the new Superpave performance-graded (PG) binder specification has been implemented in Ontario (MTO 1998). This new specification does not rate asphalt cements based on penetration at a standard test temperature. Indeed, a rigorous testing regime using a combination of the Dynamic Shear Rheometer and Dynamic Viscometer measure viscosity at medium and high temperatures to better characterize high terr~perature binder performance (Asphalt Institute 1997a). However, penetration has been included in this investigation as the vast majority of existing roads were designed using the Marshall method, not the Supeyave mix design system. Clearly, the effect of PG Binder properties would be applicable to new roads designed using Superpave.
2.6.3 Effect of Modifiers As will be discussed further in Category G, asphalt cement properties are highly temperature sensitive. As temperature of the asphalt cement increases, its stiffness decreases, thereby increasing the potential for permanent deformation, as a larger deflection of the asphalt layer is incurred under the same load. The high temperature susceptibility of asphalt cement may be reduced through the addition of polymer modifiers. Terrel and Epps (1988) list a number of modifiers in Table 3.

Table 3: Various Asphalt Cement Modifiers (from Terrel and Epps 1988) Polymer Type ~j Example Natural latex Natural rubber ~

Synthetic latexStyrene-butadiene (SBR) Rubber Block copolymerStyrene-butadiene-styrene (SBS) Reclaimed rubberRecycled Tires Polyethylene Plastic Polypropylene Ethyl-vinyl-acetate (EVA) ~- Combinationcombination I of above 2.6.4 Effect of Other Additives In addition to polymer modifiers, other additives such as liquid anti-stnppina agents or hydrated lime are used to improve the bond between the asphalt cement and the aggregate particles. Krutz and Stroup-Gardiner ( 1990) investigated the influence of moisture damage on rutting for the Nevada Department of Transportation (DOT). The. loss of asphalt cement from stripping allowed the ag'reaates to shift, causing severe rutting of the pavements analysed.
Therefore, the use of anti-stripping additives should reduce rutting by reducing the loss of asphalt concrete through moisture damage.
2.7 Category B - Mineral Aggregates 2.7.1 Effect of Source Properties Aggregate source properties include soundness, toughness and deleterious materials. These propea-ties are empirical in nature and are usually used only to evaluate local aggre~uate sources on a comparison basis.

While studying the effect of gradation on permanent deformation for the Nevada DOT, ILrutz and Sebaaly (1993) compared source properties of various aggregates. A test regime utilising standard triaxial and repeated triaxial tests revealed that rutting performance of mix gradations containing substantial amount of fine material (i.e. finer overall gradations) was directly linked to source properties.
While not targeting any individual pavement distress, Wu et. al. (1998) subjectively compared source properties to pavement performance. The Micro-Deval (toughness) and Magnesium Sulphate Soundness (soundness) tests were more strongly correlated with the subjective pavement penormance ratings compared to other source property tests such as the Los Angeles Abrasion and Freeze-Thaw Soundness tests.
2.7.2 Effect of Consensus Properties Aggregate consensus properties include coarse aggregate an~ularitv, fine aggregate angularity, flat and elongated panicles and clay content.
The NCAT National Rutting Study concluded that the effect of aggregate angularity on rutting was dependent on in-situ air voids (Cross and Brown 199?, 1991). For in-situ air voids above 2.S~lo, the angularity of the coarse aggregate (two or more crushed faces) and the National Aggregate Association uncompacted voids for the fine aggregate (now referred to as ''fine aggregate angularity"j were highly correlated with rate of rutting. If the in-situ voids were less than 2.5°~~, rutting was likely to occur regardless of aggregate properties.
Marks et. al. (1990) concluded that the percentage of crushed aggregate strongly influenced creep resistance factors. As the percentage of crushed material increased, creep resistance factor also increased. The relationship of creep to rutting is presented in Category D.
Button et. al. (1990) observed the relationship between aggregate properties and permanent deformation as the amount of manufactured (crushed) sand was replaced with rounded natural sand in the mix. The first observation was that the texture, shape and porosity of the fine aggregate were major factors related to plastic deformation. Second, permanent deformation increased significantly as the percent of rounded natural sand increased (i.e. as manufactured sand was replaced).
No information was found for flat/elongated particles or clay content at this time. However, it is well known that flat and elongated particles tend to break under compactive effort, alterinU the gradation of the mix. Clay is a compressive soil that can potentially change volume with time and moisture causing localized swell or settlement in the pavement structure.
2.8 Category C - Mix Design (Volumetric) Parameters 2.8.1 Introduction to Volumetric Parameters A comprehensive historical review (and reinterpretation) of mix design volumetrics was presented by Coree (1999). Coree divided volumetric parameters into two main categories - Primary and Secondary as detailed below.
Primary Volumetric Parameters The primary volumetric parameters are those relating directly to the relative volumes of the individual components:
~ Air Voids, V~ - the volume of air voids ~ Binder Volume, Vb - the volume of the bituminous binder ~ Aggregate Volume, V; - the volume of the mineral aggregate Figure 9, commonly referred to as a "phase diagram," displays the various volumetric components.
Air Voids, Vv Effective Binder Total Binder Volume, Vbe Volume, Vb Total bsorbed Binder Volume, Vba Volume, Vtotal Effective Aggregate Bulk Aggregate Volume, Vsb Volume, Vse rigure 9: Phase Diagram of Mix Constituents in Compacted Specimen (from Coree 1999) As shown, the bituminous binder is absorbed into the external pore structure of the aggregate such that a portion of the aggregate and bituminous binder share space. Therefore, the sum of the individual volumes (V~ + V5) is greater than their combined volume (Vh+s). This situation allows further sub-division of the primary parameters given above:

~ Effective Binder Volume, Vhe - the volume of bituminous binder external to the aggregate particles, i.e., that volume not absorbed into the aggregate ~ Absorbed Binder Volume, Vba - the volume of bituminous binder absorbed into the internal pore structure of the aggregate ~ Bulk Aggregate Volume, Vsb - the total volume of the aggregate, comprising the "solid" aggregate volume, the volume of the pore structure permeable to water but not to bituminous binder and the volume of the pore structure permeable to the bituminous binder ~ Effective Ag'Tregate Volume, VS~, - the volume of the aggregate comprising the "solid" aggregate volume and the volume of the pore structure permeable to water but not to bituminous binder.
~ Apparent Aa~Tregate Volume, V5~ - the volume of the "solid" aggregate, i.e., that volume permeable to neither water nor bituminous binder.
Secondary volumetric parameters Three additional parameters have been widely used, and at various times, have formed critical deli<m thresholds. These are the Percent Air Voids, V~, the Voids in the Mineral Aggregate, VMA, and the Voids Filled with Asphalt, VFA.
~ Percent Air Voids, V~ - simply V'- expressed as a percentage of the total volume of the mixture.
~ Voids in the Mineral Aggregate, VMA - the sure of V~~ and ~',,~ expressed as a percentage of the total volume of the mixture. This parameter is directly analogous to "porosity" in soil mechanics.

~ Voids Filled with Asphalt, VFA - the degree to which the VMA are filled with the bituminous binder, expressed as a percentage. This property is directly analogous to the "degree of saturation" in soil mechanics.
It is important to recognize that Va, VMA and VFA are highly dependent on the decree of compaction and, according to Coree (1999), secondary parameters should never be quoted without referencing the degree of compactive effort used.
The following relationships may be derived from the above definitions and Fi pure 9:
i VFA = 1 ''e * 100 ~;, _ ~'= * 100 VMA = y + I'" * 100 T[~ r t ' LE~ ~Irunl ~~lrrrrli Simple algebraic manipulation reveals that the above equations are not mutually exclusive, since:
VMA-V
~'FA = ° * 1 C)0 VMA
Coree (1999) explained that in the process of mixture design, it is frequently necessary to seek to chance the magnitude of one or more of these parameters.
For example, upon analyzing a mixture, it may appear desirable to increase the VMA
(a relatively common problem), or to manipulate the air voids. Vauious recommendations and techniques exist to achieve this. However, it is neither clear what effect such a change might have on the other parameters, nor whether that change might, in itself, compromise compliance in another direction. Indeed, no such change in any one parameter should ever be contemplated without checking the effects on the other two.

Although the interaction of volumetric properties is complex and greatly affects pavement performance, the effects of individual properties have been noted by numerous studies and are presented below.
2.8.2 Effect of Air Voids In-situ air void content has been identified as perhaps the most critical parameter affecting rutting. Furthermore, the range of air voids identified for good rutting resistance is well known. Indeed, it was observed that for the roads selected during the NCAT study, none of the 50 or 75-blow mixes displayed unacceptable rutting rates if in-situ air voids remained greater than 4%r.
Brown and Cross (1990) provide the following explanation:
An asphalt mixture with low voids acts very much like a saturated soil.
It has no shear strength. When the vc>ids are reduced to a very low level (2 to 3 percent), pore pressures terLd to build up Lender traffic, the effective stress on tire aggregate is reduced, and shear or plaszi.c floe takes place.
Therefore, pavements which retain 4~7o air voids or greater after years of traffic loading show excellent performance with respect to rutting.
Both Huber and Heiman (1987) and Kandhal et. al. (1993) concluded that pavements begin to exhibit plastic deformation when the air voids reached threshold values (usually 3~Ic or less).
2.8.3 Effect of Asphalt Cement Content Asphalt cement content refers to the amount (percentage) of asphalt binder in the asphalt mixture by weight. Effective asphalt content does not include the asphalt binder absorbed in the mineral aggregate. For rutting resistance, asphalt cement content should be relatively low to prevent shear flow under loading and elevated temperature. However, enough asphalt cement must be present to bind the aggregate particles in place.
Huber and Heiman (1987) concluded that asphalt content and voids filled with asphalt were the most basic parameters affecting rutting, while Abd El Halim et. al.
(199_5) indicated that rutting decreases with increasing asphalt content to a maximum value, after which rutting increases with increasing asphalt content.
The Montana study, while not exploring the effect of asphalt content specifically, indicated indirectly the importance of asphalt content by cautioning that mix design asphalt content could have been the oven-iding factor for that investigation (Bruce 1987 and Jennings et. al. 1988).
Cross and Brown (199?, 1991) indicated that asphalt cement content was extremely important to rutting resistance. Kandhal et. al. (1993) echoed the importance of asphalt cement content by indicating that excessive asphalt content could effectively nullify other good properties of a mix such as crushed aggregates.
2.8.4 Effect of Uradation A study of 3? asphalt concrete overlays placed over rigid pavements completed by Carpenter and Enockson (1987) indicated that the majority of rutting problems could be attributed to gradation. The tender mix phenomenon associated with a "hump" in the 0.45 power gradation chart was associated with rutting.
Furthermore, the percent passinj the No. 40 sieve and retained on the No. 80 sieve was found to influence rutting.
As previously mentioned, the effect of gradation on permanent deformation was studied by Krutz and Sebaaly (19931 for the Nevada Department of Transportation (NDOT). ,A second conclusion from this study v.~as that rutting resistance of finer gradations was influenced by binder characteristics more so than for more coarse gradations. Conversely, the performance of coarse gradations is more dependent on aggregate properties and less sensitive to binder type.
Work completed by Anani et. al. (1990) in Saudi Arabia indicated that a finer gradation of coarse portion of the aggregate (No. 4 and above) improved rutting resistance. This conclusion is in general disagreement with conventional (North American) mix design, however, as indicated by Krutz and Sebaaly (199p), finer Gradations are more sensitive to binder type than more coarsely graded mixes.
The binder used for the coarse mixes may have been different from that used for the finer graded mixes (or a different asphalt content). Furthermore, the loading condition of the selected roads was not reported. The effects of loading conditions on rutting resistance are explored in Category E.
Brown and Bassett (1990) indicated that increasing the maximum aggregate size of the mix increased the mix quality with regard to creep performance, resilient modulus and tensile strength. Each of these properties has an important relationship to rutting performance as will be presented in Category D.
2.8.5 Effect of VMA and ~'FA
The use of VMA and/or VFA to predict pavement performance has been debated for almost 100 years. Specifications for VMA and VFA were originally developed to provide a minimum asphalt content in the mix for durability and a minimum voids content for rutting.
While various researchers argue which parameter better predicts performance, most agree that increasing VMA and VFA (to maximum values) are good methods to reduce rutting. Sorne investigations concerning VMA and VFA with respect to rutting are as follows.
Anani et. al. (1990) indicated that VMA was a primary variable with regard to rutting for the surface coarse. VMA of the unrutted sections was higher than that of the rutted sections.
Huber and Heiman (1987) indicated that VFA was one of the most basic parameters affecting rutting. Increasing VFA to a maximum value decreased rutting, while Carpenter and Enockson (1987) expressed that VMA was a significant variable for rutting.
The NCAT study indicated that VMA was more significant for the base coarse than the surface coarse (Cross and Brown 1992, 1991).
2.8.6 Effect of Dust Content Coarsely Graded asphalt mixes often include a relatively high proportion of dust to increase the stiffness of the asphalt binder. This is particularly true of Stone :Mastic Asphalt (SM.A) mixes that require stiff mastics to prevent the lane aggregate panicles from moving under load. As with polymer modification, increasing the stiffness of the asphalt cement by increasing dust content reduces susceptibility to rutting (at the expense of fatigue resistance). However, mixes with too much dust may display poor adhesion between the asphalt cement and the aggregate particles (stripping).
2.8.7 Effect of Laboratory Density and Compaction One basic assumption underlying the mix design process is that prepared laboratory specimens will have the same density (and air voids) as the mix in the field after years of traffic (typically 4 percent). Insufficient laboratory compactive effort results in low in-situ air voids since primary (construction) and secondary (traffic) compactive effort will be greater. .As previously discussed, low air voids (below 3%) are a major cause of rutting in asphalt pavements. Conversely, excessive laboratory compaction leads to in-service pavements with high air void contents (10% or even higher). This situation results in excessive traffic induced compaction. Furthermore, continuous air voids are formed at high voids content, increasing permeability, which can reduce durability through accelerated ageing and/or stripping.
For the Marshall method, the MCAT study observed that stronger relationships between mix properties and rate of rutting were. found for 7>-blow mixes than with 50-blow mixes (Cross and Brown 1992. 1991). This result is expected since 7>-blow Marshall mixes are designed specifically with rutting resistance as the primary design criteria. Fifty-blow mixes are designed for lower volume roads, whose primary design criteria are likely fatigue and thermal cracking resistance as opposed to rutting.
Kandhal et. al. ( 1990 found that mixtures ~u-e generally compacted to a higi~er dejree by traffic than that provided by laboratory (Marshall) compaction. 1t was therefore recommended that laboratory compactive effort be increased for pavements designed specifically for heavy traffic.
Particle orientation under compactive effort also contributes to rutting as indicated by varyin~l performance observed with Marshall compaction, gyratory compaction and roller compaction. ILandhal et. al. ( 199;x) concluded that a Marshall compactor with rotating base and slanted foot gave the highest density overall when compared to standard Marshall and gyratory compaction. However, the gyratory compactor achieved densities greater than the standard Marshall compactor for large aggregate gradations. Additional research concerning laboratory compactor type was completed by SHRP. Studies by Harvey and Monismith (1993), and Sousa et.
al. (1991) have concluded that gyratory, rolling wheel, and kneading compaction produced specimens with significantly different permanent deformation responses to repeated shear loadin'T. This indicated that each compaction method caused a particular type of aggregate structure and binder-aggregate film. It has also been shown that fatigue behaviour of a compacted mix is influenced by the mixing and compaction viscosities of the binder (Harvey et. al. 1994).
It should be mentioned that the Superpave system has adopted gyratory compaction, as the specimens are much smaller and easy to handle.
Unfortunately, gyratory compaction does not produce the same aggregate and void structure as field compaetion, therefore, permanent deformation response of gyratory compacted specimens will not be representative of an in-service. pavement.
Finally, consistency of laboratory specimen preparation is another important consideration. A round-robin test program completed by Lai (1993) investigated the variation in laboratory compacted specimens tested using the Georgia Loaded Wheel Tester by six different laboratories. Each laboratory prepared test specimens using materials provided by Georgia DOT. Lai observed that although the variation in density among specimens prepared within each laboratory was very low, the variation among individual laboratories was very large. 'This indicated that different laboratories used different preparation techniques, which in turn affected the performance of the laboratory specimens. Indeed, rutting observed from the LWT
was significantly different among the different laboratories for the exact same mix.

~l The results clearly indicate that improvement and standardisation of laboratory preparation specifications is required.
2.9 Category D - Strength/Resistance Properties of Mix 2.9.1 Effect of Marshall Testing Like penetration, the Marshall stability and flow tests are empirical in nature.
Not surprisingly, results from Marshall tests have not yielded consistent information regarding rutting resistance. Huber and Heiman (1987) concluded that Marshall stability and flow values did not show an independent effect on rutting performance.
Similarly, the NCAT National Rutting Study concluded that Marshall recompacted mix properties (stability and flow) did not correlate well with rate of rutting (Cross and Brown 1992, 1991).
Conversely, Anani et. al. (1990) concluded that Marshall stability was generally higher for unrutted sections and was a significant variable for rutting in the base asphalt coarse.
It should be noted that Hveem stability has been directly correlated with rutting since the test assesses the shear capability of the mix. However, since the Hveem method is not used in Canada, it will not be pursued further for this project.
2.9.2 Effect of Shear Strength and Stiffness Shear properties of an asphalt pavement are achieved through both aggregate particle contact to foam a ti4~ht, load-bearing skeleton and the asphalt binder that holds the particles in place. At elevated temperatures, Alavi and Monismith (1994) concluded that the influence of the agare~ate skeleton is more pronounced than j~
binder properties. However, the influence of the binder on shear stren~th/stiffness increases dramatically at the onset of plastic failure as the in-situ air voids decrease below ?.5%.
During the NCAT study, gyratory testing machine (GTM) specimens were tested for shear properties and showed that the gyratorv shear index (GSI) had higher correlation with rutting than the Marshall stability and flow (Cross and Brown 199?, 1991). The best relationship found was between rutting and GTM
shear strength. As the shear strength decreased (GSI increased), the rate of rutting increased as well.
Kandhal et. al. X1993) also confirmed that GSI values were directly proportional to ruttin~~ performance. Pavements displayin;~ high GSl values indicated high potential for rutting.
The importance of shear strength and stiffness has been emphasised by the United States Strategic Highway Research Program (US-SHRP 1994). US-SHRP
research has identified that rutting appears to be more closely related to shear stress than normal or horizcmtal stresses. The SHRP research also referenced work by Celard (1977), who emphasised that based on the results of dynamic creep tests, the rate of permanent deformation was strongly related to shear stress. For example, Celard increased the shear stress from 0.1 MPa to 0.2~ MPa (at constant normal stress of 0.1 MPa) and observed a 100-fold increase. in the rate of permanent deformation. However, varying the normal stress did not appear to change the rate of permanent deformation.

5~
Laboratory analysis of asphalt concrete cores by Abd EI Halim et. al. (1990 indicated that increasing the shear strength of any asphalt mix can reduce surface rutting significantly.
2.9.3 Effect of Resilient Modules and Indirect Tensile Strength The resilient modules Mr, is defined by Huang (1993) as the elastic modules based on recoverable strain under repeated loads. Although asphalt pavements incur some permanent deformation after each load, if it is assumed that the load is small compared to the strength of the asphalt and is repeated a large number of times. the deformation under each load is almost completely recovered and can be considered as elastic (Huang 1993). The resilient modules is determined through the indirect tension apparatus, and is considered a non-destructive test, allowing the same sample to be used a number of times under different loading and environmental conditions. The indirect tensile strength test, however, fails the sample.
The strength tests completed by Carpenter and Enockson (1987) showed that resilient modules and indirect tensile strength bore a strong relationship to rutting for asphalt overlays placed over concrete bases.
Anani et. al. (1990) indicated that resilient modules was inversely related to rutting for both the surface and base asphalt layers. Unrutted sections generally displayed higher M~ values than rutted pavements, however, threshold values were not given in the investigation.
Abdel Nabi ( 1995] observed a linear relationship between the laboratory shear strength and indirect tensile strength of asphalt cores. This finding indirectly ~4 suggested a relationship between indirect tensile strength and rutting through the relationship to shear strength.
2.9.4 Effect of Creep Because asphalt concrete is a viscoelastic material, its properties are temperature and time dependent. One method to characterize this behaviour is through creep compliance at various times. Huang (1993) noted that at constant stress, the creep compliance is the reciprocal of Young's Modulus. Creep compliance is determined through a creep test, involving either static or dynamic creep loading.
Van de Loo ( 1974) analysed the relationship between rutting and creep testing. Data from static and dynamic creep tests indicated that mix stiffness decreased as the number of load applications increased (likely due to strain softenin~T). During the same. study, Van de Loo also developed a method of estimating rut depth based on his results, often referred to as the "Shell Method."
2.10 Category E - Pavement Structural and Geometric Design 2.10.1 Effect of Order of Rigidity of Pavement Layers Structural design of asphalt pavements is a critical component to rutting performance as the state of stress and strain under traffic loading is directly related to the structural system. Typical newly constructed pavements (asphalt concrete or PCC) are constructed such that the quality and strength of the pavement layers decreases with depth. I=or example, the asphalt concrete (or PCC) is of higher quality and strength (rigidity or stiffness) than the granular base layer(s).
which in J
turn is of higher quality and strength than the natural subgrade. Under this condition, the compressive load applied by the tires causes beam action in the asphalt layer subjecting the top of the granular base and subgrade to vertical compressive stress while the bottom of the asphalt layer is subjected to tensile stress. The actual stresses expepenced in the pavement structure are dependent on the modular ratios of the constituent layers. Huang (1993] shows that, as the modular ratio increases, the vertical compressive. stresses applied to the top of the subgrade are lowered significantly. A high modular ratio is therefore desirable to minimise rutting of the subgrade layer. Rutting of these pavements typically occurs over longer time as the ~~ranular layers and/or subgrade consolidate under the asphalt layer.
However, in the early 1980's, observations made in Ontario and Nova Scotia showed that rutting of asphalt overlays was occuu-ing after only a few years in service. It was apparent that this new type of rutting was not described using the conventional theory. Gervais and Abd El Halim (1990) used the concept of relative rigidity and field observations to explain this phenomenon. The premature rutting for these cases was a result of the low modular ratio between the overlay and the underlying mature asphalt for PCC) layer. Even after compaction, the new overlay remains relatively soft, producing a low relative rigidity between the asphalt layers.
Under this condition, they showed that the asphalt overlay was in a state of compressive stress. The compressive stresses measured were small and not considered to be the rutting failure criterion. However, the resulting strain condition revealed that high tensile strains were produced within the asphalt overlay causing lateral flow of the soft overlay material. The compressive stress condition acted to confine the deformations such that only the overlay deformed, much like a sandwich.
2.10.2 Effect of Pavement Layer Thickness Pavement layer thickness also plays a major role in determining the stress and strain distributions throughout the pavement structure. In General, increasing asphalt layer thickness causes the same effect as increasing the modular ratio of the pavement layers, that being a reduction in the vertical compressive stress applied to the underlying base, sub base and subgrade layers. This in turn reduces rutting.
Kandhal et. al. ( 1993) concluded that the underlying layer conditions (modulus and thickness) contributed to the surface rut depth in the majority of cases. This is not surprising since that investigation examined asphalt overlays on top of PCC pavements. Although not referencing the relative rigidity concept in that report, the findings of Kandhal et. al. ( 1993) appear to confirm the explanation of premature rutting presented by Gervais and Abd El Halim (1990).
2.10.3 Effect of Surface (Wearing) Course vs. Base Coarse Anani et. al. (1990) completed separate analyses for the surface (wearing) course and the base asphalt course. Regression analysis indicated that different variables were significant for the individual asphalt layers. This observation is not surprising since applied stresses and strains are. significantly different depending on location in the pavement structure.
The NCAT study also completed separate analyses for surface and base asphalt layers (Cross and Brown 199?, 1991). As with .Anani et. al. (1990), different variables were significant for the respective asphalt layers, however, the >7 strongest rutting relationships were observed in the surface layer. This is not surprising since almost all of the rutting was observed in the surface layer.
2.10.4 Effect of Pavement Alignment Abdel Nabi (1995) demonstrated that route alignment significantly influences rutting performance. Pavement sections on hills and curves often display increased rutting due to two (additive) mechanisms. First, traffic speed is reduced for these sections, thereby reducing the asphalt layer modulus and increasing rutting.
Second, the asphalt layers for these sections are subjected to sustained loading through the force of gravity. This gravity-induced creep was shown to dramatically reduce the shear strength of the asphalt, again increasing rutting under traffic loading.
2.11 Category F - Construction-Related Factors 2.11.1 Effect of Compaction and other Construction Practices Field compaction must achieve specified in-situ density and air voids to ensure adequate pavement performance. In addition to density/void specifications, field compaction must strive for consistency throughout the pavement construction.
Inconsistent compaction causes localized areas containing too little or too many air voids, allowing shear flow or traffic induced densification, respectively.
Pavement surface permeability is an area of construction that is commonly overlooked. A waterti,ht surface prevents the infiltration of moisture that can cause surface stripping of asphalt. Furthermore, infiltration of moisture into the base and subgrade layers can cause hydraulic scour under traffic loading. However. in ~s addition to a watertight surface, adequate drainage must be provided for the granular layers to prevent stripping from below. Good construction and compaction can provide both adequate drainage and a watertight surface if completed properly.
Interestingly, the affect of aggregate particle contact has not been investigated until very recently. Particle contact is essential to transmit traffic loading through the asphalt layer and into the base and subgrade layers. A study underway at the Turner-Fairbank Highway Research Center (FHWA 1998b) involves analysing particle contact in asphalt cores using computerized tomography, as known as ''CT"
scanning. Initial results indicate that cun-ent compaction practices produce aggregate skeletons for which only 15% of aggregates carry over 50%n of the applied load and 50% of the particles carry over S0% of the load. These striking results clearly indicate room for improvement concerning the in-service contact of a~are~ate particles under compactive effort.
J ~m The round-robin investigation concerning laboratory compac>ion variability completed by Lai ( 199x) also has application to field compaction. Clearly different paving contractors use different techniques to compact asphalt concrete, leading to significant variability between various sites and even along individual sites.
Field compaction has long been touted as the most significant variable toward pavement performance (including rutting), however, little resources have been allocated toward improvement in the cun-ent compaction techniques or equipment.
Carpenter (1993) reported that the nix parameters produced during the initial construction of the pavement will influence how much permanent strain occurs when the limiting voids develop. In more simple terms, initial mix properties produced during field compaction determine how much rutting occurs prior to the onset of plastic flow.
Of course, significant work has been completed by Abd El Halim in the field of asphalt compaction. Numerous field tests with the AMIR compactor have conclusively shown that improved compaction techniques reduce variability of density throughout the pavement structure, reduce permeability and improve pavement fatigue life by up to 50 percent (Abd El Halim et. al. 1990.
2.11.2 Effect of Quality ControUQuality Assurance (QC/QA) Consistency during construction is critical to pavement performance. Poor QC allows areas with varying parameters and promotes localized deformation.
Poor QA allows global permanent deformation if critical rutting variables are poor.
The NCAT study recommended that the most important QC/QA test that can be conducted during construction is to compact plant produced material in the laboratory and evaluate the air voids of the specimens (Cross and Brown 199?, 1991). This recommendation makes a strong case to provide on-site evaluation for QC/QA purposes.
Kandhal et. al. ( 199x) indicated that asphalt content measured from field cores was ~enerallv deficient from the value specified for the job-mix formula (laboratory mix design). Clearly, improved control over asphalt content is required for plant production since asphalt content is one of the primary factors governing rutting performance of pavements.

2.12 Category G - Environmental Factors 2.12.1 Effect of Temperature Asphalt cement (and therefore asphalt concrete) is highly temperature sensitive. Anani et. al. (1990) noted that, because asphalt concrete is black, solar energy is readily absorbed and then retained due to its low thermal conductivity. As the temperature of the pavement increases, the stiffness of the asphalt layers) decreases. Reduced asphalt cement stiffness allows aggregate particle movement and reorientation under traffic loading causing permanent deformation.
Therefore. a strong aggregate skeleton is required to resist rutting at elevated temperatures.
An excellent illustration of the effects that temperature can impose on asphalt binder stiffness was published by Rickards (1998). Using the Shell Bands program, asphalt binder stiffness was plotted versus temperature in response to the compactive effort of three different compaction devices as shown in Figure 10.
10000 _ _ . _.... _ _ _ -_ _ -_ fAMiR
1000 - _ , _ _ _ --Roller ~'"~""'--- ..... -~ Vibration _._ 100 ~....~.."~~- _ _ _ _ - .._.
c 4 '-'----- _ ~ _ _ __.
1 ,I _ ._ ._ >
0.1 - _ ___ _ ~ j 0.01 Mix Temperature (deg. C) Figure 10: Bitumen Stiffness vs. Mix Temperature for Three Compaction Devices (from Rickards 1998) Figure 10 clearly displays the reduction in binder stiffness with increasing temperature for each of the compaction devices. As an example, an increase in temperature of 40°C ( 110°C to 150°C) caused a 10-fold reduction in stiffness.
While such elevated temperatures are not usually encountered during normal operating conditions, extrapolation of Figure 10 suggests similar relative changes in stiffness could be experienced for more typical operating temperatures. This statement is supported by Hofstra and Kolomp (1972) who observed the significant effect that normal operatinD temperatures can have on asphalt mixes. During their investigation, a change in temperature from 20°C to 60°C reduced the modulus (stiffness) of the asphalt concrete by a factor of 60, while rutting increased by a factor ranging from ?~0 to 350 times. Clearly, temperature has a significant effect on asphalt pavement rutting.
A final example of the significant effect that temperature (and direct sunlight) have over rutting performance was noted by Anani et. al. (1990) in Saudi Arabia where rutting was significantlwreduced, or even non-existent, under brides where the pavement is shaded by the bridge deck.
2.12.2 Effect of Ageing Over time and environmental conditioning, asphalt cement loses some of its flexibility (i.e. its stiffness increases). Therefore, the oxidation of asphalt cement actually increases the pavements resistance to permanent deformation so lon';
as the bond between the asphalt cement and aggregate particles is maintained.
Kandhal et. al. (1993) explained that during early stages of a (newly constructed] pavement life cycle, rutting is directly proportional to traffic.

However, after this initial densification, the rate of rutting decreases with increasing traffic until finally a condition of plastic flow occurs and the rate of rutting again increases.
The importance of separating the rutting cycle into distinct periods was further held by Carpenter (1993). According to Carpenter, two vital criteria to judge the long term performance of a mixture are how quickly a critical rut depth is reached in the mixture, and the "rapidity" with which the mixture reaches the failure point at the onset of plastic flow. These criteria are not mutually inclusive as a mixture may reach critical rutting before the mixture becomes unstable, or it may become unstable before it reaches the critical rut depth.
2.12.3 Effect of Moisture Damage (Stripping) Stripping involves the removal of asphalt cement from the mineral aggregates through moisture infiltration. Stripping can occur on the surface of the pavement causing loss of surface aggregate (ravelling) or can occur from below the asphalt layers due to poor drainage conditions or sealing of the asphalt surface. The loss of bond between asphalt and aggregate allows the aggregate particles to move or shrift under traffic loading, promoting permanent deformation.
As previously mentioned, Krutz and Stroup-Gardiner (1990) investigated the influence of moisture damage on rutting of chip-sealed pavements for the Nevada DOT. They found that sealing of the surface with the chip-seal accelerated stripping by trapping moisture under the asphalt layers. The loss of asphalt cement allowed the aggregates to shift, causing severe rutting of the pavements analysed.
Drainage conditions of the pavements were likely insufficient in those cases.

2.13 Category H - Traffic (Load) Related Factors 2.13.1 Effect of Tire Contact Pressure (Load Magnitude) The size of the tire contact area depends on the contact pressure between the tire and the pavement surface. Huang (1993) indicates that pavement contact pressure is greater than the tire pressure for low-pressure tires, because the walls of the tire are in compression and the sum of the vertical forces due to wall and tire pressure must be equal to the force due to contact pressure. Contact pressure is smaller than the tire pressure for high-pressure tires, because the tire walls are in tension. For simplicity, most pavement designs assume that the contact pressure is equal to the tire pressure, which is consistent with the findings of Gerrard and Harnson (1970).
Kandhal et. al. (1993) have reported that tire pressures have increased substantially in recent years. Tire pressures average 661 kPa (96 psi) and 689 kPa (100 psi) in Illinois and Texas surveys, respectively. Therefore, increased pressures are applied to the pavement, which will cause increased pavement damage.
Substantial finite element modelling of the effect of tire pressure was also completed during the SHRP research. Model runs were completed with tire pressures of 690 kPa (100 psi), 1380 kPa (200 psi) and 3450 kPa (500 psi), respectively. Results indicated that rut depth increased almost linearly with increased maximum permanent strain, which was directly related to increased tire pressure (US-SHRP 1994).

2.13.2 Effect of Tire Material The concept of relative rigidity is also applicable to the modular ratio between the tire and the asphalt surface. Tires composed of different rubber/steel combinations produce different modular ratios between the tire and the asphalt surface, which affects the contact stresses. Gervais and Abd El Halim (1990) proposed that the switch from bias-ply to radial tires by the automotive industry represented a fundamental increase in rutting damage to asphalt pavements.
2.13.3 Effect of Number of Load Applications (ESAL's) Rutting usually occurs over an extended period of time with numerous load applications according to the rutting cycle outlined by Kandhal et. al. (1993) and Carpenter (1993). With each applied load, a small amount of permanent deformation is introduced within the asphalt layer. The magnitude of this deformation is dependent on the stage of rutting.
Therefore, the number of applied ESAL's is directly proportional to rutting in asphalt pavements, at least during the traffic densification stage. Many rutting models incorporate ESAL counts, whereas some investigations such as Anani et.
al.
(1990) do not incorporate traffic effects directly, considering traffic to be an uncontrollable variable.
Other researchers have convened the total amount of rutting into a rutting rate by normalizing with traffic such as Cross and Brown (199?, 1991) during the NCAT National Rutting Study as well as Kandhal et. al. (1993). Other studies by Brown and Cross (1988) and Parker and Brown (1990) indicated that expressing the rate of rutting as a function of the square root of total traffic better models pavement behaviour than other expressions such as the arithmetic sum or log of total traffic.

2.13.4 Effect of Rate of Loading Being a viscoelastic substance, the stiffness of an asphalt concrete pavement is dependent on load duration as well as temperature. Loads applied slowly cause a reduction in layer stiffness thereby increasing rutting by allowing asphalt to flow (similar effect to increasing temperature). Again, a strong aggregate skeleton is required to minimise rutting under these conditions.
Generally, the greater the speed, the larger the (asphalt concrete) layer modulus, and the smaller the strains in the pavement (Carpenter and Enockson 1987). Therefore, higher travelling speeds actually cause less rutting than lower travelling speeds (all else equal). The effect of load rate is apparent at areas of reduced speed such as intersections, hills and curves that exhibit increased rutting.
The effect of load application rate is also apparent in Figure 10 (Rickards 1998). The three separate lines in Figure 10 simulate three asphalt compactors applying different load rates to the asphalt concrete. As shown, the higher loading rate of the vibratory compactor invokes a greater stiffness response of the asphalt binder than the static steel roller or AMIR roller, respectively.
2.14 Category X - Combinations of the Other Categories Variables listed in the above eight categories have been reviewed independently.
However, many of these variables are strongly colinear and therefore work together (or against each other) to provide resistance to rutting. The interaction of aggregate angularity and in-situ air voids towards rutting was observed during the NCAT
National Rutting Study. For in-situ air voids above 2.5%, the angularity of the coarse aggregate (two or more crushed faces) and NAA uncompacted voids for the fine aggregate (aka fine aggregate angularity) are highly correlated with the rate of rutting.
If the in-situ voids were less than 2.5%, rutting is likely to occur regardless of aggregate properties (Cross and Brown 1992, 1991).
The interaction between asphalt cement and gradation towards rutting was studied by Krutz and Sebaaly (1993). They concluded that rutting performance of finer gradations is influenced by binder characteristics more so than more coarse gradations.
Conversely, the performance of coarse gradations is more dependent on aggregate properties and less sensitive to binder type.
The effect of asphalt-aggregate irneraction was also completed during the SHRP
research (US-SHRP 1994). Regression analysis of rutting induced by wheel-tracking devices displayed that asphalt-aggregate interaction accounted for upwards of 15% of the observed rutting. Finally, the interaction of volumetric properties such as air voids, VMA and VFA was well examined by Coree (1999) as presented in Section 3.3.
2.1~ Summary of Rutting Variable Relationships Table 4 summarizes the qualitative relationships between the categorized variables and permanent deformation. An entry of "Increase" indicates that rutting resistance increases with an increase in that particular variable, while an entry of "Decrease" indicates that rutting resistance is reduced with an increase in that variable (i.e. rutting increases all else equal). The inclusion of a question mark "?"
indicates that the general trend is not well defined or questionable. The use of the word "max"
indicates an upper limit, above which rutting resistance is reduced.

Table 4: Summary Table of Rutting Variables and Qualitative Relationships Category A - Bituminous Materials and Additives Chemistry Penetration Viscosity Use of Modifiers AsphaltenesSaturates Polymer Antistrip I

Decrease? ncrease Decrease Decrease Increase Increase'?

Category B - Mineral Aggregates Source Consensus Properties Properties ToughnessSoundnessDeleteriousCoarse Fine Flat/ Clay MaterialsAngularity Angularity Elongated Content Increase?Increase?Decrease?Increase Increase Decrease? Decrease'?

Category C - Mix Design Parameters GradationAir VoidsCont nt VMA VFA Co Dust Content pabtion see sectionIncrease Increase Increase Increase Increase Increase 2.8.4 (max) (max) (max) (max) (max) (max) Category D - Engineering Properties of Mix Marshall Shear Resilient Testing Modulus/ Creep StabilityFlow Stren~th/Stiffness Indirect Tensile Increase?Increase?Dramatic Increase Decrease Increase Category E - Pavement Structural Design Asphalt Asphalt Layer Layer Stiffness Thickness and Deflection Increase Increase Category F - Construction-Related Field Quality Compaction Control and Assurance (QC/QA) Increase Increase (max) Category G - Environmental Temperature Aging Moisture Dama;e Decrease Increase Decrease Category H - Traffic-Related Contact Number Rate Pressure of ESAL's of (Load) Loading (Speed) Decrease Decrease Increase Category X - Combinations of Above Categories 2.16 State-of-the-Practice: Asphalt Rutting Testers Pavement rutting testers are currently attracting much attention from the asphalt industry. While some of these devices have been used for years, the widespread adoption of Superpave in the United States (more slowly in Canada), has re-ignited the search for a device that can both separate poor and good performing mixes, and also predict the long term field performance of pavements prior to construction. As mentioned, Superpave currently is based solely on volumetrics, binder and aggregate selection criteria.
At this time, there are numerous asphalt rutting tests in use by various agencies.
Some are empirical tests, not based on engineering properties or analysis.
Examples include the French Rut Tester, the Hamburg Wheel-Tracking Device, the Asphalt Pavement Analyzer (formerly Georgia Loaded Wheel Tester) and the Accelerated Load Facility (ALF). Other rutting tests, such as the Superpave Shear Tester (SST) measure engineering properties such as the shear strength or modulus (stiffness) of an asphalt mix. It is believed that these performance-based tests hold the most promise for modelling and predicting long term performance of pavements since engineering properties can be directly related to performance.
The objective of this section is to review some of the existing devices, which will lead into the following section that discusses their benefits and weaknesses with respect to performance prediction.
2.16.1 LCPC (French) Rut Tester The LCPC Rut Tester was developed at the Laboratoire Central des Ponts et Chaussees (LCPC) in France. As shown in Figure 1 l, the device uses two reciprocating pneumatic tires with diameter of 415 mm and width of 110 mm to assess the rutting resistance of mixes. Test slabs are 500 mm long, 180 mm wide and either 50 mm or 100 mm in thickness. A standard tire pressure of 0.60 ~0.03 MPa is applied approximately 67 cycles per minute (about 1.1 Hz). One cycle consists of a forward and backward pass of the loaded wheel; therefore, 134 individual passes are completed per minute (Romero and Stuart 1998). Test air temperature of 60°C is maintained without regard to the environment where the pavement is located or the depth at which the mixture is located within the pavement structure (Huber 1999). A LCPC Rut Tester costs about $125,000 CAD.
Interestingly, the developers of the French Rutting Test do not believe that statistical correlation between rutting observed in the test and that observed in the field can be developed since the rut tester simulates extremely severe rutting conditions (Huber 1999). However, LCPC reports that roads meeting the LCPC rut tester specification do not exhibit rutting in service.
Figure 11: LCPC Rutting Tester Extensive work has been completed in Colorado using the LCPC device to correlate laboratory and field rutting performance by Aschenbrener (1994). The study investigated 33 pavement sections with satisfactory or poor performance in rutting resistance. Test slabs taken from the sites were tested with the LCPC
Rut Tester and indicated that the French specifications were too severe for Colorado conditions. To reduce the severity, the test temperature was modified based on the actual field temperatures associated with Colorado conditions. Test data was also separated into high, medium and low categories. Regression analyses yielded high correlation (R'' of 0.87 for high traffic and 0.68 for medium traffic) between field rutting and the slope of the rutting curve observed with the French device.
LCPC Tests were also completed on specimens recovered from the Westrack experiment. Good correlation (R'=69.4%) between laboratory values and rutting observed at the test track was achieved through regression analysis (FHWA
1998c).
2.16.2 Hamburg Wheel Track Tester and Couch Wheel Track Tester Esso AG developed the Hamburg Wheel Tracking Device in Hamburg, Germany in the 1970's (Romero and Stuart 1998). A solid steel wheel with a diameter of 204 mm and width of 47 mm rolls across an asphalt concrete slab immersed in water kept at 40°C or 50°C as shown in Figure 12.
Immersion of the test specimens in water allows for the simultaneous testing of rutting and moisture damage (stripping) resistance of various mixes. Test slabs are 320 mm long, mm wide and may be 40, 80 or 120 mm thick. A fixed load of 0.69 kN is applied to the wheels producing an average contact stress of 0.73 MPa (although actual contact pressure varies due to variable contact area during the test). This contact stress approximates the stress produced by one rear tire of a double axle truck.

Approximately 53 passes per minute (26 cycles per minute) are applied and the original test was performed to 9500 wheel passes. However, it was later discovered that some mixes could deteriorate due to moisture damage shortly after 10,000 passes. The number of test passes was subsequently raised to 19,200 to observe moisture damage. A Hamburg Wheel Tracking Device costs about $90,000 CAD.
',m.2~rS?~rePG n .~;cpe. ~T2BG!rcm ~, s xnGaio~ I Z, .
HhHV".iGn I
I i .~
1 //99~~
I ~~//'~~__71 Sr~P:GrtY~ Point 1 I// S
!Is'iJO~'9CS I rr.3~.l.~ppny ~1 I ~ SI~E<.=::2ø'~.tm ~~~ .
t I
_20 l t t 1 f f ,I, .-.t.--0 5000 10DOD 15000 ?DD00 Number of Wheal Pasxes Figure 12: Hamburg Wheel Tracking Tester Performance correlation between field performance and results from the Hamburg Wheel Tracking Device was also completed at the Colorado DOT by Aschenbrener (1995). Although the Hamburg stripping slope and stripping inflection point were able to distinguish between good and poor field stripping performances, the moisture conditioning system used by the device appeared too severe for rutting determination. However, regression analysis between the Hamburg device and rutting at Westrack yielded good correlation (RZ=75.6°l0) (FHWA 1998c).
The Couch Wheel Track Tester is a variation of the Hamburg test. A single solid rubber wheel with an approximate contact pressure of 950 kPa (140 psi) is used to rut an asphalt slab. As with the Hamburg test, specimen temperature is controlled through submerging the specimen in a heated water bath. The number of wheel passes is counted with a digital counter while the rutting profile is measured with a linear variable differential transducer (LVDT). The LVDT measures the rut depth at the centre of the specimen and sends the signal to a linear graphing printer which provides continuous output during the test. An automatic cut-off switch terminates the test if a specimen prematurely fails or reaches the complete test cycle of 20000 passes of the rutting wheel. From the graph of rut depth versus number of cycles, the average rutting rate may be determined from the slope of the tangent from the consolidation point (typically measured 10 minutes after the start of the test) to the stripping inflection point. The stripping inflection point (if present) indicates a change in the rate of rutting with time due to loss of bond between the asphalt binder and the mineral aggregates. If no stripping inflection point occurs, the average rutting rate is simply the slope of the tangent from the consolidation point to the 20000 cycle mark. The graph provides much additional information including the stripping inflection point, as well as the rut rate prior to, and after the inflection point (Aschenbrener 1994).
2.16.3 Georgia Loaded Wheel Tester and Asphalt Pavement Analyzer As the name implies, the Georgia Loaded Wheel Test (GLWT) was originally developed at the Georgia Institute of Technology in the mid 1980's for the Georgia Department of Transportation to test rutting resistance of asphalt mixes (Lai 1986).
Unlike the French or Hamburg devices, the GLWT assesses rutting resistance by rolling a concave steel wheel across a pressurized rubber hose placed along a test beam. The 29 mm diameter hose is pressurized to 0.69 MPa. The device operates at 67 passes per minute for 8000 cycles (16000 passes).
In 1995, the rights to commercially manufacture and market the GLWT were purchased by Pavement Technology Inc. (Prowell and Schreck 2000). Numerous improvements were introduced to the original design and the resulting device was renamed the Asphalt Pavement Analyzer (APA). Unlike the GLWT, the APA
includes a water storage tank for testing specimens under water, and is capable of testing both beam and gyratory specimens as shown in Figure 13.
t :~:'.
Figure 13: Asphalt Pavement Analyzer As with the French and Hamburg devices, good correlation between field rutting performance and the GLWT/APA has been observed. For example, regression analysis between the APA rut depth and field rutting observed at Westrack has yielded R2=79.7% (FHWA 1998c).
2.16.4 Accelerated Load Facility The Accelerated Load Facility (ALF) is a full scale wheel tracking device incorporating one half of a single truck axle travelling along a 29 metre frame over a full scale pavement test section approximately 10 metres in length. Loads between 44.5 to IOO.IkN may be applied. Unlike laboratory wheel tracking tests, the ALF
applies loads in one direction only and can impose lateral distribution of the load to better simulate truck traffic loading (wander). ALF can simulate 20 years of cumulative traffic in six months or less. The ALF is shown in Figure 14.
Figure 14: Accelerated Load Facility (ALF) 2.16.5 Superpave Shear Tester One of the major products developed during SHRP research in the US was the Superpave Shear Tester (SST). SHRP researchers identified that rutting appears to be more closely related to shear stress than normal or horizontal stresses (SHRP
1994). As previously mentioned, the SHRP research also referenced work by Celard (1977), who emphasised that, based on the results of dynamic creep tests, the rate of permanent deformation was strongly related to shear stress. During SHRP, it was anticipated that the SST would provide input to the Superpave performance-based models, although development of the performance models is not expected until 2005. The SST is illustrated in Figure 15.
Figure 15: Superpave Shear Tester Performance testing with the SST to date has produced acceptable correlation between shear properties and field rutting. At Westrack, correlation coefficients (Rz) of 0.55, 0.4 and 0.26 were observed for repeated shear at constant height, frequency sweep at constant height and simple shear at constant height, respectively (FHWA 1998c). While these values are significantly lower than those achieved with wheel tracking tests, it is important to note that the shear properties measured by the SST are not specifically meant for regression analysis, but are to be input into performance-based prediction models that have yet to be developed.
A second device known as the Field Shear Test (FST) was subsequently developed by Endura-Tec Systems as a field quality control device for Superpave (NCHRP 1998). The primary difference between the SST and FST are specimen orientation and the fact that the FST is a portable test and that the specimen is tested diametrally with the FST (similar to the indirect tensile test). Both the SST
and FST
are under investigation as simple performance tests under NCHRP 9-19.
2.17 Deficiencies with Current Testing/Modelling Practices 2.17.1 Discussion of Empirical Rut Testers Table 5 compares the characteristics of the LCPC, Hamburg, GLWT/APA and ALF devices. With the exception of the ALF, all wheel tracking devices incorporate a small rolling wheel across a prepared specimen or core of known dimension. It is known that these tests can effectively rank asphalt mixes in terms of relative rutting resistance and, as previously mentioned, they have even displayed good con-elation to observed field rutting. However, there are numerous characteristics of these tests that preclude them from accurately predicting rutting performance of field pavements. First, laboratory wheel tracking tests do not have proper boundary conditions. The test specimens are surrounded by steel molds and are resting on a steel base, which is never the case with the testing of real pavements (Romero and Stuart 1998). Furthermore, stress development in laboratory rut testers is never representative of real life conditions because the size and pressure at the test wheel are unlikely to be representative of real wheels.
Table 5: Characteristics of Rut Testers (compiled from Huber 1999; Romero and Stuart 1998; Prowell and Schreck 2000) Laboratory Full Scale Wheel Tracking Tests French Georgia-Type ALF

Hamburg LCPC GLWT APA

44500 to Wheel Load 5000 705 700 533 to 100100 (N) 700 Contact Pressure600 730 to 1500690 690 to Variable (kPa) Loading Rate (cycles per 60 to 26 to 60 33 to 45 6.3 minute) Load MechanismPneumaticSteel WheelSteel Wheel on Full Size Tire Pressurized Hose Pneumatic Load Wheel 400 (diameter)200 (diameter) 29 (hose Truck Tire diameter) Dimensions 110 (width)47 (width) (mm) Test Environment.Air Water Air Air or Air Water 500 (length)320 (length) Beam or Full Scale 300 (length) 150 Specimen 180 (width)260 (width) (diameter) 125 (width) Section Dimensions, 50 or 40, 80, Core/Gyro 9800 (length) (mm) 100 120 ~
75 (thick) (thick) (thick) specimen l Test Temperature60 40 or 50 40 49 to 60 Ambient (C) or 60 No. of Cycles30000 9500 to 8000 N/A
in 19200 S ecification Max. Allowable10 4 ( 10000 7 N/A N/A
cycles) Rut De th (mm) 000 $15,000 $130,000 Variable $125 CAD

Cost , $90,000 (minimum) CAD ($ millions) CAD CAD

Specimen size may also contribute to lack of correlation since the relative size of the wheel compared to material constituents (such as aggregates), is not consistent with in service pavements. Finally, for any test to be valid, the load applied to a specimen should always be in proportion to the specimen size (Romero and Stuart 1998). This is not the case with most of the devices with the exception of the ALF.
However, although the ALF addresses the problems of dimensional incompatibility due to its full-scale nature, the resulting properties (rut depth or rate) do not represent fundamental engineering properties that can be input into a mechanics of materials model for performance prediction. Furthermore, the ALF
is extremely expensive ($ millions) and not feasible for field QC/QA.
2.17.2 Discussion of Existing St~ear Tests Although the development of the SST and the FST represented an important step toward measuring asphalt shear properties, neither test is ideally suited for widespread implementation. The SST does provide a great deal of information with regard to mix shear properties, however, it is very expensive (approximately $250,000 USD), confined to the laboratory, and requires a great deal of training to use correctly.
While the FST is a portable device, the diametral loading condition is not representative of field loading conditions. Furthermore, Sousa et. al. (1991) have reported that diametral loading (from the indirect tension test) is inappropriate for permanent deformation characterization because the state of stress is non-uniform and strongly dependent on the specimen.

Finally, both tests require the preparation of cylindrical specimens either through gyratory compaction or coring of in-service pavements. As has been discussed at length throughout this thesis, these preparation methods are either non-representative of the mix in the field, or damage the specimen to a large degree.
The development of an in-situ test will both provide an excellent complimentary test device to existing laboratory tests, as well as better represent the performance in the field.

CHAPTER 3: REVIEW OF PREVIOUS WORK AND
ANALYTICAL MODELLING
3.1 Introduction and Chapter Overview Two previous research efforts formed the foundation for the current investigation. The first was a comprehensive laboratory investigation of asphalt shear properties and pavement rutting completed at Carleton University by Zahw (1995).
This chapter begins with a review of that investigation, followed by the results of a new study to investigate the relationships between asphalt mix characteristics, shear properties and pavement rutting, using data collected during his research.
The second underlying research effort involved the construction of a first generation in-situ shear strength test device, also at Carleton University by Abdel Naby (1995). A review of the device, known as the Carleton In-Situ Shear Strength Test (CiSSST), is provided including its main benefits and the results of his research concerning in-situ shear strength and its relation to pavement performance.
The chapter concludes by introducing an improved analytical approach to derive asphalt pavement shear properties from the surface plate loading condition developed using closed form equations and the finite element method.

3.2 Review of Previous Work - Laboratory Torsion Testing of Asphalt Concrete 3.2.1 Introduction A comprehensive laboratory study of asphalt pavement rutting and shear strength and stiffness was completed at Carleton University in 1995 (Zahw 1995).
The testing program involved the mixing, compaction and testing of over 1200 standard Marshall specimens representing a total of 58 different asphalt mixes. Mix shear strength and modulus were determined through laboratory torsion testing of cylindrical specimens to failure. The Tinius-Olsen torsion test machine is shown below in Figure 16.
Figure 16: Torsion Test Equipment at Carleton University Cylindrical Marshall specimens or cores were glued to steel plates using an epoxy and loaded horizontally into the device. All testing was completed at 25°C.
Torque and twist angle at failure were recorded by the device and the specimens failed in shear with a characteristic 45° failure suuface as highlighted in Figure 17.
Figure 17: Typical Failure of Asphalt Specimen in Torsion Test Device Permanent deformation characteristics were determined through the Shell Pavement Design method utilizing uniaxial unconfined static creep tests at three stress levels (0.1 MPa, 0.3 MPa and 0.6 MPa).
3.2.2 Deriving Shear Properties from Laboratory Torsion Tests By definition, fundamental engineering properties of materials such as tensile or compressive strength, shear strength and stiffness, elastic modulus, etc.
are unique to individual materials and not dependent on boundary conditions.
However, there are few testing procedures (if any) that directly measure fundamental properties. In most cases, a given load is applied to a test specimen and the desired fundamental property is then determined knowing the specimen dimensions. For example, compressive strength (f'~), which is a fundamental property of Portland Cement Concrete, is determined by applying an axial load to failure, and then using that failure load and the cross sectional area of the specimen to calculate f'~.. However, the measured values of the fundamental properties can be strongly dependent on the test conditions such as load rate, confining pressure, temperature etc. With PCC (and many other materials including asphalt concrete), the faster the applied load, the greater the resulting strength response.
Therefore, various standards (CSA, ASTM, etc.) have been developed so that a single set of test parameters is used to produce comparable results.
Under similar test conditions, alternative methods may be used to determine fundamental properties. For example, the Superpave Shear Test (SST) measures shear properties of asphalt concrete by applying a force across an asphalt core or gyratory specimen. Given the specimen dimensions (cross sectional area), the shear properties of the mix are easily calculated. The same shear properties of the core or gyratory specimen may be determined using a torsion test as well. Again, given the specimen dimensions, the shear properties may be determined from the applied torque. A comparison of the mechanics behind the SST and torsion test is shown in Figure 18.
While the method of force application is different, the differential elements (dA) within the specimens are subjected to shear in both cases, thereby allowing the calculation of the shear properties independent of the test method or boundary conditions.

sa F
z --~- z ~~ ..
F
~dA~
~ C&~ ~.5 ~=Gy z __ Figure 18: Determination of Shear Prof~erties from Different Test Methods As chown in Figure 1 s, the shear strength of the asphalt rni.x may be determined using Equation 1:
where:
(1) J
T = the shear stren~tlu (MPaj T = the maximum applied torque (!~'~m) c = the radita of the test specimen (mm) J = the polar moment of inertia (mm' J

bi 3.2.3 Major Findings of Laboratory Torsion Testing The work of Zahw (19>~) representec new and extensive research toward a better understandin;v of the rutting phenomenon and its underlying causes. One of the main findings was that conventional asphalt design criteria such as density alone do not provide a reliable indicator of high rutting resistance, whereas the use of shear properties better characterized the mix performance. This finding was supported by research completed during the Strategic Hiwhway Research Program (LAS-SHRP 1994). which vvas being completed concurrently by Zahw ~19~)~).
In addition to the development and verification of a shear testing f~ramem.~rk usin~T laboratory torsion testing. Zahw also ,venerated a large volume o1 dma including mix properties, measured shear properties and mix performance as determined through unconi'ined static creep tests. This database was utilized during, the current investigation to produce nevi and valuable m~~dels relating mi.x properties to shear properties, as well as shear properties to calculated rutting. The results of this anal:;sis are presented in the followin'v section.
3.3 Analysis of Laboratory Min, Shear and Rutting Database 3.3.1 Relation of >ftia Characteristics to Shear Properties Sixteen asphalt mix properties were available in the Zahw database for the analysis as listed r~clow in Table O. These properties were subseduently ~~rouped into three main ca~e«orie:; -- Asphalt Binder Properties, Mineral A'are';ate Properties and'~'Iiv Deli<Tn Properties.

Table 6: Mix Properties Available from Zahw ( 1995) Database i Asphalt Binder Mineral Aggregate ~ Mix Design j Properties Properties i Properties ~ C.'oefficient of Cniformity ~ Penetration L 25C j ~ No. of Blows with ~ ~ Coefficient of C.'urvature ~ Rind= and Ball Softening ' Marshall Hamrr,er ~ F'ercentaUe of Coarse I
point '~ ~ ~ Final Specimen Density ;
I M uteri al ' ~ hinematic Viscosity I ~ Asphalt Cement Content ~ Presence of Crushed ;
~ Penetration Index ~ I ~ Voids in the Mineral Stone Present in Mix ~~
~ Viscosity @ ?~C A~~re~aate ~ Percentage of Mineral I
~ Binder Stiffness, Sbit ~ Dust to Binder Ratio ' Filler .
Table 7 displays the measured engineering properties includinU shear stress, strain and modulus, as well as the estimated rutting from unconfined uniaxial static creep testing at 0.1, 0.3 and 0.6 MPG stress levels.
Table 7: Engineering Yroperties Available from Zahw (1995) Database Measured Shear Properties ', Estimated Rutting Yroperties*
~ Average Shear Stren~,th II ~ Rut Depth at 0. I MPG Stress i I
~ Avera'e Shear Strain ~ Rut Depth at 0.:; MPG Stress '.
I
~ Average Shear Modulus ~ Rut Depth at 0.~~ N7Pa Stress *Estimated rumn'_ prc~perues based on unec>nfined static ere~p tesunc (Shell Wcthc>d~
The first step in the analysis involved relating the asphalt mix properties to the measured laboratory shear properties. A correlation matrix was developed for all 16 variables usin~f the statistical features of Microsoft Excel. The seven mix variables yielding the Greatest cot~relation coefficients are displayed in Table b.

Table 8: Mix Properties Yielding Greatest Correlation to Shear Properties Correlation Coefficients Mix Property i Shear Strength , Shear Strain Shear Modules I (MPa) ~ ( ~o ) (MI'a>
~I Penetration @ ''~(_' -0.67 -0.80* -0.49 _ _ ~I V'iscosity C- ?SC ~ 0.44 ~~'~ 0.66** 0.1~
Coefficient of 0.~9 I negligible 0.70 I
Uniformity_ ~ _ I Presence of Crushed I, 0.41 0.?2** 0.4b I
h CCoarse Aggregate _; __ Mo. of Blows with II ~, _ j Marshall Hammer 0'07 -0.'~ 0._''_' i _ Density i _-0.~7 ~li<_'ibie 0.71 I _ Voids in Mineral (?.?1 -0.07 .=~~Qre~ate ' I " ' i _ ~
ceeYficient should have pc>sitive relationship '* ~«cffi~i~nt should have ne~auve rel,uionship examination of the correlation coefficients indicated that the effect of each mix charge-ieristic on the shear properties made raticmal sense, with the exceptio a of some of the relationships between the mix properties and shear strain.
Intuitivc:lv, a parameter that causes an increase in shear stren;~th or modules should have a ne~Tative effect cm shear strain. However, in the cases of penetration, viscosity un~i the presence of crushed coarse u;»~reUate, tl~e correlation coefficients for shear strain did not display the ci>n-ect sense (positive or neeativej. relative to the shear strength and modules, however, much lower con-elation was displayed between shear strain and many ol~ the mix properties. Therefore, as the shear strain is inherently contained within the shear modules, further <rnalysis of shear strain w.rs not completed in this investigation. A brief discussion of the resulting relationships is now presented.

~8 Until the Superpave performance-graded binder specification was developed, penetration was the pr-imarv criterion for thc~ selection of asphalt hinders for road construction in North America. Therefore, although it is an empirical measure of binder stiffness, it was not surprising to see that penetration was well correlated with rnix shear properties. Less correlation was observed with viscosity than penetratron.
hhe coefficient of unifor-mitv is a measure taken from soil mechanics to describe the shape of the gradation curve of the mineral a;;'re~aates used in eaclu mix. As indicated in previous sections, the aggregate skeleton is critical for rutting resistance, therefore. the shape of the gradation curve and its associated packins, configuration would likewise be highly cowelated with shear properties. The high cowelation with shear strength and modulus clearly indic<rted the importance of aggregate skeleton to transfer the load to the underlying layers of the pavement.
The variable refer7~ed to as "Presence of Crushed Coarse Ag~lre'~ate~~ was simply a binary choice of whether or not the coarse aggregates in the mix v, ere crushed (i.e quarried stone? or not (i.e. river '_=ravel). If floe a~~re'_ate was crushed, a value of "1'~ was assi'~nc:d. whereas a value of "0" was assigned if the aggregates were not crushed. ~~Ithou'~h it may have been desirable W have: a more descriptive measure of a~Tare~~te an~Tularim such as fractured face count, this information was not recorded durins; tire initial investigation by Z.ahw. Interestingly, the binary choice variable proved to he well correlated with shear strength and stiffness and was therefore kept in the analysis.
The number of blows with the Marshall hammer is a measure of compactive effort. Marshall mixes designed for high truck traffic applications are usually f; 9 designed as 7~-blow mixes, indicating that 75 blows with the Marshall hammer are applied per side of the specimen during mix design. Mixes designed for regular traffic levels usually- are designed with ~0 Mows per side. Therefore, the greater number of blows required to achieve the design mix properties. the stron'er the mix.
Final specimen density has long been the primary measure of mix adequacy, therefore, it was not surprising to see a hi'~In correlation between density and shear properties. Finally. voids in the mineral aQ~regate (VM.A) is a measure of the void space within the compacted mix. Volumetric properties are the foundation of mix design, therefore good correlation was expected. It should be mentioned that asphalt cement content was not selected as en independent variable, despite its well-known effect on rutting resistance. rfhis decision was largely made based on the greater con-elation observed between VMA and the shear and rutting properties when compared to trsphalt content, as well as the fact that asphalt content information is contained vyithin the VMA I~~urameter.
Based on the correlation matr7x, the dependent vanahles listed above in Table ~ should have provided the best input for regression models to explain the measured shear properties. However. some of the properties were ioi~hly conr-elated to one another, refen-ed to as collinear dependent war7ables. Such variables could not be used for regression in their cur-r-ent form. Vv'hile one option would be to simply remove the colline;~rr variables, the information associated with those variables would then be lost in the model. ;W other technique involves the use of combination variables. For example, Penetration and Viscosity @ ?>(_' were highly ccsrrelated t-0.66j and therefore could not both be incorporated into r::'~ression analysis.
However, instead of discarding the variable displaying tire lower ccmelation to the shear properties (in this case Viscosity), the two var7ables were combined into a sin;le variable referred to as Penetration-Viscosity Ratio (PVR) as shogun below, in Equation 2:
~~',R _ Peltration C '>C ~~) UL1' COS Itl' C ~~C~
A second conobination variable was also created from the original seven -Average Rate of Densification CARD), whi~:h was defined as the ratio of final specimen density to the compactive effort as expressed by the square root of the number of blows applied by the Marshall hammer as shown in Equation s:
LJ('lI S'Itv AIUL) _ - I ;1 J# ~~Ort C
It should be stressed that combining cowelated dependent variables should nc>t be completed haphazardly; the combined variable must make rational sense. In the case of PVR, the resulting correlation coefficient for shear strength was -ti.67, indic~7tin~ that ors PV'R increases, shear strength decreases. Examinin~~ the ratio itself, f'VR will increase with either an inc:rcase in penetration or a decrease in viscosity, or both. ('hereforf:, uccordin~ to the ratio, shear strength will decrease with an increase in penetration or a decrease: in viscosity ~i.c. a softer asphalt is used). This relationship makes sense since softer asphalts have a higher tendency to rut - all else equal.
Bv dividing the final density by the square root of the number of Marshall hammer blows, the ARD variable represented an average slope of the densific<ttion curve. In other words. .~RD provides an indication of th;~ amount of compactive effort needed to ~tc~ieve the final density. .-lsphalt mixes that have stron'=
a'~gre'~ate skeletons typically require more compactive effort to obtain specified density since it is noore difficult to rearrange the aggregates during compaction.
Therefore, it was expected that lower values of ARD would result in more rut resistant mixes.
T'he square root of the number of Marshall hammer blows was selected as it has been shown that field rutting is better described by the square root of traffic when compared to arithmetic or logarithmic functions (,Brown and Cross 1988. and Parker and Brown 1990).
A second con-elation matrix was developed for the resulting > selected mix variables as shown in .Appendix A. Note that this matrix, displayed both acceptably loos cowelation between the individual dependent variables. and very hi~Th cowelation between the dependent and independent variables.
Equations =I~ and ~ bclcwv were developed from multiple re'Tression analysis of the five selected mix variables to model measured shear stren«th and n~odulus.
The actual data used in the regression analyses is reported in Appendices B (mix variables] and C' (shear properties). .As shown by the coefficients of determination ( R-). the dependent variables within Equations ~ and ~ explain a higi~ degree of the variability observed in the shear modulus and stren~~th. respectively.
(i =>6U-~.7*PF'R+ '~~=rCC-' -- ~-14"C;,:,: -l~l ~s"Ah'D+1OC80*1-'MA
(R' = 0.83) i;5) =-66-U.9*YVR-s0~' CC% +(i~~'C,."" _ ''=IIC)~'~ARD-~-'~~W'-VM.=~
(R' .- 0.88) 9' where:
G (kPa) = Shear Modulus (Stiffness) at '_'~''C;
i (kPa) = Shear Strength at 25°C;
PVR (mm/Pa*s) = Ratio of Penetration (mm) to Viscosity at ?~''C (Pa*sj;
CU= Coefficient of Uniformity (D60/D10);
Ch;" = Presence of Crushed Coarse Aggregate in Mix (Binary choice of 1 for 'r es or 0 for :xloj;
ARD = A~,era~~e Rate of Densification (ratio of final mix density to the square root of the- number of blows with Marshal I hammer); and VMA = Voids in the Mineral Aggregate (~r~ ) The use of the combination variables in Equations 4 and ~ maximized the amount of inforn~ation contained per variable without introducing_, collinear dependent variables. Furthern~ore, the dependent variables utilized cover all of the major areas govennin<~ mi.x performance; bitumen properties (PVR), gradation (CU), an'~ularitv/rou;~hness of the aggregates (Chin), density and compactive effort (ARDI
and volumetric properties (~'MA).
Tables 9 and 10 display the regression statistics for the shear modulus and shear strength equations respectively. In both equations. the intercept term was statistically insignificant as indicated by the low t statistic. All other variables yielded high t statistics, indicatin~~ hi'Th siy:nificance.

y;
Table 9: Regression Statistics for Shear Modulus (Equation 4) ~ Coefficients Standard Errort Stat 1 ! I
--~ -Intercept ~'I X60 I 7~> 0.~

~I Penetration-Viscosity Ratio ? 0.?8 -9.7 (mm'/M C ?~C> .
_ i I
Coefficient of Uniformity~ ~ ~~ ='.
~., 0 ' 10 _ . .

'r 1D60/D 101 I ~ ' I

i - . _ _ _ --1 ~

Ii -~~~ 119 I Presence of Crushed Coarse Aa~reaate _ ;i I

'~. n ',, D.,ro of Tl~"c;fi~t,nn .
~

s I! [Fmal~Density/Sqrt(#hlcwv )] I -1~1 ~-,, _-_.__-_ ~ - ',, - - --Voids in I~'Iineral .A~'.~re~ute Ii 1065(1 ~' '_9sS !~ ;.O
Table 10: Regression Statistics for Shear Strength (Equation ~) Coeff icients t Stat !1 Standard Error I

I

~ -Gt; I(W.i -0.4 ~llntcrcept -_ - ' II
~ Penetration-Viscosity -C).9=I ~ -I;s.l ;~;atio 0.()6?

' (mn1'/~ ~n~ 7wC) -- ----__ - _ _ Coefficient of Uniformity~().-; ~ 10.(1 ~.0 ',1DC0/D10) ~ Presence of Crushed C:~arse 6=I6 ?b.? j ?.~
!, j , ~'~~~tc''ute Avera'le Rate of Densrinc<ition -?410 ~ (i(~~ _ .(~

', [Final Density/Sdrt(#b' ows)] j _ -___ ---- -_____ - __ ~
_ - r--- _ __ ' oids in Mineral Ayyre'~ttte~ -' The individual coefficient of detemination (R?) for each individual variahle w-as also investi'~atcd and the results are displayed in Tahle 11. As shown.
the Penetration-Viscosity Ratio (P~'R~ alone <:ould explain approximately '_'_".%~
of the ~4 variation in shear modulus and ~4~Io of the ~, ariation in shear strength.
Coefficient of Uniformity accounted for 49'~n of the shear modulus and =~~'7 of the shear strength. Interestingly, the binary choice variable - "Presence of Crushed Coarse r'~~Qreaate in Mix" -- represented a lame portion of the shear properties;
'_'3°i~ of the shear modulus and I7~i~ of the shear strength. The remaining variables, ARD
and \'MA explained less of the variation in the shear properties; however. then were hi'hlv si'nificant in hoth equations.
The results of the individual regression analyses tended to confirm the results found in the SHRP research; that binder properties conmibute ~~pproximatelv , ~~~a toward rutting resistance, while the remaining contribution cones from the a'7<yreyrates and the asphalt-aa~~reaate intera~=lion [US-SHRP 19~)~1].
Table 11: Contribution of Individual Variables Toward Shear Properties Individual Individual Coefficient of Coefficient c~f Variable I Determination for Determination for Shear l~'lodulus ~ Shear Strength (R~ ~%r ) ~ (IZ-~~c P\'R '~ ~ 1.9 ' 4~.?
~

~ -IS.t~ _- _~4.7 Coefficient of L-nifu>rmitv Crushed Coarse '_'?.S 1(.8 I' A RD ~; ~' . 7 9 . S
-~ ____ I___ ~.(~ ., -_ ~ VLVI~~
b.,~ ~ -3.3.2 Relation of Shear Properties to 1W tong Additional analyses were completed to investigate the relationship between the she~ir properties of the mix and its rutting resistance as determined throu<eh the 9s Shell procedure. Figure 19 displays the graphs of rut depth versus shear modulus of the mix for each unconfined creep test stress level. As shown, a high degree of con-elation was observed between shear modulus and rut depth with R~ values ranging from 0.60 to 0.80 using a power relationship. Similar relationships were seen for shear strength as illustrated in Figure ?0, although less cowelation was observed than with shear modules. The results of the analysis ;:learly indicated that shear strength and rnodulus are able to explain a large an;ount of the variability observed in the rutting as measured through laboratory creep tests, although shear modules appeared to be a better indicator. The rutting me>dels displayed in Fi~~ure 19 and ?0 are also listed in Table 1? below For reference, the actual shear and rutting_= data frorr~ the Zahw- ( 1990 database are attached as ,appendix C.
Table 12: Rutting Models for Shear Strength and Modules (_'reep 'host laboratory Rutting 1\-lodels ~
S tress ~--- -_ _- -.,, bevel Shear Modules @ 2~C Shear Strength C~? 2~C:
(MPa) (kl'a) ' (kPa) _ ().l Rut = 2~~*C,-o w~ iR~=0.80) Rut = I 8.?J'v-° ~>> fR=-:().(>~, Ij Rut _ -~;>Ol e'Gu o;, 1R=y0.6(i) ~I Rut = (~?.~)? '- T -~ c>,c~ 1R==t).-l70 0.
' i ___ _ __---,- ~___ ___ -l.S?0 , ~ ~_ -1.3~_ , 0.6 Rut = 3E6'~G (R'=0.7_a> Rut = 1 17W ' T (R =0.671 ', Figures 19 and ~0 contain much important infonm~tion that necessitates further discussion. First, the graphs cleurl~r~ show that as shear strength and modules increase, the amount of ruttinU experienced dining the test decreases si~nif~icantl~, particularly at him/her test stress levels. Fo.~ example, an increase in shear mod~alus from '?>00 MPa to 3000 MPa (?0'ro increasel reduced the amount of ruttin,T
from '?
mm to 1.4 mm (a s0°i'o reduction) for the 0.6 MPa test stress level.
For the lower stress levels, the graphs become flatter, indicating a reduced benefit for increased shear modules and strength for lower stress scenarios. .Although beyond the scope of this investigation, the information can potentially be used to optimize mix design selection (based on modules or stren'lthj for a desired level of p~rfounan~e (rutt.in'~
limit) and a 'liven traffic loadin' scenario (stress level). 1=urthermore, the results of quality control and assurance testing (QC/Q.A) could be checked against the ;raphs to ensure that the finished pavement will perform as specified, with possible penalty or hones implications to the contractor.

-,, O
O
O

n r O

c~ , ~ O

,~ r ~ ~ u~
~

x r~ r r t~

O o r f n ~ r ~ O
r ~ ~1 O

'~ ~ i 1' 'r v r / I~ !! _ ~

r ~ m G
r a /~ S O r r ~ ~i r r ~ ~ cc s~ i1 N ~. r' o r i ~

O

O

~~ ~ N O

a ~ ~ a~ ~
~

~

H
O .~ -r ~

~ pa ~ ! l~ i.

m ~ t~ 1 ~ r ~P

c~
P

O
f ' K tJ

N ~ O

O
'" 91 O
O ' ~ N_ nr M

t..CrJVr O O O

tb m m o ~

a_ n_ a O r ' ~ ~' O
' : ~

,Q. - ~ Q a L~
11 - ~
.-- r~
c'~

C~ ~ O ~i O n_n_ n_ ' ~. #w I I i _a e~ ~n N ~.n =- ~.~ o N ~- O
lwW~ uxda~ ~~b O

a 1 a i1 r o o~

rn N

d ,~

!

(~ ~ O
4 i1 O

i I I

a .

r' ,~ ,~ ti , :.c.

11 ~ a . y n . o :i CO

c~

J' ! ~ a ~ ~ cn a ~ ~

~ . .

~ ~ ..

v ~ ~1 G7 a t !

a 1 ~ . 4a r L

! Q U3 Q L

~ . ~ N

~ .
r, ~ .

M Q

cp Q

~' ~ !
'x m . .

a s ~ . ~ '1 II ~~ ! k ~

dt ~'~ ~ o .

r~ . Q

I _ _ _ N
cT7 cn~0 II .

C ~ .~ ~

r C~JC'~

M
cb ~C N ~

x ~

O

2 ~ Q~Qu d NII C d r--C'iC~ G

O O O ~._ CL<2 CD ~

~,, r-1 ~1 ,.... . .......... ........ .............._ ....._,__._..._._.........~...~.... ............. ....Q
.. . . ,_.._.
.

C~'W.f7 N c!) - cfj o N ~- O

(~W) u~d~Cl ~n~

3.:I Review of Previous Work - The Car leton In-Situ Shear Strength Test (CiSSST) 3.4.1 I ntroduction The concept oi~ testing the shear stren;Tth of asphalt pavement surfaces using a rotational load in the field was first conceived by Abd Cl Halim and Ahd Fl Nahi in the early l9)0's. To investi~~ate the feasii~i ity of this concept, the Carletc>n In->>itu Shear Streny~th Tesi (CiSSST) facility. shown in figure '? 1, was constructed at Carleton University ,. _ .
.. a e.~t"
. , . 4 .... _ ,.
. iy~,y', ,~.. _ ~. f ~'~~, hi~ure 21: The Carleton In-Situ Shear Strenbth 'Pest (CiSSST) h'acility The facility consisted of a cart-like chassis on small castor wheels f'or positionin~T. Force was applied via an electric motor coupled m a ~Te~u-reducer and a right-an'led aearhox. The ~~earhox transmitted the torque throu~~h a vertical drive shaft to a torque cell, which in twin was attached to a steel loading plate honded to the asphalt surface with a stoon~ epoxy resin. During the testinU procedure, torque 1 (O0 was applied until failure of the asphalt surface occurred. The failure torque v as measured with a torque cell and the failure strain was determined by measurin4; the angle of twist at failure with a protractor.
3.4.2 Deriving Shear Properties from Field Torsion Tests The laboratory torsicm specimens tested by Zahw ( 199 j had clearly defined houndarv conditions that allowed the calculation of shear properties usin'T
simple closed form equations. However, deriving shear properties from torsion tests with the CiSSST was nnuch more difficult due to the field boundary conditions. The method of testin« with the (~iSSST, known as the surface-plate method, utilircd a steel plate attached to the pavement sur-fac;. usin~~ epoxy resin. Therefore, torsion was heina applied tram a steel disc of finite dimension ( 100 mm diameter) onto a flat surface with (practically) infinite dimension, often r~:ferTed to as a half-space.
This loading conditiim is similar to the linear torsional shear stress/displacemcnt conditions shown in Fi'ure ~, and represented a completely different set of boundary conditions than the laboratory torsion test. Therefore, a different set of constitutive equations was required to determine the shear properties. The surface plate loading condition o1'the CiSSST device is illustrated in Fi~Ture 2~.

1 p~~li~cf T;>rdue ( r 1 1_u<lulll'_' ~)I~~~ ~'l;'11 - 07,11 L lll~ Il~ ~ilf~l:lvv ~7 ~ll~tll'4 ~~II~O~~>yl i ,_ i i J.
Figure 22: Loading and Boundary Conditions of CiSSST
In previous studies with the CiSSST device. Equation 6 was developed by :abdel Nabv ( I99~ i for calculatin~T the mix shear stren~~th based on the ussumytion that the failed surface formed the frustum ;>f a cone as shown in Fiy~ure ~?.
r ,.- ._ ,-_ " Jl , ~ r-;' (~) J
where:
T = the maximum applied torque (~~~m>
T = the in-situ shear str~ n~lth (MPa I
h,-= the failure depth (mm) J.

10~
r; = the upper radius of the frustum c~f the failed cone (mm) r" = the lower radius of the frustum of the failed cone (mmi 3.4.3 Main Results of Previous E~:periment The results of Abdel'~aby (199>) provided two important conclusions. First, the CiSSST was able to differentiate between the shear properties of different mixes, as well as the differences within the same mix placed in different weomc~tcies (curved sections vs. straight sections). Thi;~ indicated that the device was sensitive to changes in mix shear properties and could potentially differentiate between mixes with good or poor rutting resistance.
Second, ~~reater variation between replicate specin-~en results was observed Burin<.: laboratory testinV~ than in-situ testing. .As supported by Peck and I_ow~e (l9OOj, it was hypothesized that the coring process used for laboratory specimen preparation damaged the specimens to such an extent that a greater variation was observed during laboratory testing than that experienced during field testing.
3.4.4 Advantages of CiSSS'T Prototype The concept behind the C;iSSST facility was quite simple, yet very effective.
~o special preparation of the asphalt surface was required and the test could be completed very quickly once the epoxy resin had cured. The applied rotational force produced a state of pure shear stress without complicated bending.
tensile or compressive forces. The test was very repeatable as indicated by the low variation between tests (Abd El .~Iabi 199>) and measured actual field shear properties instead of attempting to simulate field conditions usin~~ iahc»-amrv analysis.

I~
3.~ Improved Analytical Framework to Determine Asphalt Shear Properties from the Surface Plate Loading Wethod 3.x.1 Introduction As previously presented, an equation for asphalt shear strength was developed by Abdel Naby ( 19~)~) dunin;~ initial investigations with tnE CiSSST device.
However. shear strengths calculated using Equation 6 were consistently ';renter than those determined by Zahw ( 1990 in the lalvoratory, often by as much as ~OO~i~
.
Notwithstandin47 the fact that laboratory anti field compaction methods produce different aggregate particle orientation as outlined in previous sections, it was likely that equation 6 did not fully- represent the r~oundarv conditions of the paycmcno..
therefore, the calculated shear strength of the mix was not entirely coc~'ect.
It should be stated that for comparative puyoses, this did not present a problem to the analysis of Abdel Nahy ( 1990. However, for accurate modellin;~ of ruttim_~
resistance haled un shear properties, new constitutive relationships between the suria~e plate method and field shear properties ore can -ently' under devele~pment at Curleton University by Bekheet et. al. (x()00). Although non part of' this thesis, these relationships are b.-ieflv presented in the following section for completeness.
3.,.2 Reissner-Sa~oci E'rohlern ~s outlined m Section ~.-I.?, the surf-ace plate tors;on test involves applying a torsional force throu',h a steel test plate epoxied to the asphalt surface as shown in Figure ?3. Torque is applied at a constant rate until failure of the asphalt surface in shear. The typical failure plain is shown i:n Figure ?4.

Figure 23: Load Plate Attached to Asphalt Concrete Pavement (ACI') Surface Figure 24: Induced Failure in Asphalt (_'oncrete Pavement (ACP) Surface The prob3err~ of applying a torsional moment on the surface of a half-space was developed by Reissner and Sa~oci ( 1944) and Sneddon (1946). For an elastic, l~ OJ
homogeneous and isotropic material in cylindrical coordinates yr. ~j, all stresses vanish with the exception of shear stresses ~:;4.j and Tz« as shown in Fi~~ure ?~. A
simplified form of the shear stress iLo is available when z=(l (at the surface of the halt-space) as shown in Equation 7:
4~G (7 i ~ _-W here:
G = .Shear modulus of the material a = Radius of the loading plate ~ = nnUular displacement or the loading plate (radians) r = L>istance from the centre of the loading plate (r<aj Figure 2~: Differential Elennent Shear Stresses from Reissner-Saboci Problem (from liekheet et. at. ?0()(k) I0~
It should be mentioned that lJquation 7 is only valir.i within the radius of the loading 'plate (r<a). 'rhe relationship betty-een the applied torque, T and the resulting angular displacement can be derived by integrating TL,j over the area of the loading plate as shown in Equation ~ (Bekheet et. al. ?000):
r _ 16 G~a' (S) As shown, the shear modulus may be calculated fram the applied torque and ~tn~ular displacement (twist an~~le~. No closed form solution has been developed by Bekheet for shear stren~Tth to date. While such a relationship may be developed in the future, the me>dulus (stiffness) appears to better characterir,c ruttin~~;
therefore effo?-ts were focussed on evaluatin~~ stiffness for this invstigation.
3.x.3 Finite Element '~~lodelling and Verification Although Lquation 8 provided the means to directly calculate the shear modulus of the pavement surface. the underlying assumptions of elastic.
homogeneous and isotropic conditions do not always reflect the properties of asphalt concrete. Indeed. asphalt is a visccaelastic material dependin~~ on its temperature and rate of loadin<7. Therefcn~~, to further develop the Reissner-Sago ci equations for asphalt concrete. Bel<heet et al. (200()) im.esti~ated the use of the finite element method.
Briefly, a finilc element mesh was constructed using '_'0-node, 3 dimensional bitch elements tc> simulate the Reissner-S;.yoci problems. Elastic, homogeneous and isotropic waterial propcrtic°s were lust entered into the model to compare the results of the model run uaith the dosed form solutions provided by the Reissncr-Sa'loci equations. Results of this model verification are illustrated in Fi~~ure ?(.

As shown, the results of the initial model verification were almost identical to the closed form solutions, indicating that the finite element mesh modelled the Reissner-Sagoci problem very well.
0.00 0.10 0.20 0.30 0.40 0.50 0.60 Radial distance (m) Fibure 26: Initial Finite Element Model Verification (from I3ekheet et. al. 2000) The next step. which is currently underway, will l~e to take the verified finite element mesh and apply non-linear, viscoelastic material properties that better characterize asphalt concrete. The finite clement model may then be used hot.h for ar~alyzin~ the effect of nc>n-linear, viscoelastic properties on p;duation ~.
as well as performance modellin~~ through repeated Loading.

CHAPTER 4: DEVELOPMENT OF THE IN-SITU
SHEAR STIFFNESS TEST (InSiSSTThs) ~.l Introduction and Chapter Overviev~~
Based on the material presented in the Introduction ((-'hapter 1 ) and Literature Review (Chapter '_'), it is hoped that the need for an advanced test device fco mea,;urin«
the in-situ shear properties of asphalt pavements is not only apparent, but also wau-anted and desirable. Chapter , "paved" the way for the development process by providing a sound analytical foundation for deriving shear properties from an in-situ test using the surface plate torsion method.
This chapter pr; sents the process completed durin~T the development of such a device, known as the In-Situ Shear Stiffness '1.'est (InSiSST ~"' ).
Development of the InSiSST~"f took a log=_ical, stele-wise approach. This approach commenced by first reviewin~T previous work of other reseurcher~, follow-ed by tm analysis of ctament deficiencies, which ultimately lead to the design and fabrication stales. Once fabrication was completed, a series of validation, debug~~in;g and shakedown exercises were completed to ensure ruU~7edness of the device. An initial set of field test procedures is also presented to enahle others to use the InSiSST~".
:I.2 Critical Analysis of CiSSST Prototype Deficiencies Design of the lnSiSST''" commenced with an anals-:~is of the deficiencies observed with the CiSSST prototype. Althou;h the C'iSSST represented an important first step toward in-situ measurement of shear properties and yielded much important 1 ()8 10y information, a numbe: of desi~en-related and operational deficiencies were noticed.
during its use as outlined in the following sections.
4.2.1 Chassis Design and Weight The chassis of the CiSSST device consisted of a metal cart mounted to small castor wheels. Transportation of the test device was very inefficient and dungeruus due to its large we;ght, however. once on the pavement surface, the CiSSST was relatively easy to t.~anoeuvre into position over the test plate. In total.
the CiSSST
device weighed aplaroximately ~0 kg (110 lbs.). Durin'; set-up and breakdown.
the device had to be rrtanunllv lifted into and out of the transport vehicle by at least tour operators. Since only one or two operators were required during the actual test procedure, this was clearly an inefficient use of operator- time and resources.
commodities that are in increasinUlv short .supply. Furthermore, lifting heavy equipment inherently compromises operator safety, another important consideration.
4.2.2 Stabilization of Test Device A stable test device during the application of torque was critical to the accurate measurement u1 pavement shear ~trengthlstiffness. Stabilization of the CiSSST was achieved thruu«h two methods. 'hhe first consisted of locking the castor wheels to pt~event rullin'T. The wei;7ht of the device then acted to prevent movement during testing. ~hhis method proved unsatisfactory and a secundary stabilization system w,.ts incorporated. This system consisted of six large steel stakes that were fed throu~Yh hollow tunnels welded verticaliv to each cut-~~er and midway along the sides of the test device. The stakes v'ere then driven mtu the pavement surface using a sledgehammer. Although relatively effective, problems were identified witlo this method as well. In addition to the significant operator effort required to drive the stakes into the pavement surface, the pavement surface was damaged throu~~h the use of the stakes. The most serious problem associated with this technique stemmed from the fact ~:nat the test device had to he ~rttached to the test plate prior to stabilization. Therefore, the torque cell was already attached to the test plate as the stakes were driven into the pavement and the sensitive electronics within the torque cell were subjected to intense force as the hamme;-ing was applied. This rnay have eventually damaged the torque cell - the most costly component of the device. The hammering force maid have also affected the epoxy bond between the steel loading plate and flue asphalt surface in some cases where bond failure was onserved.
4.2.3 Epoxy System Used for Loading Plate Attachment The epoxy used for CiSSST testing required a relatively long cure period. of ?~L hours at temperatures above 10"C. This required the closure of the teal site to traffic mice within a 2~1-hour period - once to epoxy the loadin~l plates and once for actual testin'u. Closing? roads to traffic at any time period increases con'estion and driver stress, as well as presenting significant safety risk to highway personnel.
Therefore, test preparation and execution must be completed in the shortest duration possible.
4.2.4 Data Collection, Control System and Available Test Program Accurate data collection during testing was limited to applied tordue only.
Furthermore, althcuugh the datalogger rcc<rrded instantaneous torque readings, the data was not accessible for un-site viewin~~ unless downloaded into a portable computer.

Angle of twist at failure was measured crudely with a protractor-instantaneous an~~le of twist was simply not possible with the CiSSST device.
Oue to this limitation, a 4Traph of applied torque versus angle of twist (stress vs. strain) w as not possible.
Control of the CiSSS~r was also limited. Only two test (rotation) speeds were availaL~le and there was no ~ a~ of monitoring for constant stress or strain conditions.
l.?.s Overall Test Device Performance To its credit, t-he CiSSST device performed extremely well Given its basic construction and control system. However, there were a number of general performance deficiencies that were ohserved durin'a testing that required attention.
The first concerned the overall ''strength" or "capacity" of the test device.
The motor and gearbox combination selected for the CiSSST was not able to provide torque sufficient to fail some pavements at all test temperatures. A limiting test temperature of 10°C was assigned to the CSSST device to achieve failure.
Secondly; the coupling used to transmit ti~r~.lue from the motor to the ~~e.u~box wus under-designed an~1 failed durin~~ one test program.
It must be ay~un emplasiscd that the CiSSST performed extr°melv wcll for the purposes for which it was designed - a preliminary, research-oriented device.
The results obtained with the CiSSST represented an all-important first step toward the development of a mainstream test facility.

11?
4.3 Design Objectives f'or InSiSSTT~'t 'Test Facility 4.3.1 Mitigation of CiSSST Deficiencies Defining the desi~~n objectives represented the nex-t step in the development process of the InSiSSTT"' facility. The. objectives presented in this section were established largely through analysis of the CiSSST prototype deficiencies in addition to other common sense objectives essential for designing a widely used test device.
4.3.2 Reasonable Cost If a test device is to log successful in any market driven economic, its cost must be reasonable as compared to its value to users. Also, it does not matter how much benefit the test will provide if the user is unable to affor~_i its cost in the First place.
Therefore, costs incur-r-ed 'by the end-user (-purchasing cost, operatin~~ and maintenance, etc.) must be justified with regard to the benefits provided by the test.
Furthermore, as the goal was to produce a platfon~~ for ti widely used test device.
these costs must he within ~r reasonable range for the average user.
4.3.3 t'ortability and Safety As road systems spurn thousands of 6;ilometres, the portability of an in-situ pavement test device to various test sites is of great importance.
Furthermore. the device must be easily mobile within indimdual test sites since numerous tests are performed to ensure statistical significance. Operator safety is another important consideration as injuries cause employers to incur loss mf productivity ~rnd increased compensation volts. Obviously injuries are also detrimental to the employees as well.

4.3.4 Number of Operators and Ease of 1_!se Employee salaries are usually the single lar=est expense that an employer will incur. Therefore, minimizing the number of operators required to perform the field test will greatly increase the attractiveness of the device to both end-users and their clients. Additional savings may be realized by developing a test that is simple to perform such that specialized trainin~~ is not required for operators.
4.3.~ :Minimal Test 'Time and Damage to Pavement Surface Minimizing the time required to perform a field test produces two substantial advantages. First, more tests may he performed for a given time period, increasing, the amount of data acquired by the researcher and the amount of money «enerated by the contractor. The second advantage concerns the disruption to traffic flov.-. As this is an in-situ test, sections of road must be closed to her-form the test.
which increases traffic con'Testion and the potential for worker injury.
Destructive pavement tests are becornin~ increasingly undesirable since the result is usually an acceleration of pavement deterioration. Tests that are non-destructive or that produce lade disturbance (semi-destructive) to the pavemern structure are favoured.
4.3.6 Correlate Results to Pavement Performance Indicators Perhaps the most important consideration when developing the InSiSST'"
facility was the need tc~ correlate the field test results to both standard laboratory values and pavement pertormance indicators such as rutting and cracking.
Achieving such ccorrelation would yield si~,nificant and immediate benefits to the three primary areas of pavement enaineerino. The first area is mix design. An in-situ shear stiffness test in conjunction will-; laboratory testing would be a powerful combination for analyzing the potential of proposed mix designs. Also, the results of such a test apparatus could be used to produce "shift" or "master" curves relating in-situ shear stren~th/stiffness to various factors such as loading rate, temperature and asphalt content to name' a few. The se:ond area is the quality control and quality assurance (QC/QA). 'Newly constructed asphalt pavements could be tested to verify acceptable construction practices through the nneasurement and comparison of in-situ strength parameters with code requirements. The final area is the long term pavement performance (LTF'P?. :vlonitoring of actual field shear stren'~th/stiffneas of pavements with time would assist in predictin~~ future pavement performance. fhhis, in turn. would allow f~~r more efficient allocation of limited rehabilitation funds and also help determine the effect oi'real world conditions, such as environmental factors, on pavement performance.
-L4 Design of InSiSST~rx' Facility :~.:L1 Introduction and Overall Design The desi~Tn or the InSiSST~~' facility was conceived based on the states desi~Tn objectives and noted CiSSST deficiencies. Whil° complete adherence to the desi~~n objectives wus the ultimate goal. trade-offs between objectives were necessary. Thererore, the lnSiSST'''' design represented an optimization of the individual design objectives into an integrated system.
~rhe completed InSiSST''''' device is shown in Fi«rres '~? through ~~). a's shown. the components are mounted to a small A-f'ram~~ trailer to provide exceptional portability. As with the CiSSST facility. the InSiSST' ~' utilises an 11>
electric motor and gearbox to produce the required torque. The motorlgearbox combination is mounted vertically on a stee°.1 platform that is attached to a positioning system that incorporates two sets of worm-screw slides working in tandem, also referred to as an "X-Y table." The top set ef slides allows positioning of the platform in tie transverse direction (with respect to the trailer orientation).
The transverse slides are in turn mounted to a second set of slides allow°in~T
positioning in the longitudinal direction. The entire positioning system is mounted to a box-tube frame occupyin~~ the space beuveen the tow bar and the axle of the trailer. The test frame is attached to the trailer frame via lour screw jacks, one at e~rch corner of the test frame. During translaortation of the InSiSST'"', the jack:: are retracted to hold the frame in the air to prevent damage. Once driven into position.
the jacks are extended to lower the test frame to the ground and then continue extending until the wei~Tht of the trailer is supported solely by the test frame. As with the positioninyT system, an electric motor is used to raise and lower the jacla.
A sin'71e motor is used to deploy all four jacks using mitreboxes and dr7veshahts.
Contri~l of the i~rck~, and positioning; slides is provided by commercially available electric motor controls. Control of the actual test procedure is provided by a laptop computer. Instantanec-pus torque and angle of twist measurements are collected on the computer during the test procedure. A large plastic storage hox is mounted to the front of the trailer to house the electronic cor~~ponent~ Finally. a ~~enerator is mounted to the rear of the trailer to provide electricity for the InSiSST'~'.
A more detailed explanation of the individual systems is provided in the following sections.

4.4.2 The Primary Force Generation System (Powertrain) After investigating alternative methods to produce the required rotational force (torque) for the test, it was concluded that, like the CiSSST, the use of a simple electric motor and gearing still represented the best choice for this application. Systems incorporating hydraulics or pneumatics would certainly produce acceptable, if not superior results. >=however, these systems were simply too expensive, at least at the concept exploration stage.
The first in-iport~mt improvement involved the vertical ali;~nment of the main drive motor and gearbox. Ttae vertical ali~Tnment saved a si'.~nificant amount of space when compared to the CiSSST facility-, which utilized a right-angled gear'nox attached to a horizontally mounted motor. ~fhe straight gearbox an-angement provided increased capacity and reduced backlash compared to right-angled ~7earboxes.
hhe overall capacity of the motor and gearbox system was increased significantly to ensure failure of all asphalt surfaces encountered (over a reasonable temperature range). The gearbox is a triple reduction unit with a final ratio of 8101:1. Therefore. 51()1 revolutions of the main drive motor arc required to turn the output shaft of t!~c gearl~c~x a single revolution. This lnu'Te reduction was needed not only to reduce the test speed to reasonaf~le levels, but also to increase the available torque to fail the ~isphalt surface. Whereas the f:_'iSSST device produced a maximum torque of approximately >08 N*rn (4>00 lbf'~in). the InSiSST'"' can apply up to 1 »0 N" m ( I ,700 lbfrin) of torque, an increilse of ewer ''00<<
.
another siy7nificant improvement over the previous desiy'n was to mount the motor directly to the <Tearbox. As mentioned earlier. the CiSSST had a doveshait 1'_'() and coupling between the motor and gearbox which failed during one suite of field testing. The direct attachment also reduced power loss between the gearbox and the motor. A final benefit was that the direct motorlgearbox coupling provided a ti~~ht seal, thereby significantly reducing_= the likelihood of infiltration of water and/or dirt into the gearing.
One disadvantage of the vertically mounted gearbox and motor was a higher centre of gravity. However. a restraint system was developed using tie sir aps to prevent movement of the gearbox and motor during transportation.
4.4.3 The Transportation System One of the greatest pr~;~blems with the CiSSST facility was its lack of portability. The integration of the facility ~a~~ith a trailer allows exceptional portability from site to site. Furthermore, the facility no loner requires lifting i>r lowering by human effort. This drastically reduces not only the potential for injury, but also the number of operators required f~~r testin~~. which will provide si'~niticant cost savings to the end-user. Another benefit of the trailer-mounted option is that any vehicle with a trailer hitch may tow th:; facility.
One disadvanta~~e of the trailer-mounted option is that the facility is subjected to a much harsher environment. such as the infiltration of water. dust and dirt.
ljowever, judicious selection of rugged and/or sealed components reduced this concern.
4.4.4 The Test Frame and Positioning System The test frame fills the space betty-een the trailer axle and the front cross bar, providing the foundation for many of the ~=ssential InSi~ST"'' systems as shown in Fi~,ures ~7 and '_'8.

l~l There are two "levels'° to the test frame and p<asitionin~ system.
The lower level houses the- lovler slidin' system, consistin' of the ~ sets of tandem worm-screw slides (i.e. ~ individual slides] mounted orthoyanally to allow movement in the longitudinal and transverse direction as shown in Figure 30. The bottom woum screw slides are mounted within I50 mm wide steel channels that run the len~~th of the test frame to prevent damage duuna transportation.
Legend:
I A - Transverse Slides B - Longltud~nal Slides C - Drive Belt D - Motor E - Test Frame U
Fibure 30: Plan View of the Lower Positioning System (Protective Steel Channels not shown) 1 ~~' The upper test frame and sliding svstern were developed to isolate the gearbox from the worm screw slides for two main reasons. First, due to the lame weight of the gearbox (136 kg or 300 lb), it could not be directly mounted to the worm screwy slides according to the manufacturers specifications for static (dead) load.
Second, if the ~Tearbox was mounted directly to the Slides, the reactionary force produced by the gearbox during an actual field test would be transmitted thr<~ugh the slides themselves. Even at low levels of applied torque, this reactionary force would 'ready exceed the manufacturers specifications for dynannic load and would likely damage the slides.
The upper level of the test frame consorts of >0 mm hollow structural sections that provide support. t~or the upper sliding system. As shown in Pi'_ure ?~), the gearbox is mounted to a steel plate 300 mm (1? in) square. This plate is mounted on rollers and slides transversely across a set of connected ~0 mm HSS beams, which arc also mounted on rollers and slide longitudinall v along tire upper test frame. The upper sliding system is attached to the worn screw slides below.
thus allowing the slides to control the movemeno_ of the ~Tearhox within the test frame.
Before a test is initiated, the connection between the upprr and lower sliding systems is removed by clampin'1 the upper :Aiding system to the upper test frame.
B~ isolating the upper and lower sliding systems, the large wei;_=ht of the 'rearhox and reactionary forces are not applied to the worm screw slides.
A dedicated controller and control pad housed in the front stora',e box control the movement of the entire positionin' system. In its current configuration, a net travel distance of 1~0 mm (f~.~) in) in either direction fro»~ the centre position is capable with the transverse slides and a total longitudinal travel distance of 6~0 mm 1_ (?>.? inj is capable with the longitudinal slides. Therefore, a total testin«
area of 0.1 my (148 in-) is available each time the test frame is lowered. If a 100 mm (~
in) diameter test plate is used, four plates can be placed inline V.'ith a minimum distance between tests of 80 mm (3.1 in). If larger load plates such as 1?~ mm (> in) and 1 >0 mm (6 in) plates are desired for lamer aggregate mixes, :~ plates should be used to provide a minimum between plate distance of 88 mrn (3.~ in) and 63 mm (~.~
;n1.
respectively. Initial analytical modelling by Belcheet et. al. (?C)00) has indicated that strains experienced outside of the test plate drop to less than one percent at a distance of 50-mm (~' in) from the outer ed~ae of the plate. Therefore. a statistically si~Tnificant number of tests can easily be performed for a sin~Tle test-frame deployment. Furtloermcwe. 'the slides incorporate sealed motors and bearing;
to be protected from the elements.
4.-t_~ The Stabilization System A stahle test platform was a critical design factor f or the InSiSST'" faci lity.
By lowering the test frame and liftin'T the trailer off its wheels usin~l,jacks, the full wei'I~t of the trailer is applied to the test frame. Stability against the rotational force applied to the teat plate is therefore achieved through frictional force between the bottom of the test frame and the pavement surface.
The test frame used in the lnSiSST' ~' facility presents a frictional condition very similar to that observed with thrust b:varinas or disk clutches called ''disk friction'. An applicable f~rn~ula for disk ti~iction may t>e derived by considering a rotating hollow shaft. For a hollow shaft whose end is bearings against a solid flat surface, the minimum torque required to beep the shaft rotating may be computed using Equation J belay (Beer and Johnson 1988).

1?4 R~ R' ( 'y M =''~~PR~-R
~~~here R, and R_ are the inner and outer radii of the shaft respectively, A9 ~s the required torque, ,uA is the coefficient of dynamic friction and P is the axial force applied to the shaft. By replacing the dynarrtic friction coefficient ,uE, with the static friction coefficient fit" Equation 9 may be used to find the lar~~est torque that may be applied to the disk prior to slippage. For this application. the test frame itself is analo~Tous to the hollow shall. The force P is applied by the Gravitational force of the trailer and test frame on the pavement surface while the couple M is applied by the motor/Gearboa durinG the test procedure:. To find the magnitude of force P
sufficient to resist the rotational reaction, "equivalent radii" were determined for the test frame. Tha test f-rame is rectanGular m overall shape as shown in Fi~lure >l.
1200mm __ ______ _.--i _ f .~i i -- ~ J~ Equiv.
j i, ~~i, ' Outar ' ~' Min. inner ', ''~ Radius 750mm ~ ~ Radius i ~! ~'~ ~ ~ / ' (600rnm) i I, ' (450mm) J
i ~ , . ;
~---~ , I
_~
_____ _ - _ '_-i_ -__- _ -- ~~_ ___ __. _._-_ 1050mm _ ~,...' Figure 31: t'lan ~- ieH of InSiSST~r"' Test Frame 1_ Based on previous testing regimes with the CiSSST device, the maximum failure torque applied to an asphalt pavement surface using a 100 mm (-t inch) diameter test plate was approximately 508 N' m (4500 lbf*in). It was reasonable to assume that the weight of the trailer will b,; evenly distributed at each corner c~f the test frame via the ;acf:in~ system. Therefore, regardless of the position of the :motor and ~~earbox within the test frame, the full frictional resistance of the interface between the test frame and the pavement surface should be mobilized assuming that the pavement surface is relatively flat.
The total contact area between the test frame and the pavement surface is 0.~?
m- (5.6 ft-) includin'; the channels that protect the positioning glides. From the centre of the test fame. the minimum inner radius is 0.-Ii m (17.7 in). The "equivalent" outer radius was then calculaed using the c=quation for the ~Irea of a ?-dimensional 17n~~ as shm~ n in Equation 10.
(~om;r~_ jumm ~,1 ~) ) W'hele:
,A = total contact area (0.5? mu);
r"",~_ = mintrnum inner radius (0.~5 rO
The resultins, equivalent outer radius eras found to be O.C~O In (''s.9 in).
L'tilizinU these radii with :'l~I = 50~ :~'~'m (=f~00 lbfwin). alr:d assuming a cc~efl~icient oi~
static friction of (>.5, the minimum wei~T1 (normal W -ce. F') applied to the test frame must be approxim~Itelv 190 N. Therefore the load applied to the test frame must be lq5 1<g (-~?8 lbs.l according to Equation ~l.

1?6 Much of the required weight is provided by the trailer and test frame, as well as the equipment necessary for operating the test facility. This includes the jacking system and the data acquisition/control system.
A static coefficient value of 0.~ was selected for the analysis to ensure a reasonable factor of safety. The actual coefficient of static friction between the test Irame and the pavement surface is likely to be 0.7 or greater as a neoprene irubb;.r) pad has been epoxied to the bottom of the test frame to increase the friction.
This s~~stem prc;ventcd movement dur-in~ testing and eliminated the need to drive stakes into floe pavement surface. ~I~h~erefore, no damage is imposed on the pavement throu;=h stabilization and no operator efl~c»'t is required.
Furthermore, by raving the stabilization completed prior to positioning and testing, the sensitive electronics of the tordue c~'~l are not subjected to unnecessary stress. To prevent the use of multiple motors. a set of custom driveshafts and mitreboxes was fabricated to connect all four ja~~ks to a single motor, ti~~=rebv reducincost ~rnd ensurin~~ th~rt the jacks do not operate independently of one another.
The jacks themselves are coated with a plastic layer to resist corrosion. and accordion-like hclle>ws cover the serews to prevent the infiltration of wetter and dirt into the ~Tearin«.
4.4.6 Epoxy System ,~s mentioned. the epoxy system used preyiouslv required 2~ hours to cure prior to testing. This limitation required two visits to the test site, an inefficient and costly method of testin~T.
During, the ir~vesti~~ttion. numerous adhesive systems were tested for suitabilia°. Most :~vstems either required cure times that were similar to the existing I?i system or did not provide suitable strength at all. For e~;ample, instant contact-type adhesives were not effective as they rely on direct conta;a between the asphalt surface and the steel plate. Due to the rouhness of the asphalt surface, this contact was not provided.
iat this time, the best performing product is a two-part epoxy system that provides adequate strength after ~ hours at room temperature. Therefore, curin~~
time at elevated temperatures, such as those experienced in the field durin'1 testin'?
will reduce this tiroc. Two hours was deemed as an acc~.ptablv short time interval as the tests themselves can be completed ivy minutes once the epoxv~ has cured.
-L4.7 The Test C:ontroUData Collection ~vstem The main drive motor is controlled by using a variable speed motto controller with a speed sensor connected directly to the motor shat-t. The speed sensor provides a closed loop system and ensures ti~at the motor does not deviate from the desired test speed. Theref~me, the InSiSS~I''~' is a strain--controlled test as the rote of displacement ttwist an,Tlei is controlled, w-Nile the resultinC7 torque lstres5) is measured by the. torque cell. The accuracy of the mover controller is ~I
revolution per minute :rpm).
The variable speed motor controller also allows thL selection of a variable test speed between zero revolutions per minute lrpml and 1 ~sQO rpm. T;rble I ~
displays pre-programmed "train rates and there associated drive motor speed. The strarn rote of 0.000> revolutions per second corresponds with the strain rate used in the Super-pave Shear rl_ester du~in~J lrequencv sw~ecp testin'T.

1~8 Table 13: Target Test Strain Rates and Associated Motor Speeds Target Strain Rate Required '\'Iotor i Speed ( revs) (rpm) 0.000; A"

().0011) ~ -18C
_ I).~)~)1~ 7~y -O).~)~)~U~ ~~= _-().00~'O - 1 '_' 1 1 1~~'i i 0.0() i~ 1 i01 _ 1).007 ' 1500;:x_ i _ _ _ __ * Strain rate i~f SS'T for frequence sv.~eep testing_ *' 1 SO(1 rpm is ttnc ma.vin~um available motor screed Torque i5 recorded with a torque ce:l similar to th;~t used with the CiSSS~I-facility, althou~,h of hi'her capacity. Both the InSiSST'h' torque cell and the motor controller have standard RS-?s~ (serial) connections for connection to a computer.
However the laptop has only a sin'~le serial connection. To overcome tl;is problem, a Universal Serial Bus fUSB) adapter box was used. 1"this device allows the connection of up to ~ individual serial connections into the adapter with a sin«l:
LISB output to the laptop. Therefore, the laptop computer is able to contrc>I
and acquire data from up to -1 individual serial devices simultaneously. .At present, only the torque cell and motor controller are attached to the laptop. The positionin~T
system also has a serial connection and nvav he controli,~d with the laptop in the future.
This centralized control and data acquisition svstvm allows the collection of instantaneous readin's at user-defined sa.nplin~ intervals. Results are saved directly to the laptop and other relevant information su:;h as test site location.

1 , c~
weather conditions, temperatures, etc. ma.y also be directly entered into a database forfuture analysis.
4.4.8 Overall System Integration All of the separate components were selected and designed to work well to'7ether as a single unit. The result is a test facility that is portable, stable, and ru~a~ed. The test requires only a single operator, no heavy lifting or complex set-up.
and can be completed rapidly. The test results are accurate and are available instantly. All of the individual components operate under the same power rcquaremcnts as provided hw the central generator.
4.4.9 Cost As the InSiSST'"' is still in its prototype form. the actual cost of the device is not representative of a final production m~ndel. 'The component costs oi~ the lnSiSST'''' totalled approximately $37.00 C.AD ($~>.000 USD). However, there are many other coats such as labour, overhead, marketing, etc. that must he factored in when determining a final purchase price. Furthermore. the component costs would likely decrease if produced in lar'Te;-quantities. E-Iowever, based on the component cost, it is estimated that <i final production model would be priced below 50.000 LISD. w-hich was identified by th:v Ncitional Asphalt Pavement :association (I<.AP.A) as being a reasonable cost fur su~_h a performance test (IvHW.A
199~d1.
~.~ Fabrication, Debugging and ''Shakedown" 'Testing All fabrication activities wore completed at Carfeton Lniyersity in the CiviE
and Environmental En~~ineeriny Laboratories. As with any Nev. or complex te~hnolo~v, a 1s0 number of interestin? challenUes were encountered during the development of the lnSiSSTTh.f device. Perhaps the most frustrating were the long delays experienced when ordering and acquiring the component parts for the InSiSSTT'''. .Although most of the components were comrnerciallv available, they were not actually fabricated or assembled until ordered. thus requiring up to '_' months to receive and in turn delaying the fabrication of the InSiSSTT'f on multiple occasions.
4.5.1 Positioning System Debugging When the p~>sitionin~z system was first installed and attached to the <_earbox, the system w-ould often "stall" while mcmm« the ~Tearbc>x bath and forth within the test frame. It first appeared that the steppin'; motors of~the positioning slides v-ere not powerful enough to move the ~enrbox. However, Moon further examination, it was discovered that the slides were not completely parallel and that the _>earbox was hitting the slides at certain spots along the slide length. Once adjusted. the positioning' svstern performed pr<aperlv.
1.~.? .lacking System I)t:inugging The jacking system itself presented no major problems durin'; development.
although the rate .~f extension and retraction of the jacks is faster than anticipated.
The addition of a year reducer to the jachin'~ motor is evpected to redact the ~,pe~d nt which the jacks rcnated, although has n~~t been completed to date.
=1.x.3 Test System Debugging Once constructed, the motor control ler and torque cell were connected to the laptop cc»nputer via the L'SI3-to-Serial ronncction box (o herein the software development. Lnfoutunately, tire test softG~~~rre conflicted with the LSB
adapter and caused the computer to 'crash" evervtime the test softy are ~-as initialed.

Isl software patch was obtained from the manufacturer of the USB adapter and the problem was solved.
4.x.4 Shakedown Testing A series of shakedown exercises were completed v-vith the InSiSST'" to ensure an acceptable amount of rubgedness prior to field testing. Most of the shakedovm testing was connpleted in and around the Carleton University campus.
Tire most challenging test of ruggedness for the InSiSST'"' was completed between July 16'h and 18'r' of 2000 when the trailer was driven from Ottawa to College Park, Maryland and then to Washington, DC for a demonstration at the Transportation Re~eurch Board. The total trip distance ~.vtis approximately 2000 hm ~ 1?~U miles). most o1 which was along mayor Canadian hi~hwuvs 1416. -101 ) and US Interstates (S0. 9j) at speeds ranging from 80 to 120 km/h (~0 to 7~ mph).
«'hile in College Park and y'v'ashin~ton, tf,~e InSiSS~T'" was towed along city streets, many of which were in poor condition d~st7laving potholes, extensive cracking.
patchin4a and rutting. The InSiSST'"' trav::rscd countless shays bumps and dips durin« the journey and there was some concern for the foealth of the electronic equipment housed in the stora«e box. ~vo~cever. no danuaUe whatsoever was observed upon return to Carlcton University, with the exception of some very thin surface rust on the metal test frame. The test frame will he soon cleaned and painted to prevent future rusting_. To ensure that the electronic eduipment does not fail in future tops, hubble pud~iina will i,e installed to protect the eduipment from shock.

I >_ 4.6 Field Test Procedure The following is an initial draft of the field test procedures for effectively and safely using the InSiSSTI"~. It is expected that more comprehensive versions of l:hese procedures will be developed with the increased use of the device.
4.6.1 Equipment Checklist Before leaving to the test site. the operator or technician should ensure that all necessary equipment is packed and in proper v~orkina order. Table 14 lists the eduipment needed for field testing.
Table I4: I~quipment Checklist ~ Sufficient epoxy for field testing and associated mixing equipment ~; ~ Sufficient loading plates (cleaned and roughened) ~~ ~ Fuel for the generator j ~ Infrared ti~erTmmeter ~ Stiff brush t.:~ elcan dust from asphalt surface ~ Callipers tot rneasurin~ the depth of failure ~ Dipstick profiler (if rutting survey desired) _ _______-_._.___- ___ ___ __-_ ~ ~ Nuclear Dcnsiv Ga~_y1e (it density survey desired)-__ .
~ laptop computer _ :L6.? Transportation Safety The InSiSS~h'"' is a trailer-mounted device that is towed behind a vehicle to w<crious test sites. Therefore, safmy must he an important consideration during the transpooation of the InSiSST'"'. Tu help ensure a safe journey, users of the InSiSST'~" should follow towinV~ safety 'Juidelines recommended by trailer manufacturers and/or aove~z~ment a«enci~°s. A comprehensive wide to towiryT
safety is produced by Sherline Products Incorporated ('iherline 1999) and the ., , 1a>
Ontario Ministry of Transportation (MTO ?000) has developed a quick checklist for trailer safety.
Prior to transportation with the InSiSSTTM, the jacking sv~stem should be completely retracted such that the test frame is in its uppermost position.
Fur-ther-more, the ,,earbox and drive motor must be secured to the test frame and.~or trailer using the ratchetin~ nylon straps error to transportation.
4.6.3 Securing the 'Pest Site Closing of a road section or lane alv:ays involves some risk. Therefore.
traffic control should oniv be car~r-ied out by trained professionals with the proper equipment. If possible, traffic control measures should be initiated <rnd completed by local or provincial transportation aQencv personnel t<> ensure the safest wcwkin~
conditions. I-~owever, if private traffic control is required (and permittedl, the contractor should ; ontact their local or provincial a'encv for appropriate tratfic control procedures and eduipment.
t.6.4 Preparation of Ya~~ement Surface and Bonding the Loading Plates Pavements arc subjected to numerous types of dir!, c»1 and other chemicals that are introducec.l and tracked by automobiles, trucks and other vehicles.
These chemicals will often adversely affect the quality of the bond between the asphalt surface and the steel loadin~~ plate, therehs~. affecting_ the test results.
Tu ensure the highest quality hand, the pavement surface should be free of deleterious substances.
.At a minimum, a stiff brush car hroorn should he utilized to remove fine particulate materials. In some cases. rt may be also necessary to '~antlv v a.sh the rrsphnlt surface with soap and water to remove rn:»-e stubborn substances. Care should he 1 i ~l taken in these cases not to dama~le or strip excess asphalt from the pavement surface.
The surface should be completely dry prior to the placement of the loadin~a plates. The epoxy must be prepared according to the manufacturers specificatir~ns to ensure maximum strength and bond. The epoxy sho~_~ld be spread evenly across the bottom of the loading plate with care such that air buhbles are not entrapped.
Enough epoxy should be used such that when the plate is placed on the asphalt surface and compressed. a small amount of epoxy is displaced along the perimeter of the loading plate. The loading plates should be placed in either a straight fine or staggered such that they will ~rll tit within the test frame when lowered.
Excess epoxy may be removed with a clean towel. If possible, a small weight such as a brick should be placed on the plate durzn« the curia' process. Pavement temperature is measured with the thermometer and the required curing time haled on manufacturers information is assessed.
:~.6.s Ruttin<t;/Densitv Surveys (Optional) While the epoxy is curing, a ruttin~_~ and/or density survey of the test secaion may be completed to provide additional i;aformation for analysis. :~ ~~rid system such as that illustrated in Fi~7ure s? can hG_ marked usin~~ a chalkline.
'hransverse and longitudinal profiles may be measured using the Dipstick or similar profiler to provide rutting and roughness data. A nuclear density ~Tauge may also be used to measure the density of the asphalt both it the whe~lpatlrs and the midlane.
,an example pattew for in-situ shear testing for research purposes (white circles) and c~rir~g ('rev circles) is also shown in Figure ;'', althou;;h the pattern shown v,-c>uld require the movement of the InSiSST'"' within the test site.

1 i1 \li~ll;m~
lmm:- ( 7u;~~r \\ Imll';nit ~ \1 I~~:~II'atl~
i O O
CO
C_~
o c~~
nl, o c_ O
O CO
Figure 32: Outline of Rutting and L)ensity Survey 1 ,6 x.6.6 lnSiSSTT~' Test Procedure Once the epoxy htis cured, the field tcstin~ with InSiSST'-" may commence.
~hhe followin~ steps should he completed in order:
Detach the InSiSST''~' from the tow vehicle and manoeuvre it over the test plates such that all test plates will be within the test frame when lowered (Figure s,).
?. Chock the tires of the trailer once in position to prevent movement of the trailer. A brick or wooden wedge work well (Fissure 34).
.attach the torque cell cable to the torque cell (Figure ,5j.
-I. Install the torque cell ;end connecting collar onto the gearbox driveshaft.
Lower the test frame to the 'round and ensure that the we°.i«ht of the trailer has been transfeurcd to the test Iramc (the trailer suspension will relax when this Occ L1 I's ).
(~. :attach the L'SB connemion to the laptop and tuc~n tf~e laptop on.
7. Turn on the slide controller.
S. Use the slide.~ontrol pad to align the torque cell ~and <~earhox) over the tia~st test plate (Ti'lure ;6).
~). :~ctivtlte the motor controller softwaro on the Iaptol~ and turn the torque cell until it is aliened with the load plate in the radial direction.
10. Lower the connectin'T collar on the torque cell into the load plate (connectin~~
collarshovn in Fi~urc _s7).
I 1. Weasure the pavement temperature directly adjacent to the load plate and record it for future ann°'vsls (Fissure s8 i.
1 ~. W'ith the torque cell tirn~ly connected to the land plt~te, activate the torque cell software. The soft~w~rt~ v ill he«in to t,Ike readm~s tit the predefined samplin«
rate and recor~:1 them in a text file. Th::re should be very little load shown for these initial rcadin<7s as the drive motor has not bee°I restarted vet.
1 ~. Press the red calibr~Ition button to calibrate the torque cell (Fieure >G). Ensure that at least one calibration reading has keen recorded in the text file.

-, lm 14. Once calibrated, the test may be initiated. To do this, select the desired strain rate from the motor controller software. As the stain is applied. the torque readings should increase until failure of the asphalt surface.
l~. When failure occurs, press the Escape key to stop the drive motor.
16. Inspect the failed asphalt surface and loading plate. Weasure the depth of failure with the callipers and record it for later an,ilvsis.
1 i. Save the test file to the hard drive.
18. Repeat steps b throu«h 16 for the re;mainin' load plates.

rigure 34: Chocking the Trailer 'fire Ia figure 33: InSiSS'C~'M Trailer over 'Test Plates IsO
Figure 3~: Attach 'Torque Cell Cable to Torque Cell Figure 36: InSiSS'I'T"' Controls (Computer not shown) 1=lU
~.~"~
Torque Cell -~_ _..~r...-~ . _ . 'best .
,.N: ~ Plate p~..a x Co nnecting .. ~- Collar ~ ..
a~ ~. ~ ~ .R
. ~ . ~, z .
.~
.. ._ :~ . ~ , Figure 37: Connecting Collar from 'Torque Cell to Test Plate Figure 38: 'Taking Pavement Temperature with IR Thermometer 1-f 1 4.6.7 Leaving the Test Site After all of the desired tests have been completed, all equipment should be collected and stored for transportation. The InSiSSTTh' test frame must be fully raised and the gearbox must be secured to ~.:he test frame and trailer usin~T
the ratcheting straps. Reattach the InSiSSTT~' to the tow vehicle as per the safety recommendations discussed previously. Ensure that the site is left clean and that the small divots removed by the InSiSSTT"' durin~T testing are sealed with a slunrv mix to prevent moisture infiltration.

CHAPTER ~: Preliminary Testing and Validation ~.1 Introduction and Overview With the lnSiSSTTM facility constructed and the base analytical models available.
the final stayTe of the investigation involved preliminary field testing for validation pw-poses. Chapter ~ first presents the results of an exercis=e to validate the linear elastic assumption made by the Reissner-Saaoci equations. lvext. the results of comparison testin' with the CiSSST and InSiSST~~'' devices are presented, includin<~ an interesun;~
observation concernin<~ the test pl<rte diameter. Chapter s concludes with a comparison of field shear properties to those observed in the laborator;..
~.2 Analytical Models vs. h'ield Test Results x.2.1 Verif ication of Linear Elastic .Assumption As reported 1~v F3ekheet et. al. (''00O). preliminary field tests were conducted usin<T the CiSSST with the objective of verifying the applicability of the linear elastic assumption used ha the Reissner-Sa~oci equations for asphalt concrete:. 1~0 this end, pilot testy mere completed at Carleton Llniversitv with a modified test procedure to accurately measure the suri~uce displacements (I3el<heet et_ al.
200U>.
The an~lulur displacement vulr.res lrr.m the tests were also compared to the expected values using Equ~rtion 8, as shou.-n in Figure ~9. While the number of data points from the field tests was low, they displayed hi'Th cowelation with Equation S
(R~=O.S6). This vesult implied that the linear-elastic ussumpuon for asphalt pavement behaviour was reas«nable in this case. How<wcr, tf~e pavement L

1-L i temperature durin; the test was close to room temperature (approximately ??"C
or 72°Fj, v hich would have contributed to the linear elast» response.
1.200 - _ __ - _._. __ 1.000 ~ ~ Mathematical Model --, E
~ 0. 800 _.
~ 0.600 -a~
U
Q 0.400 ~ 0.200 -0.000 ~ - -0.00 0.10 0.20 0.30 0.~~0 0.50 0.60 Radial distance (m) h,i~ure 39: Comparison of Field Results with Reiasner-Sal;oci Model (from Bekheet 2000) x.2.2 InSiSST'"' vs. (.'iSSST
The surface plate rruahod of testing is used for bath the CiSSST and the°, InSaSSTI"', however, a sct of comparison tests were completed to ensure that similar results were achieved. The tests were completed in a purkiny lot at Carleton L:~niversitv consisting of an IdL~ asphalt concrete mix. HLS is a standard sura~ace mix used throughout Ontario for lov- to medium traffic volumes.
Unfuntun~ttely, due to the aye of the mix c8 years), the actual mix desi_;n data was not ~~vailable for analysis.

Test plates of 9? mm i 3.6 in) diametc;r were epoxied to the pavement sunace for each test device. The plates were placed in a straight line with a minimum clear spacing of 65 mm ('_'.5 inl between each plate. Testing vvas completed the following day to ensure full cure of the epc.~xy. Results u~f the CiSSST
testin~~ are given in detail in Appendix D, and summarized in Table 1 >. One of the CiSSST
tests failed between the epoxy and the steel test plate, while the remaining 5 tests produced failure of the asphalt surface. 'The pavement surface temperature for tests C 1 through C4 was '__'9°C, while test C5 was tested at a surface temperature of ?7'C.
The averay~e shear strength calculated through Equation 6 was ?058 kPtt with a relatively high coefficient c~f variation (C'OV') of 19.?~~~.
Table 1~: COnI~aI'1S011 Of CISS~T and InSiSSTT"' Results CiSSS~T Testing -i---_. IrI~iSST Testing Ultimate Pavement Ultimate ' Pavement ~, Surface Shear Surface Shear ~

Test ~O. , Test ~O.
~I Strength rCemh. Strength' Temp.

(kPaj ~ (C) ~ (kPa) ("C'l __ y ',:1~
CI __~77> II. ' , __ __ C~ 1~ ~ ,1~~
I -~~~~
.~~) -_ ~- I', -__ IT ~~~? -._ 5 -1 _ ~ I _ ,~ _,I
_ ~r~
__ C;

_ -~~ _ ,:~>~ ~__ ,~
_ _ _ _ _ Avera~t , _?3(14 Average j ' ~'.(l5li ' ~

S'1'Dev ~'rD_ev '~ ~~~~-._ __ I

I 9 C_O_V_ .2 I 7.2 i CO V I -~

_ uation _ 6 ~ Abdel "'Liltimate Nahy Shear 1991 Stren,_tlv ~
calculated with E~.i Once CiSSST the ~
testing InSiSS~r was complete, pl4ttes were tested -however. ween the each steel test test yielded plates bond and the failure epcy-.
het As these it was test hypothesized plates that were a thin newly layer machined. of grease or oil was present c}n the steel surface, which hail compromised the hand.

1-l~
The plates were placed in an oven for 2 hours, roughened with a circular Grinder and then re-epoxied to the asphalt surface. ~~Jhen tested main, all 3 plates produce°d the desired failure in the asphalt surface.
Results of the lnSiSST testing are shoe-n in detail in Appendix E and also summarized in fhable l~. (_'nfor-tunately, the pavement temperature at the time of InSiSST'~''' testing was hither than during the CiSSST testing. The surface temperature for tests Il and I? was 3s°C, while I3 was tested at ~?°C. To compam the results of the InSiSST'"' to the CiSSST, the shear strength of the InSiSST'"
tests were calculated using Equation 6. .A.s shown in Table 1 ~. the avera~~e shear strength was ?~~0=I kPa, with a COV of 7.?'io.
A two-sample t-test was used to determine whether or not the mean valraes were statistically the same. The null hypothesis assumed that the difference between the means was zero, and the resG.ltin~r t statistic was calculated using, Equation 7 1 (Miller et. al. 1~~90):
~r~,(rn +rr. -?1 -_-_- _-- i -__ ( 1 1 ) y~ijt, -1).~i--t,y -t).~y ~' rl, T n-v here:
x,, x_ = the sample means c~ = the desired difference hetween the means (zero in this case) n,, n ~ = the number of observations s,. s~ = the sample standard deviations The resultin~~ t statistic was x.36, w hich was less than the critical t of ~.7U 7 for (n,+n,-?) _ (~ degrees of freedom at the I r~ confidence level. Therefore, the null 1~6 hypothesis was accepted and the difference between the mean values (d) was statistically zero. These results indicated that both devices were tneasurzn' the same material property. althou,h the InSiSSTT"' results were much more consistent (lower COV). however, it was expected that the InSiSSTT"' results would be lower than the CiSSST results as the asphalt temperature was <greater- and the CiSSST test had a faster loading, rate. While more testis' is clearly required. the high COV of the CiSSST tests t If~.7~~~.) could explain why the CiSSS f strengths were lower: in this ease.
x.2.3 Practical Calculation of Shear Modulus Using (:quation 8 As previously mentioned, the current an alvtical models were devcioped asaumin~ linear elastic material properties,. I-Iowever, v. he-n actually measuring material properties in the field or laboratory usin' test equipment, the resultin' data i5 often not strictly linear. Fi'Ture =IO displays a typical :_rruph produced during testing with the InSiSST~~'. Although the 'uraph is fairly linear, an ''S"
type curve is observed due to tolerances with the devrce at the beainnin~ of the test, as well as non-linear yieldin~~ of the asphalt close to the failure po;nt. Therefore, to determine the linear-elastic modulus usinU Equation ~u. the strain (resist anjle) was corrected by taking the tangent of the "S'~ curve. The eduation of thv tangent line was determined as shown in Fi~trre -II . From the resultin' .:quation, the intercept of the tangent line with the x-axis (twist an~lel was then determined and used to shift the points of the torque-twist an'le 'rraph to zero. It shoulc be noted that the tordue values were not adjusted - ie. the maximi_rm torque value observed durin', the test was used in Equation ~.

900.0 __ - _ 800.0 - -__-___- _ __.___. _ __ __.____ __._. ____- _ - __. --_ 700.0 - _____ ____ __ _.. _ _ ._ __.._.._ ____ _- __.._ _ ___ 600.0 Z 500.0 -__- _... _. ______ _ _ _ _ ___. ___-__ ___- __ .__.__ _.__ _ _ __- -.
.__ __ _ ._________.__ _ -_ - -.- ___- _ _. - - _ 400.0 >S
'0 300.0 _-. . __ _- _. ___________ _-__ .. ___ _ _-H
200.0 ___-_._ __.. .__ ___.__ -_ _-. _ 100.0 -___-- - _-. _. .. _____..
0.0 1- ~ ~--''~ _ _ 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.'18 Angular Displacement (rad) Figure 40: Typical Torque vs. Twist Angle Graph from InSiSST~''' aoo - -.-_---_--._. ~-_.__ - ----_ . --_--------- --w--__ __, 700 _~Seriesl 600 Linear (Series' ) _ _ E 600 i z 400 j -y = t 1329x - 647.96 ~ 300 200 _ t OC _ 0 0.02 O G4 G.06 0.08 0 1 0.12 0.14 Angular Displa~~ement (rad) Figure 41: Determining the Tangent of the 'Torque-Twist "S" Curve x.2.4 Asphalt Mlodulus vs. Torque Yer Unit 'Twist Using the technique pesented in the previous section, the modulus of the asphalt pavement at failure was calculated for the 3 InSiSSTr~~ tests as shown in Tahle I6. From the ;~ InSiSST''' tests producing failure of the ~isphalt ec>ncrete. the average modulu:, ~,~~~s calculated at ?08~'S kPa usin<? Ey.mtion S, with a ('O\' of 17.6~~~. Examination of the test results indicated that to is I1 and I? had nearly 1.~ 8 identical values for rr~odulus ( 18909 and 18~? 1 lcPa, respectively), while rest 1=s yielded a higher modulus of ~~O~G kPa. Unfortunately, due tc~ time and ~aeathnr constraints, additional testing with InSiSS I''~~' could not be completed prior to the completion of this thesis. However, an additional 8 tests had been completed previously with InSiSSTT~' on the same asphalt concrete. although each of those tests produced bond failure between the steel loading plate and the epoxy.
Whilc the modulus of the asphalt concrete at failure could not he calculated for these tests, there were enough data points poor to the bond failure tip calculate the linear slope of the torque-twist graph ( Fiytlre -f 1 ) for each test.
Table 16: Shear Modulus vs. Torque Per L-nit 'Twist ' i ~ Ultimate Torque :asphalt j ~tlear I'er Test Failure Unit r ~ ~lodulus ~
~~~~lL, 'Twist I 'surface I
I
No ( (kPa) I (N~'m/rad>
~
.
h ~

~ I1 18909 11100 I, ~
~_ ~

_ -- ,_.--- , ' --~

I? ' 18~'' I 13?9 I
As I
halt ' ~
i 1~ ' ~>o>f~ I I;laB
~~

-_~---___ ?U~~~ 11859 average _ ~.l,Dev 3666 1122 f C'OV 17.6 ~ 9.;
t ''~
) BFl I _ 10?'I
' _ BF? I 11961 ~

BF, _ I 1 1 ~
7' _ BFI I\;u>t i I ; ~
', I I
Epoxy-~

Load Plutc Applicable 6 BF> I l-118 I I I
lnterfare -j l I I ,1 I
B F6 _ ~_ ~, 13~3-~7~~ I I ~ 1 ' i'_' I ;8?, ~ . _.
_ _ _____.-____. :lverager 12180 ', II

S'TI)ev* 1104 CO~' ( ~'c 9.1 i' ' *includes te,t results fr~r I 1. I'_' and I

I:l~9 The resultin; mix property, referred to as "Torque Per Unit Twist (N~m/rad)'~
was calculated for bc_>th those tests vieldin~ failure in the asph~rlt concrete, as well as those producing bond failure. The results are also presented in Table 16. A
number of interesting observations were made based on this measure. First, the 'Torque per Unit Twist (TLTTj values for the tests yielding failure in the asphalt concrete v~~ere much more consistent than the ultimate shear modules :rom ~Jquation b as indicated by the low COV (9.~'if~ vs. 17.6'0). Furthermore, the Tll~h values were virtually identical for all 1 I tests, regardless of whether asphalt failure or bond failure v.vas observed.
5.2.~ Effect of Loading Plate Diameter InSiSST~r" testing was also completed wish 1''> ram (> in) plates to observe the effect of test plate diameter on calculated material properties. Detailed test results are attached in Appendix E for reference. Two r nterestin'~ results were observed with the I~> mm plates. The first concerned the shape of the torque vs.
twist an'11e ~iraph. .=~s shown in Pi'~ure -I_, the 1''~ mm plates displayed a rapid increase in torque to what appeared to he a yieldin'a point, and then a march slower increase in torque over a large increase in strain to an ultimate failure point. Upim further inspection of the data and Figure ~12, it was observed that tile linear slope of the torque-twist an<~le graph (Torque Per Unit Twist) was almost identical for both the 9? and I?5 mm plates as shown in ~I~able 17. The results of Table 17 implied that the. onset of failure was independent of plate diameter.
For the 9'' mm platen, complete f~ulure occurred at this point, whereas ultimate failure was not observed with the I?~ mm plates until a much larger strain was imposed. Tuo potential hypotheses were developed to explain this behaviour.

1 ~0 First, it was hypothesized that the initial yield point represented the failure of the asphalt-aQ~reQate interface ulon~ the failure plane, while the additional increase in torque required to mail the asphalt complete°.ly with the 1='~ mm plates was needed to overcome the ag~reaate interlock. It is not. known at this time why the same behaviour was not observed with the 92 mm test plates since the failure suuface imposed by both plates was the same shop=.~ (Figure ?~). However, the lamer failure surface of the 1?~ mm plate may have simply required additional torque to overcome the a~J~~rc<~ate interlock.

1~1 ~s ~s c~
~

;c 0 ~s o ~ , t"~ c:~

p t ,...
,.., c~
c~

a. ~~, a ... :.
a.. ~

'' ~ f \ ~ ~' v f ~

,: ~ N :W1 ~ ~' '~' ~Cn r C~ r~
~ f/
~ ' s~

r N C CO
u} 7 i n a a a, a~ a~ _ H- a~ ~ c~'~ ca ~-- f-E-~

K e~ ; x c i ~ ~ .
J 9 ._ p o 0 k ~6'p ~ i ~

cPt .

O v 9t 96, O

cr', ~

o <:i c9 -~' t t o _ '~

5'U

s~ ~

~o ~

O O o O O O O

o fl o '~ o '~ o O O O O o c~

N O CO Cfl V' (wN) anblol 1>?
Table 17: Comparison of Turque Per Unit Twist for 92n1m and l2~mm Plates _ - __ --_ Torque Per Cnit 'W vist ', ~i (N~'n~lrad >
__ _ -92 mm Test Plates ~ 12> mm Test Plates j --_ -_ __ 11100 1~1s - ,~ , 113? I1_»
I, ! 1318 Average ~ 11~~9 ' 11706 STDev I 1122 ___- _ n/a -COV ' 9.5 nla A second possible explanation of this behaviour could be the existence caf an "a'lin~' gradient" tllrou~Thout the depth of the asphalt pavement. As previously mentioned, asphalt concrete undergoes sti'fenin~ with time due to oxidation, rain and sunlight. However, it is unlikely that the stiffcnin' is consistent thrcvu<ahout the lover. It is more likely that the pavement surface is the mast stiff, and that decreasin« stiffne.5s is observed with imreasin'a deluh. Therefore, it is possible that the initial yield point observed with the I '~ mm plates represented the failure of the upper "crust", while the lar~_e amount of secondary strain was associated with the softer asphalt concrete underneath. 'This sorter i:rver wus likely not penetrated w-ith the ~~'_' mm plates.
.Althou~Th the TC?T values were independent of lead plate diameter. the resulting shear m~.>dulus values were not. Since v°irtualiy the same tordue was required to initially yield for Tail) the asphalt surface woth boll test plate dianneters, the shear rnodulus rcsultin<- from the 1 ~~ rt~m plates wns much lower than the ~W
mm test plates accr~rdin~T m Eduation 8 as shown in ~raUle 18. The aspl;alt moclulus l~
was calculated both at the initial yield point and at the ultimate. failure.
As shcwvn.
the average modulus at yield was 8610 kPa according to 'Equation S with a CO~V
of 14.8°lc. while the modulus at failure was ?'~88 kPa due tc> the large strain incurred.
The shear modulus at yield represented a 6~'~~ reduction compared to the O? mm test plate at the same temperature, reflectintT tf~e fact that the radius of the test plate is raised to the third power in Equation 8.
Table 18: Results of lnSiSST Testing with 12~ mm Plates Shear ~~~i Shear'I Shear Shear ~ Pavement I

Modulus ' Strain l~lodulus Strain ,l.en~P, Test i '~io . at Yield at 7'ieldI
at Failure at Failure I~ i ' I (kPa) (''~ ) (kPa) 1l %r) ( ' i __ _ .--().074 __ -_ 0.~'8 I7 I c~>>'?.-10-~ :'_>~)3.0~- 0 ~~
I - I=f ~p?
~

IS 77~8.~8 ().09~ . .
~
'SS
____ __.~',era~e BC~II -- O.Ot~ ~~88 0.' ~ --__ S~'Dev 127( 0.01 7.7 O.fll ~ ~ i ~__ CO~' 1.1.b 1 s.6 0.30 , _._-_-_~ __- _.____ _ _ _._ __ - __ -x.2.6 Discussion of Field Test Results and Analytical Modelling In theory. th~~ ~l~ordue Per Unit Twist (TUT) measure is a load her unit displacement, nut a stress h~cr unit strain. 'Therefore, wf~ile ii is not strictly a "stiffness" or rrtodulus. it i.<<; directly proportional to the muduius.
Indeed. for the ~l' mm test plates, the ultimate shear mudulus (in kPa1 was 1.7~ times the 1 UT
(in N'~m/rad). For the 1'~~mm Mates, the ultimate shear mc~dulua (in kPa) was 0.7~
times the TL 'T (in :sl*m/rad).

I ~-I
As previously mentioned. fundamental engineering properties should be unique to the individual material, not dependent upcm boundary conditions. or loading plate diameter in the cuwent case. The discrepancy between modulus values calculated for the 9'? and 12s mm test plates usiny~ Equation 8 appeared to indicate that further investi~lation and potential modification to Equation 8 is required prior tc»ts use to provide asphalt tnodulus. This is beyond the scope of the current investigation, however is under analysis at Carleton University at this time.
However, the TUT values appear independent of load plate diameter. Therefore, the TLT measure may be considered a fundamental en~~ineerin~ property of the asph~ilt miwand will be investigated further during future testing, with the InSiSST'"'.
~.?.7 Comparison of Field and Laboratory Results A final exercise of the verification stage was to cc>mpare the field stiffness calculated from the InSiSSTT"' results to the laboratory salues observed by Iahvr ( 199>). As shown in Table 16, the aver~y~~ shear modulus was ?08?8 hPa for the HL~s mix. Modulus values Irom the Zahw database arc attached in Appendix C.
The laboratory shear modulus values ranged from 930 kPa for a sand-asphalt mix to X700 kPa for an HL=f mix. The various Hl.s mixes yielded shear rnoduli ran~in<.~
bcmveen ?>00 and ~~00 kF'a. Therefore. tic field shear moduli were approximately C~ times ~areater then those ohserved in the lahoratory dunng the 199>
investi~ratim.
This result was rr~ost likel.~ due to a<,~in_~ of the asphalt mix. .As previously mentioned, the HL~ tested was construmed in I99?, th_;refore was 8 years old.
The labc»watorv testin;~ in 199 was completed on newly cc»npacted asphalt, therefore, it hud not been suhjected to environmental aonditionin~T and sut~sequent stiflenin« of the asphalt binder.

CHAPTER 6: C~onclnsions and Recommendations 6.1 Review of Project Objectives The phenomenon of permanent deformation orruttin~: in asphalt concrete pavements is extremely complex and has beer: the focus of ~_:oncentrated research efforts in the past decade. Although the term ~ rutting" is ofte-n used interchan;eabls with permanent defoonation, it is only one of four manifestations observed in North America. In general, rertti~a,~ is characterized by channelized depressions (troughs) that run longitudinally in the wheelpuths. However, rutting is usaally the most common formof permanent deionnation analyzed. The Strate~~ic Hi~~hwa~ Research Pro'Tram ~SIiRP) has identified that ruttin'; appears to he more closely related to sherir stresses and strains than normal or horizontal ones. Subsequently, it is important to mvesty~ate the shear properties of asphalt layers. As explained earlier, shear strength of an asphalt concrete mix is achieved through both a~~gre~aate particle contact to form a tight, load-hearing skeleton and the asploalt hinder that holds the particles in place. As :z result, the main ohjectives of this thesis ~,~ere as tollows:
1. To review the phenomenon of permanent deformation and identify its main causes, ?. To study the main factors contiibuting to the rutting phenomenon and determine the relatvonship bem-een these factors and the shear properties of the asphalt mix, and 3. To desi~~n and Build an advanced test facility to provide reliable data concernine the shear properties of the mix in the field.

1 i6 These objectives were achieved and discussed throughout the thesis. This Chapter presents the main conclusions and major findin's oV the research as related to the above mentioned objectives.
6.? Review of Permanent Deformation and Previous Investigations Permanent deformation or rutting of asphalt roads has been found to progress in three sta~7es. The first st~ye beg°.ins with continued densification of the asphalt lave:r under traffic loading. During this stage, ruttin~~ is directly proportional to traffic. The second stage involves a stable shear period drain'.: which the: rate of ruttin'a decreases with increasin; traffic until a third and final stage when a c~ mdition of plastic flay.
occurs and the rate of rutting main increases (vapid unstable shear failure).
It is the onset of rapid shear failure that is of particular interest to tl;e objectives of this investigation. In depth review of available information on the subject showed that the ruttin~l phenomenon is very complex. One of the main conclusions of this atudy v.cts that at present. there is no single independent variable that ;:aptures or predicts ru2tin~~
with a significant dct~ree of confidence. In addition, a single "deficiency' in a «iven property, such as excessive asphalt content, can nullify the over all duality obtained when other good properties. such as coarse a~_gre~~ate with lUOch tractured faced count.
are available. Another import,mt observation was the fact rutting is caused by various combinations of pavement layer instability and heavy tract,-tires. Until recently, laboratory and field ~nvesti'~ations of the rutting phenumelton did noU
address the fundamental property <rf a pa a-ement to resist ruttin'a: ,sJneup srrerc,yth.
SHRI' research has acknowledged the impui~tance of shear properties. and the now Supenpave design I~i method may soon incorporate shear propenies as important inputs toward the long-term performance of mixes in the field. However, it should be noted that Superpave shear tests are completed in the laboratory on laboratory prepared specimens or cores retrieved from constructed pavements.
These findinGs led to the consideration of data and test results reported by two previous studies on the subject completed at Carleton University. The first study was the comprehensive and intensive laboratory-testing proerar.~ carried out by Zahw to identity the influence of the rriain factors of an asphalt mix on rutting resistance. In this thesis, the author imported the data and test results reported by Zahw ~md a more rigorous analysis was perforir~ed. This step produced a set of new statistically based equations relating the most irnporiant factors affecting the rutting, phenomenon to the shear properties of the mix as shown in the following section.
The second study completed by .Abclel Naby resulted in the construction of a first Generation test device, known as the (~'arleton In-Situ Shear Stren~_th Tcst (CiSSSTj. The re°_sults of ,~bdel Maby provided uvo important conclusions. First, the CiSSST w~is able to differentiate between the shear properties of different mixes, as well as the cliffrrences within the same mix Enlaced in different '~eotnetries (curved sections vs. straiylht sections). Sc°cond, «reater variation between replicate specimen results was observed during laboratory testing than in-situ testin'.
6.3 Asphalt Min; Properties and Shear Characteristics Consideration of the extensive data collected by' Zahw ( 199>) showed that a number of mix characteristics contribute to the shear properties of the mix.
Two 1,s equations were developed to d~°scribe the shear modulus and shear strength. These equations incorporate traditional mix properties that should be recorded by any a~:encs.
However, many of these variables have been combined in such a way that all aspects of mix design (binder properties, gradation, aggregate angularity. density, compactive effort and volumetric properties] are included without the introduction of collinear variables. Equations 4 and ~ are shown below:
f4) -~6()-~.7'~f'~'I~+~~~vC(r-;-f-1='°(' -1-113>~':1IZD+1U67()"'t~'~11A
(R~ -- 0.83) ;5) --(~()-0.9~ P1'R-~U~ C't'+(o':C,,;, -~~I()*ARD+?~j,:>y,'hl.<1 (R~ = 0.88) where: G LkPa) = Shear Modulus (Stiffness) at ?~"C;
z tkPa) = Shear Stren<7th at ?~"C;
PVR (mm/Pa s) = Ratio of Penetration (mmJ to y'iscosity at ~~'C (Pa~ s);
CL = Coefficient of llniformitv (D(i0/D10);
C,,;" = Crushed Coarse A'T~~regate Present in Mix'' (Binary choice of 1 for 1''es l)r ~ f()r i\lo);
ARD = Average Rote o1 Densif~icati~:)n (ratio of- tinal mix density to the square root of the number of blov~°s with Marshal: hammer); and VMA = Voids in the Mineral Ag~Tregate ('%~) aecording to the models developed, tloe asphalt binder properties (PVR) represented ~'?~i~ of the shear modulus and =f-~~~ of the shear strength.
Ag«regat~

1~9 gradation (CL) accounted for ~9% of the shear modulus and s>°ic of the shear strength, while coarse aggregate angulantv (Cb;~) accounted for ~s~7c of the shear modulus ~~.nd 17~i'c of the shear strength. Volumetric propenies (:~RD and ~'MA) represented 8'i~
and 7 r'r of the shear modulus and 10'%'~ and ~°la of the shear strength, respectively.
These findings are in Ueneral a~'reement as research completed under the LTS-SHRP
( 1994).
6.4 Asphalt Shear Characteristics and Rutting Mix shear stren~Tt1 and modulus were hi~~hly correlated to rutuny at vanous stress levels as predicted by laboratory creep testing (Shell Metho d), althou«h mix stiffness (modulusl appeared to he a sliUhtl~ better indicator of mix performance.
Rutting models based on shear modulus and ~,tren~lth v~ere also developed ~m shown below.
Creep ~ Laboratory T Rutting Models -Stress ---- _- T _ Shear l~IcrdulusO' ~sC ' Shear Strength ta~
?sC

Level j r VlPa ~ (kl'al ~ (kPa) l _- ~.1 Rut = ?-~j=i=G-u.~a~( R~=(l.t;()) = I ~.'? ~'e-o ; n I Rut ( R-=0.O ; ) Rut = ~~~II''~i-1.U37(IZ~=~I.()~)) = ~~1,L~~. ~-0.66 (R~-().~~') RLIt ~, ~ Rut = 3Ii~6 Rut = 11791 ~' z-''s~ (R':=().67) 0 ~G~'.r'Ju (R'=0.73) I

. i 6.~ Modelling In-Situ Shear Properties A review of previous modellin~~ eflort:> as well the de°velopo~cnt of nm constitutive equations provided the follown<_ conclusions:

i The boundary conditions observed during field testing with the CiSSST
device and tine InSiSST'~'vere significantly different than those experienced in the laboratory torsion testing. During the development of the CiSSST
device by Abdel Naby, Equation 6 was developed to calculate shear strength of the mix from field torsion testing. While acceptable for that investigaticm, it has been subsequently determined that Equation 6 does not likely best characterize mix shear stren~~th. Therefore, a new analysis procedure based on the Reissner-Sagoci problem was developed by Bekheet et. al. 0'000) and adopted for this investigation. This model (Equati~an 8j directly provides shear modules haled on applied torque, and was subsequently verified for the linear elastic condition both with finit.~ element modellin<~ and field testing at room temperature.
iil Additional verification and/or modification to Equation S rrrav be completed in future investigations for cases of nom-linear behaviour such as those encountered at lover test speeds and 1-~i'her pavement temperatures.
7~he conclusions and findtn~rs obtained after achieving the first two objectives paved the way to the development of the new lnSiSS~h''~' as discussed below 6.6 Design, Development and Verification of the InSiSST~'~"r The primary objective of this investiaaticvn was the de;~elopme-nt of an advanced in-situ shear stiffness teat (InSiSST'"j for asphalt concrete pavements. The :,uccesstul completion of this objective provided the following c~mclusoms:

i j The InSiSSTT~' device was designed and developed based on a sound theory and thorou;h analysis of deficiencies observed with the CiSSST device and additional desi._~n considerations. The result is a test facility that is portable, stable, and rua;~ed. The test reduires oily a sin~7le operator, no heavy Iiftin'= or complex set-up, and can be completed rapidly. The test results are accurate and are available instantly. .~11 of the individual systems operate under the same power reduirements as provided by the central Generator. To date. the newly built InSiSS'1' fvas been towed a total of 4500 Km at speeds as high as l~'Okm/h.
Clearly, this is a testimony to the reliability and tou'ulness of the new facility.
ii ) Field verification of the InSiSSTT'' has indicated that the device accurately and repeatedly measures shear properties of the asphalt mix. Additional verification testing is reduircd, however, to further explore the efrect of test plate diameter on measured shear properties.
G.7 Recomoendations for Future Modifications to INSISST1~~' G.7.1 h:nvironmental Chamber Asphalt concrete shear strcn~Tth and stiflne5s are hi'nly dependent un temperature due to its viscoelastic nature. In its cun-ent term, tl~c InSiSST'~r can test only at the pavement temperature present in the field. Therefore, the r.sults obtained in the field must be normalized to a standard or reference temperature before comparison with other sections may be completed. A temperature mastc:;
curve may be used t -~ do this, however, the master curve nnust first be developed.
The addition of an environmental chamber t;~ the InSiSS~h'-"' test frame will alloy 16?
the development of temperature master curves, as well us reduce the cure time for the epoxy by allowing the introduction of heat prior to tl~e test.
6.7.2 Hydraulics For this concept exploration project, the use of hydraulics or pneumatics instead of the electromechanical system fo:r applying to torque was prohibitive:v expensive. However, if the lnSiSSTTM is to be developed further by a manufactur~in~ company for mainstream use, it is recommended that hydraulics or pneumatics be used fur three primary reasons. First, all mechanical systems including the jacking svstenn. positioning system and load application system may be driven from a single pump, thus reducing- the. number of components (the InSiSST'-"' uses 4 individual electric motors). Second, hydraulic systems may ~c very quickly and ac:curatelv reversed. thus allowing dynamic fc~ading of the pamment in addition to the cuwent static testing. Finally, hvdr,.mlic systems are generally more rugged and resistant to the elements than electromechanical systems.
V hen assembled in larger quantities, hydraulic systems would likely hecc~me much more cost effective than for a sin'lle prototype.
1.7.3 Shear Vane Altllou~Th the InSiSS~I-'" desi'Tn ~greatlv r~:duced the time required for testin<u over the CiSSST, the epoxy cure time remained as the overall governing factor.
7,he use of epoxies with more rabid cure times improved the test time si~;nificantlv, hovyever, an improved affixation method is recommended to remove the dependency on epoxy altogether. One such method would be the use of a shear vane, similar to those used in soil mechanics to test the shear properties o1~
clays in-situ. For newly constructed pavements, the vane could he placed on the asphalt 16~
surface behind the spreader and compacted into the pavement, thus allowing for both current and future testing. For existing pavements, a saw could be used to cut the pavement surface and tlue vane could be inserted for testin~~. Standard formulas exist for vane testing, homver, additional analysis would need to be completed t<, ensure the accuracy of the tf°_st. Comparison with laboratory values could assist that eff ort.
6.8 Recommendations f'or Further 'I'tysting 6.8.1 Additional ~ eritication 1'estin~
The initial verification testing completed during this investigation was sufficient to provide ~~eneral trends and relationships to Ensure that the InSiSS~,' ~' device was measuring the desired shear properties accurately and repeatedly.
I-fowever, significant additional testis; will be required to further observe the effects of temperature, directly compare to laboratory torsion tests and explore the phenomenon observed with the I'_'s mm test plates.
6.8.2 Long 'Term Performance One of the primary goals of in-situ ;hear testin~~ will be to ultimately model and predict rutting in the field. Extensive field testis' at numerous test cites will Lie required to provide the necessary data to der:elop such models clue to the iar~~e number of variables that corntribute to rutting (as presented in Chapter ?).
.At the time of this waiting. Carleton I'niversitv has partnered with the Ontario Ministry of Transportation (MTOj to complete field and laboratory shear testing on approximately ~ of the 16 test sections.

6.8.3 Additional Testing for QC/QA Specification Development As mentioned previously, there is a sTreat need for a simple penormance test for the Superpave asphalt naix design system, largely for use as a qu~tlitv control and duality assurance (QC/QA) test. In-situ shear testing with the InSiSST'" could potentially provide such testing as the device is well suited for rapid and accurate testing in the field us opposed to in the laboratory.
Clearly, the continuous tasting and improvements of this new testing lacilitv should ultimately provide the I>avement industry with a reliable tool that is desi~~necl:
i 1 ) To enhance new construction methods thrnu~h adeduate duality control, and (?) To improve the ability to predict long term performance of' asphalt surfaces and pavements.

References -~:~bd El Halirn, A. O., Z.ahw, M.A., and Bahgat, A.Cp. 199. Influences of Types of Asphalt Binders on the Tensile and Shear Strengths of Asphalt Mixes. Proceedings of the Canadian Technical asphalt .Association, 40: ?91-3U~1.
'~Abdel i'~abi, R. M. 1991. Influem:e of Highway Cleometrv on Distress of Asphalt Concrete Pavement.5. A Ph.D. thesis. Zagzig University, Cairo.
alavi, S. H. and Monismith, C. L. 199-I. Time and Temperature Dependent Properties ~rf Asphalt Concrete Mixes Tested as Hollow Cylinders and Subjected to Dynamic Axial and Shear Loads. Journal of the Association of Asphalt Paving Technologists, 6s: 1 ~?-I 81.
American Association of State Highway and Transportation Officials (AASHTOj.
1999.
Superpave Implementation Survey III 1998-1999. AASEITO Task Force on SHRP
Implementation and FI-1WA I_eud State Program, Washington, DC.
American Association of State Highway and Transportation Officials (AASHTO).
1993.
AASHTO Guide for Desiy~n of Pavement Structures. Wa~;hington, DC.
.-~nani, B. A., Balghunaim. F. r~., arnd AI-Hazraa. a. S. 1990, laboratory and Field Study of Pavement Rutting in Saudi Arabia. Transportation Research Record, 1?~9:
79-90.
Asphalt Institute. 1998. C'onstrurtion of Not Mix -l5phalt Pavements, Second Edition.
vlanual Series :~o. ~~ (MS-~' ~. Lexington F~~entuckv.
Asphalt Institute. 1997x. Performance Graded Asphalt Binder Specification and Testin~l.
Superpave Series No I (SP-I ). I_e,rinaton. Kentucky-* _ Invention references .Asphalt Institute. 1997b. Superpave Mix Desi~?n. Superpavc Series No. ~ (SP-?j.
Lexington, Kentucky.
.Asphalt Institute. 1991. Thickness Design - ,asphalt Pavements for Highways and Streets. Manual Series No. 1 (MS-1). Lexin~=ton, Kentucky.
.Aschenbrener. T. 1995. Evaluation of Hamburg wheel Tracking Device to Predict _'vloisture Damn'=e in Hot-Mix .Asphalt. Transportation Research Record, 149:':
19s-?Ol.
:~schenhrener, T. 199. Comparison of Results Obtained from the l.CPC Rutting Tester with Pavements of known Eield Performance. Transportation Research Record.
14 5.~: 66-73.
Aschenhrener, T., Terrel. R., and Zamora, R. 19!4. Compari;;on of Hambur~~ w' heel ~hrackina Device and the Environmental C~,mditionin~r S~,~stem to Pavements of Known Strippin~~ Performance. CDOT-D T D-R-94- I . Colorado Department of Transportation. Denver Colorado.
Beer. F.P. and Johnson Jr., E.R. t 988. Vector I~9echanics for Engineers -Statics.
Second Metric Edition. McGraw-Hill Book Company, Singapore.
Bekheet. ~'.. Hassan. O. and Ahd El Halim, A.O. ~0(l0. .~ssessin~l the Effect c>f Elastic Properties of Asphalt Concrete Pavements on Shear Stren~~th. Proceeding's o1 the Canadian Society for Civil En'~ineerin~l s'"'TrunsportuticEn Specialty Conference (CD-ROM), pp. ~~c~_~;~.
Bonse. R. P. H. and huhn, S. H. 1950. Dynamic Forces Exerted by Moving Vehicles on a Road Surface. Bulletin of the Highway R_esearct~ Board. ?s ~: 9-3?.

Brown, E. R., and Bassett, C.E. 1990. Effects of Maximum Aggregate Size on Rutting Potential and Othe,~ Properties of Asphalt-Aa~re~ate Mi:;.tures.
Transportation Research Record, 1 ? ~ 9: 107-1 19.
Brown. E.R. and Cross, S.A. 1998. A Study of 1n Place Rutting of :Asphalt Pavements.
Journal of the Association of Asphalt Pavin~l Technologists, >8: 1-39.
Bruce. B. 1987. Asphalt as a Variable - Big Tinnber Test Sections. .loumal of the Association of ,Asphalt Paving Technologists, 56: 688-710.
Button, J.W., Perdomo, D., and Ll~~tton, R.L. 1990. Influence ~>f A~Ure~~ate on Ruttin~T in Asphalt Concrete Pavements. Transportaticm Research Record. 1~~9: l~ll-l~?.
Canadian Meteorolo~~ical Centre. X000. Data from CMC web~ite at www.cmc.doe.ca Canadian Strategic Highway Research Program (~'-SHRPj. 1999. Superpave ?000:
Improved Standards for a'slew~ Millenium. C-SI-IRP Technical Br-ief'_Vo. 17.
Otta~ u.
Canadian Strategic Highway Research Program (~:'-SHRP). 1998. C-SHRP Phase IIIf Pro~;rom Proposal 1999-?001. Ottawa.
Canadian Strategic Highway Research Program (n.'-SHRP). 19~~7. C-LTPP Database User's Guide: Se~cc>nd Edition. Ottawa.
Carpenter, S. H. 1993. P~.rtnanent Deformation: Meld Evaluation.
Transportation Research Record 1=117: 13~-1=~3.
Carpenter, S.H. and Enockson. L. 1987. Field :analysis of Rutting in Overlays of Concrete Interstate Pavements in Illinois. Transportation Research Record, 1 1?6:
=16-~6.
C_'clard. B. 1977. ESSO Road Design 7-echnolo~av. Proceedin;~s of the Fourth International Conference on thi: Structural Design of Asphalt Pavements. Vol. '?, pp. 76-77.

Coree, B.J. 1999. ?-IMA Volumetrics Revisited - A New Paradi'um.
Transportation Research Board ('TRB) 78'h Annual Meeting Preprint CD-ROM. National Research Council, Washington DC.
Energy, Mines and Resources Canada. 1974. The National Atlas of Canada, Fourth Edition. Macmillan Company of Canada 1_,td. Toronto.
Cross. S. A. and Brown, E. R. 1992. A National Study of Rutting in Hot Min Asphalt (HMA) Pavements. Journal of the Associ;jtion of Asph~~lt Paving Technologists, 61: J_sj-j8?.
Cross, S. .A. and Brown, E.R. 1991. .A National Study of Ruttinyl in Asphalt Pavements.
NCAT Draft Report No. 91-8, National Centre for Asphalt ~hechnolow. .auburn University. .Alabama.
Farz-is, R.E. 199?. The Overall Challenge: A New Look." Prc>ceedin'ls of the 199?
l-ransportation As:>ociation of Canada (rl-.ACI ,annual Cenl~erence, Quebec Cit~,~.
Volume 6. pp. A? 1-A~(l.
F~:deral Highway Administration (FHV'A). 199'la. LTPP: 1995 Fear in Review.
Report FHVVA-RD-99-1 1 1. United States Depamnent of Transportation, Washington DC.
Federal Highway Administration (FHWA). I9~)r~b. DataPave ?.U CD-ROW. Kong Term Pavement 1'er-form,mce Pr~r_rum. United States Departs-gent of Transportation.
Washington DC'.
Federal Hi';hway ,Admimsiration OFLIWAj. 199Sa. L.TPP: The Neat Decade. Report FLIw"A-RD-98-IU'O. L)nited States Department of ~hranspot~tatiun.
V~'ashington. DC.
Federal Highway .Administration. 199Sb. Simul;_ttiun, lo~a'~in;~, and Mec hanks of Asphalt Pavements iSI.MAf'). Report FHV\'A-RD-98-1~_'8. United State Department of Transportation, Washington. DC.

Federal Highway Administration. 1998c. Performance of Coarse-Graded Mixes at w'estrack - Premature Rutting Final Report. United States Department of Transportation, Washing=ton DC.
Federal High~~ay Administration l~)98d. A Simple Test for Super-pave Mix Performance. Focus. December 1998 Issue. United States Department of Transportation, Washington DC.
Gcnrard, C. M. and Harnison, W. J. 1970. Circular Loads Applied to a Cross-Anistropic Haff Space. Division of .Applied Geomechanics Technical Paper No. 8.
Commonweath Scientific and Industrial Research Or~ar~ization. Australia.
Gervais, F. and Abd EL Halim, A. O. 1990. Ruttin; of Asphalt Overlays:
Analytical and Field Studies. Proceedings of the Canadia:i Technical Asphalt .Association, Vol. s~, pp. ~-J-4~, Winnipeg. Manitoba.
Goodman, S. and Abd E1 1-Ialim. .<~Ø 1997. Design. Development rind Vcril-i;ation of an Advanced In-Situ Shear Strength Test t~:~r :asphalt Concrete Pavements.
Proposal submitted to the L:nitcd States ~l~r,ansportutiun Research Board f pRB) Innovations Deserving Exploratory Analysis (IDEA) Program. C'arleton (_tniversitv, Ottawa.
Harvey. J., Ericksen, K., Sousa, J. J. and Monisnoith, C.I~. I99~. Effects of Laboratory Specimen Preparation on Aagre~~ate-Asph~~lt Structure, .\ir ~~'oid Content Measurement, and Repetitive Simple Shear Test Results Transportation Research Record, 14~=I: I13--1??.
Harvey, J. and Monismith C. L. 199,. Effects of Laboratory :asphalt Concrete Speciman Preparation Variables on Fatigue ~rnd Permanent Deformation Test Results usin<_ Strategic Highway Research Program A-0(13A Proposed Testing Equipment.
Transportation Research Record, 1417: 38~-48.
Hicks, Tyler G. 1994. Standard I--landbool; of Engineering Ca:culations. Third Edition.
McGraw-Hill Publishing Company.
Hofstra, A. and Kolomp. 197?. Permanent Deformation in Flexible Pavements under Simulated Road Traffic Condition. 3"~ Lnternational Conference on Structural Design of Asphalt Pavement, 1: 61s-671.
Huan';. ~~an~ H. 199>. Pavement .~nalvsis and Desi~_n. Prentice Hall, l~ew Jersey.
Huber, G. 1999. Methods to Achieve Rut-Resistant Durable Pavements. NCHRP
Synthesis ?74. Transportation Research Board, National Research Council.
Washington. DC.
Huber, G. A. and Hieman, G. H. 1987. Effect c;t Asphalt Concrete Parameters on Rutting Performance: A Field Investigation. Joun~al of the Association of Asphalt Pav:inv~
Technologists, sC: ~~-Ol.
Jennin~~s, P. «'., Pribanic, J. .A.. Smith. J. and Mendes, T. M. 1988.
Predicting the Peuformanee of Montana chest Sections by Physical and Chemical Testing.
Transportation Research Record, 1 171: ~~)-6>.
handhal. P. S., Cross, S. :1.. and Brown, E. R. 1993. 1-Ieavv Dutv ,asphalt Pavements in Pennsylvania: :fin Evaluation for Rutting. MCAT Report 93-'?. !rational Centre for .Asphalt Technology. .~uhurn lJniversitv, .~~labama.
Kawesf<i. D. and Nickeson, M. 1997. C-SHRP Bayesian Modelling: A Users Guide.
Canadian Strategic Ifi'_~hwav Research Program (C-SHRP~. Ottawa.

Krutz, ;~.C. and Sebaaly, P.E. 1993. The Effects of A~areoate Gradation on Permanent Deformation of Asphalt Concrete. Journal of the Association of Asphalt Pavin'_ Technolo;ists, f'_': 4s0-X73.
Krutz. N.C. and Stroup-c~ardiner, M. 1990. Relationship Bemeen Per-rrranent Deformation of Asphalt Concrete and Moisture Sensitivrtv. Transportation Research Record. ?~9: I69-1 77.
Lai, J. S. 1986. Development of a Simplified Test Method to Predict Ruttin C'haructeristics of Ashh~ilt Mix. Geor'ia DOT Project ~>03, Final Repur~.
Ge~.~ryiu Department of Transportation, GA.
Lai. J.S. I993. Results of Round-Robin Test Prouram to Evaluate Rutting of ,Asphalt Mixes Using Loaded V'heel 'Tester. Transportation Research Record. 1~ ( 7:
1='.7-1 -I.
M~rrks. ~'.J., Monroe. R.W'., and .~,dum. J.F. l~)~O). Effects of Crushed Particles in asphalt Mixtures. Transportation Research Record. 1?~~); 91-106.
Warwick, H. D. and Starhs. H. J. H. 194,1. Stresses Between l ire and Road.
Journal of the Institution of Civil En~mrmers, 16.
McGee, Ken. 1999. Summary of the Proposed ?()()' I'avcment Desi~~n Guide:
~CILRI' Project 1-37A. AASLJTO Draft Paper. An-ierican Association c>f State I-ii~lhwav and Transportation Officials. ~Vahin~ton, IoC.
Miller, L. Freund. J.E.. and Johnson, R.A. 1990. Probability and Statistics for Ian<Tineers.
Fourth Edition. Prcnticc Hall I'ublishin'. ''~~ew Jcrscv.
Ministry oFff~ransporrtation. Ontario (M~TO). ?0C)(). Quick dotes on...Pullin~~ a Tr~rilcr Safely. Available in Adobe Portable Docur:ient Format ~ PDF) throu~~h tic MT(O
wcbshe at wv~ .mt~~.fov.on.ca. Toronto.

1~~
Ministry of Transportation, Ontario. 1998. PGAC Grades Redefine Asphalt T'echnolc>'=v.
Road Talk Maaa~.ine, Vol. -I, Issue ?. Toronto.
National Cooperative Highway Research Program. 1998. Quality Control and Acceptance of Superpave-Designed Hot r~Tix Asphalt. NCHRP Report 409.
Transportation Research Board, National Research Council, Washington DC.
Nievelt. G. and Thamfld, H. 1988. Evaluation of the Resistance to Deformatron of Different Road Structures and Asphalt Mixtures Determined in the Pavement Rutting Tester. Journal of the association of Asphalt Paving Technologists, >7:
~?0-;-f~.
Parker, F. and Brown, E.R. 1990. A Studv of R'.utting of Alal-,ama Asphalt Pavements.
Final Report Project ST ?019-9, Alabama Hi~Thway Deloartment, GA.
Peck and Lowc. 1960. Moderator's Report on oe Shear Strength of Undisturbed Cohesive Soils. Proceedings of the 1960 ,'1SCE Research Conference on Shear Strength of Cohesive Soils, Colorado.
Prowell, B.D. and Schreck, R.J. ?000. Virginia's Use of Lahoratorv Wheel-Tracking as a Mix Performance Predictor. Trunspor-tation Research Board (TRB j Annual Me~tin<r Preprint CD-ROM, National Research Council, ~'ashin~~ton I)C.
Rickards. 1. 1998. AMIR - HIl'.-~C' Trials in Australia. Pioneer Rood Services Ptv. Ltd..
Melbourne, Australia.
Reissner, E. and Sagoci, H. 194=1. Forced Torsi.>nal Oscillaticvn of an Elastic Half-Space.
Journal of Applied Physics. 1i: 6>~-6>~.
Roads and Transportation Associ~~tion of Canada (RTAC ). 1980. Canadian Strategic Highway Research Progran n (C-SHRPj. Council on Hi«hw ay and Transportation Research and Development, Ottawa.

Romero. P., and Stuari. h. 1998. Evaluating Accelerated Rut Testers. Public Roads July/.August. Federal Highway Administration. V~'ashin;~~ton, DC.
Sherline Products Incorporated. l 999. Trailer Loading and Tcvving Guide.
Califomnia, USA. .Available online at vwvw.sherline.com.
Simpson, A. L., Daleiden. J. F. and Hadley, W. O. 199. Rutting Analysis frc>m a Different Perspective. Transportation Research Record 1473: 9-16.
Sneddon, N. 1946. Note on a Boundary Value Problem by Rcissner and Sa~oci.
Journal of Applied Physics, 18: 1 s0-I s?.
Sousa. J.J., Harvey, 1_.~ Deacon, .A.. and Monismith, C.L. 1991. Evaluation of Laboratory Procedures for C',ornpactin<~ .Asphalt-Aa'are~~ate Mixtures. Technical Memorandum TM-UCB-A-00_~,A--90->. Institute of Transportation Studies, University of California at Berl:elev.
Ten-el. R. L. and Epps..LA. 195. Using Additi~.es ~rnd Modifiers in Hot-I\~lix Asph;rlt.
Quality Improvement SerleS 1 14. National .Asphalt Pavement Association (N.a.PA ).
Lanham, Maryland Transportation Association of C: anada (TAC). 1907. Pavement Desran and Mana~~em~n~
Guide. Ottawa.
fhransport Canada. 1999. Transportation in Canada: 1999. 'hr.rnsport Canada Repc»-t '1I' 1319SE. Otta~ a.
~Cransportation Research Board (TRBI. ?000. Innovations Deservin~~ Exploratory Analysis (IDEA): Program announcement ~'.000/?0(ll. National Research Council, Vv'ashin~ton, DC.

United States Strategic Highway Research Program (IJS-SHRP). 1994. Permanent Deformation Response of Asphalt-A~~Tre~~ate Mixes. SHRP-A--II >. National Research Couneil, Washington, DC.
United States Strategic Highway Research Program (US-SHRPj. 199?. SHRP Product Catalo~~. National Research Council, Wtrahington, DC.
''an de Lou, P. J. I97~1. Creep Testing: A Simple Tool to Jud~7e Asphalt Mix Stability.
Journal of the Association of Asphalt Paving Technoloai~ts, 43: ?53-28=I.
Witczak. M. ?000. Personal conversation with Dr. Matt Witczak during the 1'' International World of Asphalt Pavements Conference in Sydney, Australia held February ?0-?4, ?000. Dr. Witczak is the pz~imary investigator for the flexible (asphalt] sections of the ?00~' design guide.
Wu. ~'.. Parker, F., and Kandhal, K. 1998. Aggrc.vate Tou~hness/Abrasion Resistan<:e and Durability/Soundness Tests Related to Asphalt Concrete Performance in Pavements. NCAT Report 98-4, National Centre for Asphalt Technology, Auburn Lniversity, Alabama.
*Zahw, Magdy A. 199. Development of Testin~~ Framework for Evaluation of Ruttin;
Resistance of Asphalt Mixes. A Ph.D. thesis. Carleton L!niversitv, Ottawa.
* = Invention reference C
..,. ~ . ~ :r ~-, -~ ~ ~~ ~'~
r - .~ ; .,.. I ,... _.

r C i r C
.

i y w I
~ 'J
~ ~

T ~ '~ ~ I~ '~.V', I

, N '~ ...f~ ~',0G
"' ' ~ ', ~

... M, M C N ' (~I
,: ~. ; i i r/~ ~ ~ ~
T ,. .. .!

;r j C

..

:,.

j ~: c ~---, ~ -, ~ -., ~r, ~o ~r m --~ -1 ~ ' > .: ~ j ~ _.
. c _. ... '. , .., I
i ;

_ I
.

f U

I

L ' " r" \: T l~ I~ ~ f" cr, .
~U

. , r7 T WO v: ~ f~ oG
. , ~ I
= -~

. - ~r, ~r, ~~ ~ ~ r ~r,U
: . ~ , .. I
I

-- ' ~.. '. ~ . ~ ~' I
, . ..., , j , I
~

I
I N Y I

~ ~, ,;~ c~r-~ ('fJO
. ~' / ;
- %

',, :~ i~ '~' .~ C',1"~J 00 ~ C~ G1 ,~~ ' ~~

~ d -,.. !

:J ,'/~~'. ~. ' ~ '~ ~ ~ '.
C - . J , i i i ~ '!
.r J

r~

V, U
:

U U

'~' ~ .

L C

_ W

~ L

j r~ t! " C r .
' _ _ ~ , .

:,~J v,I
J L ~ ..L r _ _ _ _ .~i .r ~ r ;/'.U :/::/: G r G

17~
Appendix Bl: Selected Variables for Mixes 1 throubh Selected 'v'ariables ~'lix _ =~w~
Penetration- Rate of Uesi'~n Viscosity CeIfICIent Crushed .
I~ensification of Luifonn7itvC_'oarse ~'W.A
DcsiynationRatio at J (1)cnsity/sdrt(#
?~C

tDE'()/~101 I~laterial (mm/P a hlLVS)) *s) --m1-1 56.~>0 1~.8; 'i 1 . 0.''~1 I O.Ia-(a m1-? ~O.?>0 ~ 1i.83 ~ 1 I 0.'_'>? ~ U.I(~0 m l-; >6.?~0 j 1 ~.8s 1 ~ 0.~_'~s 0.1 ml-=1 ~ >6.''>0 ~~' 1~.8~; 0.?~~ o.l~',0 ' I

IT71-1 I lG.~>o ~I 11,5 i ).W 1 i - -_ I 1 ( --_.
_-~ ~ . -__ ! n7?-1 >6.'_'s0 19 1 ~ 0.?6~ ! o. l I, ~ I

m'?-_' ' >G.?7O I9 I ! o.?6s ().
I (~
I ' 11T~'-~ >6.'_'~O 1() I ! 7 i o. 65 ~ 0.1!1 I

m?-4 ~ >6.'?>0 19 1 ! 0.?(6 ~ o, l ~ ~ 1 In'_'-~ 5(.?i0 19 0.?f~4 l, 0.1~~

!

ms-I _ - _ _ ~6.~~( -- 0.?> 1 0.145 ~
) I .S ~ 0 m;-~ i6.?i0 I?.SS I 0 0.''~~ 0.1~~
~ I

, ms-~ ~6.'_'>0 1_~.SS I 0 0._'~i 0.1~~8 I

m~--l ~6.?>0 1_s.S; (? 0.?i:~ 0.175 1776-~ ~6.'_'~0 1: 0.''s > 0.1 b .SS 8 () n7=1-1 _ _ o.''i1 _ , 10.714 ', _ 0.I~19 1:>.S; ', 1 m-1-~ ' 10.7I~1 I 1:>.fis I o.~'~~ 0.160 m-1- ~ I 10.714 I _~.SS 1 ' 0.~~4 I 0.17() m-1-~l , 10.714 ~, 1:~.~s 1 ' 0.~>6 ' ().1 SO

m-~-; 10.714 I 1:~.~~ I l ' 0.~'~4 o.lS

n7~-1 j 1 U.7I4 19 -- 1 _ ~ ~>7 _ I 0. I
~ I

m~-~' I 10.714 ~~, l9 ~I I 0.~'>~) ~' ().101 m ~-s 10.714 ' I ~~ I 1 o..'ti ~ ( ). l 17TH-4 10.714 1~~ 1 o.~os ' o.lsl m~-~ 10.714 i I 9 i 1 0.?(77 0.1 ~) Appendix B2: Selected Variables for Mires 6 through 1?
Selected Variables Min Penetration-. .A~,~~,.
Rrite of , Crushed ~.oeff Desi~_rn Viscosity . Dcnsification DesignationRatio at of L.Iniful-mrty Coarse (Dc.nsitvlsdrt(~
'?>(~

~D60/D10) Material (mm/Pa*s) blow)) -- -m6-1 10.71=1 1>.83 ~ 0 j 0.'_'S~' 0.148 m6-~ 10.71-1 1.83 I~ 0 ', 0.'_'i 3 0.18 ~

m(-,- 3 10.71 1 J.S 3 0 ~, 0.''J 3 I 0.168 ~ ';

rnC-~ 10.714 1 x.83 0 0.?49 j 0.178 ~, m6-~ 10.714 1 x.53 () ' 0.?47 ~ 0.158 _ _ _ -_ m ~-1 356.36=1 15.53 ii 1 0.X50 0.1-I~~a i m7-'? ' 386.36-~ 13.53 ' I ().?~1 0.1c,(?

nl7-3 356.364 1;.53 1 u.~>~ 0.170 m7-4 386. 364 I >.83 1 ().~'~4 'i 0.1 SO
' m7-~ ' 386.364__ 1 ~.8 3 1 ! (>.~> 3 ~ U.15~) m5 386.364 1 c~ ~ - _ --().?6~' _0.1 ~
I = 1 --~. ~ I .

m5-~ 386.364 19 ~ 1 ' 0.?6-~ ', 0.161 Ills- 3 356.364 I ~) ~ 1 0.?6~ '~, 0. I /
i I

m5--1 386.364 ~, 19 ~~ I U.?U6 ' 0.181 m8-~ 386.364___ 19 1 ().?6~ 0.191 ' __ ' m9-1 ! 386.364 j ~.8 ~ () -- --. ()..,49 0.148 ~

nl9-'? 386.364 ' 1.53 0 (1.261 0.t>8 ~~

m9-3 386.364 1 >.83 0 0.?~3 0.168 m9-4 ~, 356.364 I 1 >.5 3 0 I ().?l? ~ ().178 I i :11~)-~ 356. 364 1 ~.5 3 0 i ().~'W ().155 ---_ _ IllIO-1 10.71=1 ,,, ___~ ().?7~ ().I43 ~ II
!

Ill l ()-'_'10.714 I 77 ~ 1 i 0,?76 I 0. I ~
' 3 i m10-3 10.714 ~ ''~ ~ 1 0.?7~ ' 0.164 11110-4 10.714 '_'~ ~ I I 0 0 ' ?73 ' 17=1 .

- . .
~ -- _ n111-1 10.714 11.83 1 0.30? U.1=1o ' ; I

~:~~L-_ ~~.~~-+ ~.~.o, , u.,no o.ICm nlll-3 10.714 ~ U.b3 ~~ 1 I 0.307 U.170 m 1 1-4 10.714 ' L ~.8 3 1 I 0.307 0.180 m l l -s 10.714 1 x.53 I I ~~ 0.307 0.189 ~ _-ml?-1 0 -().~lo I ~).1'4~
10.714 h ' m 1?-' 10.714 I l () i 0.~'?0 ' ~:).
I 3=1 nl 1 ~- 10.714 1 1 0 I U.? I 9 ~:). I

Ill 1 '--110.714 1 1 i 0 (1. ~' 16 ' i ).
I ~-~

Appendix C1: Shear and Rutting Properties for Mixes 1 through 5 Shear Propertiesf Mix Rutti ng Propertiesf Mix o o ~

_ Mean Rut Mean Rut Mean Rut D Av. ShearAv'~. Ay. Shear siU Shear e Depth Depth Depth n strength Stalin Modulus Designation (().1 (0. (0 MPa) ~ MPa) 6 MP;z) (kPa) at Peak (kPa) _ .

__ __ _tmm> (mm? (n1m) _ _ ml-1 X81 0.?04 ?846 0.6~? 1.>7 1.479 ~~ ~

ml-~ 61? . O.? 1 ?876 0.(>46 I.?38 ~ 1. 356 I ~ ; i ml-3 641 ~ O.:s=11 ''763 0.6?~l ().974 1.907 I
~ ~ ~ i ml--1 X41 0.~ ~'70i 0.699 1.5?7 ~ ?.19 I
' m 1-> 5? 1 ~ ci.l9 _ ''69~ _0.709 1.7~'_' x.107 4 ( _ _ 117-1 691 1.).~1 ;356 ~.6~ _ 1.493 I ' 0.917 T

m~'-~' 706 ci.~0-1 ;50i 0.609 0.908 I 1.s7~
I I

m~'-; 719 ~ (!.?03 ;59~ (?.~74 0.8~~ 1.18 i rrl~'--171'_' C~.''01 3>56 0.74 ~ ().93-~ 1.?81 i i ~~, n~~-; 711 i o.~06 ;.~6; 0.75 I 1.o1 1.5 ~ __ ~

_ ' 1113-14-1-1 C'.197 ?31 ~ 1.113 ~ ?.696 ~.~'87 I ' I i -I

lTl3-? 463 b~.186 ~-193 0.887 ''.518 ?.:~61 ~ ' I ~

r1~ l-1 -fib? (l. I '~97 U.8-11 1.817 i 1.8b ~ I 87 ' I

nn--1 ~ 10 , ?6~ 1 U.764 1.82 ; '.6? I
I (.1 ~)' I

np 3-> 471 () 184 ~~s4 __U.84-1 '_'.03~ ~.5~; -J
I ~ ~

m-~-1 ;99 U.~I ?85? ~J.s83 0.99~ 1.;9>
! I

m4-'_' 6~s () ~1 '991 ~1.~77 0.9>? ' 1.x;1 ', 1 n14-; 6~(i ~i.~ ;116 O.i7 (1.9~.~ 1.()9-~
I 1 i '~

I no4-4 7()9 ' 0.~' 3J? 3 !7.61 O.8 L9 I .034 ' 17 ~

I _-I?n4-- 6-14 O.~s?_' ?7(i~ --!1.;8> ().886 l.~'98 ~~

_ _ -i m~- 801 I 0.? 36 ;471 t). >66 0.77'7 1.1 I i ~ I

m~-~ 8?4 0.x'3 ;691 0.~6 ~ 0.7=19 1.088 i I, m~-3 8 j 0.??6 379? 0.~~7 0.7?8 ().98?
', il nr-4 84i ~ 0.~?6 375> t).~~~1 0.743 . I.()1 t ~! I

mi-3 8?I O.~n7 361? 1).>64 0.76? 1.102 ~ ' I

Appendix C2: Shear and Rutting Properties for Mixes 6 through 12 Shear Pro ~erties of Mir Ruttin~T f'ro enies of Mix Mix Mean Rut ~ tMean Rut Wean Rut Design ''''~ ~-;~ Shear Av~_. Shear Av~. Shear Depth Depth Depth Strength Strain Modulus Designation (0.1 MPa) (0.3 MPa) (0.6 MPa) ( hPa) at Peak (kPa) _ (mm) _ (mmi Imm) m6-1 503 0.~1? ?438 0.864 ~, 1.175 ' '.151 a m6-~ ~ 575 IJ.? 1 ?47? 0.763 ~~ 1.171 ~'. 34?
m6-~ ' S4? 0.~' 16 ?56; 0.7?7 1i I .134 i x.573 m6-4 ~I 539 0.~ I 5 ' ?494 ~ 0.717 ; 1.108 ''.?6 3 I
m6-5 ~ 818 _0.'_' 1 I ?484 ~ _U.815 I _ 1.131 F 1.91 ~~7_ I I ?41 r 0. I 56 ~ 1548 j 0.781 ~ l .70?
-m7-? ?s4 I I;).15 1690 ! 0.761 1.387 j not m7-3 I ?69 11.15 18'_'3 0.7?4 1.648 j m7--1 !~ X38 C!. I-14 1643 0.75(i 1.67 ~ !I a~ ailahle ! _ m7-5 ??7 ~ (:x.15 3 _ 1475 I 0.794 ' 1.831 I !
m8-1 I 3~2 (,x.148 ? 169 ().759 ' 1.07? j m8-? j 340 0.15 3 ', ''~~8 0.741 i 1.087 not m8-3 351 . 0.15> ~'?64 0.7? 1.06 ,~ uvailahle j m8-4 ,7? ' 0.156 X389 (1.69? I 1.'_'S8 I m8-5 367 0.157 X340 _().708 j _ 1.711 ' l, m9-1 T _'15 I 0.1 s8 ~I I 380 1.~75~ ''.07s m9-'_' ~'i 231 I i).16 1500 C .??6 j 1.909 ~, ', I m9- 3 X40 ' O.14 3 , 1681 I .0 34 ?.1 14 ; not n,9-~ ''i< I (I 1-1; ~ 17~;; I 199 1.97 I malilable m9-> ' ' 17 0.159 I l .389 I l .988 ', n~(o-i7~~ ~ 0:33 31~~ e).~7~ _ 0.88 l.;~s__ i m10-? 737 'i (1.'_'3''3191 ().56? 0.864 1.14 ~

j n~1()-3719 (1.~'?9 3136 0.57'' (1.93 1 ;55 ~

j m1()-4 709 ().~'36 , ~ 0.661 0.9511.431 t 3054 ~

m11-1 418 ().~'~9 1837 1.'1 ~.4''I I ~.448 i ~-_ ; , , 4j~ t).~4 ' 188? 0.985 ~.?14 J x.4=11 I null-~ ' ', rr:l l-3 468 0 X46 ', I9ls ().9~6 x.074 x.681 !, m I 1-4 486 0.? ~ 1947 ().9 1.995 ' ,.684 ~

m I I 469 _'44 1934 0.95? ~.?4~' ?.661 -5 0 ' _ I m 1 197 _ 946 _ _ '- I 0.'?09 1. 5?6 i m I ~'_? ?38 !, U.~? 1 IC)1 1. 396 not not I

I m1'_'-3~''~ I (1.'1l 1058 1.;99 available availah'~e I ' m1~' 4 181 I 0.L97 931 1.575 I i ~, ~

Appendix D: CiSSST Test Results I Maximum ~ I Lpper Lower Tcst Plate ! Failure Torque Failure Failure Test No. Temperature ~ _~-~ Diameter Depth i -- ~ Diameter Diameter (de~~. C) i (lbf~in) N~'m ~, (rr_~) (n1) t,n) L (mj - - -C' ?9 331 > ;74.4 0.09? I 0.01 ~ 0.1 ~ 0.06 i -_. _~ -C ~9 360(t 406.7-~ 0.0''? ~ 0.01 ~ 0. ( 03 0.0~~, I ~
C4 ' ?9 ~I 4>0() ~,I ~0~.4s 0.0~~., ,- (.).O 1._' ~I 0.10> ~ O.U,_ _ _ i--C~ ?9 ~60i ) ~- s 19.73 ~ 0.092 0.01 I 0.11 0.0 7 ('6 ?9 ', 460(i ~ ~ 19.73 ~II 0.()~)? (>.0I 0.10; ().07 Isl InSiSST Software Test Speed (rpm)180C> Pavement 35 Temp (deq.
C) Recording (s) 1 Test Plate 92 Diameter (mm) Calibration 23929 Date: Aug-04 First Reading -85 Time 18:' 0 RAW TEST DATA

Torque Cell Angle TorqueTorque TorqueShear Output (radians) (Ib.ft)(Ib.in) (N.m.)Modulus (kPa) -42 0.0000 ? .1 12.7 1.4 27 0.0233 2.8 33.2 3.8 310.61 981 0.0465 5 1.0 6? 2.5 69.2 2864.80 7258 0.0698 81.4 2177.1 246.0 6788.08 14968 0.0931 3'1.9 446.'3.1 501.3 10436.56 20045 0.1163 497.4 5968.3 674.3 11165.24 7832 0.1396 ' 95.62347.3 265.2 3659.35 6617 0.1629 ? 65.61987.1 224.5 2655.22 6447 0.1861 ? 61 193E3.7 21 2264.39 .4 E3.8 __- _ _ _ Raw Graph Ii -Torque y.[ICI ~... ._. .. .___ _ .. .._...... .._ __.__.. ...........
_.._._.. _ _.

_-I_I O ~ -_ -- ~ ..
~ -I ~

Z ~ i I~ U ~ _ i ..r _ I i i '.
I __-_ _-.

- _- - _ --_ __ -~I - ,_- -- T I
I i I ,,, ~-~--~-.I J._JL._ .._U_L .J L J~-.J. ..~ ~._..1.J~-i i _ I_~~ '~~, ,..~ .~....~4J~IJI1.'_~~..U ~i..~ __.-,~~.-_ i..l_ iJL~
_ AngularDisplacement (rad) Processed Graph -Torque CICO C - I I
-~- Torque _ _~~---. ,o~ ~ i i, Z au'i I~ Linear (Torque) --_.__ - _-=p0 C -___..____ - _ ~_ ~ y ! , .i i~l~ ~V, ',~, I _ I~~:, [ __ _ iJC i_ - - -_ "'v= ~ __ I! C - _--T ~- __ - -, I
0 0600 O J6~G 0 CCOC C Oi _~IC =L.L!'300 J OCSU J U~,~~'~ ~ J ~~,_a" i .4ngular Uisplacernent irad) 1 S ~' InSiSST Software Test Speed1800 Pavement 35 (rpm) Temp (deg.
C;

Recording 0.75 Test PlateDiameter 92 (s) (mm Calibration23971 Date: Aug-04 First Reading-33 T ime 18:10 RAW TEST
DATA

Shear Torque Angle Torque Torque Torque Modulus Cell Output (radians)(Ib.ft) (Ib.in) (N.m.) (kPa) -27 0 0.1 1.8 0.2 0 0.0175 0.8 9.8 1.1 122.07 167 0.0349 4.9 59.3 6.7 369.91 1471 0.0524 37.2 446.1 50.4 1854.48 4230 0.0698 105.4 1264.4 142.9 3942.32 9646 0.0873 239.2 2870.8 324.4 7160.

16755 0.1047 414.9 4979.3 562.6 10350.08 21929 0.1222 542.8 6513.9 736.0 11605.66 23612 0.1396 584.4 7013.1 792.4 10933.15 8123 0.1571 201.6 2419.1 273.3 3352.21 7471 0.1745 185.5 2225.7 251.5 2775.80 I 9C10.0 U -__ -__._. ._----___ -_.
i J
C
.

h - __ ,CC _ --_ r;
CI

CI

z r -- -~ -- -:~
a r = a _.__- ___ ---_ r_1 o I
I

a- - ~
,.

o 01l - - _-I
I

~ ' I
rt __ 0 --_-CI -I ' 1 !
Ol ~ -- _- ' ---I
CI

- _~

U ~--~_-.-_~-_~~
.1_I

C
Clc~
n~J~
~nF-, Boa o~

Angular Displacement (rac d ull ~ _____-_ -r- ' ~ t Ir 1 ~-_ _._._____-__ ';
l1 I

--~_Iflr~ if ~.~.uE_f~n~1 J -f- ---_.. ~ -_ -_- __._. _ I_I
I_I

Z 51 --_--____ _ _ -_ _ _ -_ -___ - _ - __- ---41J - v,---~ 1 ~i'_.-. r '1 ? _1 Fi -_ _ _ __._ -U

a- F2 - 1 i H

_ _ ~UU

i __ .-- ---._--_-._.-_---_.-_. --__._..._ -_.._ _-._-_._-._ lJ I-i-_.T.-- ._r_-- ~-_ ___ -~.----T_~ -__, r~ U 02 CC4 0 ~6 0 CI8 0 1 O J G~ C a s Angular Displacement (rad) InSiSST Software Test Speed 1800 Pavement C:) 32 Temp (deg.

Recording 1 Test Diameterm) 92 Plate (m Calibration 23977 Date: Aug-04 First Reading-32 Time 18:20 RAW TEST DATA

Torque Cell Angle Torque Torque Torque Shear Output (radians)(Ib.ft) (Ib.in) (N.m.) Modulus (kPa) -28 0 0.1 1.2 0.1 0 0.0233 0.8 9.5 1.1 88. 7 99 0.0465 3.2 38.8 4.4 181.68 1095 0.0698 27. B 334.2 37.8 1042.00 5374 0.0931 133.6 1603.1 181.1 3748.72 15256 0.1163 377.8 4533.5 512.2 8481.01 23636 0.1396 584.9 7018.5 793.0 10941.51 8835 0.1629 219.1 2629.4 297.1 3513.54 7924 0.1861 196.6 2359.3 266.6 2758.49 7364 0.2094 182.8 2193.2 247.8 2279.40 FFd1 fl ~'1~7 1 F,d 1 q7R ~~'~ 1 R~r1 4 R R C7~

-. _._ _-__ ~_ _. - :,----.- '= a n a _ l~-- _ -__ --_, , -_ _ __-___--0 CI 0~' CI CI:1 I ) U'~ LI Ua 0 1 i I " ~' J 1 :1 O '1 r, IJ 1 _~ I
--_-_-_ ---___ Angular Displacernerrt (rad) -.
_ I _ _ -_ ~--~- ~eries1 - - _ !J0 _ --~ -___- _ ?~
ICI -LIfW.~r~~~2rIC~.~) .- L_ -_ ~ - _- __.--l ICI

u._ _.-I~I _ _ I
I
~

~ _-_ - _-ICI ~

j1 1~ - , . =+,~~ - ~ I_I:~ ' tr~
I~I -_ _ _ _.-_ -.

_ _.
1011 ~ - -_ _-~~
~

I
~
~_ ~.~_ _~.rt~._~_~~1 n r o?

r ;

Cn-..
CI
01~
L

C, ?
r d C

E

Angular Displacement (rad) V ~ i I _ . . , I
c°'~ I ? I I
'- ~ _ I ' j -il ~ ~i I I
T I ' ,~ ' ' .i i f I ,.~
I I ~ i ~ ~ I
I ~ ~ ~ . ~, ~ ~ ..,II I _ i I ~ ~ . ~ ~ , i I a ~~~ ~ -;, ~ ~\~ , ; i ~ .
U E ~ ~ _ I
o L; v c~ ~ ~ ' ~ ', i co ~ o m i ' ~ I i __,. -, _ _ -, - _ ~~~, I , _ .
-i r ~o o c r~ o , Q f ,~ , r , #i~ _u. u.~, iwN) amlml.
pn'N) antmol _.__-_--.__..._.-..__-.__.__...
O N N LCi C'~ Ln CD O CD CJ O O c'~ N O) R Q7 V ~r,'~ ~..~,~ c~ O a0 .- c7 N O ,- T- CD CD ~ N O> O d. ~
d ~ ~ I~ N c'~ c0 N N I~. 0a CV m I~ rt N I~ o~
Q; .C ~ ,y f~ cp ~ yC; I~ ~ N O CO N OW ''W - f~ W ~ 0~ N ~ O cOD
O ~. tn N ~,J f~ OJ c~ c'~ f~ O CD C'7 .- r O (n 0~ ~ O
n E ~ S~ r- C~ V C''J ~~ N N N N N N -- ~- r-d y c~ V O ~ O N cD OJ CD ~ L7 O~ ~ N ~- O O N V ~ O
O r~ a7 O ~ r C'~ ~ CO L(~ O O O <t r' V CO T- ~ Qu f~
Q7 N O Z O ~ .-- p~ of N O d O r~ m c~7 O) ~Y ~ ~ CD OW I7 r O
> :n O ~ ~ V N CD O) 07 ~ o~ I~ !~ OJ ~ m Cn 07 of Lr7 C'7 C> c~
ca V ~ _ CL r D E-~ N '~ co I~ v a~ .- c~ r~ ~r r, v c~ u; o t~ O O
''? 'r v o ~ c~ ~ r~ o~ v cfl wi r~ co .= c; v o~ .= ~
O ~ 4~ ~, ~ ar N L~ o~ I~ c'~ cn r~ cm Ln r~ ~ cc~ r; r~ N ,-o - ' ,- ,- ~ ~ v v a~ v o~ o v a~ c-> ~t <r cD N r- cC r~
f- ~- N L~ o~ r~ I~ ~ cfl r~ r, r~ o~ m co r' ~ c~ N N
O ~ C) CD ~' T ~ L~ ~ p O f~ V r V O (D f~ t17 C7 CD V
O c~ c'~ ~ f\ N Q) CD ~'~ ~ N O) Q) C'WO CO c'~ ~ CD d' tf;CD
O .fl p .- ,- ~ ~ O t1; I~ cD :V O7 f~ of CD O) O O C7 C7 (D c':N
c ~ ~.~ N V cD cD :D tt7 ~ ~ cp CO CD f~ I~ cD V N N N
O
V
v7 O r V N
O CO u~ y = cTJ Lc~ CO c'~ CD ~ ~ V' I~ O N ~ 07 O C7 CD ~ f d' CD 07 m ' ~ .~ C': CO 07 c'7 CO 07 N CD n?) N CD Q) N ~ m N II7 0~ N L'7 ~ r O N V cD a) cY7 CD o~ O cW f7 t~ O N ~ i~ a~ ~ cD o~ .
N C 'D ~ O O O O ,-- r ,- .- ,1t N N N :~ c''7 C7 c'~ ~ V V ~1' V W ;
O O O O O O O O O O O O O O O O O O O O O O
O -~ C G~ _ CJ O; ~ ~ U yf7 of O N CD ~r- Q) N of O O f~ <? O N I~ f~
~ C7 ~ _ cG ~ Q « I~. O ~ ~ ~c'7 N I~ N O .- N O c~ N r7 O ,- ~ V ., V O t.~ O O N O ~.,~ N V Q~ ,- O .- o~ O I~ .- tn V O~ CD CD ~ '_ F- U7 L ~ O ' tC7 N ~ O~ I~ Ur' '.f~ <t c'~ C~ ~ CO ~ 0~ of o'; I~ O ~
r N N N N N N N N N N N N T t--W ~ V ~ L H
t- r- ~ U

w~ I

__ _ , I _ r .__.
_ I , i I I ~

I w I I.
I

i i i . I ~, ~ ~, I I

I ~ I I

~ I
~ f I

~ I ~ I . a I

C7 _ I

C d I

I I ~ ~ I C
I N I ' '.d i1 i ~~

y ! V . , F
I I

i I R _ ' ~ d In,,i. . O V
I Ill~ W i ~ J
I I R
I
I

I C'v~ 1 ~ ~ I
I
~.

,:
I I ~_ ~ I ~I O
I.i j 3 ~ _~ ~ ,. r r ' c_a ye I R
I

I C i c'. f' 7 ' i a L:~ C

v..~,__ I ~ I n ~
~ I j I d I

I ~~~ ' I v If ~
.' i . I ;rv I

I ~ II
~ ' L' i I I
i . I_' ~

~
I
c -J
~
-~

i ~, -,- ..
i (- -'-~' . _ _ ~ , ~~
. , n _ =
{w~N) c-anbaoy (~W~N~
anb~oy O t~tf7 07-m r off~ ~~
~ ' cn O cD
~ o7 if)~ '~ V a7C') ~ c-aO ,-O .- f\
O ~ N ~ N ~~
7 LC7 ~ OJ f~
~t t!~
C a7 c0 t1~
V

p ~ O r~j(Da7 c'~f~.c') I~<f O
,~ ~ CJ N aJ
N a CD
~
f ~) r, .- = .C O ,-07 N -O O V ~ (~
o~ ~ r~ I~
c0 I~
c J
' - (n O N c~ ~ CVO ~ N r- Q7 O y 0~ f~ ~~
'. L~ O
N
O
~
~

N ~ ~ C7C)CJ N N N r . C~ C'J
o N

O
.-.

N a7 CDa) ~ CO~ ~ N G7 f~
cJ N - ~7 Q7 (D

_ ~ nj CDN O c~)O a7,-~t ~ ~- ~ O
O ~t of f~

O O V CO;'7 CDC:)O ~ N N N
~ Z wt N .- N
~ .- o O
~
~

- N~;r~o~o>rncr,ooo o o o ~r I-D

c y ~ o cno ~rcr>,- r ~r~ ~
ccJ m o ~r o a) ~
c r ~

~ 05 v co ~ cc;ai ~to~o ~t ,_ c; ai o - :~i c~ c~i o .-~
o a) ~; L~ ~a:omnr. o om~o~ 0 c~r.N
a~ or' ~n ~ o r r~N m r~o~ 0 0 o - r~ o rn a~
~ o~ o~
~
-.. N com o~mo~ o~~ a) a) - ~r ~ a) r) o~ N

I- D
H

a _ o ca N r~ ~- T 'rt ° co r~ u.~ N o c~ N N c~ ,- a) Q) Q7 tl~ OO a7 L('.~~ N L~ a) ~ c'7 L~ N CC) 0 ~ U O V ~ f~ Q7 CD o7 N O C. <'~ C V ~ ~1W C) L!7 r r .- r»n co co r' r- ~ r~ r~ r' ~ r. r' r) 0 0 ~ c~ Lnoc.--)cDo~ v r~o N ~ coo r;c.a) ~ ' C r- .-O a~ v ~'7 CDQ~c D a)N ~ N CDO)N L7O)N L')a:) ) CD N

' N ~ O r C7CD O C:~f)I~O N~tI~.O)~~
a) V

c ~ O O Op _ r- N cJN N c'~c~)C7~'7C;W
~ V' N Q O O OO O O O O O O O O OO O O C>
cC O O
L

Q

H

Q

D

_ O ~ C 1- ~ O - O C~ ~~'~O ic')ON cDc0 _ u7 ~

O ~ ~ ~ O- C7 ~ 07N C''>CON CD~ L7M f~OJCDf~.r-O'>
c O cT O O rU' t a 0 op J = , W ~ N <t~N - ~ aJ CDC:OJ N c~~ u'7V .
- O .. ~ - ~ ~~
U
~

~ O T N~ CC)N f~ a)O~O)O O OO O O c11 1~

Q w N N N cJN ~ :'7C'JC~C'7:~r N

LV ~ Q) -_ I-L ii.
f- J

Is~
InSiSST Software TEST: BF1 - BOND FAILURE
Test Speed 1800 Pavement 28 (rpm) Temp (deg.
C) Recording 1 Test Plate ter (mm) 92 ~s) Diame Calibration 23954 Cate: Jul-31 First Reading -71 Time RAW TEST DATA

Torque Cell Torque (Ib.ft)Torque (Ib.in)Torque Modulusa(kPa) Output (N.m.) (radians) -45 0.0000 0.6 7.7 0.9 919 C.0233 24.4 293.4 33.1 2744.18 5490 0.0465 137.3 1648.0 186.2 7707.27 12676 0.0698 314.8 3777.5 426.8 1 1 7 77.80 19696 0.0931 488.1 5857.8 661.8 13E~98.04 24990 0.1163 618.9 7426.6 839.1 13893.32 26303 0.13 651.3 7815.7 883.1 12184.35 26753 0.1 662.4 7949.1 898.1 1 OE~21.92 E~29 27075 0.1861 670.4 8044.5 908.9 9405.75 27356 0.2094 677.3 8127.8 918.3 8447.21 27566 0.2 682.5 8190.0 925.3 7660.70 27756 0.2F>60687.2 8246.3 931.7 7012.15 -77 0.2792 _ -0.1 _ -1.8 -0.2 -'1.39 _ -____ - -. _-_ m:j.~.. ra"e i crag, -r I ~ IrJ I _ _ _ ........._ _ . _ J . i ~ __ _ ~ _.
~ I II I~;
LI

I ~- _-t -I -......__-_.._.._....._-. _.. _____.__ _.t- ..-_ ~.lllll -~ -1~~ f I ~_ ~I,.II7 _- __.._ _ -._._.-_._. _.__-.~--.____._ _-I_I _._..~.I
+ -..---.-----___--_..

~% - -_-..- _ -._- ~ _--~: U O i I - _ _ I
_ --.
H- - ___. .-_ _ _._- -- ...y I - _I t _. _...
l I I t_ -.

~- -T___...-. ~____T...-~.._~-_-~ --~.. T
l ._~ - T. r i I_I I_I - y~' ,~ -_ l ; ~~ ,,-,~:~ 40 t ~~~ t;~~
-' ~ ~ ~ ,~
U 0 1 ,~
- ~ r dL
- L <
~ e ' L

O- CJ' O u J I
O cw O U
o i Angular lacement I
Disp (radi ___---___..__- ___-_- -_ -_-_--_ -_._W
._ _ F'roc.esserl Graph - Torque II

I I
f IJ:.J 0 -w =~t I~ - 1 _ _ - 0 0 _- - - ~ __ _.
U _ I r I y r ._ r- r, 1 , ___ _ _- __.__- _-IJ LI IJ ____-.____.._._._ ___ . _-____ _ -, _ 11_I,-=1~ _ coq 4~
IJI I C~, -- .___- -_ _ __ _._ - __--- 'l __ I
l 0 CI I I = -___- _-_ _ _.__- _-..._ ~ --. ' i i U I-I -+- -- 7..___________..___~___..-___-T_.--__~ '',,, 0 tJ001J (.1 0.00 O IJ-'7010 I CI61_IfJ IJ CI80CI U 1000 Angular Displacement [rad) j 1S?
InSiSST Software TEST: BF2 - BOND FAILURE
Test 1800 Pavement 28 Speed Temp (rpm) (deg.
C) Recording 1 Test Plate 92 (s) Diameter (rnm) Calibration 23930 Date: ,!u1-31 First -75 Time Reading RAW
TEST
DATA

Angle Torque Shear Modulus Torque (radians) Torque (Ib.in) Torque (N.m.)(kPa) Cell (Ib.ft) Output -58 0.0000 0.4 5.0 0.6 24 0.0233 2.4 29.4 3.3 274.65 5471 0.0465 137.1 1644.9 185.8 7692.88 14626 0.0698 363.3 4360.1 492.6 13594.55 22086 0.0931 547.7 6572.7 742.6 15369.81 25029 0. "~ 163 620.5 7445.6 841.2 13928.75 24918 0.1396 617.7 7412.6 837.5 1 1555.9 448 0.1629 12.9 155.1 17.5 207.27 Haw Grap h -Torque ;_; , _ _ _ ~I : .- _ ~, ~
_ - - - __ __. __-. ~_ ' _ _ I
_, , I - -_ I
,."
h.
~
~
L'I
' ~ _---- _-___~___.___.___--_. ..___ I
~. ._ il, f=I
-~___-_ i Z

-_._-.- _ ' - _ _._ _-i ~

H _ _._ _ _.___ _._-._ -~ _ _.. - _-- _ I r __ _- -..-_-._._-_-_.-._-i __:_~_ -_ -_ -,, , ~___ - ____ .._ r...._ .____,, JOU-J ~l a ",.~ .-_.___~ J ..,1:3~:
1J l ~lc,z, _ _ _ I
_, _ ___ ~
_i l;l;
l,=: l~
( ;" .
_ ~.1 E:, i Angular Displacement __ _ _ _.._ (rad) __-_- !
__-_ _ _ __ __ -___ __.

r'rocessec~ raph -Torque I
G

J _ _ _-~ J U t ~ t i r ~ - _-_ t;~l'lll I ~ -Ln~j i~ 'i~ 1' -_ -_ _ '~ ~ _ - I __~~ _--z =iJI~~_ - -_ _ -_ ~ ~ ar , o, ~ I I . ~ ' - _-_ __ -____ . __ ___ _ _,-_ _ ~_- ! '.
i I
a ~ilii ~ - _ __._ - -r"- ~._ I _ _ _ i -_ _-__ - j ___ ___I
I 1 -~ L
I I , ' - T._ _ _._.
I~~ [il;l=?i~~ i_i I=I-'i_ i_i .;I i_;_ji l _i I _J~~Ji_I f i ~J ~ -il=l i~l !
jiji~ 'i Angular Displacement (rad) ', Ins _____- ____ I~8 TEST: BF3 - BOND FAILURE
Test 1800 Pavement 28 Speed Temp (rpm) (deg.
C) Recording 1 Test Plateameter 92 (s) Di (mm) Calibration 2~~953 Date: Jul-31 First -67 Time Reading RAW
TEST
DATA

Torque Angle Torque (Ib.ft)Torque Torque Shear Cell di in) (N.m.) Modulus Output (1b (kPa) ( ans) .
ra -57 0.0000 0.2 3.0 0.3 383 0.0233 1 1.1 133.4 15.1 1 247.61 2822 0.0465 71.4 856.3 96.7 4004.84 10862 0.0698 269.9 3239.4 366.0 1 C) 100.13 34 0.()9312.5 29.9 3.4 70.01 flaw Graph -Torque _.

~, _ I

~'J 0 _ i ' _ _ _-?,~_I - _ i _~~1~ -- _ ~~

-- - - - ~__ ~

i _ , ~
1=
;
i i , -_ l _ I
'~
~;
~J

, __ _ _ '~ ', e C~
I
~

-_ -~
, I, ~~- _ _ I_~_ - T i ~
~
-O UC't ii 1 ~y~i = _ i~i i~46~ ~~ i ~r ~~~> i i i;ry31 Angular Displacement (rad) Processed Graph - Torque :1 ~,J :J I:J ~ i - _-_ -_ - ~ j -+ ~ r a _ ~ i a ~riaa ( -;a; ~_ 1__-_ __ i _ _-, ~'. 1''J ~' ' - _-.__-- -_ __ -i L 1~~U -~
F,_ I- 1 ~1 ~ ~ I -_ _ -. _----.___-- -_.__ __ jv I~I ' -- _ -.- I
n ~i~=~~~~~ f~ r~~~~~l,f~ ~j 0:10u ~~ ~m;J':l _~ 0=pr Angular Displacement (rad) 1s9 InSiSST Software TEST: BF4 - BOND FAILURE
Test Speed (rpm) 1800 Pavement 28 Ter~p (deg.
C) Recording (s) 1 Test Plateameter 92 Di (mm) Caiibration 23953 Date: Jul-31 First Reading -50 Time RAW TEST DATA

Torque Cell Outputl Torque (Ib.ft)~~bqn~ Torque Modu us (N.m.) (kPa) (radia ns) -54 0.0000 -0.1 -1.2 -0.1 59 0.0233 2.7 32.3 3.7 302.41 642 0.0465 17.1 205.3 23.2 95'x.96 4373 0.0698 109.3 1311.9 148.2 4090.45 13615 0.0931 337.8 4053.2 458.0 94 7 8.18 -53 0.1 " -0.1 -0.9 -0.1 -1 .66 Raw (>raph - I
Torque i _:_~I [l ~

- r , - _ -=.I_I ~ -_ ; '., I

= __ ' , n ur _ _ J .,S
y _ _ ' - -o r. o - : , a- : y o ' , ~ - __- _~-. II rI -- -=

_ _ _ _~~: -- _ -.__r __.-.
. -. ~_ ___ I I IJ~JIJ.J LJ~W r~..I~.I~ ~,. I IJ Iii 1 :1 i=~ ,=I I!f~IH"= I~~J,'.

- Ang ular Displacement (rad) I

Processed Graph -Torctue tIIII~ - _.
! II
t r=~1 I _ ,~I ._l ':_, ____--Li_'w, ,rrlr-1j Z ?.I_II~ I,I - _ -.
a I ~;, I-; _~=, I _ v __ I ~ _ I_I _ I,J - ~- -_- - -F _ _- 1 I ~ .,1J IJ _ - -_ _ ' 'i I' I I -~- - , _ I~ Oi~IOii G O~'00 U ~I~r=ILJ O Ui~I~!_ i il3yC~ ~'I 1UOU
Angular Displacement (rad) II

1y0 InSiSST Software TEST: BF5 - BOND FAILURE
Test Speed (rpm) 1800 Pavement Te~np (deg. C) 28 Recording (s) 1 Test Plate Diameter (mm) 92 Calibration 23956 Date: Jul-31 First Reading -60 Time RAW TEST DATA
a Torque Cell (rad'ans)Torque (Ib.ft)~ bqn~ Torque (N.m.)Moduh s Output (kPa) -52 0.0000 0.2 2.4 0.3 0 0.0233 1.5 17.8 2.0 16E~.38 848 0.0465 22.4 269.2 30.4 1258.9?

5646 0.0698 141.0 1691.6 191.1 5274.13 13889 0.0931 344.6 4135.2 467.2 9669.93 -52 0.1163 0.2 2.4 0.3 4.44 Raw Graph - Torque -,, i - _ _ .,', r:
_ r, l.,=t[oi~ __ - __. __ _ L
I - _~ ~ s N -1i~'~ i,l O' i ... i ~'1 O 1OI_I._-i _- JN _ -_ _ - 's I
U UOiii-i i1 y!~1,J 'i! ~iy~=; ii iir;~a;~ i_i t-i~1 rn 11r;
'~ r_, -, Angular Displacement (rad) f'rncesseil Graph - To«tue ~i ~" r~
I _~
-~- ~er~ 1 _ __ Liri~~r ;:~a ri s''j Z :~ C: ~=I ~~', ~ - _ - I
'_r = 1 1 E~~~- r~~~T 1 7 I
i a !~iiJ U ~~ _ _ __ __ -F:== ~-__ _~__ __~_____-__ ,_ !~i 'iio~)i~ i ~ i~i_i~ii i _i ~~~iQ i ii~~iji= i i~;ij~i _ I i-i i Angular Displacement (rad) InSiSST Software TEST: BF6 - BOND FAILURE
Test 1800 Pavement 28 Speed Ten-p (rpm) (deg.
C) Recording 1 Test Platedieter (mm)92 (s) Dia Calibration 23931 Date: Jul-31 First -69 Time Reading RAW
TEST
DATA

Shear Torque i Torque (Ib.ft)~ bq~~ Torque (N.m.)Modulus Cell Output (rad g (kPa) ns) -58 0.0000 0.3 3.3 U.4 574 0.0233 15.9 190.7 21.6 1784.19 5210 0.0465 130.5 1566.0 176.9 7324.05 13164 0.0698 327.1 3925.6 443.5 12239.59 21850 0.0931 541.9 6502.3 734.7 15205.14 26928 0.1 163 667.4 8008.7 904.9 1498a?,?

27003 0.1396 669.2 8030.9 907.4 12519.84 27491 0.1629 681.3 8175.7 923.7 10924.73 -72 0.1861 -0.1 -0.9 -0.1 -1.04 Raw Grayh e - Torqu )00 0 -~ _ y a? i I ~'.u -__- - - -. .T'.~ -.
I ~_ ~ '' S
- _ -C~',JJCi -I _.- ___. .__ _ -__ __.___...___._-_-__ as J~J.._. _ ', i_~ _ ~

1. -_-_ -__-._.___-_____--_-__..___..-.._'i._.._ ' -~-- . ~ S

L,, 1 - _ ---r-- 7-TT Y

~ i D.''~ t:~?~b ~~ ,~b ,~?y~C~~'~ ~ I

i ~w y ;~ :' O
~", L c \
~ ~

' LW .
~
' Angular Displacement (rad) Processed Graph - l o«iue .; j I I-i i-, I ICi CI + ~,~t ~ ~ ___ __ ---__--- --r X11 C) Lirie.~r ; ~ °n~~.1 j ---__ Z =~Cli~i ~~ I - - - _ _- I
a~ 4C;I!J I=I --~-,i-- - _ y; - "_ X ~+_''~a r;' i1 f~1 C I ~----~=-~ _ _ o ' f- ~' C~ f-1 C) ~, ~I U I _ I i_I - ___ -_- -.
i i CI Cl ~ ~- T _- _, CI I~~IC CI O~=C II~ i I ~i4i JC I I i lyl~l_I r I i';.30'~l ~i 10' ilj Angular Displacement (rad) In 19~
TEST: BF7 - BOND FAILURE
Test Speed (rpm) 1800 Pavement 28 Temp (deg.
C) Recording (sj 1 Test 92 Plate Diameter (mm) Calibration 23924 C: ate: Jul-31 First Reading -82 Time RAW TEST DATA

Shear Torque Cell Output Torque (Ib.ft)~~bqn? Torque (N.m.)Modules (radians) (kPa) -67 0.0000 0.4 4.4 0.5 58 0.0;?33 3.5 41.5 4,7 388.37 1608 0.0465 41.8 501.2 56.6 2344.11 9654 0.0698 240.6 2887.5 326.2 9002.8'.6 18773 0.0~?31 466.0 5591.9 631.8 13076.:38 25221 0.1163 625.4 7504.3 847.9 14038.:8 27641 0.1396 685.2 8222.0 929.0 12817.70 28236 0.1629 699.9 8398.4 948.9 11222.40 28465 0.1861 705.5 8466.4 956.6 9899.01 0 0.2094 2.0 24.3 2.7 25.28 _ r ~
~ .
' --.
~

I (1 ..
~ ,~ : f ~~ J
lJ

~

r .S

Z S
/

r_ - -__ _ _ ~, _._ r i=

~

- _ -__ __ -__ ~ __ ..

.
~ I t ~ _ -. -___..__- '~._ : 1 r SS

- -. _~~.~~ -7.-.-._-.__~-__~.--.T_ ___~-T-;-~

O' O
~
O
~
O' ~
n' W
W
CJ
W

'I Angular Displacement {rad}

Processed Graph - roaque i=i t' I;j __ .. _ __ _ I
t I If ~ J ~. cY~ ~ - - -' Llf~r 3 ~rir_. 1 , -_ I, ~ I I. CJ r _ __ I
~ ~=I I~: CJ - _._. -_ _ ,~~ ~ ~~' r',. c ~'I lJ -i ~ I
~y CI ~ , -- -- _ __ I
-' r~<
I
r~. = 1 _- ___ -_ _ ~I i=i - [1 -_ -_ , I
0 0 ~ - ___ _ I
T -T
Ij yQI,jQ ~~yl_~~>OU a Ci4r[~ r [.ierrl ~I-r_IgCil~i n 1000 Angular Displacement (rad) Raw Graph-Torque 1~>>
InSiSST Software TEST: BF8 - BOND FAILURE
Test Speed (rpm)1800 Pavement ~~deg. 28 Temp C) Recording (s) 1 Test Plate >ver 92 DiamE (mm) Calibration 23952 Drite: Jul-31 First Reading -52 Time RAW TEST DATA

Shear l Torque Cell Output Torque ~~bqn~ Torque Modulus (Ib.ft) (N.m.) (radia ns) (kPa) -35 0.0000 0.4 5.0 0.6 0 0.0233 1.3 15.4 1.7 144.26 3486 0.04135 87.4 1049.4 118.6 4907.78 11893 0.0698 295.2 3542.9 400.3 11046.44 22687 0.0931 562.0 6744.4 762.0 15771.34 27383 0.1 163 678.1 8137.2 919.4 15222.

28260 0.1396 699.8 8397.4 948.8 13091.11 -69 0.1629 -0.4 -5.0 -0.6 -6.74 Raw Graph -Torque ! _ ~ Ci r C~ C! _.. _.__._ -"s _ ~, i~7 C C ~ . -_- _-_ ~ _-Z j=,yy-i iy __-.-_-_. _____-_-________--_._ ~ _-_ a- ;-, ~ . _ =l O 0 U
,, _-~ - --.__-_ -- -_._____ _ -. _ -_-.._ v i _ ~_~~T_ _~_-T_ _~-O C 7 0 0 Ci i~ C ~ L. '_ C r:..:~ r; = r. i e~ C ! O =i ', -1 Ci 1 I r, ~, i:
i '1 C. ~ C ; y, ~
y - _ i Angular Displacement (rad) _-____--_--- _---_. _ - -__--_- _.__-___. _ ___ ._- __ -_-_.-Pruces~ed Graph --(o« due _ __ _ I
_ I iiJ t_ ~;Fr ~.-,1 -__ Q ~ - _ ___-Ivi Lir~r ;~r (~eries1j ~ - -i~

r;
r u-a o _ -~ j n~
~

~ -_ ,,,, _ ~, , i~~ ? ~; ~ --_ ~ ~ _- _ C~ 1 ~
Ci o :~,~,~r~i~ '~ _ _ _ ~~

~~~=iC -~~~

- - -~_ J~_ ___ i~i Ui~;~ ~i i=~;~i ~i ;-yijiJ JIJt~siJip i i [o~-lia =i i i=, y~ri~=~

Angular Displacement (rad) Design, Development and Verification of an Advanced In-Situ Shear Strength Test Facility for Asphalt l.oncrete Pavements ~iCHItP-IDEA Project Ss Abd El Halim Omar Abd EI Halim, Carlc.ton University, Ottawa, Canada Stephen 1~. Goodman, Canadian Strategic Highway Res~,arch Pro~~rant, Ottawa, Canada v Wael Bekheet, Carleton University, Ottawa, Canada' l~asaer Hassan. Carleton University, Ottawa. Canada IDEA CONCEPT AND PRODUCT
The StrateyTic Highway Research I'ro'~r~am (SHRP) recotmised that shear properties are an important indicator to predict ruttin~~ potential of hot-mia~ asphalt concrete {I-1MAC) pavements.
however, current methods of measuring such properties have been linuited to time consumit~;"
expensive or unrepresentative laboratory analysis. The concept of measurin<~r the in-situ shear properties of an asphalt concrete pavement layer by applying a torque directly to the surface has been initiated at Carleton University in Ottawa. Canada. 'This concept allows relatively quick measurement of in-situ shear properties with a rninimun-. amount of damage incurred by the pavement surface.
Under the current NCHRP 1DI:A project, an advanced ire-situ device has been developed and fabricated at Carleton University. Know ~n as the In-Situ Shear StrengtliiStiffness Test {lnSiSST'~'~), the device provides the rapid and accurate rncasurement of in-situ shear properties of an asphalt concrete layer. Data collected with the InSiSSTT''~ will provide input for more accurate measurement and performance modelling of in-service pavemmnt performance - the fundamental basis of the SHRI' Supetpave system.
PROJECT UPD.~'CE
InSiSST' "' Design The completed InSiSSTT"r device is presented in Figure 1. .As shown, the components are mounted to a small trailer to provide exceptional portability. The InSiSSTT~' utilises an electric motor and Gearbox to produce the torque required during the test. The motor.%gearbox combination is mounted vertically on a steel platform that is attached to a positioning system incorporating two sets of worm-screw slides working in tandem, also referred to as an ''X-Y
table." The top set of slides allows positioning of the platform in the transverse direction (with respect to the trailer orientation). The transverse slides are in turn mounted to a second set of slides allowing positioning in the longitudinal direction. The entire positioning system is ~ Professor, Department of Civil and Environmental Enoineerina. T:le: (613) X20-~ ; 89, Fax: ('613) X20-39~ 1, Email: ahalimoccs.carleton.ca C-SHRP Pro=ram Manager. Tele: (613) 736-130, Fax: (613) 736-1396, Email ~godman~a?cshrp.or~z ' 1'h.D. Candidate. Tele: (613) >20-74?l-1961, Email w~bekheet~iccs.carleton.ca visiting Professor. Department of Civil and Environmental En~ine~rin~. Tele:
(61 ~) X20-2600 Ext. 862, Fax:
(til3j ~?0-3y~1. Email: vhassan~n~ccs carl~ton.ca mounted to a box-tube frame occulayin'r the space between the tow bar and the axle of the trailer.
The test ti-ame is attached to the trailer ti~ame via four screw jacks, on~~
at each corner c~f the test ii-ame. During tr~insportation of the InSiSS'hr", the jacla are retracted to hold the frame in the air to prevent damage. Once driven into position, the jacla are extenc7:d to lower the test frantc to the 'round and then continue cxtendin', until the wei~~l~t ofthe trail.;r is supported solely Iv the Pest frame. Control of the jacla and positionin<~ slides is provided by commercially <m aii.ablc electric motor controls. Control of the actual test procedure is provided by a laptop computer.
Instantaneous torque and angle of twist uneasurcrnents arc collected oo the computer during the test procedure. A large plastic storage box is mounted n: the front of uhc trailer to house the electronic components. Finally. a gener~atur is mounted to the rear of the trailer to provide elcctricitl. for the InSiSSM'T"'.
Figure 1: The In-Situ Shear Strength Test (InSiSSTTnI) at C:arleton University Calculation of ln-Situ Shears Properties As mentioned, the hlSiSSTT"' applies a rotational load (torque) directly to the surface of an asphalt pavement. The tordue is transferred to the asphalt through a circular steel loading plate.
epoxieci to the pavement surface as shown in Figure 2. 'he asphalt is loaded to failure and the induced failure surface is semi-spherical in shape.

Figure 2: Method of Load Application for InSiSSTT"' (Side View) The loading case used by the lnSiSSTTM device is very similar to that investigated by Reissncr and Sa'~oci in the early 1940's. With a circular loading plate affixed to a linear elastic, isotropic half space, the shear modulus of the half space (asphalt concrete) may be deterwnined using the following relationship:
T-l6Grh~l;
J
V~here: T = Applied tordue G = Shear modules of the material a = Radius of the loadin~~ plate th = Angular displacement of the loading plate (radians ) The Reissner-Sagoci relationship above was also used to develop and verify a finite element model of the problem assuming linear elastic conditiuns. In future modellin;a efforts, the material properties will be altered to linear and non-linear visccrelastic properties, more representative of asphalt concrete, to obwem~c the affect ~.~n the resultiry~
stresses and strains.
Concept t'erificatio~t and Rug~~eduess Vesting Initial verification testin~T was first completed at Carleton University in July ?000 to observe the results of the InSiSSTT'~. Subseduentlv, field tests have been completed in the City of Ottawa and in the Towns of Bancroft and Petawawa. Test results are very repeatable, with coefficients of variation as low as 1.S°ro. Complete test results will he presented in the final report, which will be prepared by the end of the year.

PRODUCT PAYOFF POTENTIAL
The successful measurement of in-situ asphalt shear properties and the:
development of a mainstream test facility will yield significant and immediate benefits t~~ the three prirnar-~- areas of pavement engineering. The first area is d~.sign. Utilization of the IoSiSS~°r"r in conjunction w°ith laboratory testing would be a powerful combinaticm for analmin~.~
the potential of proposed mix desi'uns. The second area is yuulitv cc>ntrul. 'newly constructed asphalt pavements could be tested to verify acceptable construction practices throu~_h the measurement and comparison of in-situ strew<~th parameters with code requirements. The tiaal area is lonr~~-ter-nr paverru~rrt p~rfomruutcE~ (L7'PP~. Monitoring of field shear strength of pavements with time would allow periodic updating of performance models to more accurazely predict future pavement performance. This, in turn, would allo«~ for more efficient allocation <>f limited rehabilitation funds and also help determine the effect of~real world conditions, such as environmental factors.
on pavement performance.
PRODUCT TRANSFER
The potential for a simple yet extremely effective in-situ test device has already drawn significant interest from both government and private inciustrv. In addition to IDEA Program funding. the Ontario Ministry of Transportation (MTO1 ~~nd Regional Municipality of Ottawa-C'arleton have conmnitted financial and in-hind support for this investigation. Furthermore, a number of independent consultants have also expressed interest in the potential of the InSiSST.
Tu date, demonstrations of the h~SiSSTT"' have been completed for the Ontario Ministry of Transportation, the ~1CHRP IDEA Program and at the ~'~' International IZIl_EM
Conference on Reflective C.'rackin;~ in Pavements. Once the current investigation is completed, consultants and contractors will be ~uiven instruction on h~w~ to use the InSiSSTr"'.

Design, Development and Verification of an Advanced In-Situ Shear Strength Test Facility for Asphalt Concrete Pavements ~C'HRP-IDE:~ Project ~~' Abd EL Halim Omar Abd E1 Halim, Carleton University, Ottawa. Ontario, Canada?
IDE:~ CONCEPT :1ND PRODUCT
The Strategic I-Ii'~hway Research Program (SHRP) has reco~~nised shear strength as an important indicator to predict ruttin~~ potential of asphalt concret~°
pavements (ACP's1.
However, current methods of measuring the shear stren~Tth of an usphait mix have been limited to time consumin<~, expensive or unrepresemative laborat,~rv analysis.
Tl~
concept of measuring the in-.sitcr she~rr strejr~th oi~ «u asphalt con:vrete pavement layer by applying a tordue directly to the surtnce has been inAtiated at Carleton University in Ottawa. Canada. This concept allows relatively qujck measurement of in-situ shear su~en~'th with a minimum of damage incurred by the pavement surface. ,~ basic.
first 'generation prototype In-Situ Shear 'Test Facility (ISSTF) has yielc.!ed promisiny~ results relating the maximum applied torque to the shear strenygth of ~m .'MCP layer.
Wore importantly, the ISSTF has determined that shear stren'=ths achie~.ed in the field are very different than those realized in the laboratory. These findings present strop yg evidence .
that the development of such a test device is required for more accurate measurement and parfortnance modelling of in-service pavement performance - thr~ fundamental basis of the; SHRP Superpave progr~rm.
PLANNED INVESTIGATIO'_~1 The first-yreneration ISSTF consisted of an electric motor mounted to a cart-like chassis.
A series of gears and driveshafts were used to transmit rotational force to a circular loading plate epoxied to the pavement surface. A torque cell and datalo~T~ler were used to measure torque while angle of twist at failure was measured with ;~
protractor. Two test speeds were available. To prevent rotation of the test device duriry~ testing.
six steel spikes were attached to the device and driven into the pavement. The tmt-ready ISSTF
apparatus is shown in Figure 1. Test results achievcv with the device yielded in-situ shear strengths of up to 300°% greater than those achieved in paral lel laboratory testing.
Although the basic prototype revealed significant differences between in-situ and Laboratory shear strength of asphalt pavements, a number of deficiencies were noted during initial investigations. In g=eneral, the cart-based test device was cumbersome to set-up and operate, limited in capability and required much effort and time to perform tests. Still, the potential for an in-situ facility to test and model pavement performance indicators wan-anted further consideration. Therefore, a comprehc,nsive three-stage ~ This IDEA Project has not yet commenced An 18-month investigation has been planned.
' Professor, Department of Civil and Environmental Engineering. Tele: (6l3) 5?0-?600 Ext. 5789, Fax:
(61s) S~p_~9~1 investigation has been designed to continue the development of this innovative test device.
1'i'~ur~ l: First (uencratic.m ISSS'hI~ Prototype pest t~c~nfi'7uration Sta~ac 1 will focus on three main activities. The first concerns imlorovcment of evistin'~
theoretical models relating the torque applied by the test device to the shear strength of the ACP to reflect the effect of in-service confinin« pressure. The current theory is most applicable to unbound cylindrical specimens tested in the labor-at~,w.
Simulation of field boundary conditions and loading will be accomplished usiy~ the finite element technique.
An extremely desirable objective of this investigation is to correlate in-situ shear strength of asphalt pavements with~perfonnance indicators such as rutting and crackin~a.
Therefore, results of the finite element analyses will be coupled with traditional revTression analysis to construct preliminary performance models relating shear strength and pavement performance. The resulting models will be subsequently verified and calibrated using field and laboratory test results obtained in Stay_c ~ of the project.
The third activity completed in Staye 1 will concern the desiL~n of a second-generation test facility based on a critical analysis of the deficiencies obsen~ed with the original prototype. Conceptual models include a trailer-mou~ntcd or a vehicle mounted system.
Both systems would allow the device to be more easily transported to test sites and reduce the number of required operators. The trailer could be sufficiently loaded so that it does not move duriniT the test procedure. The ''vehicle-mounted" system would be similar to a corin~~ rid,, and would swin~a or slide into position from the ba~:l: of a pick-up n-ucl; or van. The weight of the transport vehicle would be utilized to stabilize the device. Both systems would eliminate the necessity for affixing the device to the pavement, which in turn would eliminate the extra operator effort and pavement damage.
To improve control and flexibility of the test procedure, the secon;l-generation prototype will utilize a computerized tordue cell and rotational extensometer combination to apply a number of different strains (i.e an~~l~ of twist) and or stress (i.e. torduc) rates to the asphalt. The computer will also record the instantaneous applied t~.n. lue and angle of twist during the test and produce a corresponding seraph. The dimensions of the failed specimen will be input and the maximum shear strength calculatco and displayed instantly. Other notable conditions such as asphalt t::mperature test locartion, sample number, etc. could be input for future reference and modelling purposes.
Additional improvement to the basic prototype will be accomplished through the use of stron;~er materials and a more powerful motor to allow testin;-~ at lower ternpcratur~s (i.e. stiffer pavements ).
Stage' will focus on the fabrication of the second-~Teneration prototype test device.
Once constructed, a regime of "shakedown" tcstiu~; will be compl~tcd to calibrate the test instruments as well as note any deficiencies in construction. Final adjustments will then be completed to prepare the device fc~r the final stag: of the rovesr igation.
The third and final stage of this investigation concerns the verification of the test device and calibration of the analytical models constructed in Stage I . These objectives will be met through a series of field test regimes designed to analyse a number of different pavements displaying varying de~rrces of resistance to rutting ~rnd cracking.
The second-<aeneration test facility will be utilised to measure in-situ shear strcr~'~th values, while core samples will also be extractccf for laboratory testin~~
.Analysis of the results will take on three priu~ary forms. 1'he first analysis phase will conGnn that the field test is repeatable and consistent. Based on results with the first ISSTF, no problems are anticipated in this regard_ !'next, the field and laboratory results will be correlated to determine if a common factor, or "multiplier", exists between thenu.
The sensitivity of in-situ shear strength to critical factors such as unix desi<~n (especially gradation andJaggre'~ate type), traffic level and temt>eratnre will also be analysed.
Finally, analysis will correlate the measured shear s'crcngth to in-~~rvice periormancc characteristics throuvTh calibration of the analytical models developed in Sta;Te 1.
Commencement of the project is expected shortly and field trials should begin by summer 1999. By concentratin~~ on actual in-situ performance. it is anticipated that more accurate and meaningful pavement performance models will result from this investigation.
PRODUCT PAYOFF POrfE:'.~TI.~~L.
The successful measurement of in-situ ACP shear p~~:-openties and the development of a main stream test facility would yield si~~niticant amt immediate benefits to transportatie~n practice. These benefits would be realized in three primary areas of pay ement eny.~ineerin~~. The first area is clesi~n .An in-situ she=ar strength test in conjunction with laboratory testin~~ would be a powertul combination fior analvzint: the potential of proposedJmix designs. Also, the results of such a test apparatus could horsed to produce "shut" or "master" curves relating in-situ shear stren~~th to various factors such as loadin~~
rate, temperature and asphalt content to name a few-.
The second area is gualito ccmtrol. Newly constructed asphalt pavements could be tested to verify acceptable construction practices through the measurement and comparison of in-situ strength parameters with code requirements. The final area is long teryn pcrmmem perforrnunce (LTPP). Monitoring of field shear strength of pavements with time would allow periodic updating of initial perlbnnance models which tin-ther assist in the prediction of future pavement performance. This. is tum, would allow for snore efficient allocation of limited rehabilitation timds and also h:;lp determine the effect of real world conditions, such as environmental factors, on pavement performance.
PRODUCT TIR.ANSFER
The potential for a simple yet extremely effective in-situ test device has already drawn significant interest from both goven~ment and priv~:,te industry. In addition to IDt?A
Program fundin~~, the Ontario Ministry of Transportation (M~hOI and Regional ;~-lunicipality° of~ Ottawa-C=arleton have committed financial and in-l<inci support for this investi~~ation. Furthermore. a number of independent consultants have also expressed interest in the potential of the test facility. To facilitate a more smooth transition into main stream use, these consultants will also be included durin;~ t1e development of the.
test facility. Their role will initially focus on technical advice; he ~wever, the consultants are expected to acquire the technolo«y upon comhl~:aion of the investigation.
Continued evaluation and assessment of market potential will he the primary i~ocus of consultants at that stare.

CA 02330431 2001-O1-08 - ". _.,_~_~...~~_~r, , DESIGN, DEVELOPMENT, AND VERIFICATION OF AN
ADVANCED IN-SITU SHEAR STRENGTH TEST FACILITY FOR
ASPHALT CONCRETE PAVEMENTS
NCHRP-IDEA Project 55 :\bd I:l I-lalim Omar:\bd EI Ilalixn, Carleton L'niversitv, Ottawa. Ontario, Canada, I'rofcssor, ' Ilepartment of Civil and hnvironmental I~n~sineerin,~.
~

'Left: (613) 3?(,~-> ~ ~q, 1'aw (613) ~?0-39>1, Ernail: cxhulim ecs.ccarletort.ea Stephen N. Goodman, n;arleton L'niversiy, Ottawa, Ontario, Canada, NI.Eng. Candidate.

Tele: (613) ~33-1961, Email s~;uoclrnunCeccs.cur/econ.cn t V'acl Bekhcct, Carlcton IJniyersity, Ottawa, Ontario, Canada, Ph.U. Candidate. r Tele: (613) >2t)-771-1961, Email 2e~bchheet@ccs.curletorr.cu 1-asser II lssan, <:arleton hn(versity, Ottawa, Ont.lrio, ( Inada, \visiting Professor, I>epartnlent c>f (-;ivll anti Ellyironnlental Engineenn:;. ;

Tele: (613) S?()-260(1 Ext. h(i?~, F<t~: (trl3) i?()-39;1, Email: ylwssun~a~ccs.curletom.ca c IDEA Concept and Product The Str3tcgic Highway' Research Program (SHRP) has recognized~a.~"r shear strength as an impor-rant indicator to predict rutting potential of ho.:-min ~
:ISphalt c.eoncrete (Ih\-L1C ) pavements.

however, current methods of measuring the shear strength , of an asphalt mix have been lim-ited to tune consuming, expensive, or unrepresentative laboraiorv"~'e~
analysis. The concept of measuring the in-situ shear strenytlr of an asphalt concrete pavement layer by applying a torque directly to the aurface has keen initiated at C;arleton~~~
L'niversiy in Ottawa, Canada.

IhIS CoIlcCpt allolt'S rel:IClvel\' (Inick meaSllI'clnellt Of Itl-SIClt 517c:1r Strcllgtl7 wlCh a IIlInImrIIn Of d.rlna'e Incurred by th-' . ~ pavement surface. A f~~isic, first generation prototype called the C:arleton In-Situ Shear Strength 'Ccst ((:ISSST) has yielded prornising results relating the maei-mum applied torque to the shear strength of an l Ib(.~\(-: T
layer. Alorc importantly. the CISSS

h:ls determined that shear strengths achieved in tllc field arc very different from those realized In the laboratory. These findings present strong evidence ' that tile devclopnlcnt of such a test .
~

device is required for more accurate measurement and performancef~'~'y modeling of in-service pavement pcrfarmanc=e--the fundamental basis of the 51-1RI' Supcrpave system.

Project Results The CISSST consisted of an electric motor mounted to a cart-like chassis as shown in Figure 1.
a series of gears and driveshafts were used to transmit rotational force to a circular loading plate eposied to the pavement surface. A torque cell and datafot;ger were used to measure torque while angle of twist at failure was measured with a protractor. Two test speeds were available. To prevent rotati<m <7f the test device during testing, r~it steel spikes were attached to the device and driven into the pavement. Test results achieved with the device yielded in-situ shear strengths of up to 300°~ greater than those achieved in parallel laboratory testing.
However, a number of deticic:ncics were noted during initial investigations with the CI;3SST.
Therefore, a comprehensive three-stage investigation was presented to the II>Er\ Program to continue the development of this innovative test device.

rrigure '1 lira (~enm-cuion GIS.SS~I~ f'rotoW pc Test Cnniv~urc ion The pr«ject w<rs initiated in February of 19')9. \\'ithin ~ta~e l, completed hetwcm February and .Iune, three main objectives were addressed. These «bjectives were defined as (i) n: carry c>ut a critical analysis or the existiniC;ISSST t<t~~ilitv and demrmine its main deticicncim, (ii) to prepare a preliminary dcsi~;n o'i a second-,t~eneratic~n shear test t,tcility;
and liii) to develop a framcvcorl< for <t set of analwical models to pre:lict paventcnt pcrf'nrmancc based on field shear cl,tt;t.
t\torc spccificall, tile first objective involved a critical evaluation of the existing Carlet~.~n pro-totvpc. It was evident that problems associated with the wui~;ht of the facility. ~ortaL,ility, smltilization, wpc of epoxy used, and lael: of accurate control and data acquisition were among the most important deficiencies of the (:ISSS'1_. The evaluation excrc~ise prop idol c list of tlcsi~Sn objectives f«r the second-t~eneratien tr.cilitv to si:nitieantlv enhance its pcrforrrtanec.
'I'ha second objective concerned the design of the new facility. 'The research leant at Carleton Gniversitv held a number of internal mectin~;s to discuss a design approach to provide an ol~timurn~coml»nation of the design objectives. 'rhc rcsultin,~ new and improved facility has been dubbed the "In,Si,SST," an acronym for In-,~iru ,~W orr Str-c~yth host.
I3riet'lv, the InSiSST ine<>rporates a trailer-mounted sytcm for simplified transportation.
stable test platform is provided through a solid testing zramc that is lowered to the grc:und via a jaelcin~ system. I'ositic>nin~; slides allow movement of the test motor and gcarbo:c in the longitudinal and transverse dire~~tions. Test measurements are recorded using a torque mounted to the rusting plate, itself epoxied ro the asphalt pavement surface_ Control and c acquisition will he handled using a laptop comp, ~ter and test software. In its current form, InSiSS'T is able to perform five replicate in-situ shear tests each time the test frame is lowe allowing more rapid testing and assuring statist ;:al significance of the results.
The third and final olective of Stage 1 was to create a framework for a set of analytical moc to ultimately predict long-term pavement performance of asphalt pavement surfaces basec results obtained with the InSiSST device. This framework consists of numerous tasks to r matey achieve the kual of performance prediction. These tasks include the use of fin element modetin~ to simulate the asphalt layer and loading conditions imposed by InSiSST, obsen~ing and determining the resulting stress and strain behavior, calibrating finite-element models with simplified closed form solutions and field data, and finally, pred ing long-term pavement performance. The process of analytical me>dclin t; for this problem be an on going process, requiring long-term performance data for model calibration not wit the scope of this investigation. Stage 1 formed the foundation of the framework by present preliminary finite-element models of the pavement surface and various loading conditio Based on the gcocl correlation butu~een the prelirninarv finite element models and closed fo solution, for a simplified case, future modeling efforts will incorporate more complex ~
coelastic material properties found in asphalt c,yncrete. :1s field lust results are gathered Stage 3 of the project, the finite element models may be letter calibrated.
Stage ? of the project commenced in July 1999 and fabrication of the second-generation p totype gust device is currently underway. Once constructed, a regime of "shakedown" testi will be cermpleted to calibrate the test instruments and mote any deficiencies in constructic Final adjustments will then be completed to prepare the device for the final stage of the inm tigationr. Completion of Stage 2 of the project is scheduled for h-larch 20(10.
The third and final stage of the investigation will consist of two sets of field testing; one <)ttawa and the other TRB selecte~J site. ~~nalvsi.s of the results will take on numerous form The first analysis phase will confirm that the field test is repeatable and consistent. Next, t:
field results will be compared to replicate laf>oratc~rv test results to ohsen~e if a common fact or "multiplier", exists between them. 'I'hc sensitivity of in-situ shear strength to critical fa torn such as mix parameters, traffic Icvel and temperature will also tie analysed. Finally, tl results will be used to calibrate the analytical models developed. The project completion da is scheduled for August ?000.
Product Payoff Potential The successful measurement of in-situ IIML:~C, shear properties and the development of a mai stream t<~st facility would yield significant and immediate benet""its to transportation practic These benefits would be realized in three primar~.~ areas of pavement engineering. The fir area is d~aihn. Utilization of the InSiSST in conjunction with laboratory testing would be powerful combination for analyzing; the potential ~ ~f proposed mix designs.
The second area quality ccrrrtrol. Newly constructed asphalt pavements could be tested to verify aceeptab construction practices through the measurement and comparison <>f in-situ strength paran eters with code requirements. The final area is ~o~t~-term pavement ~>ertormance (LTPP
1\fonitoring field shear strength of pavements with time would allow periodic updating of initi performance models to more accurately predict future pavement performance.
This. in turn _ _ _ _ _~ ~_ would allow for more efficient allocation of limited rehabilit<rtion funds and also help deter-mine the effect of real world conditions, such as environmental factors, on pavement perfor-manse Product Transfer The potential for a simple vet e~tremelv effective in-situ test device has already drawn signifi-cant interest from both government and private industry. In addition to IDE.~
Program fund-ing, the Ontario Vlinistrv of Transportation and Regional L-lunieipality of Ottawa-Carleton have committed financial and in-kind support for this investigation. Furthermore, a number of independent consultants have also etpressed interest in the potential of the InSiSS'I'. To facili-tate a more smooth transition into mainstream use, these consultants have continued to be included during the development of the test facility. Their role has initially focused on techni-cal advice: however, the consultants are e.tpected to acquire the technology upon completion of the investigation. Continued evaluation and assessment of market potential will be the primary focus of consultants at that stage Shear Vane for Asphalt Pavement Surface Testin Overview Currently, the In-Situ Shear Stiffness Test (InSiSST) facility developed at Carleton University uses a surface-plate method of testin<~ the .shear properties of an asphalt pavement layer. This surface-plate method entails epoxving a steel test plate to the pavement surface, allowing the epoxy to cure and then applying a ytorsional force to observe various shear properties c>f the asphalt mix (Figure 1 j. For ultimate strength/stiffness testing, th~, asphalt is loaded to failure and the induced failure surface is semi-spherical in shape.
Vertical shaft T
EPOXY I I I Circular disc L~CP
Figure 1: Surface-Plate Method of Load Application for InSiSSTTM (Elevation View) New Technique To improve upon the current practice, a new method of applying the torsional force has been developed. Borrowing from soil mechanics. a shear vane for asphalt concrete has been designed as shown in Figure ?. The use of a shear vane provides two primary benetits over the surface-plate method. First, because the I>lades are embedded within the asphalt layer, no epoav is required. This reduces the time required for testing from hours to minutes.
Second, the shear vane provides a defined, consistent failure plane, unlike the inconsistent faili.ire plane shown in FicYure 1. The defined failure plane allows more accurate calculation of stresses within the asphalt layer, which are critical for determining the desired shear properties.
Like traditional vanes, the vane is formed by two perpendicular blades.
However, unlike a traditional rectangular shear vane used for testing clays, the new shear vane blades are semi-spherical in shape. There were two primay reasons for selecting the semi-spherical shape. First, unlike clays, which are soft and allow a rectangular shear vane to be easily inserted, compacted asphalt concrete is relatively stiff and a shear vane of any configuration could not be simply punched or stamped into the compacted asphalt layer without damaging the asphalt or destroying the vane. Therefore, a saw must be used to cut into the asphalt to form the channels for the vane.
Since cutting blades are circular in shape, the resulting cut is semi-spherical. By matching the vane diameter and shape to the selected saw blade, a perfect match may be achieved in the asphalt pavement.

tfertical shaft T
Vane ~ ~ ~ Circul,ar disc ACP
Figure 2: New Shear Vane developed for lnSiSSTT'~ (Elevation View) The second reason concerns the stress distribution around the steel plate.
During the test. the stresses applied by the torsional force are greatest at the edge of the steel plate and reduce to zero at the centre of the plate. Due to the relatively high stiffness of asphalt concrete. rectangular blades would be subjected to extremely high stresses and would require a very large blade thickness to resist bending or breakage at the tips. For the semi-spherical blades, the blade has minimal cross sectional area at the edges of the steel plate and there is no "tip". Therefore, the blades are not subjected to the same level of stresses and can be made with less thickness.
The actual width and depth of the shear vane can be customized for the particular asphalt mix design in question to accommodate the aggregate distribution in the mix. This is important because the failure depth must be equal to,Vor larger than, the largest aggregate size to ensure that the resulting shear properties are representative of the asphalt mix and not simply the aggregates themselves. Based on the relative cutting widths and depths that can be achieved with a standard l84 mm (7.25 inch) diameter saw blade, the Following standard vane sizes are recommended, as well as the recommended standard asphalt maximum aggre~~ate sizes for asphalt mixes. Different vanes may be produced for different saw blade diameters a s well.
Vane Width (mm) Resulting Vane Suggested Asphalt Mix Depth (mm) ! Maximum Aggregate Size 150 40.6 > 24.5 nun 125 24.5 19.5 mm and ?4.5 mm 100 I 14.8 < 12.5 mrrr As mentioned above, the new vane has been designed for testing existing (or newly compacted) asphalt pavements. However. before the pavement is compacted, the asphalt is soft and the vane can be easily inserted into the mix without damaging the surrounding asphalt.
I Iowever, the vane must be made with sufficient strength to resist the large compactive force of the compaction equipment. In this case the semi-spherical shape is not required to prevent damage to the asphalt, however, it is required to prevent the large stress concentration at the vane tip that would occur with rectangular blades.

Dynamic Testing with InSiSST
Currently, the In-Situ Shear Stiffness Test (InSiSST) facility developed at Carleton University completes a static strength/stiffizess test using a surface-plate method of testing the shear properties of an asphalt pavement layer. This static test involves the application of a constant strain (rate of twist) starting from a state of no applied load, until the asphalt concrete fails. This test provides the ultimate shear strength and stiffness of the asphalt rnix for the conditions experienced durin~~, the test (temperature, etc.).
To increase the amount of information gathered with the InSiSST, dynamic testing capability will likely be added. Dynamic testing involves the application of repeated, cyclic or non-constant rate Ic>ad or strain. The Superpave Shear Tester (SST?, which is a laboraturv shear test facility, uses dwamic testing in the following manner to test asphalt concrete specimens:
Repeated Shear Testing - involving the application a specific number of repeated loads/strains, each consisting of a load/strain pulse of specific duration followed by a relaxation period (no load/strain) to test the resistance of the asphalt mix to permanent strain.
Frequency Sweep Testing - involving the application of a specific number of complete load/strain cycles, each cycle consisting of a load/strain in one direction followed by a load/strain in the opposite direction, again to test the resistance to permanent strain.
The above tests can be completed with the InSiSST device with modification.
Furthermore.. a non-constant rate of strain can be applied to the asphalt surface with the InSiSST in its current configuration.

Claims (12)

1. Apparatus for testing shear properties of a solid planar material in situ comprising:
(a) rotational drive means for applying rotational force to said solid planar material in situ, said rotational drive means being mounted on;
(b) positioning means for positioning said rotational drive means in a plurality of positions against said solid planar material without moving the apparatus, said positioning means being mounted on;
(c) anchoring means preventing movement of the apparatus over the solid planar material; and (d) rotational drive control and rotational drive parameter recording means.

2. Apparatus according to claim 1 wherein said rotational drive means comprises a loading plate anchorable to said solid planar material connectable to an electric motor capable of applying rotational force to said loading plate when said loading plate is anchored to said solid planar material.

3. Apparatus according to claim 2 wherein said loading plate comprises a plate adhesively anchorable to said solid planar material.

4. Apparatus according to claim 2 wherein said loading plate comprises shear vanes embeddable in said solid planar material.

5. Apparatus according to claim 1 wherein said rotational drive means is capable of applying repeated, cyclic or non-constant force to said solid planar material.

6. Apparatus according to claim 5 wherein said rotational drive means is capable of repeated shear testing.

7. Apparatus according to claim 5 wherein said rotational drive means is capable of frequency sweep testing.

8. Apparatus for testing shear properties of a solid planar material in situ comprising:
(aa loading plate anchorable to said solid planar material connectable to;
(b) rotational drive means for applying rotational force to said loading plate when said loading plate is anchored to said solid planar material, said rotational drive means being mounted on;
(c) positioning means for positioning said rotational drive means so that it is connectable to said loading plate when said loading plate is anchored to said solid planar material, said positioning means additionally being capable of connection serially to a plurality of said loading plates anchored on said solid planar material without moving the apparatus, said positioning means being mounted on;
(d) anchoring means preventing counter-rotational movement of said rotational drive means when said rotational drive means applies rotational force to said loading plate when said loading plate is anchored, and (e) rotational drive control and rotational drive parameter recording means.

9. Apparatus for testing shear properties of asphalt concrete pavement in situ comprising:
(a) a loading plate anchorable to asphalt concrete pavement connectable to;
(b) an electric motor capable of applying rotational force to said loading plate when said loading plate is anchored to said asphalt concrete pavement, said electric motor being mounted on;
(c) a positioning slide, said positioning slide being mounted on;

(d) a testing frame jackably anchorable to said asphalt concrete pavement, said positioning slide allowing said electric motor to be moved substantially horizontally in at least one dimension over said asphalt concrete pavement when said testing frame is anchored to said asphalt concrete pavement to allow said electric motor to be connected sequentially to more than one loading plate anchored to said pavement without moving the apparatus over said pavement; and said electric motor being operably connected to;
(e) electric motor control and electric motor parameter recording means.

10. Apparatus according to claim 9 wherein said loading plate comprises a plate adhesively anchorable to asphalt concrete pavement.

11. Apparatus according to claim 9 wherein said loading plate comprises shear vanes embeddable in said asphalt concrete pavement.

12. A loading plate for use with apparatus for testing shear properties of a solid planar material in situ comprising a plurality of semi-spherical shear vanes for embedding in said solid planar material attached to a drive shaft for applying rotational force to said plurality of shear vanes.