A
.. , It... , .r ..
rwnyflffim

_.‘?§ :14}?

”9.4%“ .
Fbwfi .
1N ..

ha:
drain

..
..
ll“
ifimrgfiwa
253., .5...
hasvduéfla

..
u¢

. .W r.
. Ptiflih

1!... {2.3.2.

ifiww

an."

.7.

u.‘ v5.3.3.1...
I... .: 11......
p 5.0....Hurfl .

 

 

gm LIBRARY”
2007 MIChIQsfl State
University

 

 

 

This is to certify that the
dissertation entitled

EVALUATION OF PORTLAND CEMENT CONCRETE
COEFFICIENT OF THERMAL EXPANSION TEST PROTOCOL
AND THE IMPACT OF CTE ON
PERFORMANCE OF JOINTED CONCRETE PAVEMENTS

presented by

SHERVIN JAHANGIRNEJAD

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Civil Engineering

 

 

 

 

”Major Professor’s Signature
6 I 14 I 197
I l I

Date

MSU is an Affirmative Action/Equal Opportunity Employer

 

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KglProleccsPresIClRC/Daleoue.indd

 

EVALUATION OF PORTLAND CEMENT CONCRETE COEFFICIENT OF
THERMAL EXPANSION TEST PROTOCOL AND THE IMPACT OF CTE ON
PERFORMANCE OF JOINTED CONCRETE PAVEMENTS
By

Shervin J ahangirnej ad

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY

Civil Engineering

2009

ABSTRACT
EVALUATION OF PORTLAND CEMENT CONCRETE COEFFICIENT OF
THERMAL EXPANSION TEST PROTOCOL AND THE IMPACT OF CTE ON
PERFORMANCE OF J OINTED CONCRETE PAVEMENTS
By
Shervin J ahangirnej ad

A number of factors affect the magnitude of Portland Cement Concrete (PCC) coefficient
of thermal expansion (CTE). The selected factors in this study were aggregate geology,
age of the specimens, and the number of heating-cooling cycles applied to the test
specimens. The effects of these variables on the measurement of CTE and the practical
implications of the variables on the performance of jointed concrete pavements were
investigated in this study.

A laboratory investigation was conducted to determine the CTE of typical
concrete paving mixtures made with coarse aggregate from eight different sources in
Michigan. The primary aggregate class included limestone, dolomite, slag, gravel, and
trap rock. Three replicate test specimens were fabricated for each mixture-age
combination. Over 700 specimens were fabricated to characterize the mechanical
properties and CTE of the concrete specimens. The test specimens were moist-cured for
3,7, 14, 28, 90, 180, and 365 days prior to testing. Mechanical properties tests including
compressive strength, flexural strength, splitting tensile strength, and elastic modulus
tests were conducted to check the quality and specification conformance of the concrete
mixtures. The CTE was determined using two test methods; the provisional AASHTO

TP60 protocol and a revised method based on the TP60 protocol. The measured average

28-day CTE values ranged from 4.51 to 5.92 us/°F (8.11 to 10.65 us/°C).

Statistical analyses were employed to evaluate the significance of the factors
affecting CTE magnitude. These analyses included analysis of variance (ANOVA),
which was performed using factorial treatment design and mixed effects models. The
results of these analyses indicated that aggregate geology, specimen age at the time of
testing, and the number of heating-cooling cycles that the Specimen is subjected to, have
a statistically significant impact on the magnitude of measured CTE.

Sensitivity analyses were conducted to investigate the practical impact of CTE
and other variables on short-term (first 72-hours) and long-term performance of jointed
concrete pavements using HIPERPAV II and M-E PDG software, respectively. The
impact of FCC CTE among other pavement design variables (slab thickness, joint
spacing, etc.) on pavement design and performance was investigated. The long-term
impacts of the factors that were analyzed statistically are also discussed in this
dissertation.

Recommendations are made to improve testing of the CTE based on the results of

all analyses and suggested CTE values for agency use in Michigan were presented.

Com/right by
SHERVIN JAHANGIRNEJAD

2009

Dedicated to my family specially to my wzfe

ACKNOWLEDGEMENTS

I would like to express my deepest gratitude to my advisor Dr. Neeraj Buch, for the
Opportunity to work on this project and for his guidance and advice during the course of
my Ph.D. studies.

I wish to thank Dr. Parviz Soroushian and Dr. Karim Chatti for their advice and
comments during this study.

Special thanks to Dr. Alexandra Kravchenko for helping me and advising me
every step of the way on statistical analyses.

I would like to thank Mr. Siavosh Ravanbakhsh for his unconditional support,
cooperation, and kindness. I owe my laboratory testing experience to his instructions.

I would like to acknowledge the financial support provided by the Michigan
Department of Transportation for this project.

Finally, I would like to express my gratitude to my family and specially to my wife who

understood me, supported me, and made sacrifices during the term of my studies.

vi.

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LIST OF TABLES ix
LIST OF FIGURES xii
CHAPTER 1
INTRODUCTION 1
1.1 Introduction 1
1.2 Problem Statement 5
1.3 Research Objectives and Plan 6
1.3.1 Research Objectives 6
1.3.2 Research Plan 6
1.4 Dissertation Organization 7
CHAPTER 2
Literature Review 8
2.1 Introduction 8
2.2 Literature on Micro-scale CTE Mechanisms in PCC 9
2.3 Literature on Variables Affecting CTE Value 15
2.4 Literature on CTE Impact on Pavement Performance 28
2.5 Application of the Reviewed Literature to the Research Plan 44
CHAPTER 3
EXPERIMENTAL PROGRAM 46
3.1 Introduction 46
3.2 Materials 47
3.2.1. Aggregates 47
3.2.2. Concrete 51
3.3 Concrete Properties Tests 53
3.4 Thermal Property Test (CTE Test) 55
3.4.1 Controlled Temperature Water Bath 55
3.4.2 Data Acquisition System 57
3.4.3 Linear Variable Differential Transformer (LVDT) 57
3.4.4 Rigid Support Frame 59
3.4.4.1 Frame Apparatus 59
3.4.4.2 Frame Calibration 60
3.4.5 Test Procedures, Data Collection, and CTE Calculations 61
3.4.5.1 Specimen Conditioning 61
3.4.5.2 AASHTO TP60 Test Procedure, Data Collection, and CTE
Calculation 61
3.4.5.3 Revised TP60 Test Procedure, Data Collection, and CTE
Calculation 63
3.4.5.4 Executed Test Procedure 64

vii

’;

CHAPTER 4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

RESULTS AND DISCUSSION 67
4.1 Introduction 67
4.2 Physical Properties of Coarse Aggregates 68
4.3 Fresh Concrete Properties 69
4.4 Hardened Concrete Properties 70
4.5 Thermal Property 73
4.6 CTE Test Variability 80
4.7 Statistical Analyses 81

4.7.1 Statistical Analysis of Factors Affecting CTE of PCC 81
4.7.1.1 Aggregate Type Effect 87
4.7.1.2 Sample Age Effect 88
4.7.1.3 Number of Heating-Cooling Cycles Effect 90

4.7.2 Comparison of AASHTO TP60 and Revised TP60 Methods 92

4.8 Impact of CTE on Performance of Jointed Concrete Pavements 96

4.8.1 Short-Term Effects Analysis 97

4.8.2 Long-Term Effects Analysis 101
4.8.2.1 Effect of CTE Variability on JPCP Performance 101
4.8.2.2 Effect of Coarse Aggregate Type on JPCP Performance ------------- 104
4.8.2.3 Effect of CTE at Different Ages on JPCP Performance ------------- 106
4.8.2.4 Effect of Number of Cycles on JPCP Performance 109
4.8.2.5 Effect of CTE among Other Design Features on JPCP Performance113

CHAPTER 5

SUMMARY, CONCLUSIONS, RECOMMENDATIONS, AND FUTURE RESEARCH

NEEDS 115
5.1 Introduction 115
5.2 Summary of the Performed Work 116
5.3 Factors Affecting Measurement of CTE and Impact of CTE on Pavement
Performance 1 17

5.3.1 Factors Affecting Measurement of CTE 117

5.3.2 Impact of the CTE on Pavement Performance 118

5.4 Recommendations 119
5.4.1 CTE Testing Recommendations 119
5.4.2 Recommended CTE Values for Agency Use in Michigan 119
5.5 Future Research Needs 121

APPENDICES

APPENDIX A. HARDENED CONCRETE PROPERTIES DATA 123

APPENDIX B. COEFFICIENT OF THERMAL EXPANSION DATA 132

 

REFERENCES 142

 

viii

LIST OF TABLES

Table 1-1. Influence of CTE on Pavement Performance

 

Table 2-1. Mean, Maximum, and Minimum CTE at Different Sites

 

Table 2-2. Predicted Distresses for Computed and Measured PCC CTE Values for

 

Typical Aggregates in Kansas

 

Table 3-1. Coarse Aggregate Types and Source Names

Table 3-2. Mineralogical and Physical Properties of the Coarse Aggregate .............

 

Table 3-3. Petrographic Composition of Slag and Gravel Aggregates

 

Table 3-4. Chemical Composition of Gabbro Aggregate

Table 3-5. Concrete Mixture Designs

 

 

Table 3-6. Summary of Material Characterization Tests

 

Table 4-1. Physical Properties of Coarse Aggregates

Table 4-2. Fresh Concrete Properties

 

 

Table 4.3 Average 28-day CTE Results

 

Table 44. Typical CTE Ranges for Common Components and Concrete

Table 4-5. Physical Properties of the FHWA Coarse Aggregates

 

 

Table 4-6. CTE Test Results for F HWA Specimens

Table 4-7. Factorial Design Table

 

 

Table 4-8. Description of Terms in the Mixed Effects Model

Table 4-9. Tests of Fixed Effects

 

Table 4-10. Factorial Design Table

 

 

Table 4-11. Description of Terms in the Model for Method Comparison

Table 4-12. Tests of Fixed Effects for Method Comparison

 

ix

23

43

47

49

50

50

51

54

68

69

74

74

79

79

82

83

87

93

94

95

Table 4-13. Input Variables for HIPERPAV II Sensitivity Analysis 98

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 4—14. Examples of HIPERPAV II Sensitivity Analysis Inputs 99
Table 4-15. Examples of HIPERPAV II Sensitivity Analysis Results 99
Table 4—16. Input Variables for M-E PDG Sensitivity Analyses 102
Table 4-17. Percent Slabs Cracked for Mixtures with Different Aggregates ------------- 105
Table 4-18. Percent Slabs Cracked Based on CTE Values for Different Ages ----------- 107
Table 4-19. Performance Criteria for Rigid Pavements 109
Table 4-20. Percent Slabs Cracked Based on CTE Values for Different Cycles --------- 110
Table 4-21. Variables for Designs with Maximum Percent Slabs Cracked 113
Table 5-1. Recommended CTE Values for Concrete Made With Different Coarse

Aggregate Used In Michigan 120
Table A-1. Hardened Concrete Properties Table for CTE 1 Mixture 124
Table A-2. Hardened Concrete Properties Table for CTE 2 Mixture 125
Table A-3. Hardened Concrete Properties Table for CTE 3 Mixture 126
Table A-4. Hardened Concrete Properties Table for CTE 4 Mixture 127
Table A-5. Hardened Concrete Properties Table for CTE 5 Mixture 128
Table A-6. Hardened Concrete Properties Table for CTE 6 Mixture 129
Table A-7. Hardened Concrete Properties Table for CTE 7 Mixture 130
Table A-8. Hardened Concrete Properties Table for CTE 8 Mixture 131
Table B-1. CTE Values for CTE 1 Mixture 133
Table B-2. CTE Values for CTE 2 Mixture 134
Table B-3. CTE Values for CTE 3 Mixture 135
Table B-4. CTE Values for CTE 4 Mixture 136
Table B-5. CTE Values for CTE 5 Mixture 137

 

Table B-6. CTE Values for CTE 6 Mixture 138

 

Table B-7. CTE Values for CTE 7 Mixture 139

 

Table B-8. CTE Values for CTE 8 Mixture 140

 

xi

LIST OF FIGURES

Figure 2-1. Estimated Typical Response to a Step Input of Temperature for the Three

 

Different Components of Thermal Dilatation, at Various Humidities h 12

Figure 2-2. Dilatation of Solid Microstructure Induced by Relaxation with Increasing

 

 

 

 

 

 

 

 

 

 

 

 

Temperature 12
Figure 2-3. Family of Thermal Dilatation Curves for Various Humidities 13
Figure 24. Variation of CTE over Time 17
Figure 2-5. Effect of Coarse Aggregate Volume in Concrete on CTE 18
Figure 2-6. Variation of CTE From 32 Aggregate Sources in Texas 19
Figure 2-7. Effect of Repeated Thermal Cycles on CTE, (a) Increase in R2, (b)

Decrease in CTE 20
Figure 2-8. Effect of Saturation on CTE on Two Oven-Dried Cores 22
Figure 2-9. CTE Variability at High Saturation Levels 23
Figure 2-10. CTE of Concrete Made with Gravel, Quartzite, Granite, Diabase, or

Basalt 25
Figure 2-11. CTE of Concrete Made with Dolomite and Different Cementitious

Materials 25
Figure 2-12. Histogram of the Mean CTE of the Specimens 27
Figure 2-13. Histogram of the OCTE 27
Figure 2-14. Effect of PCC CTE and its Variability on the M-E PDG Predicted --------- 30
Percent Slabs 30

 

Figure 2-15. Effect of CTE and its Variability on the M-E PDG Predicted Mean Joint
Faulting 30

 

Figure 2-16. Effect of CTE and its variability on the M-E PDG predicted IRI ----------- 31

xii

Figure 2-17. Difference in the Predicted Percent Slabs Cracked as a Result of the

 

dCTE 32
Figure 2-18. Difference in the Predicted Faulting as a Result of the dCTE ---------------- 33
Figure 2-19. Difference in the Predicted IRI as a Result of the dCTE 34

 

Figure 2-20. Effect of Constituent Properties on Pavement Performance Predictions --- 35

 

 

 

Figure 2-21. Examples of FSC, TSC, and CC 37
Figure 2-22. Comparison of Ratio of Cracked Slabs for Pavements with High and Low

CTE (F SC and TSC) 38
Figure 2-23. Comparison of Ratio of Cracked Slabs for Pavements with High and Low

CTE (FSC+TSC and CC) 39
Figure 2—24. Calculated and Measured PCC CTE Values (x 10'6/°F) for Kansas -------- 41
Figure 2-25. Calculated and Measured PCC CTE Values (x 10'6/°F) for Iowa ----------- 41

Figure 2-26. Calculated and Measured PCC CTE Values (x 10'6/°F) for Missouri ------ 42

 

 

 

 

 

 

 

 

 

 

Figure 3-1. Locations of the Aggregate Sources 48
Figure 3—2. Schematic of the Test Setup 55
Figure 3-3. Complete Test Setup 56
Figure 3-4. Controlled Temperature Water Bath 56
Figure 3-5. Data Acquisition System 57
Figure 3-6. DaqViewTM Software Channel Setup Screen 58
Figure 3-7. Spring-Loaded LVDTS 58
Figure 3-8. Rigid Support Frame 59
Figure 3-9. Side View (Left) and Plan View (Right) of the Rigid Support Frame -------- 60
Figure 3-10. A Typical Revised TP60 Method Graph 64
Figure 3-11. Three Typical Heating-Cooling Cycles 65

 

Figure 3-12. A Typical Concrete and Water Temperature Graph 66

xiii

Figure 4-1. Average 28-day Compressive Strength Values 71

 

 

 

 

 

 

Figure 4-2. Average 28-day F lexural Strength Values 71
Figure 4-3. Average 28-day Elastic Modulus Values 72
Figure 4-4. Average 28-day Split Tensile Values 72
Figure 4.5 Average 28-day CTE Results 73
Figure 4-6. CTE 1 (Limestone Concrete) Coefficient of Thermal Expansion ------------- 75
Figure 4-7. CTE 2 (Gravel Concrete) Coefficient of Thermal Expansion 75

Figure 4-8. CTE 3 (Dolomitic Limestone Concrete) Coefficient of Thermal Expansion 76

 

 

 

 

 

 

 

 

 

 

 

Figure 4-9. CTE 4 (Slag Concrete) Coefficient of Thermal Expansion 76
Figure 4-10. CTE 5 (Dolomite Concrete) Coefficient of Thermal Expansion ------------- 77
Figure 4-11. CTE 6 (Gabbro or Trap Rock Concrete) Coefficient of Thermal

Expansion 77
Figure 4-12. CTE 7 (Dolomite Concrete) Coefficient of Thermal Expansion ------------- 78
Figure 4-13. CTE 8 (Dolomite Concrete) Coefficient of Thermal Expansion ------------- 78
Figure 4-14. Variability Histogram 80
Figure 4-15. Histogram of the Residuals 85
Figure 4-16. Normal Probability Plot of the Residuals 85
Figure 4-17. Side-By—Side Box Plots of the Residuals for Age Factor 86
Figure 4-18. Average CTE for Various Aggregates as a Function of Test Specimen Age
Based on 3 Replicates 89
Figure 4-19. Average CTE for Various Aggregates as a Function of Number of Cycles
Based on 3 Replicates 91
Figure 4-20. HIPERPAV 11 Performance for CTE 1 Designs 100
Figure 4-21. HIPERPAV II Performance for CTE 8 Designs 100
Figure 4-22. Percent Slabs Cracked for CTE Test Variability 104

xiv

 

Figure 4-23. Percent Slabs Cracked for Mixtures with Different Aggregates ------------ 105

Figure 4-24. CTE] Cracking Performance with CTE Values at Different Ages --------- 108
Figure 4-25. CTE2 Cracking Performance with CTE Values at Different Ages --------- 108
Figure 4-26. CTES Cracking Performance with CTE Values at Different Ages --------- 109

Figure 4-27. CTEl Cracking Performance with CTE Based on Different Cycles ------- 111
Figure 4-28. CTE2 Cracking Performance with CTE Based on Different Cycles ------- 112

Figure 4-29. CTES Cracking Performance with CTE Based on Different Cycles ------- 112

XV

 

CHAPTER 1

INTRODUCTION

1.1 Introduction

The coefficient of linear thermal expansion is defined as the change in unit length per
unit change in temperature. It is referred to as the coefficient of thermal expansion (CTE)

or “or” in this dissertation. The general form of the CTE equation is expressed as

CTE = [£]/ AT
L0

in which AL is the length change, L0 is the original length, and AT is the temperature

change. CTE is usually expressed in microstrain (106) per degree Celsius (ue/°C) or

microstrain per degree Fahrenheit (us/0F).

The CTE of Portland Cement Concrete (PCC) is an important parameter — among
other factors — in the development of thermally-induced stresses and deformations in
concrete pavements during the first 72 hours after paving and over the design life. For
example, CTE affects mean crack spacing in continuously reinforced concrete pavements
(CRCPS). Another example is curling stress in the x direction of a finite slab expressed as

EaAT
- ———(Cx + va)

csx — 2
2(1—v )
in which ox is curling stress in the x direction, or is the CTE of concrete, E is the elastic

modulus of concrete, AT is the temperature differential between top and bottom of the

 

slab, v is the Poisson’s ratio of concrete, and CK and Cy are correction factors for a finite

slab.

CTE also influences thermal deformations in slabs which in turn impact crack
width in CRCPS, joint spacing (slab length), joint sealant reservoir design, and joint
movement in jointed plain concrete pavements (J PCPs). Joint opening for example is
affected by CTE. It is expressed as

AL = CL(0IAT + s)
in which AL is the joint opening, or is the CTE of concrete, C is the adjustment factor due
to slab-subbase fi'iction, L is the joint spacing, AT is the temperature range, and 8 is the

drying shrinkage coefficient of concrete.

The above-mentioned stresses and deformations cause different types of distresses

in pavements which consequently affect pavement performance. Mallela et al.1 have
listed pavement distresses that are affected by CTE (Table 1.1).

Another distress which is related to CTE is the blow-up in concrete slabs at a joint
or crack. This type of distress is due to insufficient space for slab expansion. All the
mentioned distresses can potentially affect another aspect of pavement performance

which is the quality of ride measured by the International Roughness Index (IRI).

”2, CTE is an important factor in

According to the “Guide to Concrete Overlays
design and construction of bonded concrete overlays of concrete pavements. The PCC
overlay and existing pavement should move as one structure in order to achieve desired
performance. The unified movement of the two layers will reduce the shear stress at the

interface and is achieved if the CTE of the overlay is similar to the CTE of the existing

pavement.

 

Table 1-1. Influence of CTE on Pavement Performance (After Mallela et 111.1)

 

Pavement Distress

Role of CTE

 

Early-age or premature random cracking
due to excessive longitudinal Slab
movement

High CTE causes excessive longitudinal
movement that, if restrained by restraint
forces, leads to cracking.

 

Higher curling (temperature-induced)

 

 

 

 

 

Mid-panel fatrgu e cracking stresses lead to higher mid-panel cracking.
Higher corner deflections due to negative

Faulting curling, which is a function of temperature
gradients and CTE.

. . Excessive joint opening and closing causes

Jornt spallmg joint sealant failure and joint spalling.
Crack load transfer efficiency is affected

Punchouts in CRCP by spacing and width of cracks which are

impacted by CTE.

 

 

Based on the above discussion, quantification of the concrete CTE is a necessary

step in concrete pavement design.

The “Mechanistic-Empirical Pavement Design Guide”3 (M-E PDG) allows for the

input of CTE at three levels (quality of data). Level 1 for CTE determination employs

direct measurement in accordance with a provisional test protocol developed by the

American Association of State Highway and Transportation Officials (AASHTO)

designated as AASHTO TP60, "Standard Test Method for the Coefficient of Thermal

Expansion of Hydraulic Cement Concrete

. Level 2 for CTE calculation uses a weighted

average of the constituent CTE values based on the relative volumes of the constituents

and is expressed as

aconcrete = (aaggregate

XV

aggregate) + (a paste x Vpaste)

in which a is the CTE of the constituent and V is the volumetric proportion of the
constituent in the PCC mixture. Level 3 for CTE estimation is based on historical
averages.

The greatest potential for error is associated with Level 3 data quality, because
PCC materials vary considerably. Realistic data for the types of materials being used in
concrete mixtures are rarely available and, if they are available, they are likely to be
based on specific PCC mixtures.

The importance of CTE in pavement design and performance warrants the choice of an

appropriate method for CTE determination for concrete pavement projects.
Quantification of concrete CTE based on Level 1 is the most reliable method especially

for concrete pavement mixtures that are composed of locally supplied aggregates.

1.2 Problem Statement

The recently completed M-E PDG uses CTE as one of the parameters to characterize the
thermal properties of PCC paving mixture. CTE of the concrete pavement is one of the
key inputs for computing response parameters such as curling stresses and joint
movement for JPCP as well as crack spacing and width for CRCP. These response
parameters in turn influence performance prediction (cracking, faulting, punch-outs, and
Currently, the Michigan Department of Transportation (MDOT) does not call for the
determination of CTE for the design of concrete pavements. However, CTE has a
significant bearing on the computation of concrete pavement response and performance
prediction in the new M-E PDG. For the successful implementation of the new design
procedure the determination of CTE is necessary. Details about CTE testing and
reporting the values should also be described. Finally, the effect of CTE on pavement

performance Should be investigated.

1.3 Research Objectives and Plan

1.3.1 Research Objectives

The objectives of this research were:

Quantification of the concrete CTE composed of different aggregate types used
by MDOT

Evaluation of the AASHTO TP-60 test protocol

Studying variables that affect the direct measurement of concrete CTE

Investigation of effects of CTE on pavement performance

1.3.2 Research Plan

The research Objectives were achieved by executing the following plan:

Development of a test matrix representing the various mixtures for the state of
Michigan and different test conditions

Execution of the test matrix by casting the required concrete samples for CTE and
concrete quality control tests

Reporting the results obtained from the testing phase

Statistical and practical significance analyses of the test results

Conducting sensitivity analyses by using mechanistic-empirical prediction models
(HIPERPAV II and M-E PDG software)

Presenting the findings of the conducted analyses

Recommendation of input ranges for the execution of the new design guide

 

1.4 Dissertation Organization

A review of the literature regarding CTE of concrete, variables affecting the CTE, and the
effect of CTE on pavement performance is presented in Chapter 2 of this dissertation.

In Chapter 3, the experimental program is presented in detail which includes
information about testing equipment and various tests carried out on concrete samples.

The results are presented and discussed in Chapter 4. This chapter includes testing
results and the results of statistical, practical significance, and sensitivity analyses and
discussions.

Chapter 5 includes the summary, conclusions, recommendations (test method and
CTE values), and future research needs.

Appendices and references are presented at the end respectively.

CHAPTER 2

LITERATURE REVIEW

2.1 Introduction

Several research investigations have been conducted to study the concrete CTE at micro-
scale levels. A limited number of studies have been conducted to identify variables that
have an influence on the magnitude of CTE, to evaluate CTE variability, and to
investigate the effects of PCC CTE on concrete pavement performance. A number of
these limited studies only investigate the variables that affect CTE and/or CTE
variability. Some other papers only study the effect of CTE on pavement performance,
and there are some studies that examine both aspects.

The literature review presented in this chapter is divided into three sections. The
first section summarizes the studies regarding micro-scale mechanisms of concrete CTE.
The second section summarizes information from literature investigating the impact of
test variables on the magnitude of CTE and/or the CTE variability. The third section
summarizes the impact of CTE on pavement performance of j ointed concrete pavements.

At the end of this chapter, the findings from literature review and their application

to this research plan are presented.

2.2 Literature on Micro-scale CTE Mechanisms in PCC

The CTE of concrete depends on thermal deformations in hardened cement paste (HCP)
and in aggregate. The thermal deformations in concrete are tied with its moisture

conditions. The CTE of aggregates is not influenced by changes in relative humidity

(RH) as it is in the case of HCP according to Meyerss. The thermal expansion (dilatation
or dilation) mechanism in HCP has been the subject of several studies some of which are
presented in this section.

Since the micro-scale CTE mechanisms take place in HCP pore system, this pore
system and the water held in it are reviewed before discussing the moisture conditions.

Pores in HCP can be classified as capillary pores and gel pores based on their
size. The diameter of the capillary pores range between 10 and 10000 nanometers while
the diameter of the gel pores range between less than 0.5 and 10 nanometers. Larger
capillary pores can hold free (bulk) water while smaller capillary pores can hold water
with surface tension forces acting on the water. Larger gel pores can also hold water with

large surface tension forces acting on it while smaller gel pores contain adsorbed or

interlayer water with no surface tension acting on the water (Mindess et al.6).

There are three moisture states in HCP; dry, saturated, and partially saturated. In
. dry (bone—dry) state, HCP does not contain any evaporable water. The term “evaporable
water” is applied to all classes of water except chemically-bound water. In fully saturated
state, all types of pores are filled with water. In partially saturated state, capillary pores
are partially filled with water (partially-bound or surface adsorbed water) and gel pores

are partially or completely filled with water depending on the RH of HCP.

Different mechanisms that cause thermal deformations in HCP and the relations

between these mechanisms and various moisture states are discussed below.

Bazant7 differentiated three components in thermal dilation of HCP. These
components are pure thermal dilation, thermal shrinkage or swelling, and hygrotherrnic
dilation.

Pure thermal dilation is due to the CTE of HCP constituents (solid particles and
different classes of water e.g. free water, adsorbed water, etc.). Dry pastes experience an
immediate deformation. For saturated and partially saturated pastes, the immediate
deformation is followed by a delayed deformation opposite to the direction of the
immediate deformation.

For example, rising temperature causes immediate expansion of layers containing
adsorbed water. It is then followed by gradual flow of the adsorbed water out of the layer
— causing contraction of the layer — until the pressure equilibrium is achieved. This
behavior is illustrated in part (a) of Figure 2-1. Letter “h” in this figure refers to RH of
HCP, and “a” is volumetric strain.

Thermal shrinkage or swelling is caused by redistribution of water between gel
pores and capillary pores. Dry pastes do not undergo this type of deformation. For
saturated and partially saturated pastes, this type of deformation occurs in opposite
direction to the pure thermal dilation.

For instance, an increase in temperature results in higher chemical potential in gel
water than that of capillary water. In order for the system to be in equilibrium, water from

gel pores move into capillary pores and the result is the shrinkage of the gel pores

10

(Helmuthg). This mechanism is classified as a delayed deformation. This mechanism is
shown in part (b) of Figure 2-1.

Hygrothermic dilation is associated with change in the RH of the pores at constant
water content. Dry pastes and saturated pastes do not experience this type of deformation.
This phenomenon takes place in specimens that are partially saturated where air-water
menisci exist'in capillary pores. A change in RH results in a change in capillary tension
which leads to deformation of HCP.

For example, an increase in temperature causes expansion of the capillary water
which increases the radius of the meniscus and subsequently increases the RH. The
increase in temperature also causes reduction of the tension at the air-water interface

which again leads to an increase in RH. These two mechanisms are illustrated in Figure

2-2 (Grasley and Lange9). Another possible reason for a raise in RH is the thermal
shrinkage or swelling mechanism explained before. An increase in capillary water
content (supplied by gel pore upon temperature increase) increases the meniscus radius
and RH.

This mechanism is recognized as an immediate deformation by Sellevold and

Bjrzmtegaard10

while Baiant7 identifies a delayed deformation component in addition to
an immediate deformation component in this mechanism. This type of dilation is

presented in part (c) of Figure 2-1.

11

STEP INPUT OF

 

 

 

 

 

 

 

 

 

 

 

 

  
 

I TEMPERATURE
To IL
I
\ h =10 (a) PURE THERMAL
uu'n
\ h = 0.7 DI A ON
h = 9'9
0 a1.
5 h = 0.0 t
’(b) THERMAL
h = 0.7 SHRINKAGE
h = 1.0
E A h = 0.7 _‘
(c) chRoTHERMIc
muuAnou

h=0.4ORh=O.85J_

 

h=0.00Rh=1.0 t

.-

 

Figure 2—1. Estimated Typical Response to a Step Input of Temperature for the
Three Different Components of Thermal Dilatation, at Various Humidities h

(Baiant7)

RH=80%

RH= 75%

Increasing
temperature

==9

 
 

Figure 2-2. Dilatation of Solid Microstructure Induced by Relaxation with
Increasing Temperature (Grasley and Lange9)

12

Combination of the three dilation mechanisms explained above may demonstrate
the total thermal deformation behavior of HCP. Figure 2-3 shows an estimation of the
combined behavior of HCP with different RH values. Letter “h” in this figure refers to

the RH, and “a” is volumetric strain.

T

h=0.7
h=0.4
11:1.0

h

0.0

 

Figure 2-3. Family of Thermal Dilatation Curves for Various Humidities (Baiant7)
It can be seen form the above figure that dry paste has the lowest thermal dilation,
saturated paste has slightly higher delayed dilation (but close to that of the dry paste), and

partially saturated paste at RH of 70% has the highest dilation.

l3

Other researchers have also compared the thermal deformations of pastes with

these three different moisture conditions. Powers and Brownyard11 stated that the thermal
coefficients for dry and saturated pastes are about equal and are less than the coefficient

of a partially saturated paste. The same observation was reported by Bonnell and
Harper12 . Helmuth8 also reported higher CTE values for partially saturated pastes

compared with fully saturated pastes. Meyers5 stated that the thermal coefficient is at a

minimum in dry and saturated states and that the coefficient is at maximum when the RH
of paste is near 70%. Sabri and Illston13 reported a maximum deformation at 75% RH,

and similar minimum deformations for dry and saturated states. Yeon et a1.14 stated that
the maximum CTE of cement paste — and also concrete — occurs at RH values about 70%

to 80%.

14

2.3 Literature on Variables Affecting CTE Value

In this section studies (or part of studies) that investigate the variables affecting CTE are
presented. Brief descriptions of experimental plans are presented where available. Results

and discussions of each study are also presented.

In a laboratory study conducted by Alungbe et al.15 in 1992, effects of aggregate
type, water to cement ratio, curing, and specimen condition on the magnitude of CTE
were investigated. Three types of aggregate were investigated as part of this study.
Porous limestone, dense limestone, and river gravel. Three combinations of water to
cement ratio and cement content were studied as well as two curing durations (28 and 90
days). Another variable was the specimen condition with two levels, water-saturated, and
oven-dried. A length comparator was used to measure the length changes of specimens.
Specimens were square prisms with these dimensions: 3 in.><3 in.><11.25 in.

In this study, the concrete samples with porous limestone as coarse aggregate, had
a CTE that ranged from 5.42 to 5.80 ps/°F (9.76 us/°C to 10.44 us/°C); concrete samples
produced from dense limestone had a range of 5.82 to 6.14 ua/° F (10.48 ue/°C to 11.05
Its/0C), and concrete samples made of gravel coarse aggregate had a CTE range of 6.49 to
7.63 Ire/(T (11.68 Its/0C to 13.73 us/OC). A statistical analysis (factorial design) was used
to study the effect of different variables on CTE magnitude. Based on statistical analysis
results, the authors concluded that aggregate type affects the CTE value, but water to
cement ratio and cement content have “no effect” on the CTE. The water-saturated
specimen had lower CTE values compared to oven—dried samples. There was no

significant difference between samples with different curing durations in water-saturated

IS

specimens. However, the CTE values of the 28-day cured specimens were higher than the

value of 90-day cured samples in oven-dried specimens.

Won16 at the University of Texas, Austin evaluated the effect of coarse aggregate
content and 32 different aggregate types on the CTE and the effect of sample age, rate of
heating-cooling cycle, and size of specimen on measured CTE values. As part of this
study, he suggested improvements to the AASHTO TP60 method.

The paper stated that the accuracy and repeatability of this test procedure greatly
depends on the stability and accuracy of the displacement readings at 50 and 122 °F (10
and 50 °C). As an alternative, it was suggested that the correlation between temperature
and displacement changes he used for determination of CTE which results in a repeatable
and more accurate CTE test procedure. The testing apparatus and specimen conditioning
is the same as in TP60, but the temperature and linear variable differential transformer
(LVDT) displacement readings are recorded every minute. The CTE calculation method
is also different from the TP60 method and is based on a regression analysis between
temperature and displacement readings. It was stated that with the revised procedure, the
difference between heating and cooling CTE values is smaller than the difference based
on TP60 method resulting in a more accurate and repeatable method in comparison with
AASHTO TP60 method.

The effect of concrete age was also investigated. Concrete cylinders were tested
over a period of 3 weeks and it was found that the age of concrete had little effect on

CTE for up to three weeks. This is illustrated in Figure 2-4.

16

CTE (mlcrostralnIC)

 

o 2 4 6 8 10 12 14 16 18 20 22
Time (days)

Figure 24. Variation of CTE over Time16

The effect of the rate of heating and cooling was studied. Two difi‘erent rates were
applied on a specific test specimen. The CTE value of the slow rate (0.93 °F/hr or 1.67
°C/hr) was found to be 5.82 ua/°F (10.48 WC), and for the fast rate (14.83 °F/hr or 26.7
°C/hr) it was found to be 5.87 ua/‘T (10.57 118/°C). It was concluded that the rate of
heating and cooling has little effect on the CTE value.

For the effect of the coarse aggregate content on the CT'E value, the experimental
results indicated an almost linear relationship between the %volume of coarse aggregate
in the PCC mixture and the resulting CTE. It was concluded that there is a 0.025 ue/°F
(0.045 pe/°C) change in the measured CTE per percent change in the coarse aggregate

volume as shown in Figure 2-5.

For the effect of aggregate type on the CTE value, the CTE of concrete Specimens

made from coarse aggregate obtained from 32 producers in the state of Texas were

measured. The results indicated that concrete specimens fabricated using the limestone

aggregate sources had CTE values of about 4.44 us/°F (8.0 us/°C) with a variability of

0.4 ua/°F (0.72 us/° C) whereas, concrete specimens fabricated with gravel as coarse

aggregate had a CTE range of 4.50 rte/°F to 7.20 us/‘T (8.10 us/°C to 12.96 rte/°C). It

was concluded that this variability is attributed to the different geological make up of the

gravel sources. It was stated that this difference in variability between limestone and

gravel might explain the better performance and less variability in the performance of

PCC pavements made with limestone coarse aggregate versus more variability in

performance of the pavements made with gravel aggregates as illustrated in Figure 2-6.

CTE (microstrain/C)
on

 

 

~ . . , _ i. y = -0.0456x + 10.76
""~-...._R2=0.9963

 

 

 

 

 

l

 

0 25 50

Coarse Aggregate Volume (%)

Figure 2-5. Effect of Coarse Aggregate Volume in Concrete on CTE16

18

14

 

 

12
Gravel Limestone

.L
O

 

 

 

CTE (microstrain/C)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2-6. Variation of CTE From 32 Aggregate Sources in Texas16

Kohler et al.17 at the University of California, Davis qualitatively investigated the
effects of aggregate geology, number of thermal cycles and soaking time on the
magnitude of CTE. The CTE test was conducted on 74 core samples obtained from four
regions within the state of California. The testing was done in accordance with the
revised AASHTO TP60 protocol proposed by Won“. The overall range of CTE was
between 4.5 and 6.7 rte/°F (8.10 and 12.06 rte/°C).

In order to study the effect of the number of heating-cooling cycles, 74 cores were
analyzed. Samples were subjected to three heating-cooling cycles. It was found that the
third cycle produced better coefficient of determination (R2) values for the regression
analysis used to calculate CTE values and that the difference between heating and

cooling cycles was reduced in the third cycle. Also, in 76% of the cases, the third cycle

 

resulted in lower CTE values than the first cycle values. The CTE of the third cycle was

found to be on average 0.15 ue/°F (0.27 us/°C) lower than the first cycle CTE value.
Figure 2-7 shows the effect of repeated thermal cycles on R2 and CTE values. It was

stated that this improvement in R2 value and the difference between heating-cooling
cycles improved the confidence in the results and it was an indication that the concrete
had reached a stable condition regarding pore water.

To quantify the effect of concrete saturation on CTE value, three cores were oven-
dried overnight. Two of the cores were tested for CTE immediately after air cooling, and
the third one was soaked for 96 hours. The dry cores showed a reduction of the difference
between heating and cooling cycles during the first 10 to 15 hours (Figure 2—8). The
saturated core showed a constant CTE value for heating and cooling cycles during the 9

cycles to which the core was subjected (Figure 2-9).

 

 

 

 

 

 

 

a) 0.99950 b) 6.0 L L ,,
uRising
0.99900 , 5.9 , (heating)
E: IFalllng
E 5.8 7 -
0.99850 1; CLOOI'UQ
7r 3 5.7
0.99800 E
E 5.6
0.99750 . o 5 5
0.99700 5.4
1 2 3 1 2 3
Thermal Cycle Thermal Cycle

Figure 2-7. Effect of Repeated Thermal Cycles on CTE, (a) Increase in R2, (b)
Decrease in CTEl7

The geographical variability was assessed by testing cores from four California

Department of Transportation districts. Northern area (District 2) aggregates were

20

probably sourced from alluvial or glacial deposits. A mix of sandstone and basalt rocks
was evident. Southern area (District 11) aggregates were predominantly granitic. Coastal
area (District 4) and valley area (District 10) aggregates were predominantly sandstone.
The average CTE of District 2 was 6.3 ua/°F (11.34 Its/0C), for District 11 the average
was 5 .5 us/°F (9.90 Its/°C), District 4 had an average CTE value of 5.2 ue/‘T (9.36
ua/°C), and the average CTE value of the District 10 was 6.4 ua/°F (11.52 Its/°C). Table

2-1 shows the CTE values at different sites. It was concluded that the geographical l

variability is probably associated with variability in aggregates of different mineralogical L
composition.
Tran et a1.18 evaluated the effects of mixture properties on concrete CTE as part

of a study that also evaluated pavement performance sensitivity to CTE values. Twelve
standard PCC mixtures were prepared using four major aggregate types including
limestone, sandstone, syenite, and gravel. Three types of cementitious materials (cement
only, cement and 20% fly ash, and cement and 25% GGBFS) were used with each
aggregate type. Twelve additional cement paste samples were prepared by wet-sieving
the standard PCC mixtures. For each mixture, three replicates were fabricated and the
samples were tested at 7 and 28 days.

Range of the CTE values was approximately 5 to 7 us/° F (9 to 12.6 pe/° C). PCC
mixtures made with limestone and syenite had the lowest average CTE of about 5.2 ua/°F
(9.36 118/0C). PCC mixtures made with gravels had the highest average CTE value of
approximately 6.9 us/°F (12.42 ps/° C). The wet-sieved cement paste samples from all

mixtures had similar CTE values with an approximate average of 6.6 usf’ F (11.88 Its/°C).

21

After conducting a multi-factor analysis of variance (AN OVA) on mixture
properties, it was reported that the aggregate type had a pronounced effect on the CTE. It
was also reported that the type of cementitious materials did not influence the CTE of
concrete. The fully saturated PCC and cement paste samples showed no significant

difference in their respective CTE values at 7 and 28 days.

9

 

 

o Rising
:1 Falling by

G)

 

 

 

?

.——D—#— 4:, *

 

 

 

U'I

 

CTE(microstraln/°F)
O)

A

 

 

 

(JO
q

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE (microstrain/°F)
O)

 

 

 

 

I . .
0 10 20 30 4o 50

Time (hours)

Figure 2-8. Effect of Saturation on CTE on Two Oven-Dried Cores”

22

 

 

 

 

 

CTE (microstrain/°F)
0:

 

 

 

3 I I I
97% 98% 99% 100% 101%

Saturation (%)

 

Figure 2-9. CTE Variability at High Saturation Levers"

Table 2-1. Mean, Maximum, and Minimum CTE at Different Sites].7

 

District Site Postmiles Nr.of Mean CTE Min/Max CTE Range
cores (arr-10‘) (arr-10'“)
4 SCL-85-N 13.90-15.17 6 5.22 4.73 / 5.68 0.95
4 SCL-85-S 13.52-15.52 12 5.08 4.46 / 6.07 1.61
4 SOL-80-E 18.46-34.34 12 5.38 4.63 / 6.24 1.61
4 SON-lOl-N 50.52-51.79 5 5.14 4.50 / 5.60 1.10
4 SON-lOl-S 50.84-53.02 7 5.18 4.80 / 5.62 0.82
10 SJ-580-E 5.02-8.88 10 6.35 6.12 / 6.57 0.45
10 SJ-580-W 5.35-8.70 9 6.48 6.21 / 6.69 0.48
2 SHA-S-N 37.85-39.91 6 6.29 6.23 / 6.39 0.16
2 SHA-S-S 29.53-31.71 3 6.28 5.96 / 6.69 0.73
I l lMP-86-S 23.50-29.56 4 5.48 5.43 / 5.53 0.10

 

An investigation was conducted by Naik et al.19 to quantify values for CTE of
concrete in order to support the implementation of the M—E PDG program in Wisconsin.
Coarse aggregate from 15 sources were used in the fabrication of concrete

specimens. Glacial gravel from six sources and dolomite from 5 sources were used.
Quartzite, granite, diabase, and basalt each from one source were also used. CTE values

were obtained according to the AASHTO TP60 test protocol. Three replicate specimens

23

at the age of 28 days were tested. The cementitious materials proportion of the mixture
design included 70% type I cement and 30% class C fly ash. In another part of the study,
the effect of cementitious materials on CTE of concrete was evaluated. Four mixture
designs were considered. Each mixture design included a different source of dolomitic
aggregate. Cementitious materials used were cement, cement plus fly ash (two different
mixtures), and cement plus ground granulated blast furnace slag (GGBFS).

As shown in Figure 2-10, the concrete made with quartzite showed the highest
‘ CTE value of 6.8 118/OF (12.2 118/0C). The lowest CTE values were those of concrete
made with diabase, basalt, and granite ranging from 5.2 to 5.3 Its/°F (9.3 to 9.5 118/°C).
Concrete made with glacial gravels from six different sources had CTE values between
5.4 and 5.9 us/°F (9.7 and 10.7 us/°C) and the CTE range for concrete made with
dolomite from five different sources (Figure 2-11) was relatively uniform, between 5.8 to
6.0 118/0F (10.4 to 10.8 118/0C).

According to this study, the types and sources of cementitious materials had a
negligible influence on the concrete made with dolomite. The CTE was influenced very
little (0.0 to 0.1 us/°F or 0.0 to 0.2 tie/0C) by the source of cement and class C fly ash, the

use of fly ash versus GGBFS, and the use of cement versus cement plus class C fly ash.

24

 

-r 6.8

 

 

 

 

 

 

 

 

 

 

'6

E + 6.6
° 8

If L »—-—-~ 6.4
'5 E -. 6.2
E v

.3 5 __dr 6.0
50 g —r 5.8
o 3.

0 x

E DJ

D

65

N

 

 

 

 

 

 

le1- leZ- G\I3- le4- GMS— GVIG- Qtz- Gnt- Dbs- Bst-
c1-f1 01-f1 c1-f1 c1-f1 c1-f1 c1-f1 c1-f1 c1-f1 c1-f1 c1-f1

Figure 2-10. CTE of Concrete Made with Gravel, Quartzite, Granite, Diabase, or
Basaltl9

 

-r 6.8

 

 

 

 

Expansion (10"I°C)

 

28-Day Coefficient of Thermal

 

 

Umf Um} mm} UmS UmS Wm¢ Wm¢ UmS Mm5
c1-f1 c1-f1 c2-f1 c1-f1 c1-s c1-f1 c1 c1-f1 c1-f2

Figure 2-11. CTE of Concrete Made with Dolomite and Different Cementitious
Materialsl9

Mallela et al.1 presented the effect of aggregate type on CTE value in a paper that
discussed the practical significance of CTE variability on the performance of concrete
pavements. The CTE results were based on hundreds of cores obtained from LTPP study

sections throughout the United States.

25

AASHTO TP60 test protocol was used to measure and calculate CTE values of
the specimens. A total of 673 cores representing hundreds of pavement sections
throughout the United States were tested and analyzed. The predominant aggregate type
in each specimen was identified using different methods including optical microscopy.
The general range of CTE values for the tested specimens in this study was between 5
and 7 ue/°F (9 and 12.6 116/°C). It was observed that concrete made with igneous
aggregate generally had lower average CTE than concrete made of sedimentary
aggregate. It was also observed that with some exceptions, the variability (standard
deviation) of the measured CTE was higher for concrete made with sedimentary

aggregate than the concrete made with igneous aggregate.

In a paper by Tanesi et a1.20 CTE variability was examined in order to be
incorporated in the other part of the project in which the effect of CTE variability on
concrete pavement performance was investigated. The AASHTO TP60 method was
described and possible sources of CTE variability were mentioned. Specimen induced
variability (moisture state and temperature gradient within the concrete specimen,
specimen inhomogeneities) and equipment induced variability (LVDT sensitivity, power
fluctuations, and frame calibration) as well as the intrinsic equipment limitations were
mentioned as possible sources of variability among CTE values.

Since 1996 the Federal Highway Administration (FHWA) has tested over 1800
core samples from various LTPP sections throughout the country. The CTE values
ranged from 4.5 us/°F to 7.5 rte/0F (8.1 11st to 13.5 Its/0C) as shown in Figure 2-12.
Approximately 150 specimens were tested multiple times to determine the repeatability

of the test procedure. The average variability amongst replicate samples was reported to

26

be 0.4 ua/°F (0.72 ua/°C) as presented in Figure 2-13. SCTE is the maximum difference

between CTE test results performed on the same specimen.

 

 

 

   
   

 

35
30
g 25
3 20
5
3 15
S
E 10
5
o
4 4.5 5 5.5 6 6.5 7 7.5
Mean CTE (x1o*3 inlinI°F)
Figure 2-12. Histogram of the Mean CTE of the Specimens20
20
18 Average acre: 0.4 x 10*5 in/in/°F l
15 __ Standard deviation SCTE = 0.4 x 10‘6 r
g 14 - in/in/°F ~
3 12 ‘ Median 5CTE = 0.3 x 10'6 in/in/°F
c 10 T
3 8 ‘
.. 5 -
u. 4 _
0 4

 

 

o 0.1 0.2 0.3 0.4 0.5 0.7 0.9 1.2 1.5 2 2.5
8CTE (x1o‘5 inlinloF)

Figure 2-13. Histogram of the tier}:20

27

2.4 Literature on CTE Impact on Pavement Performance

In this section studies (or part of studies) that investigate the impact of CTE on pavement
performance are presented. Results of the analyses and relevant discussions are also
presented.

As it was mentioned before, the CTE of PCC is a resultant of the HCP CTE and
aggregate CTE. It has been suggested that if these two CTE values differ too much (more
than 3.0 ue/°F [5.5 Its/°C]), a large temperature change (more than 122 °F [50 0C]) may
cause differential movement and breakage of the bond between aggregate and

surrounding HCP leading to cracking in concrete (N eville21).

Mallela et a1.1 qualitatively investigated the practical significance of CTE
variability on the performance of concrete pavements. The authors used the M-E PDG
software for this investigation.

It was stated that the PCC CTE affects joint spacing, joint load transfer, curling
stresses, and comer deflections in JPCP, which in turn affect the transverse cracking,
joint faulting, and smoothness. It was also stated that the interaction of CTE with other
design features and site conditions plays a Significant role in the extent of effect that CTE
has on pavement performance. For example, higher CTE values coupled with a high
temperature climate, is more detrimental than a climate with low temperatures. Similarly,
the effect of CTE is more pronounced in pavements with larger joint spacing than the
ones with shorter joint Spacing. So, two types of sensitivity analyses were carried out on
JPCP performance. In one analysis, only the effect of CTE on performance was
investigated and in the other analysis, the interaction effects of CTE with other PCC

design factors were studied.

28

In the first sensitivity analysis, a representative design with only the CTE being
the variable was assumed. Three levels of CTE investigated were mean, mean plus two
standard deviations, and mean minus two standard deviations for each aggregate type.
The effect of PCC CTE on percent slabs cracked, faulting, and IRI is shown in Figures 2-
14, 2-15, and 2-16 respectively. It was observed that CTE affects cracking, faulting, and
IRI, but the CTE effect is more pronounced on predicted cracking than on mean joint
faulting. It was also stated that in general, the higher the CTE, the poorer the pavement
performance, and that the aggregate type has the largest influence on CTE value. Finally,
the higher the variability in the measured CTE (for each aggregate type; aggregate source
information is not known from the paper), the more unpredictable the pavement
performance.

The critical design inputs and site conditions used to investigate the interaction
effect of CTE and design factors were PCC flexural strength (500 and 750 psi) and elastic
modulus (co-varied with flexural strength), transverse joint spacing (15 and 20 ft), and
climatic conditions (wet freeze and dry-no freeze). Three levels of CTE (4.5, 5.5, and 7.0
us/‘T or 8.10, 9.90, and 12.60 us/° C) were also considered in the analysis. It was found
that in general, higher CTE values resulted in higher joint faulting, slab cracking, and
roughness. Larger joint spacing and concrete strength increased the effect of CTE on joint
faulting due to higher curling deflections (higher modulus of rupture relates to higher
elastic modulus in the strength relationships used in M-E PDG). In wet freeze climate,
higher joint faulting values were observed. Larger joint spacing and lower concrete
strength resulted in amplified effect of CTE on transverse cracking. The CTE effect on

IRI was more sensitive to concrete flexural strength.

29

80 ..__ -

 

70~
so. ._

401 - Ioc '

10 — f 3
01 J l-

 

Percent Slabs Cracked

 

 

 

 

 

 

 

 

 

 

 

g6” gt? 96.5” ‘5‘” {\‘é .001{\ «$5 ‘00“ (4;? 1,06%
000 90 {‘00 (9‘0 90 O o\0 0’9 0% 6'9
Y‘ Q Q 06‘ 0 (90°

Figure 2-14. Effect of PCC CTE and its Variability on the M-E PDG Predicted
Percent Slabs Crackedl

0.20 7% —— ...___.____ _ -.---__ . ___

0.18 J —
A 0.16 i _
0.14 “l —
0.12
0.101
0.08 ~
0.06 ~
0.04 ~

0.02

 

In

 
 
 
 
  
  

 

 

 

 

Mean Joint Faulting (

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2-15. Effect of CTE and its Variability on the M-E PDG Predicted Mean
Joint Faulting‘

30

 

“at...

IRI (in/mile)

 

 

 

 

 

 

 

2
t.
:‘3
1‘
51

fl

 

Figure 2-16. Effect of CTE and its variability on the M-E PDG predicted IRIl

In a paper by Tanesi et al.20

the impact of CTE variability on the predicted
performance of concrete pavements was documented (as part of a project that also
investigated the sources of CTE variability). Based on sensitivity analysis using M-E
PDG program the effect of the CTE variability on slab cracking was found to be
significant. The higher the CTE, the greater is the effect of variability. As an example, a
difference of 2.0 ue/°F (3.6 Its/0C) between minimum and maximum measured CTE
values for the same specimen with an average CTE value of 4.0 118/0F (7.2 ua/°C) would
result in 8% difference in the predicted percent of slabs cracked, but the difference would
be 65% if the average CTE value were 6.5 118/OF (11.7 usl° C). Figure 2-17 shows the

effect of CTE variability on predicted percent slabs cracked. dCTE in Figures 2-17,

through 2-19 is same as the BCTE.

31

Colorbar: change in Percent Cracking

 

60
1.75:
1.51 50
of I l ,
: 125: x=5 . 40
E Y: 0.9 ' , »
“to 1 Level= 6.3579
; : 30
E 0.75 ..... ;
g _ 20
0.5 l ‘
0.251 i 10
0‘ I 0
4 4.5 5 5.5 6 6.5 7

CTE (x10'6 in/in/OF)

Figure 2-17. Difference in the Predicted Percent Slabs Cracked as a Result of the
dCTE20

The same effect mentioned above can be seen on the predicted faulting of

concrete pavements as shown in Figure 2-18.

32

dCTE (x1061n/in/0F)

Colorbar: change in Faulting in in.

 

1.8
1.6 0.1
1.4
0.08
1.2‘ X: 5
, Y: 0.9 .
l _ .. . .
1 ‘ Level— 0.028494 . f f. 0.06
0.83 x—6.5
Y: 0.5
0.6 Level: 0.02415 0'04
0.4]
0.02
0.2‘
0 0
4 4.5 5 5.5 6 6.5 7

CTE (x10‘6 in/in/OF)
Figure 2-18. Difference in the Predicted Faulting as a Result of the dCTE20

The impact of 5CTE on the International Roughness Index (IRI) was also

documented. The effect is similar to the one of the percent slabs cracked case. For the

same example mentioned above, the IRI for the first case (a difference of 2.0 ue/°F or 3.6

lie/°C for a specimen with average CTE value of 4.0 Ila/0F or 7.2 ue/°C) the difference in

IRI is 33 inch/mile, while for the second case it is 113 inch/mile. Figure 2-19 illustrates

this effect. The authors concluded that the CTE test variability leads to significant

discrepancies in the predicted IRI, percent slabs cracked, and faulting.

33

Colorbar: change in IRI (in/mile)

X: 5 X.
Y: 0.9 ..,,
:i'

Level: 23.4067 ’effig
I
X: 6.5
Y: 0.5
Level: 23.9157

dCTE (x1061n/in/0F)

 

4 4.5 5 5.5 6 6.5 7
CTE (x10'6 in/in/OF)

Figure 2-19. Difference in the Predicted IRI as a Result of the dCTE20

Tran et a1.18 evaluated pavement performance sensitivity to CTE values as part of
a study that also evaluated effects of mixture properties on concrete CTE.

In order to investigate the practical significance to pavement design, a sensitivity
analysis using M-E PDG software (version 1.0) was conducted. Three levels of joint
spacing and four levels of CTE values corresponding to cement paste, PCC mixtures with
sandstone, gravel, and limestone/syenite were considered. The results are showed in
Figure 2-20. It was concluded that the type and proportion of coarse aggregate influences

the joint spacing design and subsequently, the pavement performance.

34

 

0.5

 

 

 

 

 

 

g 0.4
I; 0.3
C
E 0.2 -
3 \
u_ 0.1 , ,
0.0
5.2 6.4 6.6 6.9
(PCC w/ (PCC w/ (Paste) (PCC w/
Limestone or Sandstone) Gravel)
Syenite)
CTE (1o‘/°l=)
(a) Faulting versus CTE for three joint spacings
100
.3 80
E ‘50 in 15 ft
E 40 I.“ 20 ft_
0 20
0
5.2 6.4 6.6 6.9
(PCC w/ (PCC w/ (Paste) (PCC w/
Limestone or Sandstone) Gravel)
Syenite)
CTE (10‘/°F)

(b) Cracking versus CTE for three joint spacings

 

 

 

 

5.2 6.4 6.6 6.9
(PCC w/ (PCC w/ (Paste) (PCC w/
Limestone or Sandstone) Gravel)
Syenite)
cTE (io‘PF)

(C) IR] versus CTE for three joint spacings

Figure 2-20. Effect of Constituent Properties on Pavement Performance
Predictions18

35

Kohler and Kannekanti22 studied the influence of PCC CTE on the cracking of
the JPCP. One hundred and four in-service highway sections in California were selected
for this study and cores obtained from these sections were tested in the University of
California Pavement Research Center (UCPRC) laboratory.

The CTE testing protocol followed in the University of California study was

based on the recommended amendments to the AASHTO TP60 protocol proposed by

Won“. The main features of this testing procedure are summarized in Won’s paper. The
CTE of each section was generally determined from the CTE results of at least two cores.
The total number of tested cores was 185. The CTE values ranged from 4.8 to 6.7 rte/°F
(8.6 to 12.0 ire/°C).

Three types of cracking levels were incorporated in the data collected by
California Department of Transportation. First stage cracks (FSC) are cracks that do not
intersect and divide the slab into two or more large pieces; third stage cracks (TSC) are
interconnected cracks that divide the slab into three or more large pieces; comer cracks
(CC) are diagonal cracks that meet both a longitudinal and transverse joint within 6 feet
and over 2 feet at the same slab comer. Figure 2-21 shows examples of these cracking
levels. Slab cracking is reported as a percentage based on the number of slabs exhibiting
these cracking levels over the surveyed distance (0.1 to 1.5 miles for a given
homogeneous section). Ratio of the severely cracked sections to total number of sections
was calculated (k). The severity limit used for FSC and TSC was 10% and for CC, 5%

limit was used.

36

 

 

Figure 2-21. Examples of FSC, TSC, and CC22

The ratio of cracked slabs mentioned above was computed for two groups of
sections. One group consisted of sections with CTE values lower than an arbitrary limit
and the other group included sections with CTE values higher than the arbitrary limit.
Two arbitrary limits were considered. These limits were 5.7 ue/°F (10.26 ue/°C) and 6.0
lie/°F (10.80 lie/°C). This ratio (k) versus pavement age in years was then plotted (Figure
2-22 for FSC and TSC, Figure 2.20 for FCS+TSC and CC). It was stated that the
cracking trends for low and high CTE pavements were drastically different when the
CTE limit was 5.7 Ila/°F (10.26 Ila/°C). For all types of cracks, pavements with high CTE
developed more cracks over time than the pavements with low CTE. It was concluded
that if low CTE is specified, it can reduce cracking over the life of JPCP and longer

lasting concrete pavements can be expected.

37

Ratio of cracked slabs (k)

 

FSC

 

 

 

 

 

 

 

 

 

 

 

 

0 10

Pavement Age (years)

 

 

 

o CTE >5.7

o CI‘E >60
CTE >5.7
------- CTE >6.0

 

- CTE <=5.7
.. CTE <=6.0
CTE <=5.7

------- CTE <=60

 

 

Ratio of cracked slabs (k)

 

 

 

 

 

 

 

 

 

 

 

 

 

Pavement Age (years)

 

 

 

o CTE >5.7

o CI‘E >6.0
CTE >5.7
------- CTE >60

 

I Cl‘E <=5.7

a CTE <=6.0
CTE <=5.7

------- CTE <=6.0

 

 

Figure 2-22. Comparison of Ratio of Cracked Slabs for Pavements with High and

Low CTE (FSC and TSC)22

38

'F—i—n—j

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

90 v o
FSC+TSC
80 --— —
8
U)
E
(I)
g
c...
o
.9
$2
0 10 20 30 40 50
Pavement Age (years)
0 CTE >5.7 l CTE <=5.7
o CTE >60 0 CTE <=6.0
CTE >5.7 CTE <=5.7
------- CTE>6.0 ------'CTE<=6.0
90
80 ‘—~ CC —---—- --—--
A 70 -«-—— ~ — “—7- -~ LL# - #—
if:
E 60 _ LH_____ W -———~ -—-—-- ——
U) 50 __.__._. EAL... .. _ __ .—
8 5 *
E 40 n.-,,__,,,,‘,, ‘ -_-_ .-.. . 9.- _-___ . __
(H 30 ----‘
o
.9. 20
52 10 ——
O ._

 

 

 

0 10 20

Pavement Age (years)

 

 

 

 

 

o CTE >5.7 - CTE <=5.7

o CTE >6.0 o CTE <=6.0
CTE >5.7 CTE <=5.7
------- CTE>6.0 -------CTE<=6.0

 

 

1‘ t

Figure 2-23. Comparison of Ratio of Cracked Slabs for Pavements with High and
Low CTE (FSC+TSC and CC)22

39

Hossain et a1.23

investigated the effect of the hierarchical input levels of CTE on
the performance of jointed concrete pavements using the M-E PDG program. The CTE
data was obtained from in-service pavements in Kansas. CTE results from LTPP projects
in Iowa, Kansas, and Missouri were also reviewed.

After an overview of the AASHTO TP60 test procedure and describing the
hierarchical input levels used in M-E PDG, input levels 1 and 2 for CTE were presented.

For input level 1, two cores were retrieved from a PCC pavement in Kansas. The
CTE values were 5.4 and 5.5 p8/°F (9.8 and 9.9 118/°C). The CTE values from the LTPP
database were a result of testing on 51 cores and ranged from 4 to 7.1 ue/°F (7.2 to 12.8
118/0C). The lowest 10% (4.3 us/‘T or 7.8 paf’C) and highest 10% (6.5 its/”F or 11.7
us/°C) mean values were used in the sensitivity analyses in this study. The CTE values
for Iowa retrieved from LTPP database ranged from 4.4 to 7.6 ue/°F (8.0 to 13.8 ue/°C)
based on 62 cores. Also, the CTE values for Missouri were between 4.1 and 11.0 ua/°F
(7.3 and 19.8 pus/°C).

Level 2 CTE values were calculated from an equation suggested by M-E PDG.
The values needed for the calculation were extracted from the LTPP database. For
Kansas aggregates, CTE of dolomite, gravel, limestone, and sandstone were calculated
and compared to measured values (Figure 2-24). The calculated CTE values were 11 to
19% higher than the measured values. For Iowa, dolomite and limestone CTE values
were calculated. The calculated values were 10 to 14% higher than the measured ones
(Figure 2-25). The aggregates available in Missouri were dolomite, limestone, a

combination of dolomite and limestone, and sandstone. The calculated values, except for

40

dolomite, were 13 to 30% higher than the measured values (Figure 2-26). For dolomite,
the calculated values were 25% lower than the measured values.

In the same study, in order to investigate the effect of CTE on PCC pavement
performance, six in-service jointed plain concrete pavement projects were selected.
Three levels of CTE (average of the highest 10% based on LTPP data, CTE based on a
recently built project, and average of the lowest 10% based on LTPP data) were used in

the M-E PDG design analysis.

 

8.000
7.500
7.000 <7
6.500

6.000 <
5.500 -
5.000 -
4.500

4.000 <

1 I Calculated l

I Measured

a pee

 

 

 

Dolomite Gravel Limestone Sandstone

Coarse Aggregate Type
Figure 2-24. Calculated and Measured PCC CTE Values (x 104%) for Kansas23

6.200 , , - , 7
6.100 ~ , —~~
6.000
5.900 - , ~
5.800 . , W,
5.700 .

5.600 TA”
5.500 r ~
5.400 . , ,.

1 i
’ ' El Calculated

, 7— . i
gIMeasured 3

a pee

 

   

Dolomite Limestone

Coarse Aggregate Type

Figure 2-25. Calculated and Measured PCC CTE Values (x 10'6/°F) for Iowa23

41

8.600

 

 

 

 

 

8.000
7.400 r__.__’

3 6.800 ICalculated

a.

a 6.200 IMeasured :
5.600 —J
5.000
4.400

Dolomite Limestone Limestone 8. Sandstone
Dolomite
Coarse Aggregate Type

Figure 2-26. Calculated and Measured PCC CTE Values (x 10'6/°F) for Missouri23

The effect of CTE on IRI is that higher CTE would result in higher IRI. For
example with an increase in CTE from 4.3 to 6.5 ua/°F (7.8 to 11.7 ue/°C), IRI increased
from 114 to 135 inch/mile. By studying other pavements, it appeared that the effect of
CTE on IRI is more pronounced in pavements with thinner slabs or lower strength. It was
also found that the CTE does not affect the predicted IRI for pavements with widened
lanes and tied PCC shoulders.

Faulting was found to be sensitive to CTE values. A combination of high cement
factor and higher CTE values would result in higher faulting.

The effect of CTE on percent slabs cracked was found to be very significant. For
example, with a CTE value of 4.3 ue/°F (7.8 lie/°C), the percent slabs cracked was 0.2%
while for a CTE value of 6.5 ua/°F (11.7 usf’C), the percentage increased to 2% which is
a tenfold increase for a 50% increase in the CTE value.

Another part of the study was to investigate the hierarchical input levels on PCC
performance. For this purpose, one project was selected. Four types of coarse aggregates
(dolomite, limestone, gravel, and sandstone) were studied. Level 1 (measured) and level 2

(calculated) inputs were investigated. Table 2-2 shows the results. The design using

42

calculated CTE values failed (based on design reliability of 90%) for IRI and/or percent

slabs cracked for all aggregate types and for gravel with measured CTE value. Faulting

was relatively unaffected for all aggregate types and both input levels.

Table 2-2. Predicted Distresses for Computed and Measured PCC CTE Values for

 

 

 

Typical Aggregates in Kansas23
Calculated a pcc (/°F) Measured a pcc (/°F)
Coarse
Aggregate IRI Faulting % Slabs IRI Faulting % Slabs
Type (in/mi) (in) Cracked (in/mi) (in) Cracked
1343* 9.5*
Dolomite Failed 0.035 Failed 120.5 0.021 1.0
147.0 20.4 131.1 7.0
Gravel Failed 0'042 Failed Failed 0'0” Failed
Limestone 12:33! 0.03 l 5.4 122.0 0.024 1 .5
1755* 46.7*
Sandstone Failed 0'055 Failed ”20 0'024 1'5

 

* one sample

43

2.5 Application of the Reviewed Literature to the Research

Plan

According to sections 2-2 and 2-3 of this literature review it was found that several
factors affect the magnitude of the CTE of concrete. These factors include test-related as
well as specimen-related factors.

Test-related factors affecting the CTE could be induced by LVDT sensitivity,
power fluctuations, frame calibration, intrinsic equipment limitations, number of heating-
cooling cycles, and test method used to determine the CTE.

Specimen-related factors that affect the CTE could stem from coarse aggregate
type, age of the specimen, moisture state and temperature gradient within the concrete
specimen, and specimen inhomogeneities.

Based on the literature review, available laboratory resources, funding, time
frame, and funding agency’s demands, factors investigated in this research plan included
coarse aggregate type, specimen age, number of heating-cooling cycles, and test methods.

The reviewed literature did not investigate the effects of aggregate type, age (from
3 to 365 days), and number of heating-cooling cycles together in a factorial design the
way it was conducted in this study. The effects of these factors were studied by
employing a methodical statistical approach not seen in the reviewed studies. The effects
of some of these factors on CTE testing were also subjects that had not been investigated
in most studies.

One factor that was not investigated in this study was the moisture state of the
concrete. There are a couple of reasons for not considering this factor. The first reason is

that the current testing procedure for concrete CTE is AASHTO TP60 and in this method,

44

the concrete specimen is in tested in saturated state. The second reason is that moisture
conditions affect the HCP more than they affect the concrete. Generally, 60% to 80% of

the volume of concrete is consisted of aggregates therefore the CTE of aggregate affects

the concrete CTE more than that of the other constituents such as cement. Yeon et al.14

reported that the difference between maximum and minimum CTE values in cement paste
(associated with difference in RH) was about 10% to 12% while little difference (3%)
was observed between maximum concrete CTE at 70% - 80% RH and CTE of concrete at
100% RH.

According to section 2-7 of this literature review it was found that CTE of
concrete affects different aspects of pavement performance such as cracking, faulting,
roughness, and crack spacing and width. Based on the available sources and limitations,
sensitivity analysis factorials were developed in this research plan to investigate the

early-age stresses and long-term transverse cracking in JPCPs.

45

CHAPTER 3

EXPERIMENTAL PROGRAM

3.1 Introduction

This chapter provides details regarding materials that were used in different concrete
mixtures, sample fabrication process, curing, tests used for determining fresh and
hardened concrete properties, and CTE test protocols.

Coarse aggregate from different sources were used in concrete mixtures. One type
of fine aggregate form one source was used in all mixtures. Mixture design was a typical
MDOT design and the samples were prepared according to ASTM standards.

Two types of tests were conducted in this experimental program; quality control
tests and CTE test. The quality control tests were conducted in order to check the quality
and conformance of the supplied concrete. These tests included fresh and hardened
concrete mechanical properties tests. The CTE test was the most important test in this
program and was conducted in a way so that its results could be used in two different

methods of CTE calculation. This will be explained later in this chapter.

46

3.2 Materials

3.2.1. Aggregates

Fine aggregate used in all mixtures was 2NS sand from Schlegel pits. Coarse aggregates
were 6AA class and their sources are presented in Table 3-1. The 2NS and 6AA are

MDOT grading classes for fine and coarse aggregate respectively and their sieve analysis

requirements can be found in “2003 Standard Specifications for Construction”24. Figure
3-1 shows the locations of the various aggregate sources within the state of Michigan.
Mineralogical composition and physical properties of coarse aggregate are summarized in
Table 3-2. For aggregates with no mineralogical information available, petrographic or
chemical compositions are summarized in Tables 3-3 or Table 3-4, respectively.

Table 3-1. Coarse Aggregate Types and Source Names

 

 

 

 

 

 

 

 

 

Mixture ID Primary Aggregate Class Aggregate Source, County
CTE 1 Limestone Pit # 71-47, Presque Isle
CTE 2 Gravel Pit # 19-56, Clinton
CTE 3 Limestone Pit # 75-5, Schoolcrafi
CTE 4 Slag Pit # 82-19, Wayne
CTE 5 Dolomite Pit # 49-65, Mackinac
CTE 6 Gabbro Pit # 95-10, Ontario
CTE 7 Dolomite Pit # 58-11, Monroe
CTE 8 Dolomite Pit # 91-06, Cook

 

 

 

 

 

47

 

- 358-11
a 91-06

Figure 31-]. Locations of the Aggregate Sources (After Sutter et ah“)

48

 

.m-m 03m... 5 02.808 3 02:89:00 32825 .1. .m-m 28H 5 08.88.. g 00280800 0202803 *

$28.80 00388.? n U< H .5380 058% 8-030 u Own—O ._.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

$2 038200 w 5.0
. . 2828 Been . . . . e as o
S m mew N 8388 00 new bum 00 :8 EMS on N mm o em 0 3 no 0. ~ D h Eb
0:88 08 5.800
56 Sad 6:80me 88:0 858 608550: a... 0580 o 2.0
68—02me 88.3 BE: .0580
. . 6:528 858 . . . . a 56 e
me o mmn m 8.800 00 8388 >an 8 :5 Ems a O we o we 0 3 mo 0. _ D m Eb
0.5808 8803 383 00 536:0»
. . 30% 833.80 88% 2: .0305
wk. N mam N .8 0.8 00 8% 08 833.80 880 * wflm 0 who
05 chew 08 833.80 83063 22.
and owed 0:988: 808% 00m 93 8 =8 2&3 v0.0 cod 0500 has 0000885 m who
end End 32 a. 6280 N who
x52: 0:988:
v: wwwm 859% 00m a 5 £7.80 080555 who Ed 303 was 008885 _ who
55 0305 x80 8 .0305 00 cup.
650 awed 608 £00828
“9.. .680 858860 25 86m 9
Ewe? .3 ex. 0652 SEE 235:

 

 

 

 

 

 

 

 

 

 

 

A340 8 830m 80.3.0 Saweuwwaa 08an 05 me 83800.5 .833.— 000 Somme—0852 .Nrm 03:.

49

Table 3—3. Petrographic Composition of Slag and Gravel Aggregates*

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

M‘I’Bm RISES—F3... AggregateTwe M8221?”
Igneous/Metamorphic 54
Dense Carbonates 35.4
Absorbent Carbonates 4.7
Non-Friable Sandstone 1.0
CTE 2 Gravel Friable Sandstone 1.7
Siltstone 0.6
Shale + Coal 0.1
Clay Ironstone 0.5
Chert 2
Vesicular Particles 85.7
CTE 4 Slag Dense Particles 10.8
Glassy Particles 3.3
Magnetic Particles 0.2

 

 

* These petrographic testing and analyses were conducted by MDOT.

Table 34. Chemical Composition of Gabbro Aggregate (After Sutter et al.25)

 

 

 

 

 

 

 

 

 

 

 

. Primary . Oxide/Element % by
Mixture ID Rock Type Ox1de/Element Weight
MgO 8.44
A1203 18.61
SiOz 45.53
CTE 6 Gabbro
S 0.02
CaO 11.81
Fe203 13.13

 

50

 

3.2.2. Concrete

The concrete used in the fabrication of cylindrical and prismatic test specimens was

supplied by a local ready mix supplier. This ensured that all specimens needed for a given

mixture were produced fiom a single batch, thereby reducing experiment variability.

The concrete mixture design used in the fabrication of the test specimens was a

typical paving mixture design used by MDOT (designated as P1). Individual mixture

designs are summarized in Table 3-5.

Table 3-5. Concrete Mixture Designs (lbs/yd3)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

* This mixture design also included 94 lbs/yd

of Fly Ash. TAEA=Air Entraining Agent

Ingredients CTE 1 CTE 2 CTE 3 CTE 4 CTE 5 CTE 6 CTE 7 CTE 3*
Cement 564 564 560 560 560 573 560 376
Water 259 259 250 252 275 258 242 155
Coarse Agg. 1740 1760 1838 1575 1908 1774 1715 1942
Fine Agg. 1360 1360 1338 1348 1260 1230 1330 1444
ABAT, (fl, 01) 10.0 10.0 7.5 7.5 7.5 7.5 7.5 28.0
3

Concrete specimens (except for CTE 8) were prepared at the Michigan State

University (MSU) Civil Infrastructure Laboratory (CIL) according to the ASTM C 192

“Standard Practice for Making and Curing Concrete Test Specimens in the Laboratory”.

CTE 8 specimens were field-prepared specimens from an actual paving project in

Michigan.

Concrete specimens fabricated for elastic modulus, compressive strength, and

tensile strength tests were cylindrical specimens 4 inches in diameter and 8 inches in

51

 

length (height). The dimensions of the concrete beams (square prisms) fabricated for
flexural strength test were 4 x 4 x 14 inches. The samples made for CTE test were
cylinders 4 inches in diameter and 7 inches in length (height).

At least three replicate samples were fabricated for each test. Over 700 specimens
were fabricated to characterize the mechanical properties and CTE of the concrete paving
mixtures. Thermocouples were embedded in the center of designated specimens to
monitor concrete temperature for the CTE tests. All specimens were de-molded 24 hours
after fabrication and were cured at 100% relative humidity and 23 °C temperature in an
environment chamber until the time of testing. CTE specimens were placed in a
limewater bath as required by the test protocols. Once the specimens were de-molded and
cured for an appropriate time, various tests were conducted to assess the properties of

interest.

52

 

 

3.3 Concrete Properties Tests

Fresh and hardened concrete mechanical properties tests and thermal property test were
conducted on concrete specimens. The fresh and hardened concrete mechanical
properties tests were conducted in order to check the quality of the supplied concrete and
its conformance to specifications. CTE test was conducted in a way so that its results
could be used in two different methods of CTE calculation. This will be explained in the
next section.

The fresh concrete tests included slump, air content, unit weight, and temperature
tests conducted according to American Society for Testing and Materials (ASTM)
standards.

The hardened concrete mechanical properties tests included compressive, tensile,
and flexural strength tests in addition to elastic modulus test. These tests were also
conducted according to ASTM standards.

The CTE test was conducted according to AASHTO TP60 and is described in the
next section.

The material characterization tests performed on the concrete samples and their

frequency are summarized in Table 3-6.

53

 

Table 3-6. Summary of Material Characterization Tests

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Test ASTM Measured No. of Frequency
Test Name
Attribute Designation Property Specimens of Testing
Concrete
Slump C 143 workability
P . Total air
ro ertres
ofrlzresh Air content C 231 cortirtenltl Of One per Once
C es batch
oncrete concrete
Unit weight C 138 Unit weight
Temperature C 1064 Temperature
Compressive
strength* C 39
Mechanical Flexural Concrete 1, 3, 7, 14,
Properties strength C 78 strength 28,90,365
of days after
Hardened Split tensile specimen
Concrete strength C 496 Three fabrication
replicates
Elastic Concrete for each
modulus C 469 stiffness test/batch
AASHTO Linear unit 3a 79 143 28:
Coefficient of TP60 length 90’ 180’
Thermal . 365 days
thermal change/unit
Property . . afier
expansron change in .
Revised temperature SPCF‘mF’“
TP60 fabrrcatron

 

 

 

 

* Compressive strength was determined by the same apparatus and specimen used to determine the

modulus of elasticity.

54

3.4 Thermal Property Test (CTE Test)

CTE test was conducted according to the AASHTO TP60 “Standard Test Method for the

Coefficient of Thermal Expansion of Hydraulic Cement Concrete”4. The CTE test
apparatus consists of a temperature controlled water bath, a rigid frame to support the test
specimen, an LVDT to record the change in specimen length, and a data acquisition

system for continuous data collection. Figures 3-2 and 3-3 illustrate the CTE test setup.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Thermocouple \.
~ > ,__;|3-- r__
LVDT . . .
Data Acquisition
Power S stem
Slipply y
_ ; LVDT
, g
H— ’ 3:; h I.
Fixture r 1
and {3.2-.191 Invar J, =- 6
SpCCian E 1 ‘7}?- -: 4"? ROd {,’=3.‘
Controlled Temperature Water Bath Computer

Figure 3-2. Schematic of the Test Setup

3.4.1 Controlled Temperature Water Bath

Three “Programmable Refrigerating/Heating Circulators” by PolyScience were used in
this experiment. The temperature range for these circulators is -25 to +150 °C with

temperature stability of i001 °C. Figure 3-4 shows a Model 9612 circulator.

55

Computer

Data Acquisition

LVDT Power Supply
Water Bath

LVDT

Fixture

 

Figure 3-3. Complete Test Setup

  

<— Temperatnre Controller

Water Reservoir
Lid

Figure 3-4. Controlled Temperature Water Bath (PolyScience26)

56

3.4.2 Data Acquisition System
A Personal Daq/3000 data acquisition system by IOtech was used to read and record the
length changes of the specimens measured by LVDTS and also read and record water

bath temperatures measured by thermocouples. This system is shown in Figure 3-5.

 

Figure 3-5. Data Acquisition System (IOtech27)

The software used with this system was DaqViewTM which allows the user to save
the data in text format among other file formats. A screen shot of the channel setup is

illustrated in Figure 3-6.

3.4.3 Linear Variable Differential Transformer (LVDT)

Three “GHSD 750-050 Spring-Loaded DC-LVDT Position Sensors” by Macro Sensors
were used to measure the length changes of concrete specimens subjected to temperature
cycles. These LVDTS have a nominal range of i0.050 in. from null position and full scale

output of O to $10 V DC. Figure 3-7 shows the LVDTS.

57

  

[Priq‘r’iew - ['AOVIEWZDAQ [Peisonnlliaq3("100[261‘l'l 66}]
file Edit Qata window Qevice Llelp

. 10 14 ‘ng i 2
El .11 £1 P

 

 

    

§_

 

 

:- Analog & Scanned Digital inputs

 

 

 

 

 

 

 

 

 

E g. l ChannelUrt Yes v F% Pg 3% . ‘1
1 SEE] HEB q) ‘
CH Un ] Type I Polarity ] Label | Units 1 Reading 5;;
P1 0 Yes T 2 . Bipolar” _ TEMP-W3 ‘13
P1 1 Yes _T Bipolar TEMP-W1 'C 1
P1 2 Yes xi .. ,_ jBlpOlar LVDT-225 _Mili|n_
P1 4 Yes . (xi Bipolar“ LVDT-128 7min.
P1 6 Yes“ 111 ‘ Bipolar LVDT-3-COMmln.
P1 7 Yes _T . Bipolar TEMP-W2 'C
CJCIUU-Elm Yes CJC Bipolar .CJCUU-EIO ‘C
013101-02] Yes, 7 CJCW Bipolar . E‘JCUi-UZ ‘C _
011107-071 Yes CJC Bipolar [311307-07 ‘C v

 

 

 

 

 

 

Figure 3-6. DaqViewTM Software Channel Setup Screen

 

Figure 3-7. Spring-Loaded LVDTs (Macro Sensors”)

58

3.4.4 Rigid Support Frame

3.4.4.1 Frame Apparatus

Rigid support flames were fabricated based on AASHTO TP604 appendix X.1

“Specimen Measuring Apparatus”. Figure 3-8 shows the rigid support flame. The circular
base plate is made of stainless steel and has a diameter of 10 inches. Three semi-spherical
support buttons equally spaced around a 2 inch diameter circle are placed on the base
plate. The flame height is 10 inches and the vertical rods are made of Invar (a nickel-iron
alloy with very low CTE) in order to minimize the effect of flame length changes on the
measurements. The side view and plan view of the rigid support flame are shown in

Figure 3-9.

Invar Vertical Rods

Semi-Spherical Supports
Base Plate

 

Figure 3-8. Rigid Support Frame

59

 

 

 

Invar Vertical Rods

Base Plate @ O

Semi-Spherical Supports O O
A

Figure 3-9. Side View (Left) and Plan View (Right) of the Rigid Support
Frame (After AASHTO TP604)

 

 

 

 

 

 

l'—'l n A 1'

 

 

 

3.4.4.2 Frame Calibration

Expansion of the frame during the test must be accounted for in the CTE calculation. A

method for determining a correction factor (CF) based on AASHTO TP604 appendix X.2
“Reference Test for Determination of Correction Factor” was employed.

In this method, a calibration specimen with known CTE, having the same
measurements as the concrete specimens (cylinders with 4 in. x 7 in. dimensions) is

subjected to CTE test. A stainless steel specimen (grade 304) was used in this process.

The C1: is defined as
CF = ALF / LCS / AT
in which AL]: is the length change of the frame, LCS is the length of calibration specimen
in room temperature and AT is the temperature change. AL}: is expressed as

ALF = ALA - ALM

60

 

and is the difference between the actual calibration specimen length chance (AL A) and
the measured calibration specimen length change (ALM). AL A is defined as
ALA=LCS >< ac x AT

in which the ac is the known CTE of the calibration specimen.

3.4.5 Test Procedures, Data Collection, and CTE Calculations

Two test procedures were followed; AASHTO TP604 procedure and a revised procedure

based on TP60 suggested by Won16 which is going to be called “Revised TP60”
hereafter. Specimen conditioning and heating-cooling cycle are the same in both
procedures. Data collection and CTE calculation however, are different.

The CTE test was performed in a way that it satisfied data collection requirements
for both procedures. The relevant parts of data were then used to calculate the CTE based

on each method.

3.4.5.1 Specimen Conditioning

In order to condition the specimens, they were submerged in a saturated limewater bath
for at least 48 hours at 73 °F (23 °C). Conditioning was verified by weighing surface-
dried specimens at 24-hour intervals. The conditioning criterion was a weight increase of

less than 0.5 % between two successive weighings.

3.4.5.2 AASHTO TP60 Test Procedure, Data Collection, and CTE Calculation
In this procedure, after the concrete specimen is conditioned, it is subjected to heating,

cooling, and temperature equilibrium segments during which the temperature of the

61

surrounding water and displacement of the specimen are monitored. A summary of the
procedure is followed.

The first step is to place the flame in the water bath and the specimen in the
flame. LVDT is then mounted on the flame. Silicon grease is applied to the support
buttons and the tip of the LVDT to prevent any sticking to the concrete specimen. The
water bath is cooled down to 50 °F (10 °C) and is kept at this temperature until thermal
equilibrium of the specimen has been achieved. The thermal equilibrium is denoted by

“consistent readings of the LVDT to the nearest 0.00001 in. taken every ten minutes over

a one-half hour time period” (AASHTO TP604). The temperature and displacement
readings are recorded as initial readings. The water bath is then heated to 122 °F (50 °C)
and the same process is followed to record the second readings. The system is cooled
again to 50 °F (10 °C) and the final readings are recorded.

Recorded readings mentioned above are used in the following computational
sequence in order to calculate the CTE value.

The CTE of the specimen is calculated by averaging the heating and cooling cycle
CTE values providing that the difference between these values does not exceed 0.3 ue/°F
(0.5 118/°C). If the CTE values differ more than the specified value, the test has to be
repeated until this criterion is satisfied. The CTE equation is expressed as the following.

CTE = (CTEHEATING + CTECOOLING) / 2

CTE of the heating or cooling cycle is defined as the actual length change of the

specimen (AL ACTU AL) divided by the initial length of the specimen (LINITIAL) over the
temperature range (AT) as denoted by the following equation.

CTE HEATING or COOLING = (ALACTUAL / LINITIAL) / AT

62

AL ACTU AL is defined as the summation of the measured length change of the

specimen (ALSPECIMEN) and the length change of the measuring apparatus

(ALAPPARATUS)-

ALACTUAL = ALSPECIMEN + ALAPPARATUS

The length change of the measuring apparatus is defined as
ALAPPARATUS = CF X LINITIAL X AT

in which C1: is correction factor of the flame (explained in section 3.4.4.2).

3.4.5.3 Revised TP60 Test Procedure, Data Collection, and CTE Calculation

The Revised TP60 procedure is similar to AASHTO TP60 method with regards to
specimen conditioning, heating and cooling segments. In this method however, the
temperature, specimen displacement, and time are recorded every minute during heating
and cooling segments. A summary of the procedure follows.

After conditioning and placement of the specimen in the flame inside the water
bath, the temperature of the bath is set to 50 °F (10 °C) and temperature, specimen
displacement, and time are recorded every minute. The bath temperature is kept at 50 °F
(10 °C) for an hour. The same procedure takes place when the temperature is raised to
122 °F (50 °C) and back to 50 °F (10 °C).

Recorded readings are then used to plot a temperature versus displacement graph
in which heating and cooling data are plotted separately. A typical graph such as the one

explained is shown in Figure 3-10. A linear regression analysis is applied between

temperature and displacement for the range of 59 °F (15 °C) to 113 °F (45 °C). Won16

stated that temperature gradients inside the specimen are not consistent below and above

63

 

this range and therefore are not considered in the regression analysis. The R2 of the
regression analysis for heating and cooling should be greater than 0.999. CTE of each
segment is then calculated using the regression equation. The difference between CTE of
each segment must be less than or equal to 0.15 p8/°F (0.3 tie/°C). If the CTE values
differ more than this specified value, the test has to be repeated until this criterion is
satisfied.

Temperature (°F)

41 50 59 68 77 86 95 104 113 122

0.0025 1 I 1 ‘ 1 I ‘ L
l A Cooling I Heating —Linear (Cooling) —Linear (Heatingfl

 

 

0.0020 4

0.0015 -

0.0010 ~

0.0005 ~

Displacement (in.)

0.0000 ~

-0.0005 ~

 

 

'0-0010 I I I I I I I I
5 10 15 20 25 30 35 40 45 50

Temperature (°C)

 

Figure 3-10. A Typical Revised TP60 Method Graph

3.4.5.4 Executed Test Procedure

A test procedure was designed in order to satisfy both AASHTO TP60 and Revised TP60
requirements. Specimens were subjected to heating and cooling cycles between 50 °F and
122 °F (10 °C and 50 °C) and the thermal equilibrium time was one hour between the

segments. Several tests with different thermal equilibrium time periods were conducted to

64

 

establish an optimal time period. Temperature, specimen displacement, and time were
recorded continuously throughout the testing period. Relevant parts of the recorded data
were used for CTE calculation based on each method.

Specimens were subjected to at least three heating-cooling cycles (Figure 3-11).
This way, if one cycle was not suitable for CTE calculations due to problems with test
conditions (specimen and LVDT misalignment, lack of proper seating of the specimen,

etc.), the replicate cycles could be used for CTE calculations.

 

0.0035

0.0030 -

0.0025 -

0.0020 —

0.0015 a

0.0010 -

Displacement (In.)

0.0005 ~

0.0000 1

 

-0.0005

 

 

-0.0010 . . . . . . .
0 200 400 600 800 1000 1200 1400

Time (Min.)
Figure 3—1 1. Three Typical Heating-Cooling Cycles
Thermocouples were inserted in water bath to record the water temperature.
Thermocouples were also embedded in a number of concrete samples to monitor the
specimen temperature during the CTE test. It was found that both the specimen and water

follow similar temperature signature, however, the concrete specimen lags the water

65

temperature by approximately 10 minutes. Figure 3-12 shows a typical time-temperature

graph.

 

60 122

501

§

 

I
00
0\

Temperature (°C)
L»)
o
l.
\1
Temperature (°F)

 

 

 

~68

10 ~50

0 1 1 1 1 1 32
0 50 100 150 200 250 300

Time (Min.)

 

 

 

I Concrete Temp. 0 Water Temp.

 

Figure 3-12. A Typical Concrete and Water Temperature Graph

66

CHAPTER 4

RESULTS AND DISCUSSION

4.1 Introduction

Results of the laboratory experimental program described in Chapter 3 are documented in
this chapter. The summarized results in this chapter include physical properties of
aggregate, flesh and hardened properties of concrete, and coefficient of thermal
expansion of concrete fabricated using various aggregate types.

Furthermore, the statistical and operational (practical) impact of test variables on
the magnitude of CTE is presented in this chapter. Two methods of CTE calculation were

compared using statistical analysis. The impact of CTE on pavement design and its

implications are also discussed.

67

 

4.2 Physical Properties of Coarse Aggregates

The absorption capacity and specific gravity tests (ASTM C127) were conducted on

sampled aggregates. For each aggregate type, the aggregate sample was divided into four

batches and tests were conducted on each batch. The results were then averaged. The

summarized results are presented in Table 4-1.

CTE 8 specimens were field-prepared specimens from an actual paving project in

Michigan (supplied by the MDOT) and therefore no laboratory test information was

available for this mixture.

Table 4-1. Physical Properties of Coarse Aggregates

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

- S ecific Gravi (SD)**
Mixture Airgge Pit AC * p ty
.1. B 1k
ID Class Number (SD) Apparent u 1 Bulk Dry
SSD
. 1.13 2.655 2.591 2.552
[ CTE] “mm” “'47 (0.04) (0.004) (0.033) (0.052)
2.77 2.762 2.637 2.566
/ CTE2 Gravel ”'56 (0.11) (0.011) (0.010) (0.011)
Dolomitic 0.69 2.698 2.668 2.649
LCTE 3 Limestone 75'“ (0.04) (0.002) (0.004) (0.005)
3.47 2.490 2.393 2.329
[ CTE4 Slag 82‘” (0.12) 0.0440 (0.008) (0.037)
. 0.68 2.787 2.753 2.735
/ CTE 5 DOIOm‘te 49"” (0.02) (0.001) (0.001) (0.002)
Gabbro 0.21 2.928 2.916 2.91
LCTE 6 (Trap Rock) 95‘") (0.01) (0.005) (0.005) (0.006)
. 3.13 2.769 2.628 2.548
K CTE 7 DOIOm‘te 58‘” (0.27) (0.009) (0.006) (0.012)
/ CTE 8 1 Dolomite 91-06 N/A N/A N/A N/A

 

" AC = Absorption Capacity (%). ** SD = Standard Deviation. T SSD = Saturated Surface-Dry.

68

 

 

 

 

4.3 Fresh Concrete Properties

Fresh concrete properties results conducted according to aforementioned standards (Table
3.6) are shown in Table 4-2. The target slump was 3 :1: 0.5 in. and the target air content
was 6.5 i 1.5 %. It should be noted that these tests were conducted to make sure that the

concrete tested in laboratory is not significantly different from the concrete used in field

 

for paving. However, strict conformance to field parameters was not the goal of this

study.

 

CTE 8 specimens were field-prepared specimens from an actual paving project in
Michigan and no flesh concrete properties test information was available for this mixture.

Table 4-2. Fresh Concrete Properties

 

 

 

 

 

 

 

 

 

 

Mixture ID Test Parameter
Slump, inches Air, % Unit Weight, pcf" Temperature, °F
CTE 1 3.0 6.0 147.0 54
CTE 2 4.0 5.2 149.4 70
CTE 3 6.0 6.0 145.0 70
CTE 4 3.0 5.8 145.4 77
CTE 5 3.8 4.2 150.8 75
CTE 6 4.0 5.0 152.0 61
CTE 7 3.0 4.9 148.2 69
CTE 8 N/A N/A N/A N/A

 

 

 

 

 

 

 

* pcf = pounds per cubic foot

 

69

4.4 Hardened Concrete Properties

Hardened concrete properties tests were conducted on laboratory cured specimens at 1, 3,
7, 1 4, 28, 90, and 365 days after casting. Average 28-day test results for all mixtures are
summarized in Figures 4-1 through 4-4. Figures 4-1 and 4-2 show the average 28-day
compressive and flexural strengths respectively. Dashed lines in these figures represent
target strengths suggested by MDOT. Figures 4-3 and 4-4 show 28-day elastic modulus
and split tensile strength. The error bars in figures 4-1 through 4-4 represent the standard
deviation of test values among three replicates. The data used for developing these
summary graphs are presented in Appendix A.
All concrete batches conformed to average 28-day compressive strength of 3500
psi. The 28-day flexural strength requirement of 650 psi was met for all mixtures except
for mixture 3. Two specimens of this mixture met the 650 psi requirement (658 and 666
psi). The third specimen had flexural strength of 612 psi which brought the average to
645 psi. However, this mixture met the requirement at 90 days of age.
The hardened concrete property tests were conducted for two reasons:
| To evaluate the quality of the concrete
' To be used as level 1 inputs in M-E PDG and HIPERPAV 11 software

In general, the delivered concrete met the required specified strengths.

70

 

 

28-day compressive strength, psi

 

 

CTE] CTE2 CTE3 CTE4 CTE5 CTE6 CTE7 CTE8

Mixtrn'e ID

Figure 4-1. Average 28-day Compressive Strength Values

1 000

 

900 —
800 —
700 —
600 —
500 ~
400 —
300 —
200 -
100 — p

0 ' . ~ 1

28—day flexural strength, psi

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I

l

I

CTE] CTE2 CTE3 CTE4 CTE5 CTE6 CTE7

Mixture ID

Figure 4-2. Average 28—day Flexural Strength Values*

* There was no flexural specimen for CT E 8.

 

 

 

o. 9
O O
I

:5
O
l

28—day elastic modulus x 10‘, psi
N t.»
'o o

p—i
O
l

 

 

 

.0
O
l

CTE] CTE2 CTE3 CTE4 CTE5 CTE6 CTE7 CTE8
MixtureID

Figure 4-3. Average 28-day Ela stic Modulus Values

 

700

8
O
1

  

28-day split tensile strength, psi

 

 

CTEl CTE2 CTE3 CTE4 CTE5 CTE6 CTE7 CTE8*
Mixtm'e ID
Figure 4-4. Average 28-day Split Tensile Values

* There was only one split tensile specimen for CTE 8.

72

4.5 Thermal Property

The CTE test was conducted on laboratory cured specimens at 3, 7, 14, 28, 90, 180, and

365 days after casting. CTE was calculated based on both the AASHTO TP60 and the

Revised TP60 methods. Figure 4.5 shows average 28-day CTE results for each aggregate

type based on AASHTO TP60 method. The numbers shown above each aggregate type

are typical CTE ranges as reported in the M-E PDG3 . Average 28-day results based on

both methods are presented in Table 4.3.

 

 

 

 

 

11'5 _ pa/°F 3.4—5.1 6.0-8.7 51-59 445.3 5.1-6.4 6'4

11.0 — (pa/°C (6.1-9.2) 10.8—15.7 (9.2-10.6) (7.9-9.5) (92-115) 2 6.1

10.5 — _ __ 58

10.0 — _ 55
5,9 3(5) 5 . 5.2 a
3 ' — 4.9 1
III 8.5 — III
F n — 4.6 p.
U 8.0 1' U

.0 2 4.3

7.5 — g.

7.0 — .8 ‘ 4-0

"m — 3 7

6.0 ..... 53“ — 3.4

5.5 1 ’4 . 3.1

 

 

 

 

 

 

 

 

 

   

 

CTE 1 CTE 3 CTE 2 CTE 4 CTE 6 CTE 8 CTE 7 CTE 5

Mixture ID

Figure 4.5 Average 28—day CTE Results

The CTE results for each mixture at all ages are presented in Figures 4-6 to 4-13.

The data used for computing the summary information are presented in Appendix B. The

error bars in these figures show the standard deviation of the test specimens based on

three replicates. The dashed lines show the typical CTE ranges for concretes made with

that particular aggregate according to M-E PDG3. These ranges are shown in Table 4.4.

73

 

Table 4.3 Average 28-day CTE Results

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

28-da CTE, e/°F a/°C
Mixture ID y 11 (p. )
AASHTO TP60 Revised TP60
CTE 1 4.54 (8.18) 4.51 (8.12)
CTE 2 5.84 (10.52) 5.82 (10.47)
CTE 3 4.51 (8.11) 4.53 (8.16)
CTE 4 5.71 (10.27) 5.72 (10.29)
CTE 5 5.92 (10.65) 5.99 (10.79)
CTE 6 5.41 (9.73) 5.44 (9.79)
CTE 7 5.90 (10.62) 5.94 (10.69)
CTE 8 5.87 (10.57) 6.00 (10.80)
Table 44. Typical CTE Ranges for Common Components and Concrete3
Material Coefficient of Concrete Coefficient of
T Thermal Thermal Expansion (made
ype Expansion, 10'6/°F from this material), 10‘6/°F
Aggregates
Marbles 2.2-3.9 2.3
Limestones 2.0-3.6 3.4-5.1
] Granites & Gneisses 3.2-5.3 3.8-5.3
Syenites, Diorites,
Andesite, Basalt, Gabbros, 3.0-4.5 4.4-5.3
Diabase
[Dolomites 3.9-5.5 5.1-6.4
Blast Furnace Slag 5.1-5.9
Sandstones 5.6-6.7 5.6-6.5
Quartz Sands & Gravels 5.5-7.1 6.0-8.7
[ Quartzite, Cherts 6.1-7.0 6.6-7.1
Cement Paste (saturated)
w/c = 0.4 to 0.6 1 10-11 --
Concrete Cores
Cores from LTPP pavement -6 -6 -6
sections, many ofwhich N/A 40*10 - 55*10 — 72*10

were used in calibration

 

 

(Min - Mean - Max)

 

74

 

 

 

7.5

 

 

 

 

 

 

 

 

 

 

 

 

14 28 180

Time, days

90

 

 

—AASHTOTP60 [:1RevisedTP6O — - -Minimum - - - -Maximum

 

 

Figure 4-6. CTE l (Limestone Concrete) Coefficient of Thermal Expansion

 

 

 

 

 

 

 

 

 

17.5 ~ _ 9.5

16.5 —

15.5 -' """""""""""""""" '1 8.5

14.5 —
U 13.5 — — 7.5 m
g, 12.5 - a
=': 11.5 - - 6.5 ..
E 10.5 — ' E.
U 9.5 - ’ 5'5 U

8.5 ~

7.5 _ 4.5

6.5 — — 3.5

5.5 ~

4.5 4 2.5

 

 

14 28 90 180 365

Time, days

 

 

 

_ AASHTO TP60 [:3 Revised TP60 — - - Minimum - - - - Maximum

 

Figure 4—7. CTE 2 (Gravel Concrete) Coefficient of Thermal Expansion

75

 

 

13.5 7.5

p—n
p—s
(I!
l
l
O\
U!

1
U:
U!

CTE, uePC
CTE, u£I°F

 

 

 

Time, days

 

F-AASHTOTP6OZIRevisedTP6O — ' -Minimum‘ ' ' 'Maximum

 

Figure 4-8. CTE 3 (Dolomitic Limestone Concrete) Coefficient of Thermal

 

 

 

 

 

 

 

 

 

Expansion
13.5 v e v—vv e vv—ve vvv‘ 7.5
1

12.5 1 ‘

11.5 ; i 6.5

10.5 5 7 - - -‘
g L 5.5 g
1 9.5 " _ _ _ _ 1
m“ ‘ m“

8.5 1 .
8 r 4.5 S

7.5 7‘1

6.5 ~ 1 3.5

5.5 ‘

4.5 1 e e L 2.5

3 7 14 28 90 180 365
Time, days

 

L— AASHTO TP60 1:1 Revised TP60 — - - Minimum - - - - MaximumJ

 

Figure 4-9. CTE 4 (Slag Concrete) Coefficient of Thermal Expansion

76

13.5

 

12.5 -
11.5 ~
10.5 -
9.5 ~
8.5 7

CTE, paPC

7.5 A
6.5 -
5.5 7
4.5 -

 

 

 

3 7 14

 

28

Time, days

 

90

 

 

180

 

365

7.5

 

—— 2.5

 

—AASHTOTP60E:RevischP60— ' 'Minimum- - - -Maximum

 

Figure 4-10. CTE 5 (Dolomite Concrete) Coefficient of Thermal Expansion

 

13.5
12.5 e
11.5 -
10.5 ‘

CTE, u£/°C

 

 

28

Time, days

90

180

 

365

7.5

 

 

 

!— AASHTO TP-60 III Revised TP60 — - - Minimum - - - - Maximumj

 

 

Figure 4-11. CTE 6 (Gabbro or Trap Rock Concrete) Coefficient of Thermal

Expansion

77

 

 

 

 

 

 

 

 

 

 

13.5 , v 7.5

12.5 e

11.5 e ------------------------------- ” 6-5

10.5 e
g — 5.5 9‘;
=- 9.5 fl __ - I _ - _ _ - _ _ . 1
“'5 8.5 _ ‘5
5 — 4.5 S

7.5 e

6.5 . - 3.5

5.5 -

4.5 e _. — er . v 2.5

3 7 14 28 90 180 365
Time, days

 

_AASHTOTP60i:IRevisedTP60— - -Minimum- - - Mammy]

 

Figure 4-12. CTE 7 (Dolomite Concrete) Coefficient of Thermal Expansion

 

 

 

 

 

 

 

 

 

13.5 17 'v‘" 77 7 7.5

12.5 7

11.5 i 6.5

10.5 7
g 5.5 g
1 9.5 7 1
H“ - m“
5 8.5 4.5 5

7.5 7 ‘

6.5 7 ‘ 35

5.5 7

4.5 7 2.5

7 14 28 90 180 365
Time, days

 

_ AASHFO TP60 :1 Revised TP60 — - - Minimum - - - - Maximum

 

Figure 4-13. CTE 8 (Dolomite Concrete) Coefficient of Thermal Expansion

78

F HWA is conducting a study to determine the variability in CTE test results as
well as variability in different CTE devices. As part of this research, a number of
specimens were tested at MSU CIL according to AASHTO TP60 protocol.

A Grey Cast Iron (grade 410), a concrete specimen made with gravel coarse
aggregate, and a concrete specimen made with limestone coarse aggregate were shipped
to the laboratory. Physical properties of the coarse aggregates including specific gravity
(SG), absorption capacity (AC), and moisture content (MC) are presented in Table 4-5.

Table 4—5. Physical Properties of the FHWA Coarse Aggregates

 

 

 

Coarse Aggregate SG AC, % MC, %
Gravel 2.655 0.83 0.1
Limestone 2.710 0.38 0.1

 

 

 

 

 

 

The test plan required each specimen to be tested twice on two consecutive days.
The CTE test results are reported in Table 4-6.

Table 4-6. CTE Test Results for F HWA Specimens

 

 

 

 

 

 

 

Material CTE, 118/°F CTE, 118/°C
6.33 1 1.39
Steel 401
6.34 1 1.41
Gravel 6.83 12.29
concrete 6.84 12.32
Limestone 4.06 7.31
Concrete 4.05 7.29

 

 

 

 

 

At the time of writing this dissertation, no information about the results from
other laboratories is available, and therefore no additional analysis can be conducted at

this point.

79

4.6 CTE Test Variability

The test variability (5CTE) was determined by subtracting the measured CTE within a

batch (i.e. for a given mixture ID and sample age) from the batch mean (deviation).

Figure 4-14 illustrates the frequency histogram of 5CTE- Approximately 97% of the data
have a SCTE which is between $0.3 118/0F (i0.5 1.13/°C). The coefficient of variation

ranges between 2.5% and 6% among different mixtures. These results are based on
AASHTO TP60 test method.

The test variability could stem from a variety of factors including variable internal
relative humidity in the sample at the time of testing, non-uniform temperature
distribution in the sample, inhomogeneities in the specimen at the time of fabrication,

LVDT sensitivity, power fluctuations, and possible frame induced errors as reported by

 

 

 

 

 

Tanesi et 31.20.
5cmtlfi/°C
wommwommzemmmowmmowom
E$$$$$$$44323222§2232
90 [Li lllllllllllll ll
807 77
70- -0.5 ua/°C 63 0.5 u8/°C
5607 -O.3 113/°F 55 - 54 0.3 118/°F
.250-
340— 3135 37
L.
E1304 25 21
20~ 81415 ,- 13
10—1023 5201
O‘Ar—r-T-r "-‘r—_l"-"
mam—omwvoxo~mr\~mo<roomxo—~
emmmmm~~~oqoo~~mmmmmv
qqqqqqqqqqqoooooooooo
5CTEu’elg/F

Figure 4—14. Variability Histogram

80

4.7 Statistical Analyses

Factors that impact the magnitude of the measured CTE of concrete are discussed and
statistical analysis on their significance is described in this section. A comparison
between two different methods of CTE calculation (using statistical tools) is also

presented here. All the analyses were performed on the AASHTO TP60 test results.

4.7.1 Statistical Analysis of Factors Affecting CTE of PCC

Factors that potentially affected the magnitude of the CTE in this experiment were
aggregate geology, sample age, and number of heating-cooling cycles. To investigate the
impact of these factors and their interactions on the magnitude of CTE, a “factorial
treatment design” was employed. In a factorial treatment design, one factor — for example
aggregate type - is tested over one or more other factors, for example sample age and
number of heating-cooling cycles. Each factor has several categories called “levels”. For
example, levels of the sample age factor are 3, 7, 14, 28, 90, 180, and 365 days. The
design factorial included: 1) aggregate geology with eight levels; 2) sample age with
seven levels and 3) number of heating-cooling cycles with three levels. There were three
replicate samples for each combination of aggregate geology and sample age. Table 4-7
shows the factorial design table.

Aggregate geology, sample age, and number of cycles were considered as fixed
effects. In fixed effects, different levels of factors are reproducible. In other words if the
experiment was to be repeated, the levels could be duplicated. Individual sample ID
(replications) within each aggregate geology and sample age was considered as a random

effect because for each combination of aggregate type and sample age, a sample ID was

81

assigned randomly to the three replicates from the same batch of concrete. Additionally,
the number of cycles were treated as repeated measurements of individual samples since
multiple measurements (cycles) were taken on each of the replicate samples.

Table 4-7. Factorial Design Table*

A D
Aggregate Heating-Cooling ge, ays Total

Geology C 3 14 28 90 180 365

CTE]

CTE2

 

T 2 1 24 24 24 24

* Three replicates for each combination of factors were considered totaling the CI‘E values to 3 x 165 =
495.

Because the experiment included both fixed and random effects, the statistical
analysis was performed using “mixed effects” model. The following general linear model

describes the relationships between factors:

82

y = p + Agg_Type + Age + Agg_TypexAge + Sample_ID(Agg_Type Age) + Cycle +
Agg_Typenycle + Agenycle + Agg_TypexAgenycle + e
Description of each term in this model is presented in Table 4-8.

Table 4-8. Description of Terms in the Mixed Effects Model

 

 

 

 

 

Term Description
y Response variable which is the CTE value
11 Overall mean
Agg_Type Fixed effect of the aggregate geology
Age Fixed effect of the sample age at the time of testing

 

Interaction effect of the aggregate geology and sample

Agg_TypexAge age

 

Sample_ID(Agg_Type Age) Random effect of the replications

 

 

 

Cycle Fixed effect of the cycle numbers

Agg_Typex Cycle Interaction effect of the aggregate geology and cycle
numbers

Agenycle Interaction effect of the sample age and cycle numbers

 

Interaction effect of the aggregate geology, sample age,

Agg_TypexAgenycle and cycle numbers

 

6 Random experimental error

 

 

 

 

The statistical significance of the factors and their interactions was evaluated by
using the analysis of variance (ANOVA) method.

The first step in the ANOVA method is checking of assumptions for the model.
The assumptions require the residuals (difference between predicted and measured
values) to be: 1) independent; 2) normally distributed; and 3) having a common variance

for all levels of each factor.

83

The independence assumption was evaluated based on the method by which the
individual samples were chosen for testing. This assumption was satisfied because the
samples were randomly selected for CTE measurement for each aggregate type and age.

The normality assumption was checked by inspection of the distribution of the
residuals (histogram) and normal probability plot of the residuals after elimination of the
outliers. The residual histogram should show a symmetric bell shaped distribution. The
normal probability plot of the residuals should not deviate significantly from the straight
line showed on the graph. Unusually high or low residuals were checked to determine
whether those residuals were outliers or not. Raw data (measurements) of such residuals
were checked to detect any equipment malfunction during that particular test cycle. The
histogram and normal probability plot of the residuals for the entire dataset are illustrated
in Figures 4-15 and 4-16, respectively.

The homogeneity of the variances was assessed by examination of variances of
the residuals for all levels of each factor. Side-by—side box plots were employed to help in
the identification of the largest and smallest variances. For this assumption to be valid,
the ratio of the largest variance to the smallest variance should not exceed 3, 5, or 10.
These limits are different rules of thumb and “5” was selected for this analysis.

Side-by—side box plots of the residuals for different levels of “age” variable are
shown in Figure 4-17. In this figure the median is represented by the horizontal line
inside the box. The top and bottom of the box represent the 3rd quartile (75th percentile)
and the lst quartile (25th percentile), respectively. The distance between these two is the
interquartile range (IQR). Whiskers are drawn to capture the minimum and maximum

observations.

84

120

 

Count

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

n‘ .I. .' '. I.. I. . I... .I. .I. ,.,I ...' . ' '
-0.138 -0.113 -0.088 -0.063 -0.038 ~0.013 0.013 0.038 0.063 0.088 0.113 0.138
Resid
Figure 4-15. Histogram of the Residuals
1.5E—01 7 a
— Normal Parameters aunt/3
Mean (Mu) O /
1.0E-017 Std Dev (Sigma) 0.052117 ’
5.013-02- /
IE
-5.0E-027
A“
-1 .013-01 e ,,/;W
./' f
/.
-1.513-01-.° /° . . . . . . .fi .
0.1 l 5 10 25 50 75 90 95 99 99.9

Normal Percentiles

Figure 4-16. Normal Probability Plot of the Residuals

85

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.2
0.1 -
1'2
3 0 ~ + 4 +
n:
-0.1 -
-0.2- . . . . . . .
3 7 14 28 90 180 365
Age, days

Figure 4-17. Side-By-Side Box Plots of the Residuals for Age Factor

Based on the availability of the data for different comparisons and analyses,
several datasets were used throughout the statistical analysis process. For each dataset,
the assumptions were checked and appropriate adjustments were made to the analyses
where necessary.

After checking the assumptions, the significance of the factors, and their
interactions were investigated. The Least Significant Difference (LSD) method was used
for all-pairwise comparisons with confidence level of 95% (01 = 0.05). A main or
interaction effect was considered significant if p-value (level of significance) was less
than 0.05.

The results from the mixed effects analysis are summarized in Table 4-9. The

impact of aggregate geology, sample age and number of heating-cooling cycles on CTE

86

was found to be significant (p-value less than 0.05). Furthermore, the interactions
between: 1) aggregate geology and sample age; and 2) aggregate geology and number of
heating-cooling cycles were found to be statistically significant. The significant effects
are highlighted in Table 4-9.

Table 4-9. Tests of Fixed Effects

 

 

 

 

 

 

 

 

 

 

 

 

Effect F Value p-value
Aggregate Type 383.73 <.0001
Age 33 <.0001
Aggregate TypexAge 1.57 0.0338
Cycle Number 123.09 <.0001
Aggregate Typenycle Number 5.47 <.0001
Agenycle Number 1.42 0.1624
Aggregate TypexAgenycle Number 1.14 0.303

 

4.7.1.1 Aggregate Type Effect

According to the Table 4-9, aggregate type and its interactions with age and cycle

number are significant. This implies that concrete samples made with different types of

coarse aggregate (different geologies) have different CTE values. This finding is in

agreement with the results of other published literature (references 1 and 15 through 22).
Different mineralogical composition among aggregate types is the reason for

difference in aggregate CTE values. As a result, the CTE of concrete is impacted by the

type of its constituent aggregate.

87

4.7.1.2 Sample Age Effect

To investigate the effect of sample age at the time of test on CTE, the interaction effect of
aggregate type and sample age was examined. This implies that the effect of age on CTE
was investigated within each aggregate type. The variation in CTE as a function of
sample age for all aggregate types is illustrated in Figure 4-18. In this figure, within each
aggregate type, identical letters refer to replicate ages that are not significantly different
from each other at a confidence level of 95%. The column bars represent average CTE.

Some column bars have two letters assigned to them. In the all-pairwise
comparison, each level of age is compared to all other levels. For example, in CTE 1, 3-
day CTE was compared to all other levels (showed by letter “a”). When 3-day CTE is
different fiom 90-day CTE, this comparison is finished and a new comparison starts with
7-day CTE (denoted by letter “b”). Within each of the aggregate geologies, different
column bars having the same letter indicate a group of ages which are statistically
similar. If a column bar has two letters, it belongs to either one of the groups, but not both
groups.

To study the impact of sample age within an aggregate type, an all-pairwise
comparison using the LSD method was used. Based on the LSD analysis, it was found
that for most aggregate types the magnitudes of CTE measured at the end of 28 days were
lower than the magnitudes measured at the end of 90 to 365 days and that the difference

was statistically significant.

88

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

salad ‘am afimAv
.r,
F
m ‘-
(U
m
(U
m
m
N
m
(U
(U
(B
(U
m
(U
:1L72‘Lfrttr'm'ii I 3
1-1135511"?! 1' Jr
P] ’ '. ,2 43.8. r .V-i'a'
I
1n. v:
m N
v—t '—

3431i ‘31:) afiuJaAv

89

 

CTE 2 CTE 3 CTE 4 CTE 5 CTE 6 CTE 7 CTE 8
Mixture ID

CTE 1

 

E190 M180 E3365!

I14 I28

5‘7

E13

 

Figure 4-18. Average CTE for Various Aggregates as a Function of Test Specimen Age Based on 3 Replicates*

* Following column bars with the same letter within a CTE mixture type are not significantly different from each other.

Wittmann and Lukas29 have shown an increase in the CTE of saturated concrete
specimens around room temperature up to 56 days of age. Rudeloff and Sieglerschmidt30

have also reported increase in CTE of wet-cured specimens up to 90 days. Meyers5 have
shown that at 100% RH, the CTE of cement pastes 12-18 months old and 15-16 years old
were higher than the CTE of pastes 0-6 months old.

The effect of age on CTE is associated with proceeding hydration and change in

the composition of the cement paste which in turn affects the CTE of concrete.

4.7.1.3 Number of Heating-Cooling Cycles Effect
A sample from each’CTE mixture design was subjected to three heating and cooling
cycles to study the effect on CTE. In addition, since each mixture design was using a
different coarse aggregate source, the heating and cooling effect could be analyzed by
aggregate type. The variation of CTE values as a function of heating-cooling cycles is
illustrated in Figure 4-19. In this figure, within each aggregate type, identical letters refer
to cycle numbers that are not significantly different from each other at a confidence level
of 95%. The column bars represent average CTE.

In all cases (except for CTE 6) the CTE determined at the end of the first heating-
cooling cycle was higher than the CTE determined at the end of the second and third
heating-cooling cycles and the difference was statistically significant. These conclusions

are based on analyses similar to age effect analyses.

Kohler et al.17 have also studied the effect of three consecutive heating-cooling
cycles on the magnitude of concrete CTE and have shown a decrease in CTE values with

additional heating-cooling cycles.

90

Reasons for this effect could be lack of a stable condition regarding pore water

(Kohler et al.17), specimen and LVDT misalignment, and lack of proper seating of the

specimen.

 

Average CTE, uc/°C
Average CTE, ucl°F

 

 

 

Mixture ID

 

| I Cycle 1 El Cycle 2 I Cycle 3

Figure 4-19. Average CTE for Various Aggregates as a Function of Number of
Cycles Based on 3 Replicates*

* Following column bars with the same letter within a CT E mixture type are not significantly different
from each other.

The statistical analysis also showed that for most mixture types, the CTE values
based on second and third cycles were not statistically different from each other. In order
to investigate whether additional cycles produce the same results as cycles two and three,
all the specimens 365-day old were subjected to six heating-cooling cycles. The same
statistical analyses mentioned before were employed in order to compare the CTE values

based on each of these six cycles. It was found that CTE value based on cycle one is

91

statistically higher than CTE values based on all other cycles and that CTE values based
on cycles 2 through 5 were not statistically different from each other. The results of this
analysis indicate that in a test with three heating-cooling cycles, the CTE measurements
taken during cycles two and three probably correspond to a more stable condition as
opposed to CTE value based on only one heating-cooling cycle.

CTE 6 mixture (with gabbro or trap rock coarse aggregate) showed little or no
difference in measured CTE between cycles. By referring to Table 3.2 of the previous
chapter, it can be seen that this aggregate has the lowest absorption capacity (0.21%) and
highest bulk specific gravity (2.910) among different aggregate types, which is an
indication of a denser aggregate structure and possibly a factor in the stable thermal

behavior of this mixture.

4.7.2 Comparison of AASHTO TP60 and Revised TP60 Methods

A factorial treatment design similar to one described in previous section was used to
evaluate the effect of calculation method on magnitude of the CTE. The factors in this
model were aggregate type and age with the same levels described earlier, plus method of
CTE calculation with two levels (AASHTO TP60 and Revised TP60 methods). There
were three replicate samples for each combination of aggregate geology and sample age.
Table 4-10 shows the factorial design table.

Similar to the previous model, the aggregate geology, sample age, and calculation
method were considered as fixed effects and individual sample ID (replications) within

each aggregate geology and sample age was considered as a random effect.

92

Table 4-10. Factorial Design Table*

Aggregate CTE Calculation Age, Days
Geology Method

Total
14 28 90 180 365

AASHTO T P60
Revised TP60
AASHTO TP60
Revised TP60
AASHTO TP60
Revised TP60
AASHTO TP60
Revised TP60
AASHTO TP60
Revised TP60
AASHTO TP60
Revised TP60
AASHTO TP60
Revised TP60
AASHTO TP60
Revised TP60

Total 14 16 16 16 16 16 16

CTE]

CTE2

CTE3

CTE4

HO\O\\I\I\I\I\I\I\I\I\I\I\I\I\I\I

y—n

0

 

* Three replicates for each combination of factors were considered totaling the CI'E values to 3 x 110 =
330.

The following general linear model describes the relationships between factors in
the “mixed effects” model:
y = p + Agg_Type + Age + Agg_TypexAge + Sample_ID(Agg_Type Age) + Method +
Agg_TypexMethod + AgexMethod + Agg_TypexAgexMethod + e

Description of each term in this model is presented in Table 4-11.

93

Table 4-11. Description of Terms in the Model for Method Comparison

 

 

 

 

 

Term Description
y Response variable which is the CTE value
p Overall mean
Agg_Type Fixed effect of the aggregate geology
Age Fixed effect of the sample age at the time of testing

 

Interaction effect of the aggregate geology and sample

Agg_TypexAge age

 

Sample_ID(Agg_Type Age) Random effect of the replications

 

 

 

Method Fixed effect of the method
Agg_Typex Method Interaction effect of the aggregate geology and method
Agex Method Interaction effect of the sample age and method

 

Interaction effect of the aggregate geology, sample age,

Agg_TypexAgex Method and method

 

 

 

 

e Random experimental error

 

The statistical significance of the factors and their interactions was evaluated by
using the analysis of variance (ANOVA) method. The ANOVA assumptions were
checked and found to be satisfactory for the data set used in the analysis.

After checking the assumptions, the significance of the factors, and their
interactions were investigated. The confidence level for this analysis was 95% (a = 0.05).
The results from the mixed effects analysis are summarized in Table 4-12. The impact of
calculation method on CTE was found to be significant (p-value less than 0.05). This
implies that the two methods produce different results with confidence level of 95%. The
interactions between the method effect and other effects were not statistically significant.

The significant effects are highlighted in Table 4-12.

94

Table 4-12. Tests of Fixed Effects for Method Comparison

 

 

 

 

 

 

 

 

Effect F Value p-value
Aggregate Type 421.57 <.0001
Age 50.82 <.0001
Aggregate TypexAge 7.01 <.0001
Method 1 1.55 0.0009
Aggregate TypexMethod 1.32 0.2469
AgexMethod 0.97 0.4460
Aggregate TypexAgexMethod 0.62 0.9617

 

 

 

 

 

An analysis of the least square means for these two methods showed that the
Revised TP60 method resulted in higher CTE values in comparison with the AASHTO
TP60 method. The estimated difference was 0.023 1.13/°F (0.041 118/°C). By referring to
sections discussing the practical impact of age and number of cycles on pavement
performance (4.8.2.3 and 4.8.2.4), it can be seen that even differences a few times larger
than 0.023 118/°F (0.041 118/°C) do not seem to have practical impact on pavement
performance. Therefore, this very small difference does not have any practical

significance in concrete pavement design and performance.

95

4.8 Impact of CTE on Performance of Jointed Concrete

Pavements

CTE affects both thermally induced stresses within the pavement and joint movement. As
a result, premature cracking, mid-panel and corner cracking, faulting, and joint spalling .
can occur, as mentioned in Chapter 1. In order to minimize the occurrence of these
distresses, the CTE value among other variables must be considered while designing a
pavement.

Sensitivity analyses were conducted to investigate the impact of CTE on short-
terrn (first 72-hours) and long-term performance of J PCP using HIPERPAV II and M-E
PDG software, respectively. HIPERPAV 11 software is designed to evaluate the impact of
pavement design, PCC mixture design, construction methods and environmental
conditions on the early-age behavior of concrete pavements. M-E PDG software is a tool
for the design of new and rehabilitated pavement structures based on mechanistic-
empirical principles.

The performance parameter of interest for the short-term effect was ratio of the
maximum tensile stress in concrete. slab to PCC tensile strength. The performance
parameter of interest for the long term effect was transverse cracking. Tables 4-13 and 4-
16 summarize the inputs for sensitivity analyses used to investigate the short-term and
long-term effects of CTE on the performance of JPCPs.

All the analyses were performed on the results from AASHTO TP60 test method.

96

4.8.1 Short-Term Effects Analysis

A factorial was developed to investigate the effects of CTE on the early age of the
concrete pavement. The details about the variables and levels used in the sensitivity
analysis (for each type of mixture) are summarized in Table 4-1 3. In addition, CTE
values for each type of aggregate were maximum, minimum, and average measured
values. Levels of splitting tensile strength were maximum, minimum, and average 28-day
values obtained from laboratory testing. Actual mixture design proportions were used as
inputs. Modulus of rupture (MOR) and elastic modulus were obtained from laboratory
test data. The initial PCC mixture temperature was assumed to be 75 °F for all mixture
designs.

All HIPERPAV 11 runs (1728 runs) were performed and the generated results
were analyzed. Ratio of maximum tensile stresses developed in cross sections to tensile
strength of the section was calculated for all combinations of variables (Max. Stress /
PCC Strength in Table 4-12). Since the HIPERPAV 11 analyses were performed on
paving mixture designs consistent with MDOT standard specifications, the potential of
early-age cracking (within 72 hours after construction) was minimal (maximum stress to
PCC strength ratio < 1). Minimum and maximum ratios were associated with CTE l and
CTE 8 mixtures respectively.

Two examples of the sensitivity analysis are presented here. Input variables for
these examples (CTE 1 and CTE 8 mixtures) are summarized in Table 4-14. The results
of the analyses for these examples are presented in Table 4-15, Figure 4-20 and Figure 4-

21. In each of these examples, combinations of variables with minimum and maximum

97

ratios were selected as designs 1 and 2 respectively. These ratios were used to compare

the two designs within each mixture type.

Table 4-13. Input Variables for HIPERPAV II Sensitivity Analysis

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Factor Variable Levels Remarks
Joint Spacing 12' , 15' , 20' 3 Levels
- . 1.25" (for 9" slab) .
Desr
gn Dowel Diameter 15,, (for 14,, slab) F 1xed
Slab Width 12' , 14' 2 Levels
PCC Slab Thickness 9" , 14" 2 Levels
CTE _ 3 Levels based on
measured values
Compressive
- Measured values
PCC Strength
Modulus of Rupture - Measured values
Elastic Modulus - Measured values
Materials Splitting Tensile _ 3 Levels based on
Strength measured values
Granular Base .
B Base Type (D G AB) Frxed
ase
Base Thickness 4" Fixed
Subbase Type Sand Fixed
Subbase
Subbase Thickness 16" Fixed
Subgrade Soil Type A-7-6 (Fine) Fixed
. . Single Coat (SC),
Constructron Currng Method Double Coat 2 Levels
Environmental Climatic Region Lansing Fixed

 

98

 

Table 4-14. Examples of HIPERPAV II Sensitivity Analysis Inputs

 

 

 

 

 

 

 

 

 

 

 

 

. Slab Joint Slab Splitting
Mrlrgure Design Thickness Spacing Width Coat Tensile CTE/013° F
(in) (a) (8) Strength (psi)
1 14 12 so 28-day (516) “$352
CTE 1 '
2 14 14 so Min (471) MESS)”
1 14 12 sc Max (587) $318536
CTE 8 '
. Mu6n
2 9 14 sc Mln(471) (11.46)

 

 

Table 4-15. Examples of HIPERPAV II Sensitivity Analysis Results

 

 

 

 

 

 

 

 

 

 

 

Mixture Desi Max. Stress Time PCC Strength Max. Stress / PCC
11) g“ (psi) (hr) (psi) Strength
1 28.7 43 284.6 0.1008
CTE 1
2 83.1 43 287.4 0.2891
1 28.6 42 182.8 0.1565
CTE 8
2 84.7 50 147.9 0.5727

 

By examining Figure 4-20, it can be seen that for design 1, the curves

representing slab stresses and PCC strength are farther away from each other, which is an

indication of better performance as compared to design 2 in which the curves are closer

to each other. This observation can also be seen from last column in Table 4-12 which

shows a lower ratio of Max. Stress / PCC strength for design 1. It can also be seen that

the CTE of PCC is one of the factors affecting the performance (lower CTE is associated

with better performance). A similar behavior can be seen for CTE 8 designs (Figure 4-

21).

99

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A 350

m 44%?” '—
30 300 ..___ , 4’;:4_4.

5 250 — e __.,//«r

e //

m 200 — 4’

a //

a 150 . ——e // - e

3’; 100 77”

2 A

s 0 ”W

Time (hr.)
— PCC Strength CTE 1 Design 1 —Critical Stress CTE 1 Design 1
— PCC Strength CTE 1 Design 2 —Critical Stress CTE 1 Design 2

 

 

 

 

Figure 4—20. HIPERPAV 11 Performance for CTE 1 Designs

 

350

1
1
1
1
1
1
1

1
1
1
1
1
1
1
1
1
1
1
1
1
1

 

1
1
1
1
1
1
1
1
1
1

150 7

 

100 77-

Tensile Stress and Strength (psi)

 

 

 

 

M
O o
1

3O 40 50 60 70
Time (hr.)

O
p—a
O
N
O

 

'- PCC Strength CTE 8 Design 1 -—-Critical Stress CTE 8 Design 1
— PCC Strength CTE 8 Design 2 —Critical Stress CTE 8 Design 2

 

 

 

Figure 4-21. HIPERPAV II Performance for CTE 8 Designs
In general, pavements made with these approved mixture designs did not show

any practical sensitivity to the magnitude of the measured PCC CTE in this experiment.

100

4.8.2 Long-Term Effects Analysis

Four sets of analyses were conducted to investigate the effects of CTE on the long term

performance of JPCPs:

The effect of CTE test variability on cracking performance of the concrete
pavement

The effect of CTE of concrete made with different aggregate types on cracking
performance of the pavement

The effect of CTE at different ages on cracking performance of the concrete
pavement

The effect of CTE based on different numbers of test cycles on cracking
performance of the concrete pavement

The effect of CTE — among other design features — on cracking performance of

the concrete pavement

The results of the above-mentioned M-E PDG analyses and the sensitivity of

percent slabs cracked to the CTE value (reported in following sections) are in agreement

with other literature that investigated theses effects (Mallela et al.', Appendix K of the

M-E PDG3, Tanesi et al.20, and Hossain et al.23).

4.8.2.1 Effect of CTE Variability on JPCP Performance

In order to investigate the practical significance of CTE test variability on pavement

performance, M-E PDG software was used. A pavement with 15 ft. joint spacing and 10

in. slab thickness was modeled. A 4 in. thick granular base on a 16 in. thick sand subbase

was selected. The considered subgrade was a fine A-7-6 soil which is a clayey soil with

101

greater than 35% passing the #200 sieve, minimum liquid limit of 41, and minimum
plasticity index of 11 (AASHTO M 145 “Standard Specifications for Classification of
Soils and Soil-Aggregate Mixtures for Highway Construction Purposes”). Other inputs
(for each type of mixture) are summarized in Table 4-16.

Table 4-16. Input Variables for M-E PDG Sensitivity Analyses

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Factor Variable Description
Joint Spacing 15'
Edge Support Tied Shoulders
Design Dowel Diameter 1.25"
Dowel Spacing 12"
Slab Width 12'
PCC Slab Thickness 10"
CTE Different Measured Values for
Different Types of Analyses
PCC
28-day Compressive .
Strength 3500 ps1
Unit Weight 150 pcf
Materials Base Type Granular Base (Dense Graded
Base Aggregate Base [DGAB])
Base Thickness 4"
Subbase Type Sand
Subbase
Subbase Thickness 16"
Subgrade Soil Type A-7-6 (Fine)
Environmental Climatic Region Lansing

 

 

 

 

 

102

Two types of concrete mixtures with low and high CTE values were considered
for this analysis. Mixture design “CTE A” was composed of limestone as the coarse
aggregate. The two levels of CTE values for this design were 4.60 and 4.91 118/°F (8.28
and 8.84 [ME/°C). The difference between these two CTE levels was 0.31 ue/°F (0.56
1.13/°C). Mixture design “CTE B”, with dolomite as the coarse aggregate, had CTE values
of 5.87 and 6.18 ue/°F (10.56 and 11.12 118/°C). The difference between these two CTE
values was also 0.31 118/°F (0.56 118/°C). Each pair of CTE values for CTE A and CTE B
mixture designs were actual measured CTE values that were different from each other
due to test variability. The test results were explored in order to find actual measurements
with identical difference and the resulting CTE values with 0.31 us/°F (0.56 118/°C)
difference were selected from the test results.

To study the exclusive effect of CTE on pavement performance, the mixture
design and concrete strength were held constant (CTE 1 mixture design and 3500 psi 28-
day compressive strength). Figure 4-22 illustrates the effect of CTE test variability on the
cracking performance of JPCP. The results of this analysis show that the pavement
constructed with CTE B mixture (with higher CTE), exhibited greater levels of transverse
cracking (percent slabs cracked after 30-year changes from 35.9% to 57.3%) as compared
to the pavement constructed with CTE A mixture (with lower CTE), even though the
CTE variability (the difference between two CTE levels) was identical. Percent slabs

cracked after 30 years changed from 1.1% to 2.6%. A similar observation was reported

1.20. This observation warrants the determination of CTE for paving

by Tanesi et a
mixtures and adjustment in joint design for example, by reducing the slab length for more

thermally active mixtures to accommodate the thermal activity of the concrete mixture.

103

60
55 — r
50 — ,x’
451 /

40 - r

35 .. f .'
30 1 / x
25 e r o

20 _ / .'

15 — x’

10 ... f, ”

 

Percent slabs cracked, %

 

 

 

,. , h.
‘F I l l 1 #1 I —T l
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32
Pavement age, years

{—CTE A-4.60 —CTE A-4.91 - - CTE B-5.87 — -CTE 8-6.110

 

 

 

 

Figure 4-22. Percent Slabs Cracked for CTE Test Variability

4.8.2.2 Effect of Coarse Aggregate Type on JPCP Performance

The same structural design that was described for the CTE variability was considered in a
sensitivity analysis to investigate the effect of aggregate geology on the cracking
performance of J PCPs. Average CTE values for each aggregate type were used. The
results from the M-E PDG outputs are presented in Table 4-17. In this table, the percent
slabs cracked for different mixture designs at four points in pavement’s life are presented
(5, 10, 20, and 30 years). Figure 4-23 illustrates these results for three commonly used
aggregate types (limestone, dolomite, and gravel) since the practical significance was the
subject of this analysis. The results of this analysis show that the percent slabs cracked

varies with aggregate geology. This implies that design features (contraction joint spacing

104

for example) should be modified to accommodate thermal properties of concrete

mixtures.

Table 4-17. Percent Slabs Cracked for Mixtures with Different Aggregates

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

' P em tA ,Years

M1331” Aggregate Type CTE, us/°F CTE, 118/”C av en ge
5 10 20 30

CTE] Limestone 4.75 8.55 0.1 0.2 1.1 2.8
CTE2 Gravel 5.93 10.67 1.9 8.4 31.4 56.1
CTE3 Dolomitic Limestone 4.56 8.21 0 0.1 0.6 1.6
CTE4 Slag 5.81 10.46 1.4 6.2 24.6 47.6
CTE5 Dolomite 6.01 10.82 2.4 10.3 36.4 61.5
CTE6 Gabbro (Trap Rock) 5.48 9.86 0.5 2.4 10.6 24.7
CTE7 Dolomite 6.00 10.80 2.3 10.0 35.8 60.9
CTE8 Dolomite 5.96 10.73 2.1 9.1 33.2 58.1

60 7
3
8a 50 s
'6
0
x
8 40 —
L:
U
3 30 _
.93
m
E 20 7
O
8’
n- 10 7

0

CTE] CTE2 CTE5
Mixture ID

 

 

I5yrs. CI 10 yrs. [320 yrs. I30 yrs.

 

 

Figure 4-23. Percent Slabs Cracked for Mixtures with Different Aggregates

105

 

4.8.2.3 Effect of CTE at Different Ages on JPCP Performance
For the analysis regarding the impact of CTE of concrete measured at different ages on
pavement performance, the same pavement cross section and aggregate types mentioned
previously were considered in order to evaluate the practical or operational significance.
For each of the aggregate geologies, average CTE values at 14, 28, and 90 days were
considered. Practicality was considered in selection of these levels. The results of the M-
E PDG analysis are shown in Table 4—18. The difference in percent slabs cracked
(between three age levels) ranged from 0.3% to 31.8% after 30 years.

Figures 4-24 through 4-26 illustrate the M-E PDG results for the previously
mentioned commonly used aggregate types (limestone, dolomite, and gravel). In these
figures, the “cracking performance ratio” is the ratio of the percent slabs cracked to the

percent slabs cracked for good and normal pavements at each point in the pavement’s

life.

Khazanovich et al.31 have established the percent slabs cracked during rigid
pavement life for three levels of performance (good, normal, and poor). Table 4—19
presents these criteria. The “performance criteria limit” shown in figures 4-24 through 4-
26 represents the cracking limit for a good or normal pavement. If the cracking
performance ratio is above 1, it indicates poor performance while a ratio below 1 is an
indication of good or normal performance.

The overall effect of sample age on percent slabs cracked, based on these results,
does not seem to be significant (operationally) for this typical design. Based on this
limited analysis, measurement of the CTE at 28 days seems to be adequate for pavement

design.

106

Table 4-18. Percent Slabs Cracked Based on CTE Values for Different Ages

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mixture Age, CTE, CTE, Pavement Age, Years Comments
ID Days WP WC 5 10 20 30
14 4.52 8.13 0 0.1 0.5 1.4 Thereisamaximum of
CTEI 28 4.54 8.18 0 0.1 0.6 1.4 :l‘jggr'firgggg
90 4.77 8.59 0.1 0.3 1.1 2.9 YearSformismixture-
14 5.87 10.57 1.6 7.2 27.9 51.9 There isamaximum of
CTE2 28 5.84 10.52 1.5 6.7 26.2 49.7 :1“fo $33,333:?)
90 5.96 10.73 1.5 9.1 33.2 58.1 Years formsmixm’e-
14 4.50 8.1 0 0.1 0.5 1.3 Age at thetirne oftesting
CTE3 28 4.51 8.11 0 0.1 0.5 1.4 :fiezt'fg/f'gggsw'm)
90 4.56 8.21 0 0.1 0.6 1.6 “301‘“-
14 5.56 10.01 0.7 3.0 13.3 29.7 Thereisamaximumof
CTE4 28 5.71 10.27 1.1 4.6 19.3 39.9 zlzffig'ifjggeeg?
90 6.01 10.81 2.4 10.3 36.4 61.5 yearSforthismixture-
14 5.86 10.55 1.6 7.0 27.3 51.1 Thereisamaximumof
CTE5 28 5.92 10.65 1.9 8.2 30.8 55.4 iggggifiggfi
90 6.11 10.99 3.1 13.0 27.3 67.7 YearSfor‘hismixme-
14 5.49 9.89 0.8 2.4 10.9 25.3 Age at thetime Oftesting
CTE6 28 5.41 9.73 0.4 1.9 8.7 20.7 hasasmall(1.2%) effect
on % slabs cracked.
90 5.47 9.85 0.5 2.3 10.3 24.1
14 5.92 10.65 1.9 8.2 30.8 55.4 There isamaximum of
CTE7 28 5.90 10.62 1.8 7.8 29.6 54.0 :1:£f::::§:iteli'3/ti
90 5.96 10.73 2.1 9.1 33.2 58.1 YearSfor‘hismim’e-
14 5.66 10.19 0.9 4.0 17.1 36.3 Thereisamaximum of
CTE8 28 5.87 10.57 1.6 7.2 27.9 51.9 259% differencein'y"
slabs cracked after 30
90 6.02 10.84 2.5 10.5 37.0 62.2 YearSforthismixmre-

 

107

 

 

 

b—i
l

Cracking Performance Ratio

 

 

 

 

 

   

Pavement Age (Years)

 

 

- 14-day EZI 28-day l:l 90-day - - Cracking Performance Limit

Figure 4-24. CTE] Cracking Performance with CTE Values at Different Ages

 

5

 

 

 

Cracking Performance Ratio

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pavement Age (Years)

 

- l4-day [:1 28-day :1 90-day - - Cracking Performance Limitj

 

Figure 4-25. CTE2 Cracking Performance with CTE Values at Different Ages

108

 

 

 

Cracking Performance Ratio

 

 

 

 

 

 

 

 

 

 

 

 

 

Pavement Age (Years)

 

 

 

 

 

 

 

-14-day 1:1 28-day :1 90-day - - Cracking Performance Limit

 

 

Figure 4-26. CTE5 Cracking Performance with CTE Values at Different Ages

Table 4-19. Performance Criteria for Rigid Pavements

 

Pavement Age (years)

 

 

 

 

 

 

 

 

 

 

Performance Measure Boundary between:
5 10 20 30
Percent Slabs Cracked Good and Normal 1.25 2.5 5 7.5
Normal and Poor 2.5 5 10 15

 

4.8.2.4 Effect of Number of Cycles on JPCP Performance

The effect of the number of cycles on pavement performance was investigated based on

the same cross section and aggregate types mentioned previously. For each of the

aggregate geologies, average CTE values for cycles 1, 2, and 3 were considered. The

results of the M-E PDG analysis are shown in Table 4-20. The difference in percent slabs

cracked (between 3 number of cycles levels) ranged from 0.4% to 8.4% after 30 years.

109

 

 

Table 4-20. Percent Slabs Cracked Based on CTE Values for Different Cycles

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mixture Cycle CTE, CTE, Pavement Age, Years
Comments
II) Number 113/°F 118/°C 5 10 20 30
1 4.84 8.71 0.1 0.3 1.4 3.7 Number of cycles has
CTE l 2 4.78 8.60 0.1 0.3 1.1 3.0 a small effect on this
mixture (1.0%).
3 4.74 8.53 0.1 0.2 1.0 2.7
1 5.97 10.75 2.1 9.3 33.8 58.8 There is a maximum
of 4.8% difference in
CTE 2 2 5.89 10.60 1.8 8.0 30.2 54.7 % Slabs cracked after
3 5.90 10.62 1.8 7.8 29.6 54.0 30 years-
1 4.61 8.30 O 0.2 0.7 1.8 Number of cycles has
CTE 3 2 4.54 8.17 0 0.1 0.6 1.5 a very small effect on
this mixture (0.4%).
3 4.51 8.12 0 0.1 0.5 1.4
1 5.88 10.58 1.7 7.4 28.4 52.6 There is a maximum
of 5.8% difference in
CTE 4 2 5.80 10.44 1.4 6.0 24.1 46.8 % Slabs cracked afier
3 5.80 10.44 1.4 6.0 24.1 46.8 30 Years-
1 6.07 10.93 2.8 11.9 40.3 65.3 There is a maximum
of 6.5% difference in
CTE 5 2 5.99 10.78 2.3 9.8 35.1 60.1 % Slabs cracked after
3 5.97 10.75 2.1 9.3 33.8 58.8 30 Years-
1 5.50 9.90 0.6 2.5 11.2 25.9 Number of cycles has
CTE 6 2 5.48 9.86 0.5 2.4 10.6 24.7 a small effect on this
mixture (1.2%).
3 5.48 9.86 0.5 2.4 10.6 24.7
1 6.09 10.96 2.9 12.4 41.5 66.5 There is a maximum
of 8.4% difference in
CTE 7 2 5.96 10.73 2.1 9.1 33.2 58.1 % slabs cracked am
3 5.94 10.69 2.0 8.6 32.0 56.8 30 Years-
1 5.96 10.73 2.1 9.1 33.2 58.1 There is a maximum
CTE 8 of 2.7% difference in
5.94 10.69 2.0 8.6 32.0 56.8 % slabs cracked afier
3 5.92 10.66 1.9 8.2 30.8 55.4 30 Years-

 

110

 

Figures 4-27 through 4-29 illustrate the M-E PDG results for the previously
mentioned commonly used aggregate types (limestone, dolomite, and gravel). In these
figures, the “cracking performance ratio” is the ratio of the percent slabs cracked to the
percent slabs cracked for good and normal pavements at each point in the pavement’s
life. Description of the performance criteria can be found in the previous section

The number of cycles does not seem to have a practical effect on pavement
performance. It can be seen, however, that generally, the percent slabs cracked based on
the first cycle is higher than those based on other cycles. In addition, it can be observed
that the results based on the second and third cycles are closer to each other. It is
recommended that the first heating-cooling cycle be used for conditioning the sample and

the design CTE be computed based on the second or third cycle.

 

--————-—_———————_—-—_—__—_—-

i

 

Cracking Performance Ratio

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

O
1

5 10 20 30
Pavement Age (Years)

 

 

 

- Cycle 1 [:1 Cycle 2 1: Cycle 3 - - Cracking Performance Limit

 

Figure 4-27. CTE] Cracking Performance with CTE Based on Different Cycles

111

 

Cracking Performance Ratio

 

 

 

 

 

 

 

30
Pavement Age (Years)

 

- Cycle 1 1:] Cycle 2 :1 Cycle 3 - - Cracking Performance Limit1

 

Figure 4-28. CTE2 Cracking Performance with CTE Based on Different Cycles

5

 

Cracking Performance Ratio

 

 

 

 

 

 

 

 

Pavement Age (Years)

 

ECycle 171:1 chef“: Cycle 3 - - Cracking Performance Limfl

 

Figure 4-29. CTE5 Cracking Performance with CTE Based on Different Cycles

112

 

 

 

Cracking Performance Ratio

 

 

 

 

 

 

 

 

20 30
Pavement Age (Years)

|- Cycle 1 1:] CyCle 2 :1 cyele 3 4 -7CrackingiPerformance Lirm—'t1

 

Figure 4-28. CTE2 Cracking Performance with CTE Based on Different Cycles

 

Cracking Performance Ratio

 

 

 

 

 

 

 

Pavement Age (Years)

 

- Cycle 71 1:1 Cycle 2 1:1 Cycle3 — - Cracking Performance pm

 

Figure 4-29. CTE5 Cracking Performance with CTE Based on Different Cycles

112

4.8.2.5 Effect of CTE among Other Design Features on JPCP Performance
A factorial similar to one presented in Table 4-13 was developed to investigate the effects
of CTE among other design features on the long term performance of the concrete
pavement. The only difference was that the splitting tensile strength was not used because
it is not an input in M-E PDG software and that the curing method was fixed (curing
compound). CTE values for each type of aggregate were maximum, minimum, and
average measured values. Actual mixture design proportions and strength were used as
inputs. Initial two-way average annual daily truck traffic (AADTT) was considered to be
6000.

All M-E PDG runs (288 runs) were performed and the generated results were
analyzed. Table 4-21 shows designs with combinations of variables that produced
maximum percent slabs cracked at 30 years for each mixture design.

Table 4-21. Variables for Designs with Maximum Percent Slabs Cracked

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mifigure T'ln'sc'ljh)ess Sigirilr'rg ‘31:; GEES/53°F Perceztt 3S(l)a$sefr:acked
(in) (ft) (ft)
CTE 1 9 20 12 Max 5.12 (9.21) 7.8
CTE 2 9 20 12 Max 6.42 (11.55) 100
CTE 3 9 20 12 Max 4.92 (8.86) 99.5
CTE 4 9 20 12 Max 6.07 (10.92) 80.1
CTE 5 9 20 12 Max 6.30 (11.33) 100
CTE 6 9 20 12 Max 5.56 (10.00) 99.9
CTE 7 9 20 12 Max 6.15 (11.07) 83.3
CTE8 9 20 12 Max 6.13 (11.04) 100

 

 

113

It can be seen from the results shown in this table that CTE is one of the factors
that affect the long-term performance of concrete pavements. It also shows that higher
CTE values, thinner slab, and longer joint spacing lead to increased percent of slabs

cracked over the age of a pavement.

114

CHAPTER 5
SUMMARY, CONCLUSIONS,
RECOMMENDATIONS, AND FUTURE RESEARCH

NEEDS

5.1 Introduction

In this chapter a summary of research conducted during the term of this project is
presented followed by findings and conclusions regarding CTE measurement and factors
affecting it. Impact of CTE on the performance of j ointed concrete pavements is also
presented. Recommendations regarding CTE testing procedure and future research needs

are also presented in this chapter.

115

5.2 Summary of the Performed Work

Eight typical MDOT concrete paving mixtures with coarse aggregate from different
sources were subjected to various tests including CTE and quality control tests. Over 700
concrete specimens were fabricated for these tests. At least three replicate specimens
were fabricated for each test for a given test date.

Variables included in this investigation were aggregate geology, age of the sample
at the time of testing, and number of heating-cooling cycles applied to the samples. CTE
tests were conducted in a way that the results could be used to calculate CTE with two
different methods. Details about the experimental program and the test method are
documented in Chapter 3.

Results from the laboratory investigation are presented in Chapter 4. Statistical
analyses were performed on these results in order to evaluate the significance and impact
of the variables on the magnitude of the CTE. Statistical analysis was also employed to
compare two different CTE test methods.

Sensitivity analyses were conducted in order to examine the practical significance
and impact of CTE on the structural design and on the performance of j ointed concrete
pavements.

Recommendations were developed in order to improve the CTE test and CTE

values were suggested for agency use in Michigan.

116

5.3 Factors Affecting Measurement of CTE and Impact of CTE

on Pavement Performance

5.3.1 Factors Affecting Measurement of CTE

The following conclusions were based on the laboratory investigation, the statistical
analyses of the dataset, and the sensitivity analyses for practical significance:

I The magnitude of the measured CTE varied with aggregate geology. The
measured CTE magnitudes for the various aggregate geologies compared
favorably with the published values.

I Statistical analysis showed that:

1) Magnitude of the measured CTE was significantly (statistically)
influenced by the age of the sample at the time of testing.

2) Magnitudes of the measured CTE at the early ages (3, 7, 14, 28 days)
were significantly (statistically) different than the magnitudes
determined at the end of 90, 180, and 365 days.

3) Operationally, the impact of this difference on transverse cracking (as
computed by the M-E PDG software for 14, 28, and 90 days) was not
found to be significant.

' The number of heating-cooling cycles in CTE test affected the magnitude
of CTE. The CTE values calculated based on the first cycle were higher
than the values calculated based on second and third cycles. Statistically,
the CTE values based on second and third cycles were not different from

each other.

117

 

I The coefficient of variation for the data set ranged from 2.5% to 6%.
Approximately 97% of the data set had a 5cm between i 0.3 1.1ch (0.5
118/0C). It was observed that, generally, concrete with higher CTE values
was more sensitive to variability compared to concrete with low CTE
values.

I The AASHTO TP60 and the Revised TP60 methods generated slightly
different results. An analysis of the least square means showed an
estimated difference of 0.023 ua/oF (0.041 118/°C). This very small
difference does not seem to have any practical significance in concrete

pavement design and performance.

5.3.2 Impact of the CTE on Pavement Performance

Sensitivity analyses were used to investigate the impact of CTE value on short-term and
long-term cracking performance of jointed concrete pavements. The factorials used in
these analyses included different levels of CTE, slab thickness, joint spacing, and slab
width among other variables.

The results of theses analyses indicated that the CTE of PCC was one of the
factors — among other factors — that affected pavement performance (PCC tensile strength
in the first 72-hours and transverse cracking during the design life). For each mixture
design, higher of the three investigated CTE values were always associated with poor
performance.

It was also observed that, thinner slab, longer joint spacing, and higher CTE

values resulted in increased percent of slabs cracked over the age of a pavement.

118

 

5.4 Recommendations

5.4.1 CTE Testing Recommendations

The following recommendations are suggested for conducting CTE testing and obtaining
the results:

I It is suggested that the test specimens should be subjected to three cycles of
heating and cooling. The test data fi'om cycles two or three should be used for the
calculation of the CTE. The first cycle should be regarded as a conditioning cycle.

I Based on the operational significance analysis it is suggested that CTE measured
at the end of 28 days can be used as an input for jointed concrete pavement
design.

I Automation of the temperature and displacement readings with a rate of one
reading per minute (as recommended by the Revised TP60 method) makes the
testing process and CTE determination more reliable.

I Thermocouples can be inserted in the concrete specimens in order to monitor the
temperature change and also thermal equilibrium of the specimens.

I The optimal thermal equilibrium time period in this experiment was established to
be one hour. Laboratories are recommended to establish this time period based on

the concrete mixtures being tested for CTE.

5.4.2 Recommended CTE Values for Agency Use in Michigan

Based on the statistical and operational analyses, it is found that determination of the

concrete CTE at the age of 28 days is practically feasible and therefore is suggested for

119

 

pavement projects. These analyses also indicate that the CTE results based on second or

third cycle should be used in pavement design procedure.

Table 5-1 lists the CTE values for each coarse aggregate tested (and used in

Michigan), based on the above recommendations:

Table 5-1. Recommended CTE Values for Concrete Made With Different Coarse

Aggregate Used In Michigan

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

28-Day CTE 28-Day CTE
Mixture Primary Aggregate Pit (118/OF ) (118/0C )
ID Class Number Cycle Cycle Cycle Cycle
#2 #3 #2 #3
CTE 1 Limestone 71-47 4.54 4.55 8.17 8.18
CTE 2 Gravel 19-56 5.84 5.84 10.51 10.52
CTE 3 Dolomitic Limestone 75-05 4.51 4.47 8.12 8.04
CTE 4 Slag 82-19 5.69 5.73 10.24 10.31
CTE 5 Dolomite 49-65 5.91 5.91 10.65 10.64
CTE 6 Gabbro (Trap Rock) 95-10 5.40 5.38 9.72 9.68
CTE 7 Dolomite 58-11 5.93 5.77 10.67 10.38
CTE 8 Dolomite 91-06 5.90 5.87 10.63 10.57

 

120

 

 

5.5 Future Research Needs

There are a number of experimental and analytical investigations that can potentially

enhance and add to the current understanding about CTE of PCC. These potentially

beneficial studies are introduced here:

Back to back PCC test sections subjected to environmental and vehicle loading
can be constructed in order to investigate the effect of CTE and other variables on
pavement performance in a partially-controlled and continuously-monitored
setting. Series of these test sections can be the subject of a particular study. For
example, the effect of aggregate type, mixture design, etc. can be investigated.
Information about the magnitude of the design CTE and the CTE of existing
concrete pavements and new pavement projects (from testing) can be acquired
from agencies in order to analyze the validity of the assumptions made during the
design process, potential discrepancies, and relevant implications for future
design and rehabilitation.

CTE of different coarse aggregates can be tested along with the CTE of cement
paste or mortar in order to investigate the accuracy of the M-E PDG input Level 2
equation and to evaluate the practical implications of using Level 2 values instead

of level 1 (actual testing).

121

APPENDICES

122

APPENDIX A

HARDENED CONCRETE PROPERTIES DATA

123

Table A-1. Hardened Concrete Properties Table for CTE l Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTEl Agemays)

1 3 7 14 28 90 365
Compressive
Strength (psi)
Sample #1 2289 3907 4918 521 1 5416 6622 6476
Sample #2 2490 4033 4980 4093 5125 6241 6734
Sample #3 2869 4223 3958 5227 4847 6162 6320
Average 2549 4054 4619 4844 5129 6342 6510
Elastic
Modulus
(Psi)
Sample #1 2720277 3845841 4384180 4246291 4565108 5044775 4876878
Sample #2 1599140 3570805 4250277 4083412 4303844 4960549 5024210
Sample #3 3219544 3969268 4095948 4458081 4598727 5005437 5155223
Average 2512987 3795305 4243468 4262595 4489226 5003587 5018770
Split Tensile
Strength (psi)
Sample #1 256 433 491 500 482 596 634
Sample #2 241 434 445 440 581 569 666
Sample #3 254 437 479 570 485 611 602
Average 248 435 472 503 516 592 634
Flexural
Strength (psi)
Sample #1 469 745 675 747 818 793 827
Sample #2 454 822 674 913 850 841 845
Sample #3 543 679 699 763 836 840 841
Average 499 748 683 808 835 825 838

 

 

 

 

 

 

 

 

 

124

Table A-2. Hardened Concrete Properties Table for CTE 2 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 2 Age (Days)
1 3 7 14 28 90 365
Compressive
Strength (psi)
Sample #1 2051 3909 3685 4148 5292 5818 5814
Sample #2 2223 3770 3940 4388 4779 5584 5887
Sample #3 2098 3599 3745 3630 4824 5179 4631
Average 2124 3759 3790 4055 4965 5527 5444
Elastic
Modulus (psi)
Sample #1 3504811 3894648 4165865 4436156 4746978 4693725 5765917
Sample #2 3377411 4226295 4272793 4740568 4884807 5134287 5177605
Sample #3 2944210 4134072 4352554 4557317 5038801 5082407 5594764
Average 3275477 4085005 4263737 4578014 4890195 4970139 5512762
Split Tensile
Strength (psi)
Sample #1 258 389 412 446 502 512 568
Sample #2 283 444 445 461 534 493 594
Sample #3 250 417 428 466 471 519 629
Average 264 417 428 458 502 508 597
Flexural
Strength
(psi)
Sample #1 396 502 614 676 695 890 827
Sample #2 454 504 630 655 691 694 845
Sample #3 429 484 709 475 691 751 841
Average 426 497 651 602 692 778 838

 

125

 

Table A-3. Hardened Concrete Properties Table for CTE 3 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 3 Age (Days)
1 3 7 14 28 90 365
Compressive
Strength (psi)
Sample #1 2097 3283 3171 3767 4097 4707 5310
Sample #2 2131 2889 3377 3736 3695 4780 5601
Sample #3 2246 3159 3459 3701 4109 5183 5852
Average 2158 3110 3336 3735 3967 4890 5588
Elastic
Modulus (psi)
Sample #1 3007745 3825496 4247803 3988223 4474795 5143779 5260887
Sample #2 3307029 3902588 4161331 4148092 4603508 4769219 4987492
Sample #3 3060070 3793156 4091292 4388647 4626630 4866031 5256010
Average 3124948 3840414 4166808 4174987 4568311 4926343 5168130
Split Tensile
Strength (psi)
Sample #1 295 367 415 414 502 482 488
Sample #2 236 382 451 456 489 505 536
Sample #3 263 332 436 452 477 514 588
Average 265 360 434 440 489 501 537
Flexural
Strength (psi)
Sample #1 414 543 673 670 658 783 758
Sample #2 414 585 626 677 666 792 962
Sample #3 479 602 620 655 612 740 787
Average 436 577 639 667 645 772 836

 

 

 

 

 

 

 

 

126

 

Table A-4. Hardened Concrete Properties Table for CTE 4 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 4 Age (Days)
1 3 7 14 28 90 365

Compressive

Strength (psi)

Sample #1 2430 4041 4664 4728 5056 6483 6665
Sample #2 2491 4039 4239 5226 5216 5924 7087
Sample #3 2518 3811 4345 5015 5235 6146 6578
Average 2480 3964 4416 4990 5169 6184 6777
Elastic
Modulus (psi)

Sample #1 3446678 4467453 4252860 4525155 4549298 4870545 5288507
Sample #2 3487046 4230876 4345819 4567307 4729983 5078483 5037622
Sample #3 3532711 4211250 4541989 3823767 4694448 4973960 5246620
Average 3488812 4303193 4380223 4546231 4639640 4974330 5190916
Split Tensile
Strength (psi)

Sample #1 297 401 475 510 509 518 627
Sample #2 334 365 405 524 518 584 655
Sample #3 323 390 465 538 493 550 557
Average 318 385 449 524 507 551 613
Flexural
Strength (psi)

Sample #1 694 605 680 791 784 756 945
Sample #2 715 594 689 767 881 882 878
Sample #3 520 618 692 717 827 774 842
Average 643 605 687 758 831 804 888

 

 

 

 

 

 

 

127

 

Table A-5. Hardened Concrete Properties Table for CTE 5 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 5 Age (Days)
1 3 7 14 28 90 365
Compressive
Strength (psi)
Sample #1 2308 2800 3337 3760 4165 5015 5776
Sample #2 2281 3167 3205 4030 3849 4940 5816
Sample #3 2240 3138 3271 4019 4089 4863 5708
Average 2276 3035 3271 3936 I 4035 4939 5767
Elastic
Modulus (psi)
Sample #1 2770536 3319209 3694044 4181223 4773375 5431180 6054583
Sample #2 2738093 3551937 3531425 4203011 4462819 5033959 6054626
Sample #3 2485543 3182731 3816278 4430976 4720948 5243949 5784749
Average 2664724 3351292 3680582 4271736 4652381 5236363 5964653
Split Tensile
Strength (psi)
Sample #1 261 395 490 482 509 469 558
Sample #2 256 394 464 472 500 525 565
Sample #3 254 383 489 530 524 475 533
Average 257 391 481 494 51 1 490 552
Flexural
Strength (psi)
Sample #1 415 556 543 638 727 667 947
Sample #2 427 462 612 662 727 715 923
Sample #3 404 529 701 728 740 713 879
Average 415 516 619 676 731 699 916

 

 

128

 

Table A-6. Hardened Concrete Properties Table for CTE 6 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 6 Age (Days)
1 3 7 14 28 90 365
Compressive
Strength (psi)
Sample #1 2314 3297 3716 4778 5322 4950 5572
Sample #2 2335 3540 3914 4802 5437 5190 5235
Sample #3 2315 3469 4077 4164 4616 4962 6019
Average 2321 3435 3902 4581 5125 5034 5609
Elastic
Modulus (psi)
Sample #1 3509646 4603350 5113319 5474624 5486636 5319340 5750771
Sample #2 4180986 4603342 4791762 5326848 5496331 6061649 6044256
Sample #3 4142140 4875414 5204367 4995331 5183808 5224772 6570346
Average 3944258 4694035 5036483 5265601 5388925 5535254 6121791
Split Tensile
Strength (psi)
Sample #1 318 386 342 391 484 520 566
Sample #2 300 401 392 452 515 546 604
Sample #3 318 314 467 518 499 512 579
Average 312 367 400 454 500 526 583
Flexural
Strength (psi)
Sample #1 437 521 712 682 738 763 924
Sample #2 481 541 614 715 746 788 842
Sample #3 487 492 573 892 710 798 864
Average 468 518 633 763 731 783 877

 

 

 

 

 

 

 

 

129

 

Table A-7. Hardened Concrete Properties Table for CTE 7 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 7 Age (Days)
1 3 7 14 28 90 365
Compressive
Strength (psi)
Sample #1 2389 3752 4685 4587 6032 6242 6772
Sample #2 2359 3757 4442 4811 5531 6357 7122
Sample #3 2380 4107 4168 4418 5913 6455 7344
Average 2376 3872 4432 4605 5825 6352 7079
Elastic
Modulus (psi)
Sample #1 2727454 3101702 3865715 4290294 4564800 4603709 5020650
Sample #2 2583834 3367417 4060420 4186334 4512137 4796854 5143097
Sample #3 2531341 3405353 3873462 4069158 4367202 4681462 4984448
Average 2614210 3291491 3933199 4181929 4481379 4694008 5049398
Split Tensile
Strength (psi)
Sample #1 282 407 476 559 587 604 675
Sample #2 293 456 430 464 538 662 693
Sample #3 282 436 453 473 557 611 698
Average 286 433 453 499 561 626 689
Flexural
Strength (psi)
Sample #1 423 577 735 784 804 802 916
Sample #2 463 519 642 801 887 783 988
Sample #3 482 599 662 775 770 820 892
Average 456 565 680 787 820 802 932

 

130

 

Table A-8. Hardened Concrete Properties Table for CTE 8 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A D
CTE8 ge( MS)

7 14 28 90 365
Compressive
Strength (psi)
Sample #1 3785 4234 4811 5839 6451
Sample #2 3643 4442 5094 6032 6635
Average 3714 4338 4953 5936 6543
Elastic
Modulus (psi)
Sample #1 3930757 4119392 4758942 5249832 5526363
Sample#2 4170204 4187470 4663753 5137098 5661545
Average 4050480 4153431 4711347 5193465 5593954
Split Tensile
Strength (psi)
Sample #1 380 435 489 478 632

 

131

 

APPENDIX B

COEFFICIENT OF THERMAL EXPANSION DATA

132

Table B-1. CTE Values for CTE l Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 1, Age (Days)
8/°F e/°C
'1 ('4 ) 3 7 14 28 90 180 365
AASHTO TP60
Sam 1e #1 4.56 4.54 4.43 4.57 4.81 4.99 5.02
p (8.21) (8.17) (7.98) (8.23) (8.66) (8.98) (9.03)
Sam le #2 4.63 4.52 4.52 4.64 4.59 4.98 5.12
p (8.34) (8.14) (8.13) (8.35) (8.27) (8.97) (9.21)
Sam 1e #3 4.43 4.47 4.60 4.42 4.74 5.08 4.78
p (7.97) (8.05) (8.28) (7.96) (8.53) (9.15) (8.60)
Avera e 4.54 4.51 4.52 4.54 4.77 5.02 4.97
g (8.17) (8.12) (8.13) (8.18) (8.59) (9.03) (8.95)
Revised TP60
Sam le #1 4.54 4.62 4.54 4.53 4.92 5.08 5.09
p (8.17) (8.31) (8.17) (8.15) (8.86) (9.15) (9.16)
Sam le #2 4.70 4.51 4.64 4.61 4.61 5.01 5.07
p (8.46) (8.12) (8.36) (8.30) (8.30) (9.03) (9.13)
Sam 1e #3 4.56 4.63 4.56 4.41 4.70 5.15 4.87
p (8.21) (8.33) (8.21) (7.93) (8.46) (9.27) (8.77)
Avera e 4.60 4.59 4.58 4.51 4.81 5.08 5.01
g (8.28) (8.25) (8.25) (8.12) (8.66) (9.15) (9.02)

 

133

 

Table B-2. CTE Values for CTE 2 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 2, Age (Days)
e/°F ( 91°C)
'1 " 3 7 14 28 90 180 365
AASHTOTP60
Sam H1 5.90 5.86 5.85 5.88 6.06 6.22 6.11
1’ (10.63) (10.54) (10.53) (10.58) (10.91) (11.19) (10.99)
Sam M2 5.62 5.74 5.95 5.86 5.78 6.42 6.32
9 (10.12) (10.34) (10.71) (10.55) (10.40) (11.55) (11.38)
Sam W3 5.56 5.29 5.82 5.79 6.04 6.11 5.92
1’ (10.01) (9.52) (10.47) (10.43) (10.87) (11.01) (10.65)
Avera e 5.70 5.63 5.87 5.85 5.96 6.25 6.11
g (10.25) (10.13) (10.57) (10.52) (10.73) (11.25) (11.01)
RevisedTP60
Sam le #1 6.03 5.82 5.86 5.94 6.14 6.34 6.19
D (10.85) (10.48) (10.55) (10.69) (11.05) (11.41) (11.15)
Samlem 5.60 5.71 5.88 5.70 5.80 6.38 6.18
p (10.07) (10.28) (10.58) (10.26) (10.44) (11.49) (11.13)
s [#3 5.61 5.39 5.94 5.81 6.13 6.06 5.95
am“ (10.09) (9.71) (10.69) (10.46) (11.03) (10.90) (10.72)
A 5.74 5.64 5.89 5.82 6.02 6.26 6.11
wage (10.34) (10.15) (10.61) (10.47) (10.84) (11.27) (11.00)

 

134

 

Table B-3. CTE Values for CTE 3 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 3, Age (Days)
e/°F e/°C
'1 ('1 ) 3 7 14 28 90 180 365
AASHTO TP60
Sam 1e #1 4.35 4.41 4.52 4.47 4.65 4.74 4.74
p (7.83) (7.94) (8.14) (8.04) (8.37) (8.53) (8.54)
Sam le #2 4.20 4.10 4.43 4.42 4.40 4.67 4.92
p (7.57) (7.39) (7.98) (7.96) (7.91) (8.41) (8.86)
Sam le #3 4.52 4.42 4.55 4.63 4.6 4.63 4.64
p (8.14) (7.95) (8.18) (8.33) 5(8.36) (8.34) (8.35)
Avera e 4.36 4.31 4.50 4.51 4.56 4.68 4.77
g (7.85) (7.76) (8.10) (8.11) (8.21) (8.43) (8.58)
Revised TP60
Sam le #1 4.49 4.38 4.46 4.56 4.75 4.81 4.79
p (8.08) (7.88) (8.03) (8.22) (8.55) (8.66) (8.62)
Sam le #2 4.18 3.95 4.46 4.35 4.45 4.66 4.85
p (7.53) (7.11) (8.03) (7.84) (8.01) (8.39) (8.74)
Sam le #3 4.42 4.52 4.60 4.68 4.73 4.70 4.72
p (7.96) (8.13) (8.28) (8.43) (8.51) (8.46) (8.49)
Avera e 4.36 4.28 4.51 4.53 4.64 4.73 4.79
g (7.86) (7.71) (8.11) (8.16) (8.35) (8.51) (8.62)

 

135

 

Table B-4. CTE Values for CTE 4 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 4, Age (Days)
e/°F e/°C
'1 (p ) 3 7 14 28 90 180 365
AASHTO TP60
Sam le #1 5.60 5.72 5.74 5.75 5.92 6.04 6.19
p (10.08) (10.30) (10.33) (10.34) (10.66) (10.88) (11.14)
Sam 1e #2 5.57 5.43 5.42 5.55 6.03 6.03 6.21
p (10.03) (9.78) (9.75) (10.00) (10.86) (10.86) (11.18)
Sam le #3 5.68 5.66 5.52 5.81 6.07 5.95 5.95
p (10.23) (10.18) (9.94) (10.46) (10.92) (10.72) (10.70)
Avera e 5.62 5.60 5.56 5.70 6.01 6.01 6.11
g (10.11) (10.09) (10.01) (10.27) (10.81) (10.82) (11.01)
Revised TP60
Sam 1e #1 5.53 5.59 5.67 5.79 6.02 6.12 6.28
p (9.96) (10.06) (10.21) (10.43) (10.84) (11.01) (11.31)
Sam 1e #2 5.54 5.48 5.44 5.46 5.98 5.92 6.15
p (9.98) (9.87) (9.79) (9.83) (10.76) (10.66) (11.07)
Sam 1e #3 5.85 5.86 5.42 5.89 6.21 5.92 5.88
p (10.53) (10.54) (9.76) (10.60) (11.17) (10.65) (10.59)
5.64 5.64 5.51 5.71 6.07 5.99 6.11
Average
(10.15) (10.16) (9.92) (10.29) (10.93) (10.78) (10.99)

 

136

 

Table B-5. CTE Values for CTE 5 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 5, Age (Days)
rue/°F (us/°C)
3 7 14 28 90 180 365
AASHTO TP60
Sam I e #1 6.02 5.89 5.93 5.80 6.03 6.22 6.29
p (10.84) (10.59) (10.67) (10.44) (10.85) (11.20) (11.33)
Sam le #2 5.75 5.63 5.77 5.99 6.12 6.30 6.39
p (10.35) (10.14) (10.39) (10.78) (11.01) (11.33) (11.51)
Sam 1e #3 5.92 5.83 5.88 5.96 6.17 6.02 6.09
p (10.66) (10.50) (10.58) (10.72) (11.11) (10.84) (10.96)
Avera e 5.90 5.78 5.86 5.91 6.10 6.18 6.26
g (10.62) (10.41) (10.55) (10.65) (10.99) (11.12) (11.26)
Revised TP60
Sam le #1 6.01 5.82 5.81 5.88 6.17 6.33 6.30
p (10.82) (10.48) (10.46) (10.59) (11.10) (11.39) (11.35)
Sam le #2 5.72 5.70 5.77 6.04 6.19 6.25 6.31
p (10.29) (10.26) (10.38) (10.88) (11.15) (11.25) (11.36)
Sam le #3 6.05 5.85 5.82 6.05 6,17 6.19 6.14
p (10.89) (10.53) (10.48) (10.90) (11.10) (11.15) (11.05)
Avera e 5.93 5.79 5.80 5.99 6.18 6.26 6.25
g (10.67) (10.42) (10.44) (10.79) (11.12) (11.26) (11.25)

 

137

 

Table B-6. CTE Values for CTE 6 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 6, Age (Days)
118/°F (rte/°C)
3 7 14 28 90 180 365
AASHTO TP60
Sam 1e #1 5.39 5.47 5.41 5.35 5.49 5.56 5.72
p (9.71) (9.85) (9.75) (9.63) (9.88) (10.00) (10.30)
Sam le #2 5.53 5.47 5.51 5.49 5.43 5.48 5.74
p (9.95) (9.84) (9.91) (9.89) (9.77) (9.87) (10.32)
Sam 1e #3 5.55 5.48 5.55 5.36 5.50 5.19 5.26
p (10.00) (9.87) (9.99) (9.66) (9.89) (9.34) (9.47)
Avm e 5.49 5.47 5.49 5.40 5.47 5.41 5.57
g (9.89) (9.85) (9.89) (9.73) (9.85) (9.74) (10.03)
Revised TP60
Sam 1e #1 5.29 5.36 5.31 5.43 5.50 5.69 5.72
p (9.52) (9.65) (9.56) (9.78) (9.90) (10.25) (10.29)
Sam le #2 5.48 5.43 5.47 5.41 5.34 5.43 5.71
p (9.87) (9.78) (9.84) (9.74) (9.62) (9.77) (10.28)
Sam le #3 5.51 5.43 5.46 5.46 5.53 5.21 5.37
p (9.92) (9.77) (9.82) (9.83) (9.95) (9.39) (9.67)
Avera e 5.43 5.41 5.41 5.44 5.46 5.44 5.60
g (9.77) (9.73) (9.74) (9.79) (9.82) (9.80) (10.08)

 

138

 

Table B-7. CTE Values for CTE 7 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 7, Age (Days)
us/°F (us/°C)
3 7 14 28 90 180 365
AASHTO TP60
Sam 1e #1 5.85 5.95 5.98 5.91 5.95 6.15 6.15
p (10.53) (10.71) (10.77) (10.63) (10.71) (11.07) (11.07)
Sam 1 e #2 5.97 5.92 5.82 5.93 5.95 6.03 6.25
p (10.74) (10.66) (10.47) (10.67) (10.72) (10.85) (11.25)
Sam le #3 6.02 5.95 5.96 5.87 5.98 5.72 5.82
p (10.84) (10.71) (10.72) (10.57) (10.76) (10.30) (10.48)
Avera e 5.95 5.94 5.92 5.90 5.96 5.97 6.07
g (10.71) (10.69) (10.65) (10.62) (10.73) (10.74) (10.93)
Revised TP60
Sam 1e #1 5.86 5.91 5.97 5.95 5.99 6.16 6.22
p (10.54) (10.63) (10.74) (10.70) (10.78) (11.09) (11.20)
Sam le #2 5.89 5.97 5.85 5.87 5.93 6.02 6.20
p (10.60) (10.74) (10.52) (10.57) (10.67) (10.84) (11.17)
Sam 1e #3 6.16 6.00 6.06 5.82 6.12 5.78 5.92
p (11.09) (10.80) (10.91) (10.48) (11.01) (10.41) (10.66)
Avera e 5.97 5.96 5.96 5.88 6.01 5.99 6.12
g (10.74) (10.73) (10.73) (10.59) (10.82) (10.78) (11.01)

 

139

 

 

Table B-8. CTE Values for CTE 8 Mixture

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CTE 8, Age (Days)
e/°F( a/°C)
'1 p 7 14 28 90 180 365
AASHTO TP60
Sam le #1 5.97 5.71 5.86 6.02 6.13 6.37
p (10.42) (10.28) (10.54) (10.84) (11.03) (11.46)
Sam le #2 5.58 5.46 5.87 6.04 6.13 6.28
p (10.04) (9.83) (10.56) (10.87) (11.04) (11.31)
Sam le #3 5.81 5.81 5.89 6.00 6.00 6.10
p (10.46) (10.46) (10.60) (10.81) (10.81) (10.97)
Avera e 5.73 5.66 5.87 6.02 6.09 6.25
g (10.31) (10.19) (10.57) (10.84) (10.96) (11.25)
Revised TP60
Sam le #1 5.89 5.79 5.98 6.16 6.26 6.44
p (10.61) (10.43) (10.77) (11.09) (11.26) (11.60)
Sam 1e #2 5.60 5.45 5.98 6.05 6.06 6.23
p (10.08) (9.82) (10.77) (10.90) (10.90) (11.22)
Sam 1e #3 5.89 5.92 6.03 6.11 6.09 6.20
p (10.61) (10.65) (10.86) (10.99) (10.96) (11.17)
Avera e 5.80 5.72 6.00 6.11 6.13 6.29
g (10.43) (10.30) (10.80) (10.99) (11.04) (11.33)

 

140

 

REFERENCES

141

 

REFERENCES

. Mallela, J ., Abbas, A., Harman, T., Chetna, R., Liu, R., and Darter, M.,

Measurement and Significance of the Coefficient of Thermal Expansion of
Concrete in Rigid Pavement Design. Transportation Research Record: Journal of
the Transportation Research Board No. 1919, Washington DC, 2005.

. Dale Harrington, “Guide to Concrete Overlays”, National Concrete Pavement
Technology Center, Second Edition, 2008

. ARA Inc., ERES Consultants Division, Guide for Mechanistic-Empirical Design
of New and Rehabilitated Pavement Structures. Final Report, National
Cooperative Highway Research Program, 1-37A, March 2004.

. AASHTO, Standard Test Method for the Coefficient of Thermal Expansion of
Hydraulic Cement Concrete. TP60-00, 2000.

. Meyers, S. L., Thermal Expansion Characteristics of Hardened Cement Paste and
of Concrete. Proceedings of Highway Research Board, 30, 1950.

. Mindess, S., Young, J. F., Darwin, D., Concrete. Second Edition, Prentice Hall,
2003.

. Baiant, Z., P., Delayed Thermal Dilatations of Cement Paste and Concrete Due to
Mass Transport. Nuclear Engineering and Design, Vol. 14, 1970.

. Helmuth, R., A., Dimensional Changes of Hardened Portland Cement Pastes
Caused by Temperature Changes. Proceedings of the Highway Research Board,
40, 1967.

. Grasley, 2., C., Lange, D., A., Thermal Dilation and Internal Relative Humidity
of Hardened Cement Paste. Materials and Structures, RILEM, 40, 2007.

142

10.

11.

12.

13.

14.

15.

16.

17.

18.

Sellevold, E., J ., Bjentegaard, 0., Coefficient of Thermal Expansion of Cement
Paste and Concrete: Mechanisms of Microstructure Interaction. Materials and
Structures, RILEM, 39, 2006.

Powers, T., C., Brownyard, T., L., Studies of the Physical Properties of Hardened
Portland Cement Paste. Journal of the American Concrete Institute, Vol. 18, No.
5, 1947.

Bonnell, D., G., R., Harper, F ., C., The Thermal Expansion of Concrete. National
Building Studies, Technical Paper No. 7, HSMO, 1951.

Sabri, S., Illston, J ., M., Immediate and Delayed Thermal Expansion of Hardened
Cement Paste. Cement and Concrete Research, Vol. 12, 1982.

Yeon, j., H., Choi, S., Won, M., C., Effect of Relative Humidity on the
Coefficient of Thermal Expansion of Hardened Cement Paste and Concrete.
Proceedings of the 88'” Transportation Research Board Annual Meeting,
Washington DC, 2009.

Alungbe, G., Tia, M., and Bloomquist, D., Effects of Aggregate, Water/Cement
Ratio, and Curing on the Coefficient of Linear Thermal Expansion of Concrete.
Transportation Research Record: Journal of the Transportation Research Board
No. 1335, Washington DC, 1992.

Won, M., Improvements of Testing Procedures for Concrete Coefficient of
Thermal Expansion. Transportation Research Record: Journal of the
Transportation Research Board No. 1919, Washington DC, 2005.

Kohler, E., Alvarado, R., and Jones, D., Measurement and Variability of
Coefficient of Thermal Expansion for Concrete Pavements. Proceedings of the
86th Transportation Research Board Annual Meeting, Washington DC, 2007.

Tran, N., H., Hall, K., D., James, M., Characterization of Coefficient of Thermal
Expansion of Concrete Materials to Support Implementation of the Mechanistic-

Empirical Pavement Design Guide. Proceedings of the 87'“ Transportation
Research Board Annual Meeting, Washington DC, 2008.

143

 

19. Naik, T. R., Chun Y., Kraus R. N., Investigation of Concrete Properties to

20.

Support Implementation of the New AASHTO Pavement Design Guide. Final
Report. WHRP 06—14. Wisconsin Department of Transportation, 2006.

Tanesi, J., Kutay, M., Abbas, A., and Meininger, R., Effect of CTE Variability on
Concrete Pavement Performance as Predicted using the Mechanistic-Empirical
Pavement Design Guide. Proceedings of the 86" Transportation Research Board
Annual Meeting, Washington DC, 2007.

21. Neville, A., M., Properties of Concrete. Fourth Edition, John Wiley & Sons, Inc.,

22.

23.

24.

25.

26.

27.

28.

1996.

Kohler, E., Kannekanti V., Influence of the Coefficient of Thermal Expansion on
the Cracking of Jointed Concrete Pavements. Proceedings of the Sixth RILEM
International Conference on Cracking in Pavements, Chicago, Illinois, 2008.

Hossain, M., Khanum, T., Tanesi, J ., Scheiber, G., and Montney, R., The PCC
Coefficient of Thermal Expansion Input for the Mechanistic-Empirical Pavement
Design Guide. Proceedings of the 85th Transportation Research Board Annual
Meeting, Washington DC, 2006.

Michigan Department of Transportation, 2003 Standard Specifications for
Construction, Lansing, Michigan, 2003.

Sutter, L., Van Dam, T., Peterson, K., Mineral Characterization and Cataloging
of Quarried Aggregate Sources Used in Michigan Highway Construction. Volume
11: Materials Characterization Database, Michigan Department of Transportation,
June 2003.

WWW.polyscience.com

WWW.iotech.com

WWW.macrosensors.com

144

 

29. Wittmann F., Lucas, J ., Experimental Study of Thermal Expansion of Hardened
Cement Paste. Materials and Structures, 7, No.40, 1974.

30. Rudeloff, Sieglerschmidt, Deutscher Ausschuss f. Eisenbeton, No. 23, 1913.

31. Khazanovich, L., M. Darter, R. Bartlett, McPeak, T., Common Characteristics of
Good and Poorly Performing PCC Pavements. Repot No. FHWA-RD-97—131,
1998.

145