MDOT RC-1525

CSD-2008-11





ULTRA-HIGH PERFORMANCE
CONCRETE FOR MICHIGAN BRIDGES
MATERIAL PERFORMANCE – PHASE I

FINAL REPORT – NOVEMBER 2008
















CENTER FOR STRUCTURAL DURABILITY
MICHIGAN TECH TRANSPORTATION INSTITUTE










(This page intentionally left blank)


Technical Report Documentation Page
1. Report No.
Research Report RC-1525
2. Government Accession No. 3. MDOT Project Manager
Roger Till, P.E.
4. Title and Subtitle
Ultra-High-Performance-Concrete for Michigan Bridges
Material Performance – Phase I
5. Report Date
November 13, 2008
7. Author(s)
Dr. Theresa M. Ahlborn, Mr. Erron J. Peuse, Mr. Donald Li Misson
6. Performing Organization Code
MTU
9. Performing Organization Name and Address
Center for Structural Durability
Michigan Technological University
1400 Townsend Drive
Houghton MI 49931-1295
8. Performing Org Report No.
CSD-2008-11

12. Sponsoring Agency Name and Address
Michigan Department of Transportation
Construction and Technology Division
PO Box 30049
Lansing MI 48909
10. Work Unit No. (TRAIS)

11. Contract Number
2003-0063
11(a). Authorization Number
Auth 21 Rev.1
15. Supplementary Notes 13. Type of Report and Period
Covered
Final Report
14. Sponsoring Agency Code

16. Abstract
One of the latest advancements in concrete technology is Ultra-High Performance Concrete (UHPC). UHPC is
defined as concretes attaining compressive strengths exceeding 25 ksi (175 MPa). It is a fiber-reinforced, densely-
packed concrete material which exhibits increased mechanical performance and superior durability to normal and
high strength concretes. UHPC has great potential to be used in the bridge market in the United States. However, to
gain acceptance by designers, contractors, and owners this material needs to be tested according to American
Society for Testing and Materials (ASTM) International and American Association of State Highway Transportation
Officials (AASHTO) standards, and new practices must be developed.
The focus of this research was to investigate how the age at which UHPC undergoes a steam (thermal) treatment
affects some mechanical and durability properties. Four mechanical properties (compressive strength, modulus of
elasticity, Poisson’s ratio, and flexural characteristics) and properties related to durability (chloride ion penetration
resistance, freeze-thaw durability, and coefficient of thermal expansion) were investigated. The testing was
conducted with differing curing conditions and at different ages to examine how these factors influence each of the
measured properties. Specimens, independent of age at thermal treatment, yielded compressive strengths of over 30
ksi, modulus of elasticity values in excess of 8000 ksi, and a Poisson’s ratio of 0.21. Flexural characteristics were
dependent on curing regime. Testing consistently validated that UHPC had negligible chloride ion penetration, a
high resistance to freeze-thaw cycling (durability factor of 100), and coefficient of thermal expansion values similar
to that of normal strength concretes for both ambient cured and thermally treated specimens. Additional results
revealed UHPC’s autogenous healing properties while undergoing freeze-thaw cycling, low variability between
batches, and the reproducibility of results between different U.S. laboratories.
Lastly, recommendations were developed for future testing of UHPC durability properties and a preliminary life-
cycle cost comparison showed that the low life-maintenance costs of UHPC can offset higher initial costs, especially
as the use of UHPC in the U.S. increases and the initial cost of the material decreases.
17. Key Words:
Ultra High Performance Concrete, UHPC, Bridge Materials,
Compressive Strength, Modulus, Poisson’s Ratio, Flexure, Rapid
Chloride Penetration, Freeze-Thaw, Coefficient of Thermal
Expansion, Life Cycle Cost
18. Distribution Statement
No restrictions. This document is
available to the public through the
Michigan Department of
Transportation.
19. Security Classification (report)
Unclassified
20. Security Classification (Page)
Unclassified
21. No of Pages
181
22. Price
Report RC-1525










(This page intentionally left blank)


Ultra-High-Performance-Concrete for
Michigan Bridges
Material Performance – Phase I


Submitted by the Michigan Tech
CENTER FOR STRUCTURAL DURABILITY
A Michigan DOT Center of Excellence


Submitted to:




Final Report – November 2008


Submitted by:



Michigan Technological University
Civil & Environmental Eng. Dept.
1400 Townsend Dr.
Houghton, Michigan 49931
Fax: 906/487-1620

Dr. Theresa M. Ahlborn, P.E.
Associate Professor and CSD Director
906/487-2625
tess@mtu.edu

Mr. Erron J. Puese and Mr. Donald Li Misson
Former Graduate Research Assistants












(This page intentionally left blank)


ACKNOWLEDGEMENTS

This project was financially supported by the Michigan Department of Transportation in
cooperation with the Federal Highway Administration. The authors would like to thank the
members of the Michigan Department of Transportation (MDOT) Research Advisory Panel
(RAP), including Project Manager Mr. Roger Till, P.E., for their guidance, suggestions, and
patience throughout the course of the project. The authors would also like to acknowledge the
contributions of Mr. Chris Gilbertson, P.E., Research Engineer, for oversight of the experimental
studies; Ms. Kari Klaboe, undergraduate research assistant for assistance with the preliminary
cost-benefit study, and Mr. Charles Mott, MTTI Operations Manager, for technical editing of the
final report.





DISCLAIMER

The content of this report reflects the views of the authors, who are responsible for the facts and
accuracy of the information presented herein. This document is disseminated under the
sponsorship of the Michigan Department of Transportation in the interest of information
exchange. The Michigan Department of Transportation assumes no liability for the content of
this report of its use thereof.










(This page intentionally left blank)
i

Abstract
One of the latest advancements in concrete technology is Ultra-High Performance
Concrete (UHPC). UHPC is defined as concretes attaining compressive strengths exceeding 25
ksi. It is a fiber-reinforced, densely-packed concrete material which exhibits increased
mechanical performance and superior durability to normal and high strength concretes. UHPC
has great potential to be used in the bridge market in the United States. However, to gain
acceptance by designers, contractors, and owners this material needs to be tested according to
American Society for Testing and Materials (ASTM) International and American Association of
State Highway Transportation Officials (AASHTO) standards, and new practices must be
developed.
The focus of this research was to investigate how the age at which UHPC undergoes a
steam (thermal) treatment affects some mechanical and durability properties. Four mechanical
properties (compressive strength, modulus of elasticity, Poisson’s ratio, and flexural
characteristics) and properties related to durability (chloride ion penetration resistance, freeze-
thaw durability, and coefficient of thermal expansion) were investigated. The testing was
conducted with differing curing conditions and at different ages to examine how these factors
influence each of the measured properties. Specimens, independent of age at thermal treatment,
yielded compressive strengths of over 30 ksi, modulus of elasticity values in excess of 8000 ksi,
and a Poisson’s ratio of 0.21. Flexural characteristics were dependent on curing regime. Testing
consistently validated that UHPC had negligible chloride ion penetration, a high resistance to
freeze-thaw cycling (durability factor of 100), and coefficient of thermal expansion values
similar to that of normal strength concretes for both ambient cured and thermally treated
specimens. Additional results revealed UHPC’s autogenous healing properties while undergoing
freeze-thaw cycling, low variability between batches, and the reproducibility of results between
different U.S. laboratories.
Lastly, recommendations were developed for future testing of UHPC durability
properties and for a future design code, and a preliminary life-cycle cost comparison showed that
the low life-maintenance costs of UHPC can offset higher initial costs, especially as the use of
UHPC in the U.S. increases and the initial cost of the material decreases.

ii











(This page intentionally left blank)

iii

Contents
Abstract ............................................................................................................................................ i 
Contents ......................................................................................................................................... iii 
List of Figures ................................................................................................................................ vi 
List of Tables ............................................................................................................................... viii 
1.0  Introduction to Ultra-High Performance Concrete (UHPC) ............................................... 1 
1.2  Objectives .............................................................................................................. 3 
1.3  Scope ..................................................................................................................... 4 
2.0  Review of UHPC ................................................................................................................ 5 
2.1  UHPC Composition............................................................................................... 8 
2.2  Types of UHPC ................................................................................................... 10 
2.3  Applications of UHPC ........................................................................................ 10 
2.4  Mechanical Properties ......................................................................................... 15 
2.4.1  Compressive Strength ................................................................................... 15 
2.4.2  Modulus of Elasticity and Poisson’s Ratio ................................................... 17 
2.4.3  First-Crack Flexural Strength and Flexural Toughness ................................ 20 
2.4.4  Thermal Treatment ........................................................................................ 21 
2.5  Durability Improvements .................................................................................... 22 
2.5.1  Chloride Ion Penetration ............................................................................... 23 
2.5.2  Freeze-Thaw Testing ..................................................................................... 25 
2.5.3  Coefficient of Thermal Expansion ................................................................ 27 
2.5.4  Additional Durability Research ..................................................................... 29 
2.6  Other UHPC research .......................................................................................... 29 
3.0  Methodology ..................................................................................................................... 33 
3.1  Introduction ......................................................................................................... 33 
3.2  UHPC Mixing Procedure .................................................................................... 34 
3.3  Casting Specimens .............................................................................................. 39 
3.4  Curing Regimes ................................................................................................... 40 
3.5  Specimen Preparation and Test Procedures ........................................................ 41 
iv

3.5.1 Compressive Strength ...................................................................................... 42 
3.5.2 Modulus of Elasticity and Poisson’s Ratio ...................................................... 45 
3.5.3 Flexural Strength .............................................................................................. 46 
3.5.4 Rapid Chloride Penetration .............................................................................. 48 
3.5.5 Freeze-Thaw Cyclic Testing ............................................................................ 53 
3.5.6 Coefficient of Thermal Expansion ................................................................... 57 
4.0  Results and Discussion ..................................................................................................... 63 
4.1  Compression Strength ......................................................................................... 64 
4.1.1  Results ........................................................................................................... 64 
4.1.2  Statistical Analysis and Discussion ............................................................... 66 
4.1.3  Air-Cured Compressive Strength Growth over Time ................................... 70 
4.2  Modulus of Elasticity and Poisson’s Ratio ......................................................... 72 
4.2.1  Results ........................................................................................................... 73 
4.2.2  Statistical Analysis and Discussion ............................................................... 76 
4.2.3  Compressive Stress and Modulus of Elasticity Relationship ........................ 81 
4.3  Flexural Strength Testing for First Cracking ...................................................... 84 
4.3.1  Results ........................................................................................................... 85 
4.3.2  Statistical Analysis and Discussion ............................................................... 87 
4.3.3  Flexural Toughness ....................................................................................... 89 
4.4  Rapid Chloride Penetration Test ......................................................................... 94 
4.4.1  Results ........................................................................................................... 95 
4.4.2  Discussion ..................................................................................................... 95 
4.5  Freeze-Thaw Cyclic Testing ............................................................................... 98 
4.5.1  Results ........................................................................................................... 98 
4.5.2  Discussion ..................................................................................................... 99 
4.6  Coefficient of Thermal Expansion .................................................................... 108 
4.6.1  Results ......................................................................................................... 108 
4.6.2  Discussion ................................................................................................... 110 
4.6.3  Study of water absorption ........................................................................... 113 
5.0  Conclusions of the Experimental Studies ....................................................................... 115 
5.1  Compression Strength ....................................................................................... 116 
5.2  Modulus of Elasticity and Poisson’s Ratio ....................................................... 116 
v

5.3  Flexural Strength Testing for First Cracking and Toughness ........................... 117 
5.4  Rapid Chloride Penetration Test ....................................................................... 118 
5.5  Freeze-Thaw Cyclic Testing ............................................................................. 118 
5.6  Coefficient of Thermal Expansion .................................................................... 119 
6.0  Preliminary Life Cycle Costs of a UHPC Superstructure ............................................... 121 
6.1  Bridge Components ........................................................................................... 121 
6.2  Construction ...................................................................................................... 122 
6.3  Maintenance ...................................................................................................... 124 
6.4  Preliminary Life Cycle Costs ............................................................................ 126 
6.5  Conclusion of the Preliminary Life Cycle Cost Analysis ................................. 128 
6.6  Future Work ...................................................................................................... 129 
7.0  Recommendations, Implementation and Future Work ................................................... 131 
7.1  Recommendations for UHPC Testing Procedures ............................................ 131 
7.1.1  Compression Testing ................................................................................... 131 
7.1.2  Modulus of Elasticity and Poisson’s Ratio ................................................. 131 
7.1.3  Flexural Strength Testing for First Cracking and Flexural Toughness ....... 132 
7.1.4  Rapid Chloride Penetration Test ................................................................. 132 
7.1.5  Freeze-Thaw Cyclic Testing ....................................................................... 133 
7.1.6  Coefficient of Thermal Expansion .............................................................. 133 
7.2  Draft U.S. Design Recommendations for UHPC .............................................. 134 
7.3  Implementation Activities ................................................................................. 140 
7.4  Suggested Future Work ..................................................................................... 141 
References ................................................................................................................................... 145 
A  Appendix A – Experimental Test Data ........................................................................... A-1 
B  Appendix B – CTE Test Procedure Modifications ......................................................... B-1 



vi

List of Figures
Figure 2-1: UHPC Example: Sherbrooke Footbridge (Resplendino and Petitjean 2003) ............ 11 
Figure 2-2: UHPC Footbridges ..................................................................................................... 13 
Figure 2-3: UHPC Construction Examples................................................................................... 14 
Figure 2-4: UHPC Girder Testing ................................................................................................. 30 
Figure 2-5: Mars Hill Bridge in Wapello County, Iowa (Lafarge 2006b) .................................... 31 
Figure 3-1: Doyon Mixer .............................................................................................................. 35 
Figure 3-2: Turning Point of UHPC ............................................................................................. 37 
Figure 3-3: Impact Table Measurement of UHPC’s Flow ............................................................ 39 
Figure 3-4: Michigan Tech’s UHPC Thermal Treatment Cure Chamber ................................... 41 
Figure 3-5: Reid Surface Grinder ................................................................................................. 42 
Figure 3-6: End Perpendicularity Set-up ...................................................................................... 43 
Figure 3-7: Baldwin CT 300 Compression Testing Machine ....................................................... 44 
Figure 3-8: Compressometer and Extensometer ........................................................................... 46 
Figure 3-9: ASTM 1018 Loading Configuration .......................................................................... 48 
Figure 3-10: Epoxy-coated UHPC Specimens for RCPT ............................................................. 50 
Figure 3-11: ASTM C 1202 Specimen Preparation Setup............................................................ 50 
Figure 3-12: UHPC Specimen Undergoing ASTM C 1202 Testing ............................................ 51 
Figure 3-13: MTU 80-specimen Freeze-Thaw Chamber (Procedure B) ...................................... 54 
Figure 3-14: Testing the Fundamental Transverse Frequency of an UHPC Specimen ................ 55 
Figure 3-15: Length Change Measurement of an UHPC Freeze-Thaw Specimen ....................... 56 
Figure 3-16: Epoxy-coating CTE Specimen ................................................................................. 59 
Figure 3-17: Pine CTE Specimen Test Frame and Water Bath .................................................... 60 
Figure 3-18: UHPC Specimen in Water Bath Undergoing CTE Testing ..................................... 61 
Figure 4-1: Mean Compressive Results for All Ages and Curing Regimes ................................. 66 
Figure 4-2: Compressive Stress Gain over Time for Air-Cured Specimens ................................ 71 
Figure 4-3: Typical Stress-Strain Curve for Calculating the Modulus of Elasticity ..................... 74 
Figure 4-4: Mean Modulus of Elasticity Results for All Ages and Curing Regimes ................... 75 
Figure 4-5: Mean Poisson’s Ratio Results for All Ages and Curing Regimes ............................. 75 
Figure 4-6: Regression Model for Modulus of Elasticity vs. Compressive Strength ................... 82 
Figure 4-7: Mean Values of Compressive Stress and Modulus of Elasticity for Air-Cured
Specimens ......................................................................................................................... 83 
Figure 4-8: Mean First-Crack Flexural Stress for All Curing Regimes ........................................ 87 
Figure 4-9: Load Deflection Curve for Elastic-Plastic Material (ASTM C 1018 Figure X1.1) ... 90 
Figure 4-10: Typical Load Deflection Curve for Flexural Specimens ......................................... 91 
Figure 4-11: Surface Staining of UHPC Specimen after ASTM C 1202 Test ............................. 97 
Figure 4-12: Effects of Freeze-Thaw Cycling on the Average Relative Dynamic Modulus of
UHPC Samples ............................................................................................................... 101 
Figure 4-13: Cracks in Air Cured UHPC Specimens Following Freeze-Thaw Testing ............. 102 
Figure 4-14: Average Resonant Frequencies of UHPC Freeze-Thaw and Side-Study Specimens
......................................................................................................................................... 104 
Figure 4-15: Resonant Frequencies of UHPC Specimens after Freeze-Thaw Testing ............... 105 
Figure 4-16: Resonant Frequencies of UHPC Specimens after Side-Study Testing .................. 106 
vii

Figure 4-17: Typical Bell Shaped Resonant Frequency Output of Air-cured UHPC Specimen
(Frequency in Hz) ........................................................................................................... 107 
Figure 4-18: Typical Skewed Resonant Frequency Output of an Air-cured UHPC Specimen Six
Months after Freeze-Thaw Testing (Frequency in Hz) ................................................... 107 
Figure 4-19: Average CTE Values for Air-cured UHPC Specimens ......................................... 111 
Figure 6-1: Target Cost of UHPC ............................................................................................... 128 

viii

List of Tables

Table 2.1: Comparison of UHPC Material Properties to Other Concrete Classifications .............. 7 
Table 2.2: Composition of a Typical UHPC Mix ........................................................................... 8 
Table 3.1: Ductal® Mix Proportions for 0.65 ft
3
batch ................................................................ 35 
Table 3.2: Typical and Adjusted Mixing Procedures ................................................................... 36 
Table 3.3: Flow Domain Classifications of Freshly Mixed UHPC .............................................. 39 
Table 3.4: Chloride Ion Penetrability Based on Charge Passed (ASTM C 1202) ........................ 52 
Table 4.1: Experimental Test Matrix - Specimens Tested per Curing Regime ............................ 63 
Table 4.2: Compressive Stress Test Results ................................................................................. 65 
Table 4.3: Statistical Results for Compressive Strength Testing .................................................. 68 
Table 4.4: Combined Compressive Stress Results ....................................................................... 69 
Table 4.5: Statistical Results for Combined Compressive Strength Testing ................................ 69 
Table 4.6: Modulus of Elasticity and Poisson’s Ratio Test Results ............................................. 73 
Table 4.7: Statistical Results for Modulus of Elasticity Testing .................................................. 77 
Table 4.8: Combined Modulus of Elasticity Results .................................................................... 78 
Table 4.9: Statistical Results for Combined Modulus of Elasticity Testing ................................. 78 
Table 4.10: Statistical Results for Poisson’s Ratio Testing .......................................................... 79 
Table 4.11: Combined Poisson’s Ratio Results ............................................................................ 80 
Table 4.12: Statistical Results Poisson’s Ratio Testing ................................................................ 80 
Table 4.13: Flexural Stress, Deflection and Maximum Load Results .......................................... 86 
Table 4.14: Corrected First-Crack Flexural Strength Hypothesis Testing ................................... 88 
Table 4.15: Typical Toughness Values (ASTM C 1018 Figure X1.1) ......................................... 90 
Table 4.16: Experimental Toughness Indices and Residual Strength Factors .............................. 92 
Table 4.17: Michigan Tech Rapid Chloride Penetration Summary Data ..................................... 95 
Table 4.18: Graybeal (2006a) Rapid Chloride Penetration Summary Data ................................. 96 
Table 4.19: Effects of Freeze-Thaw Cycles on UHPC ................................................................. 99 
Table 4.20: Change in Resonant Frequency of UHPC Specimens after Testing Completed ..... 105 
Table 4.21: Coefficient of Thermal Expansion (CTE) Test Summary ....................................... 110 
Table 4.22: Comparison of Some Published UHPC CTE Data .................................................. 113 
Table 6.1: Bridge Component Unit Costs .................................................................................. 122 
Table 6.2: Construction Activities Unit Costs ........................................................................... 122 
Table 6.3: Estimated Construction Costs ................................................................................... 123 
Table 6.4: Unit Costs of Maintenance Activities ....................................................................... 124 
Table 6.5: Bridge Girder Maintenance ....................................................................................... 125 
Table 6.6: Bridge Deck Maintenance ......................................................................................... 126 
Table 6.7: Costs of Control and UHPC Bridges, 2007 $ ............................................................ 127 
1

1.0 Introduction to Ultra-High Performance Concrete (UHPC)
Concrete has been one of the most widely used building materials because of its
compressive strength, resistance to water, and its ability to be easily formed and placed
according to need. While normal strength concrete, NSC, has long been able to achieve
compressive strengths of 3,000 – 5,000 psi , issues with deterioration and an increasing desire to
build larger and more robust structures with smaller members has driven researchers to explore
ever stronger and more durable concrete materials. Today, high-performance concrete, or HPC
(10,000 – 12,000 psi compressive strengths), with embedded steel reinforcement replaces normal
strength concrete in many structural applications. However, as concrete structures begin to be
constructed in ever more aggressive environments, durability in addition to strength must be
considered as a principal design concern.
Research over the past decade has yielded a new classification of highly resilient
concrete, called reactive powder concrete (RPC), with compressive strengths comparable to that
of some steels. Now labeled and classified as ultra-high performance concretes (UHPC), these
materials address many of the durability performance deficiencies associated with both NSC and
HPC. Ultra-High Performance Concrete (UHPC) is one of the latest advances in concrete
technology and it addresses the shortcomings of many concretes today: low strength to weight
ratio, low tensile strength, low ductility, and volume instability. In addition to achieving high
compressive strengths in excess of 25,000 psi (sometimes greater than 30,000 psi), UHPC is also
nearly impermeable. This very low permeability allows UHPC to withstand many distresses
normally associated with NSC and HPC such as freeze-thaw deterioration, corrosion of
embedded steel, and chemical ingress.
2

The implementation of UHPC in bridge construction around the world has sparked new
research investigating the potential utilization of UHPC in the U.S. bridge industry. The higher
strengths afforded to UHPC could allow increased girder spans while maintaining similar or
smaller cross-sectional areas. Costs may be reduced as the lower span to depth ratio of UHPC
bridges require less embankment fill while providing more aesthetically pleasing profiles.
Increased span lengths mean fewer support structures such as piers which can lead to improved
safety when traveling under overpasses and lower environmental impact in water crossings.
Additionally, beam spacing can be increased allowing for faster construction times, lower
transportation costs, and increased material efficiency.
Overall, the greatest impact of UHPC materials may lie in the improved durability of
concrete structures. The need for a structural material to perform in harsh environments is a
reality whether the structure is a local bridge subjected to the constant winter salting, or a bridge
support pier enduring the constant harsh freezing and thawing of the Straits of Mackinac. The
improved durability of UHPC may lead to lower bridge repair costs and less downtime due to
repair construction. UHPC bridges or structures constructed in aggressive environments may
remain structurally safe for generations. Also, bridges and buildings that were all but thought
impossible may now be realized. Additionally, longer lasting structures minimize the impact on
the environment. Cement production is a leading contributor to industrial process-related
emission sources (Hanle et al. 2004). While UHPC requires higher cement quantities than
normal concretes, the amount of cement used in the lifetime of a UHPC structure may be far less
than the amount used for several lifetimes of a NSC or HSC structure. Similarly, UHPC requires
much less maintenance than its concrete counterparts and in turn fewer materials are required for
repair or rehabilitation.
3

Despite these apparent benefits, material properties of UHPC need verification testing to
substantiate proprietary claims on strength and durability. Similarly, results from UHPC
research abroad must be validated here in the U.S. and tested according to American Society for
Testing and Materials (ASTM) and American Association of State Highway Transportation
Officials (AASHTO) standards where appropriate. A UHPC research initiative by the Federal
Highway Administration’s (FHWA) produced the first substantial UHPC research in the U.S. in
late 2006 (Graybeal 2006a), but there is need for additional testing and inter-laboratory
confirmation of some of the tests. Moreover, the effects of curing regimes and specimen age on
the mechanical and durability properties of UHPC require a more thorough investigation.
1.2 Objectives
The primary objective of this research is to present the history of ultra-high performance
concrete and to evaluate some material properties for potential use in durable highway structures.
The goals necessary to accomplish this objective are outlined below:
• Characterize some UHPC material properties and build upon previous research at
Michigan Tech and throughout the U.S. Properties include compressive strength,
modulus of elasticity, Poisson’s ratio, flexural first-crack strength, freeze/thaw behavior,
chloride permeability, and coefficient of thermal expansion.
• Consider the impact that different curing regimes had on the above mentioned properties.
The age at thermal treatment for curing varied from 3, 10, and 24 days, and included a
baseline case of ambient-cure.
• Conduct a preliminary life cycle cost comparison between a typical prestressed concrete
bridge built using standard building materials and the same bridge built using UHPC.
• Identify the impacts of UHPC material behavior on bridge design and construction.
4

• Develop recommendations for testing mechanical and durability properties of UHPC
evaluated in this research project.
1.3 Scope
To better understand UHPC and its potential impact on the transportation industry,
previous research and testing data regarding mechanical and durability properties of UHPC was
compiled and synthesized. In the area of ultra-high performance concretes, Europe has led the
way and produced substantial research about its material properties and durability. However,
recently the U.S. also began investigating this new material and in late 2006 FHWA published
the first large scale report on UHPC (Graybeal 2006a). Research is continuing at many
universities and a summary of past and current research related to selected UHPC properties is
presented herein.
Additionally, testing of properties was performed to analyze the effects of curing regime
and cure time, age of specimen, physical distress, and ionic transport. The results were then
compared to previous research to further characterize the material behavior. Moreover, due to
the unique nature of the material, suggested UHPC test procedures and methods were also
developed for the various tests. These suggested procedures can be used for further research on
UHPC or as a foundation for developing U.S. specifications for UHPC material and durability
testing.
Currently there is a large amount of research being pursued across the globe involving
many different types of UHPC materials. However, the only UHPC that is commercially
available in the U.S. is Ductal
®
, a product of Lafarge, Inc. It was for this reason that Ductal
®

was the only UHPC considered. More specifically, Ductal
®
BS1000 was used throughout this
research program.
5

2.0 Review of UHPC
UHPC is a new family of concretes which exhibits superior mechanical and durability
properties over traditional normal strength concrete (NSC) and high performance concrete
(HPC). This review incorporates and condenses the current body of information related to
UHPC material behavior and current applications, and serves to provide a basis for
understanding UHPC durability. The majority of the information on UHPC comes from sources
outside the United States that have been published since the mid-1990’s. However, the U.S.
Federal Highway Administration (FHWA) Turner Fairbank Laboratory recently completed an
extensive material property characterization study on a proprietary UHPC material. In addition
to the research completed at FHWA, several universities are conducting research on UHPC
behavior including Georgia Tech, Iowa State, Ohio University, and Virginia Tech. Many of
these research projects are funded through the state transportation departments (DOT) including
Virginia, Georgia, and Iowa DOT’s. Currently, there is no design code for UHPC in the U.S.,
but several other codes have been developed in Europe (AFGC 2002) and Japan (JSCE 2006).
While this literature review will cover many of the known properties of UHPC, only a few
studies have been conducted using accepted U.S. procedures and standards.
In the early 1990’s two separate French contractors, Eiffage Group and Boygues
Construction, with the help of Sika Corporation and Lafarge Corporation, respectively,
developed two different UHPC’s which exhibit similar properties (Harris 2004). Eiffage Group
with Sika Corporation created BSI
®
which is noted as being coarser than other UHPCs
(Jungwirth and Muttoni 2004), and the partnership between Boygues and Lafarge produced
Ductal
®
.
6

Coming on the heels of continued developments in high performance concrete (HPC), the
development of UHPC materials have benefited from both improved aggregate gradations and
the use of a high-range water reducer, or superplasticizer. UHPC was first developed as a
reactive powder concrete (RPC) with compressive strengths ranging from 29 to 116 ksi. These
high strengths were the products of improving homogeneity by eliminating coarse aggregates,
optimizing the granular mixture, and improving microstructure of cement paste by heat treatment
application (Richard and Cheyrezy 1995). Although UHPC use of non-continuous steel fibers
does not aid in increasing compressive strength, fibers do aid in improving UHPC’s ductility and
tensile strength. Table 2.1 compares some of UHPC’s properties to HPC and NSC.
7

Table 2.1: Comparison of UHPC Material Properties to Other Concrete Classifications
Material Characteristics NSC HPC UHPC
Maximum Aggregate Size, (in) 0.75-1.00 0.38-0.50 0.016-0.024
w/c Ratio (water/cement ratio) 0.40-0.70 0.24-0.35 0.14-0.27
Mechanical Properties NSC HPC UHPC
Compression Strength, (ksi) 3.0-6.0 6.0-14.0 25.0-33.0
Split Cylinder Tensile Strength, (ksi) 0.36-0.45 - 1.0-3.5
Poisson's Ratio 0.11-0.21 - 0.19-0.24
Creep Coefficient, Cu 2.35 1.6-1.9 0.2-0.8
Porosity 20-25% 10-15% 2-6%
Fracture Energy, (k-in/in
2
)
0.00057-
0.00086
- 0.057-0.228
Young's Modulus, (ksi) 2000-6000 4500-8000 8000-9000
Modulus of Rupture 1
st
crack, (ksi) 0.4-0.6 0.8-1.2 2.4-3.2
Flexure Strength - ultimate, (ksi) - - 3.0-9.0
Shrinkage -
Post Cure
40-80x10
-5

Post Cure <1x10
-5
,
No Autogenous
Shrinkage After
Cure
Coefficient of Thermal Expansion
(per °F)
4.1-7.3x10
-6
- 7.5-8.6 x10
-6

Ductility - - 250 Times > NSC
Durability Characteristics NSC HPC UHPC
Freeze/Thaw Resistance
10%
Durable
90%
Durable
100% Durable
Chloride Penetration (coulombs
passing)
> 2000 500-2000 < 100
Air Permeability (k) at 24 hrs and
40
o
C, (in
2
)
4.65x10
-14
0 0
Water Absorption at 225 hours,
(lb/in
2
)
4x10
-3
5x10
-4
7.1x10
-5

Chloride ion diffusion coefficient
(by steady state diffusion), (in
2
/s)
1.55x10
-9
7.75x10
-10
3.1x10
-11

Penetration of Carbon / Sulfates - - None
Scaling Resistance, (lb/ft
2
)
Mass
Removal
>0.205
Mass
Removal
0.016
Mass Removal
0.002
Note: Table and information adapted from Kollmorgen (2004), Hartmann and Graybeal (2001),
O’Neil et al. (1997), Russell (1999), Mamlouk and Zaniewski (1999), Mindess et al. (2003),
Mehta and Monteiro (2006), Aitcin (1998)

8

2.1 UHPC Composition
While considered a relatively new material, UHPC consists mostly of the same
constituents as normal strength concrete such Portland cement, silica fume, water, and quartz
sand. However, it also includes finely ground quartz, steel fibers (0.008 in. dia. x 0.5 in. long),
and superplasticizer. While other constituents have also been investigated, including carbon
nanotubes (Kowald 2004), most UHPC mixes consist of these basic elements. The combination
of these components creates a dense packing matrix that improves rheological and mechanical
properties, and also reduces permeability (Schmidt and Fehling 2005). A breakdown of the basic
constituents of a typical UHPC is shown in Table 2.2.
Table 2.2: Composition of a Typical UHPC Mix
Constituent lb/yd
3
% by Weight
Premix Portland Cement 1,180 - 1,710 27 - 38
Silica Fume 385 - 530 8 - 9
Ground Quartz 0 - 390 0 - 8
Fine Sand 1,293 - 1,770 39 - 41
Metallic Fibers (8.00 x10
-3
in. dia. by 0.500 in.) 245 - 320 5 - 8
Superplasticizer 20 - 30 0.5 - 1.0
Water 260 - 350 5 - 8
Water/Cementitious Material Ratio
(silica fume content is considered a cementitious
material and included in this ratio)
0.14 - 0.27 --
Note: Information condensed from the following references: Hartmann and Graybeal (2001),
Blais and Couture (1999), Dugat et al. (1996), and Richard and Cheyrezy (1996).


Portland cement is the primary binder used in UHPC, but at a much higher proportion
rate than in NSC or even HPC. The very low water to cementitious materials ratio prevents all
the cement from hydrating. After thermal treatment, unhydrated cement grains exist in the
matrix and act as particle packing material. Cement with high proportions of tricalcium
aluminate (CA
3
) and tricalcium silicate (C
3
S), and a lower Blaine fineness are desirable for
9

UHPC, as the CA
3
and C
3
S contribute to high early strength, and the lower Blaine fineness
reduces the water demand (Mindess et al. 2003). Despite the large amount of particles left
unhydrated, an RPC with a water-to-cementitious material ratio of 0.20 would reach
discontinuous capillary porosity when 26% hydration of cement has occurred (Bonneau et al.
2000). The addition of silica fume fulfills several roles including particle packing, increasing
flowability due to spherical nature, and pozzalonic reactivity (reaction with the weaker hydration
product calcium-hydroxide) leading to the production of additional calcium-silicates (Richard
and Cheyrezy 1995).
Quartz sand with a maximum diameter of 0.024 in. is the largest constituent aside from
the steel fibers. Both the ground quartz (4.0 x 10
-4
in.) and quartz sand contribute to the
optimized packing. Additionally, the most permeable portion of a concrete tends to be the
interfacial transition zone (ITZ) between coarse aggregates and the cement matrix (Mehta and
Monteiro 2006), and therefore, the elimination of coarse aggregates aids in improving the
durability of UHPC. This zone is the area around any inclusion in the cementitious matrix, and
is where the cement grains have difficulty growing because of the presence of a large surface
which impedes crystal growth. Silica fume (the smallest component in UHPC with a diameter of
0.2 μm) helps fill this region, and because it is highly pozzolanic, aids in increased strength and
reduced permeability. Reduction of the ITZ zone increases the tensile strength and decreases the
porosity of the cementitious matrix (Mindess et al. 2003). By reducing the amount of water
necessary to produce a fluid mix, and therefore permeability, the polycarboxylate
superplasticizer also contributes to improving workability and durability.
Finally, the addition of steel fibers aids in preventing the propagation of microcracks and
macrocracks and thereby limits crack width and permeability. For this particular application of
10

UHPC, straight high carbon steel fibers with a diameter of 0.008 in. and length of 0.5 in. are
used. This is the largest particle in the mix and is added at 2 percent by volume to the mix.
Because of its size relative to the other constituents, it reinforces the concrete on the micro level
and eliminates the need for secondary reinforcement in prestressed bridge girders (Graybeal
2005). The choice and quantity of this fiber was chosen because of its availability, use in
previous research, and likelihood that it will be used in the structures industry; specifically
bridges. Other fiber types (polymers, organic, etc) and geometries (crimped, hooked, etc) are
available, but were not investigated herein.
2.2 Types of UHPC
In Europe, there is a heavy push to develop many new and innovative types of UHPC
materials. Several that have already been developed include Ductal®, BSI®, and CEMENTEC
(Ahlborn et al. 2003) which are marketed by Lafarge, Eiffage Group, and Laboratoire of Central
des Ponts et Chausses of France, respectively. Ductal® has been promoted in North America by
the Lafarge North America group and is the brand of UHPC studied in this report. While the
various UHPC materials differ slightly in composition, and many new UHPC materials are in the
process of being developed, a basic understanding of UHPC material behavior and its potential
implementation remains a priority for the U.S.
2.3 Applications of UHPC
As UHPC is being developed, the proper market has yet to be discovered to utilize its
increased strength, durability, and flexural capacity. To date this versatile material has been used
in artwork, acoustical panels, precast elements, pedestrian bridges, and a few highway bridges.
Utilization of UHPC in the U.S. has been limited, but its international roots have led to many
different applications in Canada, Europe, Asia, and Australia. While many of the applications
11

have been related to the transportation industry, more and more uses for this innovative material
are being discovered to not only reap the benefits of its strength, but also UHPC’s durability.
Only a brief overview of UHPC functions in the world are presented here, however, more
detailed investigations of these uses can be found in other sources (Behloul and Cheyrezy 2002a
and 2002b; Kollmorgen 2004; Schmidt and Fehling 2005).
Development of UHPC began in the early 1990’s, and in 1997 the first structure made of
UHPC, the Sherbrooke pedestrian bridge, was constructed in Quebec, Canada. The 197 foot
long structure is a post tension open space truss (Figure 2-1). Six match cast segments compose
the main span. Among many other benefits, the enhanced mechanical properties of UHPC
allowed for the use of a deck top of only 1.2 in. thick (Semioli 2001). To develop an
understanding of how UHPC works in actual applications, a long term monitoring program was
also implemented on the bridge to monitor deflections and forces in the prestressing tendons.

Figure 2-1: UHPC Example: Sherbrooke Footbridge (Resplendino and Petitjean 2003)

In 1997 UHPC’s durability received a test when it was used to replace steel beams in the
cooling towers of the Cattenom power plant, in France. The environment is extremely corrosive
and UHPC was chosen because of its durability properties with the expectation of reduced or
eliminated maintenance. Three years later an AFGC-SETRA working group visited the site and
12

under a normal layer of sediment no deterioration of the UHPC was noted (Resplendino and
Petitjean 2003).
Other transit applications include footbridges constructed in South Korea, Japan, France,
and Germany. The Footbridge of Peace in Seoul, South Korea (Figure 2-2a and Figure 2-2b), is
an arch-bridge with a span of 394 ft, arch height of only 49 ft , and a deck thickness varying
anywhere between 1.2 in. and 4 in. (Brouwer 2001). In Japan, the Sakata-Mirai footbridge
(Figure 2-2c) was completed in 2002 and demonstrated how a perforated webs in a UHPC
superstructure can both reduce weight and be aesthetically pleasing (Tanaka et al. 2002). France
utilized UHPC’s fire resistant capabilities and high load carrying properties to construct an
aesthetically pleasing yet, highly fire resistant footbridge (Figure 2-2d) at a Chryso Plant in
Rhodia (Behloul and Cheyrezy 2002a). Most recently, the Gärtnerplatz Bridge was completed in
Kassel, Germany (Figure 2-2e) (Fehling et al. 2008).

13


(a)
(b)
(c)
(d)

(e)
Figure 2-2: UHPC Footbridges (a) Footbridge of Peace in Seoul, South Korea at
night (Behloul and Cheyrezy 2002a); (b) and during the day (Lafarge in Searls 2007); (c)
perforated hollow UHPC bridge girder (Tanaka et al. 2002); (d) fire resistant UHPC
footbridge in Rhodia, France; (e) Gärtnerplatz Bridge - Kassel, Germany

14

In 2001 the Bourg Les Valence Bridge in France was the 1
st
vehicle bridge constructed
using UHPC. It spans approximately 145 feet with two equal spans consisting of 5 π-shaped
prestressed elements. The π-shaped elements were connected by casting UHPC in situ
(Resplendino and Petitjean 2003). Additionally, the Shepards Creek Bridge in New South
Wales, Australia used UHPC to carry four lanes of traffic over a skewed (16°) single span of 49
ft. while reducing the dead weight by over half (Rebentrost and Cavill 2006).
UHPC made the transition to the United States in 2001 with the construction of the roof
of a clinker silo (Figure 2-3a) in Joppa, Illinois (Perry 2003). The 24 wedge-shaped precast
panels with a thickness of 0.5 in. covered the 58 ft. diameter silo. Utilizing UHPC saved time
and labor as the roof was constructed faster and with fewer workers than the two companion
metal roofed silos. Continuing in the cement industry, UHPC has since been used to create
columns with a smaller cross section in a cement terminal in Detroit, Michigan (Figure 2-3b)
which allows for five more feet of truck width clearance for the three loading bays (Lafarge
North America 2006a).
(a)
(b)
Figure 2-3: UHPC Construction Examples (a) UHPC panels on Joppa clinker silo
(Behloul and Cheyrezy 2002a); (b) 54 ft UHPC columns in Detroit

15

2.4 Mechanical Properties
Characterization of the mechanical properties is imperative to the efficient design and use
of UHPC. The following sections discuss the basic mechanical properties.
2.4.1 Compressive Strength
One of the most noticeable assets of UHPC is its high compressive strength. Perry and
Zakariasen (2003) demonstrated that UHPC is capable of reaching compressive strengths of 25-
33 ksi. This was supported by Kollmorgen (2004) with research showing a compressive strength
of over 28 ksi. The increase in compressive strength, over NSC or HPC, can be attributed to the
particle packing and selection of specific constituents, and thermal curing of UHPC. When
undergoing a 48 hour thermal treatment of 194°F at 95 percent relative humidity, Graybeal
(2005) showed an increase of 53 percent over non-thermally cured specimens of the same age.
This increase in compressive strength may allow UHPC to get a foot hold in the long span and
low span-to-depth ratio market segments which have been dominated by steel; creating choices
for designers and owners.
Testing UHPC with traditional standards is difficult because of its high compressive
strength. In the United States the standard size for a concrete cylinder is 6 x 12 in.; however, a
28 ksi cylinder of this size would require almost 800 kips to break. If the load rate of 35 psi per
second specified by ASTM C 39 Standard Test Method for Compressive Strength of Cylindrical
Concrete Specimens was followed, the same cylinder would take over 13 minutes to bring to
failure instead of the normal 2 to 6 minutes with NSC. The size of compression machine and the
length of the test may prove to be a barrier for production use in the U.S. Kollmorgen (2004)
showed that there was no size effect for UHPC for cylinders as small as 3 x 6 in., and suggested
that a 3 x 6 in. cylinder be used for standard testing of UHPC. In addition, Graybeal and
16

Hartmann (2003) showed that increasing the load rate to 150 psi per second does not affect the
results and greatly reduces the time required to complete a compression test. For a 3 x 6 in.
cylinder, the total load required to break the cylinder is decreased to approximately 200 kips and
the test duration reduced to just over three minutes.
Kollmorgen (2004) investigated the mechanical behavior of thermally treated UHPC at
different ages and with different sized specimens. Three cylindrical, two cube, and two
prismatic geometries were used to complete the testing. Specimens were cured under ambient
conditions for three days before being demolded, then tested or thermally treated. Thermal
treatment included a six hour ramp up to 194°F at 100 percent relative humidity, a 48 hour hold
period, and a ramp down over night. Over 240 compressive specimens were tested at various
ages and with different geometries. Specimens were tested before thermal treatment (3 days
after mixing), and after thermal treatment (7, 14, 28, and 56 days after mixing). Three different
cylindrical (4 x 8 in., 3 x 6 in., and 2 x 4 in.) and two different cube (3.94 x 3.94 x 3.94 in. and 2
x 2 x 2 in.) geometries were tested in compression. The average compressive stress exceeded 8.5
ksi for specimens tested 3 days after casting before thermal treatment and over 28 ksi for all
specimens undergoing thermal treatment regardless of age. It was shown that the age and size
effects were minimal on compression specimens, and a 3 x 6 in. cylinder was recommended for
use for compression testing.
Graybeal (2005) conducted a material characterization study prior to performing full
scale tests on AASHTO Type II girders made of UHPC. This characterization study included
defining mechanical and durability properties, as well as the long term stability under various
loading and environmental conditions. To reduce the time before the specimens could be
demolded, an accelerator was added to the mix and the specimens were demolded at
17

approximately 24 hours after casting. Specimens were then subject to one of the four curing
treatments of air, steam, delayed steam, and tempered steam treatments. Specimens undergoing
air curing treatment remained at standard laboratory atmospheric conditions until the time of
testing. Steam treatment began within 4 hours of demolding and consisted of a 2 hour ramp up
to 194°F at 95 percent relative humidity, followed by a 44 hour hold, and a 2 hour ramp down to
atmospheric conditions. Delayed steam treatment was similar to steam treatment except it
commenced on the 15
th
day after mixing. Tempered steam treatment was similar to steam
treatment except the temperature was limited to 140°F.
Graybeal (2005) tested nearly 1000 cylindrical and cubic compression specimens which
underwent one of the four curing treatments and yielded an average 28 day compressive stress of
18.3, 28.0, 24.8, and 24.8 ksi for air, steam, delayed steam, and tempered steam, respectively.
Several recommendations were made based on the research; 3 x 6 in. cylinders can be utilized
for compression testing and the load rate for compression testing can be increased to 150 psi per
second.
2.4.2 Modulus of Elasticity and Poisson’s Ratio
The modulus of elasticity is a material dependent property which is often described as a
mathematical relationship between stress and strain. Typically when the value is given for
concrete, it is referencing the elastic portion of the compressive stress-strain curve up to 40
percent of the ultimate compressive strength (0.40
c
f ` ) as specified in ASTM C 469 Standard
Test Method for Static Modulus of Elasticity and Poisson’s Ratio of Concrete in Compression.
The slope of the elastic portion of the stress-strain curve is the modulus of elasticity. The
modulus of elasticity is used in design calculations to predict deflection behavior of the element
so the design can often satisfy the specified limit states. Because testing the modulus of
18

elasticity is time consuming, and requires additional testing jigs and software to determine;
efforts have been undertaken to develop a relationship between modulus of elasticity and
compressive strength.
ACI Committee 318 (ACI 2005) presents an equation which relates the 28 day
compressive strength (f`
c
) of normal strength concrete to the modulus of elasticity (E
c
), for
concrete with a unit weight (
c
w ) of 90 to 155 pcf. The ACI 318-05 equation, is shown as
Equation 2.1
c c c
f w E ` * 33 *
5 . 1
= (psi) Equation 2.1
However, HPC has a much greater compressive strength than normal strength concrete.
ACI Committee 363 (ACI 1997) produced a relationship for higher strength concrete with
compressive strengths from 3,000 to 12,000 psi. The equation is shown below.
6
10 * 0 . 1 ` * 000 , 40 + =
c c
f E
(psi) Equation 2.2
The previous two equations do not apply as the high compressive strength of UHPC lies
above the range of applicable compressive strengths. Additionally, the Interim
Recommendations for UHPC, which was requested by the Association Française de Génie Civil
(AFGC) and published by Service d’études Techniques des Routes et Autoroutes (SETRA), does
not have an equation relating compressive strength to modulus of elasticity (AFGC 2002). The
document states that the material property should be determined in the later design stages by
conducting a test. However, a relationship is given from the work conducted at the Cattenom
nuclear power plant on 196 cylindrical specimens with diameters of 7 cm (2.76 in.). The
equation is shown below after it was converted to English units.
( )
3
` * 000 , 262
ATT c
f E =
(psi) Equation 2.3
19

Where: f`
ATT
= Compressive stress of UHPC after thermal treatment.
Research conducted by Sritharan et al. (2003) at Iowa State University on five 3 x 6 in.
cylinders produced an equation which took the following form.
ATT c
f E ` * 000 , 50 =
(psi) Equation 2.4
Where: f`
ATT
= Compressive stress of UHPC after thermal treatment.
Work completed by Kollmorgen (2004) at Michigan Tech resulted in an equation relating
compressive strength of thermally treated specimens to modulus of elasticity. The modulus of
elasticity was determined using local deformation transducers (LDT) made out of strips of
phosphorus bronze with strain gauges attached to each side of the metallic strip. Hinges were
glued to the specimens at a set gage distance and the LDT installed. Twenty-four cylindrical
specimens of 2 x 4 in., 3 x 6 in., and 4 x 8 in. were used to determine the following relationship
with an applicable range of 5 to 30 ksi.
( )
14 . 3
` * 000 , 351
ATT c
f E =
(psi) Equation 2.5
Where: f`
ATT
= Compressive stress of UHPC after thermal treatment.
Graybeal (2005) developed yet another relationship using a total of 148 specimens
undergoing one of four different curing regimes. As previously noted, the curing treatments
were air, steam, delayed steam, and tempered steam. Two parallel solid rings with a gage
distance of 2 in. were solidly attached to the specimens. The upper ring held three LVDTs which
end bears on the lower ring. The relationship was shown to apply to UHPC with compressive
strengths between 4 and 28 ksi, and any of the aforementioned curing treatments.
c c
f E ` * 200 , 46 =
(psi) Equation 2.6
20

Poisson’s ratio (ν) is defined as the relationship of the transverse strain (
trans
ε ) divided by
the longitudinal strain (
al longitudin
ε ) as shown in the equation below.
al longitudin
trans
ε
ε
ν =
Equation 2.7
This ratio is also used to relate the shear modulus, G, and modulus of elasticity, E.
) 1 ( * * 2 ν + = G E
Equation 2.8
SETRA (AFGC 2002) gives a Poisson’s ratio of 0.2 if it is not determined by direct
testing. This value is also the widely accepted value for Portland cement concrete of normal
strength.
2.4.3 First-Crack Flexural Strength and Flexural Toughness
ASTM C 1018 Standard Test Method of Flexural Toughness and First-Crack Strength of
Fiber-Reinforced Concrete (Using Beam with Third-Point Loading) is used to evaluate the first
crack strength and flexural toughness of portland cement concrete. However, no standards are
available for UHPC but ASTM C 1018 can be adapted. Small prismatic specimens are loaded at
the third point to create a region of constant moment in the specimen. The applied load and
resulting deflection are recorded to be used in determining the first-crack strength and post crack
flexural toughness. The first-crack strength is a useful indicator of the tensile strength of UHPC,
however it can overestimate the tensile strength when small scale prisms are utilized (Graybeal
2005). Flexural toughness is calculated as the area under the load deflection curve and is an
indication of the energy absorption capabilities.
Research by Cheyrezy et al. (1998) showed that UHPC was capable of reaching a
flexural strength as high as 7.0 ksi and had a toughness of 250 times that of normal strength
21

concrete. Perry and Zakariasen (2003) showed that UHPC had flexural strengths ranging from
5.0 – 7.0 ksi which confirmed Cheyrezy’s findings. Dugat et al. (1996) reported average
modulus of rupture values of 3.2 ksi and an ultimate flexural strength of 4.6 ksi. Graybeal and
Hartmann (2003) attributed the increase in the flexural behavior of UHPC to the particle packing
and the addition of fibers which hold the cement matrix together after cracking has occurred.
UHPC exhibits ductility because as the specimen begins to microcrack the small scale fibers
reinforce the matrix causing smaller, less damaging cracks to form.
Kollmorgen (2004) conducted flexural testing on 58 specimens. The specimens had two
different geometries to determine if a size effect existed on small scale prisms. Testing was
conducted on 2 x 2 x 11.25 in. and 3 x 3 x 11.25 in. prismatic specimens with 9 in. spans and
loading applied at the third points. A constant displacement rate of 1.50 x 10
-4
in/sec, at the
testing machine head, was used to test both specimen geometries. Average values of first-crack
strengths, maximum loads, and toughness values, based on AGFC (2002) were reported.
Graybeal (2005) tested 71 flexural specimens utilizing the procedure outlined in ASTM C
1018, which controls the rate of deflection of the prism. As previously noted the specimens
underwent one of four curing treatments, and utilized five different geometries/loading
configurations. Specimens had span lengths of 6 in., 9 in., 12 in., and 15 in. with a cross section
of 2 x 2 in. and a 12 in. span with a 3 x 4 in. cross section. Corrections were applied to calculate
a more representative tensile strength from the first-crack strength. Ultimate load and toughness
values based on the procedure outlined in ASTM C 1018 were reported.
2.4.4 Thermal Treatment
Due to the very low water-to-cementitious material ratio in UHPC, the full hydration
potentials of the cement and silica fume are never reached. However, improved performance has
22

been observed after thermally treating UHPC using combinations of heat, steam, and pressure
treatments (Loukili et al. 1998; Kollmorgen 2004; Graybeal 2006). The thermal treatment
appears to allow continued hydration of the portland cement and pozzolanic reaction of the silica
fume (Gatty et al. 1998; Cheyrezy et al. 1995). Loukili et al. (1998) noted that after treating
UHPC in 194°F water, up to 65% of the cement is hydrated (compared to 48% before treatment).
In addition to improved mechanical properties, Graybeal (2006) observed improved durability
characteristics including increased resistance to chloride penetration and abrasion. These
findings indicate that the full promises of UHPC’s benefits are not only realized because of
particle packing, but also due to the method of curing.
2.5 Durability Improvements
Concrete durability has become an ever more important aspect in the design of structural
concrete. While compressive strength has long been the standard for determining the quality of a
concrete, more and more research is focused on investigating the durability aspects of concretes.
Aitcin (1998) defines durability of concrete as “the resistance of concrete to the attack of
physical or chemical aggressive agents”. The American Concrete Institute, or ACI, further
details the durability of concrete as that which is able to resist weathering, chemical attack,
abrasion, or other processes of deterioration (ACI 2002). In general, the durability of a concrete
can be summarized as the capability of a concrete to continue performing its designed functions
while maintaining its dimensional stability in a given environment. Concrete can experience
deterioration from either physical attack (abrasion, freezing and thawing, fire, or salt
crystallization) or chemical agents (alkali-silica reaction, chloride ingress and corrosion of
embedded steel, sulfate attack, or delayed ettringite formation). All of these issues can lead to
additional durability problems or build upon already existing problems. Generally, outside of
23

poor material selection leading to internal attack (high chloride content in cement paste or alkali-
aggregate reaction) or poor construction practices, high permeability in a concrete is the main
cause of durability failures (Mindess et al. 2003; Mehta and Monteiro 2005). On the other hand,
UHPC has an extremely low water/cement ratio and a densely packed matrix that may contribute
to a very low permeability.
2.5.1 Chloride Ion Penetration
Chloride ion migration through a concrete by means of capillary absorption, hydrostatic
pressure, or diffusion (Stanish et al. 2000) is one of the most problematic durability issues
associated with low permeability concretes. Mehta and Monteiro (2005) define permeability as
the ease with which a fluid under pressure flows through a solid. A concrete with high
permeability is, therefore, much more susceptible to chloride ingress which eventually leads to
corrosion of reinforcing steel. Once chloride ions reach embedded steel, corrosion can take place
through an electro-chemical reaction that expands the steel up to 600%. Steel corrosion is such a
large problem that a 1991 FHWA report on the status of reinforced concrete bridges linked
corrosion as a cause of distress for a majority of cases (Mehta and Monteiro 2005).
However, previous research demonstrated that UHPC exhibited almost no permeability
and was not susceptible to chloride ingress. The very low water/cement ratio and densely packed
matrix of UHPC contribute to permeability results even lower than HPC. Permeability testing
demonstrated that UHPC has an oxygen permeability of less than 1.6 x 10
-15
in.
2
which is on the
extremely low end of testing (AFGC 2002), while O’Neil et al. (1997) reported water absorption
of 7.1 x 10
-5
lb/in.
2
. HPC on the other hand had an air permeability of 1900 x 10
-15
in.
2
and water
absorption of 49.7 x 10
-5
lb/in.
2
(O’Neil et al. 1997). Cheyrezy et al. (1995) used mercury
24

intrusion to demonstrate that the porosity of an RPC is less than 9% in volume for the pore
diameter range of 1.48 x 10
-7
in. to 3.74 x 10
-3
in.
Another method to determine whether a concrete is susceptible to chloride ingress uses
an applied electric potential across a specimen load cell to determine concrete’s conductance
(ASTM C 1202). Bonneau et al. (1997) reported less than 10 Coulombs passing (over a six hour
period) through UHPC specimens (negligible chloride ion penetrability) that were water cured at
varying times and temperatures. In the U.S., additional research by Graybeal (2006a)
demonstrated that UHPC had negligible chloride ion penetration when thermally treated and only
very low penetration when not thermally treated. While Graybeal (2006a) demonstrated that the
steel fibers did not contribute to a short circuit effect during UHPC testing, Toutanji et al. (1998)
revealed that adding 0.75 in. polypropylene fibers increased the permeability of concrete and
adding shorter fibers 0.50 in. reduced the permeability of the concrete. Furthermore, the addition
of silica fume greatly reduced the conductivity of the specimen. However, the reduction was not
proportional to the amount of silica fume added (Toutanji et al. 1998). Therefore, results from
rapid chloride penetration testing of UHPC should reflect these claims and demonstrate UHPC’s
high resistance to chloride penetration.
Similarly, the French recommendations report that UHPC has an electrical resistivity of
2878 kW/in, a rate of reinforcement corrosion less than 0.39 μin/yr, and only surface corrosion
of steel fibers when exposed to corrosive chemical conditions (AFGC 2002). Work by Schmidt
et al. (2003) supported claims of high resistance to aggressive agents such as de-icing salts,
carbonization, and chloride ion attack. The highly dense structure of the composite and the
reduction in pore volume restricts aggressive chemicals and water from entering UHPC’s
cementitious matrix, thus preventing deterioration. The Japan Society of Civil Engineers
25

acknowledges this characteristic by mandating only a 0.79 in. cover for prestressed strands used
in UHPC (JSCE 2006). Australian publications also reported chloride diffusion on the order of
31 x 10
-12
in.
2
/s to have replaced collapsed weir covers exposed to salt-water spray at the Eraring
Power Station with UHPC panels having a life expectancy greater than 100 years (Rebentrost
and Cavill 2006).
2.5.2 Freeze-Thaw Testing
Another mechanism of concrete distress that hinges on a concrete’s permeability is
freeze-thaw durability. Typically, concrete specimens exhibit distress, and eventually
deterioration, in the form of cracking, spalling, and disintegration as freeze-thaw cycling persists.
Freezing and thawing of saturated concrete occurs regularly in northern climates and over a
period of time can disintegrate both cement matrix and aggregates. Several mechanisms are
believed to be at work in this process including hydraulic pressures developed due to expanding
ice (water expands 9% when frozen) (Powers 1945), osmotic pressure (Powers and Helmuth
1953), water expulsion (Litvan 1972), or the movement of water towards frozen water to form
ice lenses (Collins 1944). While some or all of these theories may be at work in a specimen, the
methods to avoid freeze-thaw damage are more regularly accepted. The two methods of
producing freeze-thaw resistant concrete are: entraining air voids that allow pressure to dissipate
or using a sufficiently low w/c ratio to reduce capillary porosity which reduces the amount of
freezable water (Pigeon and Pleau 1995). The freeze-thaw durability of UHPC comes from its
highly impermeable matrix that effectively eliminates capillary porosity (Bonneau et al. 2000).
Most research to date has revealed that UHPC has a durability factor of 100 or greater (no
deterioration of specimens after 300 freeze-thaw cycles) (Bonneau et al. 1997). In fact, research
by Lee et al. (2005) and Graybeal (2006a) demonstrated that UHPC actually increased its
26

relative dynamic modulus, or RDM (ratio of squared resonant frequency at the end of freeze-
thaw testing to squared resonant frequency before testing – revealing amount of deterioration in
specimen), and gained mass as the freeze-thaw cycles continued. A mass increase of 0.2% after
125 cycles was documented by Graybeal and Tanesi (2007) and led to a further study of mass
change in UHPC specimens submerged in water. Normally, concretes lose mass due to material
spalling and experience a decrease in relative dynamic modulus as micro-cracking occurs.
In spite of this, compiled research on autogenous healing (sometimes referred to as self
healing) revealed that concrete can heal when exposed to water after or during deterioration
(Jacobsen and Sellevold 1996). Jacobsen and Sellevold also demonstrated that high strength
concretes can recover from damage due to freeze-thaw if submerged in water after testing.
Concretes that lost up to 50% of their RDM values, gained nearly all of it back after being
submerged in water for three months. However, the submerged specimens regained only 4-5%
of their compressive strengths after losing nearly 22-29% of it due to freezing and thawing.
Granger et al. (2007) investigated self-healing cracks in a UHPC material and determined that
precracked specimens stored in water regained stiffness upon reloading while precracked
specimens stored in air did not regain stiffness. The methods for self healing may come from the
gathering of debris in the cracks, hydration of unhydrated cement particles in low w/c concretes,
or precipitation of calcium carbonate (CaCO
3
) (Edvardsen 1999; Granger et al. 2007). This
phenomenon was further investigated by Graybeal (2006a) on air treated and steam treated
UHPC specimens not subjected to distress. Untreated UHPC specimens that were submerged in
water, but not subjected to freeze-thaw cycling, gained 0.25 – 0.35% mass compared to 0.09 –
0.18% mass gain for specimens remaining in an ambient air environment. Additionally, the
untreated UHPC specimens submerged in water increased in compressive strength by 12%
27

relative to the compressive strengths of specimens remaining in air. On the other hand, the
compressive strengths of the steam treated specimens submerged in water increased by only 3%
relative to their counterparts that remained in an ambient air environment (Graybeal and Tanesi
2007).
2.5.3 Coefficient of Thermal Expansion
Temperature fluctuations that are not in the freezing range can also play an important role
in structural design. The coefficient of thermal expansion (CTE) of a material is the strain in the
material per a given change in temperature. The coefficient of thermal expansion needs to be
considered when more than one type of material is used in construction or when member
expansion could lead to overstressing under large strains. Thermal strains of concrete structural
elements can lead to cracking and, therefore, susceptibility to chloride ingress or freeze-thaw
damage. Given UHPC’s unique material composition, establishing an appropriate CTE value for
UHPC is extremely important before utilizing it with other structural materials.
Because concrete is a composite material, its coefficient of thermal expansion value
correlates to the proportions (volume) of its constituents (Walker et al. 1952). As concretes
generally contain a majority coarse aggregate by volume, the CTE value largely depends on the
CTE of the aggregate (Mindess et al. 2003). Other important factors when considering a
material’s CTE are the w/c ratio, specimen age, and moisture content (Mindess et al. 2003).
Some of these factors can greatly affect the recorded CTE value of a material. Partially saturated
specimens can have CTE values as much as 1.8 times higher than fully saturated specimens
(Emanuel and Hulsey 1977), and the CTE of cement pastes can increase as much as 25% as the
cement fineness increases from 590 ft
2
/lb to 1300 ft
2
/lb (Mitchell 1953). However, dry
specimens exhibit only slightly higher CTE values on the order of 1.17 times larger than fully
28

saturated specimens (Emanuel and Hulsey 1977). Emanuel and Hulsey used these factors and
developed the following equation to estimate the CTE of a concrete:
[ ]
C T M A P S FA FA CA CA
f f f α β α β α β α = + +
Equation 2.9
where: CTE of concrete
CTE of cement paste
CTE of coarse aggregate
CTE of fine aggregate
correction factor for temperature alternations
(1.0 - controlled, 0.86 - outside)
correction fact
C
S
CA
FA
T
M
f
f
α
α
α
α
=
=
=
=
=
= or for moisture
correction factor for age
proportion by volume of paste
proportion by volume of fine aggregate
proportion by volume of coarse aggregate
A
P
FA
CA
f
β
β
β
=
=
=
=

The CTE value can be very important where differential heating occurs or where a
variety of materials are used. However, little research is published about the coefficient of
thermal expansion of UHPC, with most coming from manufacturers. The Japanese specification
for UHPC design states that the coefficient of thermal expansion can change substantially and
depends on the moisture content. CTE values of 7.5 x 10
-6
/°F are given by the specification for
specimens after thermal treatment (JSCE 2006). Values of approximately 8.3 x 10
-6
/°F have
been document by Graybeal (2006a). These CTE values are slightly higher than those of most
normal strength concretes which are on the range of 4.1 x 10
-6
/°F to 7.3 x 10
-6
/°F (Mindess et al.
2003). However, the French Recommendations for UHPC (AFGC 2002) use CTE values from
European UHPC suppliers of 6.6 x 10
-6
/°F for Ductal
®
and 5.8 x 10
-6
/°F for BSI
®
which are
more in line with typical concretes.
To date, the coefficient of thermal expansion of UHPC has not been tested in the United
States.
29

2.5.4 Additional Durability Research
Despite these findings, the body of research on UHPC durability, especially in the United
States, is limited and incomplete. Variables such as specimen age at time of thermal treatment
and procedure variations (ASTM C 666 - Procedure A vs. Procedure B) have not been
considered. Current research has provided information about the durability properties of UHPC,
but has seen little done in the way of showing that these results can be replicated in other
laboratories or that variability between batches is low. Relatively few statistical analyses of the
test results have been reported due to the small sample sizes tested.
The research reported here will address several of these issues including the age of
thermal treatment, procedure variations, and inter-laboratory repeatability. These findings will
help provide the body of UHPC research in the U.S. with a more comprehensive and thorough
understanding of UHPC and its distinctive material properties. Additionally, areas of potential
future research will be addressed so that continual progress can be made towards providing
stronger, more sustainable, and aesthetically pleasing bridges and concrete structures.
2.6 Other UHPC research
Several state Department’s of Transportation and a few universities have been
investigating this relatively new material, however, the main material property tests have been
conducted by FHWA at their Turner-Fairbank Highway Research Center (TFHRC). The FHWA
material property characterization research was completed in 2006 (Graybeal 2006a) and
provided a broad, yet basic, understanding of how UHPC performs under ASTM and AASHTO
testing methods. Both mechanical and durability properties of a UHPC were investigated. Also,
full scale testing of two 80 ft prestressed AASHTO Type II UHPC girders (Figure 2-4a) was
completed in 2001 (Hartmann and Graybeal 2002). While these girders did not efficiently take
30

advantage of the UHPC properties by minimizing the cross-sectional area of the beams, they
provided valuable information about how UHPC can be cast in U.S. precast facilities. Current
research at the TFHRC is investigating an optimized girder/deck configuration for use on short
span road bridges (Graybeal et al. 2004). Initially, four 33 in. deep, 70 ft span girders with a
double-tee (or π-shaped) cross-section with an integrated 8 ft wide, 3 in. deep UHPC deck were
tested for both long-term and loading effects (Figure 2-4b). However, testing revealed problems
with the girder shape, including transverse loading and fiber distribution, and a new design has
been proposed (Keierleber et al. 2007). Results from the follow-up investigation are pending.
(a)
(b)
Figure 2-4: UHPC Girder Testing (a) Testing of a UHPC I-girder (Hartmann and
Graybeal 2002) (b) UHPC π-girder at the TFHRC (Keierleber et al. 2007)

Using the research from FHWA and collaborating with Iowa State University, the Iowa
Department of Transportation and Lafarge North America constructed the first UHPC bridge in
the United States in Wapello County, Iowa in 2006 (Figure 2-5). The 110 ft simple span Mars
Hill bridge, replacing the 73-year-old dilapidated truss bridge, was comprised of three 110 ft
Modified Iowa Bulb-Tee prestressed UHPC beams (Graybeal 2006b). In addition, no mild steel
31

reinforcement was used (steel use was limited to the steel fibers in the UHPC matrix and to the
prestressing tendons). To monitor the behavior of the beams and bridge, a 2-year performance
monitoring program was implemented.

Figure 2-5: Mars Hill Bridge in Wapello County, Iowa (Lafarge 2006b)

Additional UHPC research at the Virginia DOT and Virginia Polytechnic Institute &
State University characterized some of the structural aspects of UHPC including punching shear
of thin UHPC plates (Harris and Roberts-Wollmann 2005) and horizontal shear between bridge
decks and beams (Banta 2005). Both studies demonstrated that current AASHTO and ACI
design standards could be used for design of UHPC members for punching shear and horizontal
shear.

32











(This page intentionally left blank)

33

3.0 Methodology
3.1 Introduction
The objectives of this chapter are to outline the materials and equipment used to mix, cast
and cure UHPC. The areas investigated were compressive strength (ASTM C39), modulus of
elasticity and Poisson’s ratio (ASTM C 469), and flexural properties (ASTM C 1018), rapid
chloride penetration testing (ASTM C 1202), freeze-thaw cycling (ASTM C 666), and
coefficient of thermal expansion (AASHTO TP 60-00). Some ASTM standards were modified
slightly to facilitate use for UHPC. Modifications are discussed herein. Additionally, studies
were performed for the freeze-thaw and coefficient of thermal expansion tests specimens, to
determine the impact of water absorption on UHPC specimens under testing.
While previous research by Graybeal (2006a) showed a strong intra-batch correlation in
the durability data (i.e. – test results correlated well when specimens were from the same batch),
an inter-batch correlation is desirable to demonstrate repeatability of the batching and curing
process. Therefore, one sample for each test and corresponding test age was often cast in each
batch. For example, only one thermal treatment cured specimen for the 28-day coefficient of
thermal expansion test was cast in Batch A, and Batch B similarly had only one thermally cured
specimen for the 28-day coefficient of thermal expansion test. While this may have introduced
some variability due the fact that there was only one sample of a test age for each test, the
homogeneous nature of UHPC materials and batching processes allowed for accurate
comparisons between specimens from different batches (as shown by low coefficient of variation
results in upcoming sections). In addition, to ensure that any variability present did not greatly
affect the results, three or four compression specimens for various age testing were cast in each
batch to check the variability between each batch. If the compression cylinders did not achieve
34

proper strength, the batch would be considered to have had an error during the mixing or curing
process.
3.2 UHPC Mixing Procedure
The Ductal
®
product used in this project was purchased from Lafarge North. Enough
material to make about three fourths of a cubic yard was supplied in one shipment. The
shipment included thirty-three 80 lbs. bags of premix, five 40 lbs. boxes of steel fibers, and five
gallons of Chryso
®
Fluid Prema 150 superplasticizer. UHPC can be mixed with several different
fiber types, but high carbon/high strength steel fibers were chosen as these were consistent with
other research being conducted using Ductal
®
for precast bridge elements (Kollmorgen 2004;
Graybeal 2005). The straight steel fibers measured 0.008 in. in diameter x 0.5 in. long. Three
bags were damaged in transport and used for preliminary testing to ensure previous procedures
outlined at Michigan Tech were still valid.
Mixing UHPC requires special equipment and procedures to develop consistency in
batching, casting, and curing in a timely fashion. A high shear capacity mixer along with
vibratory table and steam cure chamber capable of maintaining 100% humidity at 194°F is
required (Kollmorgen 2004). The UHPC was batched according to procedures given by a
Lafarge North America Technician and more detailed description of procedures can be found in
Kollmorgen (2004). Mixing UHPC in Benedict Laboratory at Michigan Technological
University used a Doyon BTF-060 planetary mixer (Figure 3-1).
35


Figure 3-1: Doyon Mixer

Batches of up to 0.65 cubic feet (ft
3
) in volume were mixed using the amounts shown in
Table 3.1. This mix design was followed for all batches, except for one batch which was cast
without fibers to determine unreinforced mechanical properties. In that case, the batching
weights were kept consistent but the steel fibers were eliminated.
Table 3.1: Ductal® Mix Proportions for 0.65 ft
3
batch
Constituents 1.0 yd
3
0.65 ft
3

Premix 3700 lb. 96.3 lb.
Water 219 lb. 5.7 lb.
Superplasticizer 51 lb. 1.32 lb.
Steel Fibers 263 lb. 6.84 lb.
w/c ratio 0.20 0.20

UHPC was packaged as a blended premix that included the proprietary proportions of
Portland cement, silica fume, ground quartz, and quartz sand already combined. This premix
was added to the mixer first and then disturbed with dry mixing to break up any clumps in the
36

material. The other constituents (water, superplasticizer, and steel fibers) were added at the
appropriate times during the mixing sequence (Peuse 2008, Misson 2008). The superplasticizer
was added in two stages; half blended with water and added near the beginning of the mix, and
the other half added after the turning point (Table 3.2). The turning point of UHPC was defined
as the time at which all of premix and water, and half of the superplasticizer are completely
mixed so that the UHPC begins clumping together and falling from the sides of the mixing bowl,
Figure 3-2 (Kollmorgen 2004). The turning point was determined by the mixer attaining peak
amperage. After the turning point was reached, the remainder of the superplasticizer was added
followed by the slow addition of steel fibers.
Table 3.2: Typical and Adjusted Mixing Procedures

Typical Mix
Time
Mixer Speed
Mix Time for
Premix age 2
mo.
Mix Time for
Premix age 11
mo.
Start Mixing 0:00 1 0:00 0:00
Addition of Water + ½ Superplasticizer 2:00 1 2:00 4:00
First Speed Increase 4:15 4 4:15 6:15
Second Speed Increase 6:00 5 6:00 8:00
Turning Point 11:00 5 11:00 20:00
Addition of ½ Superplasticizer 12:00 5 12:00 22:00
Addition of Fibers 14:00 3 14:00 24:00
Reduce Speed 16:00 1 16:00 27:00
Stop Mix 18:00 1 18:00 30:00


37


Figure 3-2: Turning Point of UHPC

Several issues with the prescribed procedures developed during the mixing process
including internal temperature and mixing times. During mixing, it is suggested by Lafarge
North-America that the material’s internal temperature remain below 86°F to ensure the longest
pot life of the material. However, several of the batches exceeded this temperature during
mixing and attained mix temperatures up to 95°F. There is no known way of controlling this
internal temperature; rather it is an implicit measure. This was not viewed as a significant
problem however, as consistency of the material rheology and compressive strengths were still
achieved.
Another concern was that mixing time increased for batches made at later times in the
research program. Previous research (Graybeal 2006a) indicated that mix times increased as
premix age increased. As the premix used for this testing aged from two to eleven months, the
corresponding mixing time was also longer. Table 3.2 indicates typical and adjusted timing. The
typical mix procedure called for a total mixing time of 18 minutes and for the time from the
38

addition of water to the turning point to be approximately 8-9 minutes. However, for the older
premix the total mix time increased to nearly 30 minutes and the average time from the addition
of water to the turning point averaged 15-16 minutes.
While specific reasons for the delay in mix time were not investigated, all of the older
premix bags used for specimen batching had 1 in. or larger sized clumps of the premix material
present before mixing. The premix clumping may have inhibited the ability of the water and
superplasticizer to adequately wet all of premix in the shorter mixing time used previously.
Also, as the beginning of the mix procedure was adjusted to break apart these clumps, the time
between the addition of water and the turning point is considered a better comparison than
contrasting total mix times. The increase in mix time may lead to higher water evaporation and
therefore poor rheological properties. However, in spite of these changes, the UHPC material
obtained mix properties, compressive strengths, and durability properties comparable to that of
other research.
After mixing was completed, the rheology of the UHPC mix was tested for consistency.
An adjusted ASTM C 1437 Standard Test Method for Flow of Hydraulic Cement method was
used so that the recommendations outlined in Ductal
®
reference T006 (Operating Procedure –
Flow Test) were followed (20 impacts compared to the ASTM specified 25 impacts). UHPC was
first placed on the impact table in a short steel cone which was then lifted off slowly to allow the
concrete to flow evenly about the table. Four measurements of the diameter of the flow were
taken at equally spaced locations (Figure 3-3), and then the impact table was dropped 0.5 in.,
twenty times. The same four diameter dimensions of the UHPC flow were measured again and a
domain classification of the mix assigned according to Table 3.3. All of the batches mixed were
39

classified under Domain B, so all specimens were cast according to Ductal
®
reference T002
Cylinder and Prism Preparation for Domain B mixes.

Figure 3-3: Impact Table Measurement of UHPC’s Flow

Table 3.3: Flow Domain Classifications of Freshly Mixed UHPC
Domain A Domain B Domain C
Stiff Fluid Highly Fluid
Average Flow Measurements
after 20 blows
< 200 mm 200 mm - 250 mm > 250 mm

3.3 Casting Specimens
Cylinders and beams were cast on a vibrating table capable of reproducing the specified
0.020 in. amplitude set by Ductal
®
reference T006. Cylinder molds were held on the table and
filled in two equal lifts. Beam molds were placed on the table and filled from one end, and the
mix was allowed to flow to the opposite end to fill the mold. After the molds were filled, they
remained on the vibrating table for an additional 30 seconds to consolidate the mix. Upon
removal from the table, the cylinders were sealed with a fitting cap, and an acrylic plastic top
sealed with weather-strip covered the beam molds. These sealed tops prevented moisture loss
40

from the fresh specimens. The specimens were placed on a bench in the casting room and
allowed to remain at ambient conditions for three days until being demolded.
During batching, mineral oil was used to prevent the concrete specimens from sticking to
the steel prism molds as well as the plastic cylinder molds. In the plastic cylinder molds,
however, overuse of mineral oil resulted in more air bubbles appearing on the sides of
specimens. Moreover, when no mineral oil was used to lubricate the forms, the sides of de-
molded cylinders had a smooth, almost glassy, finish. Specimens with both types of surface
conditions were tested, yet the phenomenon appeared to have had no effect on the material
properties analyzed.
3.4 Curing Regimes
The curing regime was a primary variable in this study. Specimens were either air cured
at ambient laboratory conditions (Air), or thermally treated (TT). Air-cured specimens for
testing at 3-days were demolded two to three hours before testing to allow ample time to prepare
and test the three day specimens in the allotted time by ASTM C 39 which was 3 days ± 2 hours.
Air-cured specimens tested at later ages were also demolded at this time and placed on a
laboratory shelf in ambient conditions until being tested.
The TT-cured specimens were subjected to a 48-hour, 100% humidity, steam treatment at
194°F upon demolding at 3-days. The thermal treatment began with a 6-hour ramp up period to
194°F and 100% humidity followed by a 48-hour hold at the elevated temperature and relative
humidity. At the end of this time, the environment was allowed to ramp down to lab conditions
over 6 hours by opening the outer lid of the cure chamber (Figure 3-4). All specimens were
removed after the 60 hour cure process and allowed to return to ambient temperature before end
grinding or testing occurred. Specimens destine for delayed thermal treatment (DTT) or doubly-
41

delayed thermal treatment (DDTT) were demolded at 3-days and placed on a laboratory shelf in
ambient conditions until starting the thermal curing process at days 10 and 24, respectively.

Figure 3-4: Michigan Tech’s UHPC Thermal Treatment Cure Chamber

3.5 Specimen Preparation and Test Procedures
ASTM and AASHTO standards were used as a baseline for investigating the properties
of UHPC. However, several sections of the procedures were adjusted in an effort to maintain the
integrity of the curing practices. Also, due to the unique nature of UHPC, a clear understanding
of how the specimens were handled and tested is necessary to properly interpret and apply the
results. Therefore, it is essential that the materials, equipment, specimen preparation, and testing
procedures are outlined so that these tests may be reproduced and properly understood.
Recommended specimen preparation and testing procedures for UHPC are summarized below
for each experimental test conducted.
42

3.5.1 Compressive Strength
Specimen Preparation – Compressive Strength
The traditional methods of preparing the ends of cylinders for compressive testing by
using unbonded caps (neoprene pads) or sulfur caps greatly exceeded the allowable stress of 12.0
ksi stated ASTM C 1231 and C 617, respectively, and therefore could not be used for testing
UHPC. Currently, the best alternative is to grind the ends of the cylinders using a surface
grinder. Figure 3-5 shows the Reid surface grinder used to prepare specimens for this research.
Specimens were prepared one day before their scheduled testing time with the exception of the
three day Air-cured cylinders as they were demolded and end-ground a few hours before testing.

Figure 3-5: Reid Surface Grinder

Cylinders were held in place using a v-shaped jig which kept the specimens
perpendicular to the grinding wheel. After several passes of the wheel, the ends were checked
for perpendicularity using the procedure set forth by Ductal
®
reference T009 Operating
Procedure Cylinder End Preparation. The cylinder end planeness limit of 1° corresponds to the
limit set by ASTM C 39 which states that neither end of the cylinder can depart from
43

perpendicularity by more than 0.5°. End perpendicularity was measured using a digital dial gage
as seen in Figure 3-6 utilizing the referenced procedure (Peuse 2008). All cylinders were ground
within the 1° tolerance with the initial grinding.


Figure 3-6: End Perpendicularity Set-up

Specimen Testing – Compressive Strength
ASTM C 39 – Standard Test Method for Compressive Strength of Cylindrical Concrete
Specimens was used as the baseline for compression testing. However, testing a 6 x 12 in.
UHPC cylinder after thermal treatment would potentially require a compression machine to have
a capacity of approximately 800 kips. A machine of this size is not typically used in the precast
industry or testing laboratories. Research by Kollmorgen (2004) and Graybeal (2005) showed
that the size effect of 3 x 6, 4 x 8 or 6 x 12 in. cylindrical specimens was not statistically
significant regarding the compressive strength. Hence, the smaller 3 x 6 in. cylinder was
44

chosen, affording a test machine capacity of about 250 kips which is readily available in testing
labs and at precast plants in the United States. Tests performed on 3 x 6 in. cylinders were
conducted on a Baldwin CT 300 (Figure 3-7). This hydraulic load frame has a capacity of 300
kips and was operated by manual controls. Data was externally collected by DASYLab Version
8.0 (2004).

Figure 3-7: Baldwin CT 300 Compression Testing Machine

A slight modification to ASTM C 39 was made to make the testing of UHPC more
practical, namely the increase of the load rate applied to the specimen. The current standard sets
the load rate at 35 ± 7 psi per second which would dictate that a specimen of UHPC could take
up to 15 minutes to break. This lengthy time period would be unacceptable for the time required
to break specimens for production use. Ductal
®
reference T001-Operating Procedure
Compressive Test suggests a load rate of 150 psi per second which reduces the length of a test to
approximately three minutes and also falls within the allotted time requirement of ASTM C 39.
Additionally, Graybeal and Hartmann (2003) conducted research showing there was no
45

significant difference in UHPC compressive strength when changing the load rate from the
ASTM standard to the recommended UHPC load rate. For this reason, and the fact that previous
research completed at Michigan Tech used the manufacturer prescribed load rate, a load rate for
compression testing of 150 psi per second was used. Load data was recorded at 5 Hz until
ultimate failure of the specimen.
3.5.2 Modulus of Elasticity and Poisson’s Ratio
Specimen Preparation – E
c
and ν
The modulus of elasticity and Poisson’s ratio were conducted on 3 x 6 in. cylindrical
specimens. Specimen ends were prepared as described for compression testing one day before
their scheduled testing time with the exception of the three day Air-cured cylinders as they were
demolded and end-ground a few hours before testing. All specimens remained in the laboratory
at ambient conditions after demolding until the time of testing, with the exception of the time
that thermal treatment was applied.
Specimen Testing – E
c
and ν
The testing process followed ASTM C 469 – Standard Test Method for Static Modulus of
Elasticity and Poisson’s Ratio of Concrete in Compression, except the load rate was increased to
150 psi per second as mentioned for the compression testing. Additionally, companion
compression cylinders were not tested to determine the maximum applied load (0.40
c
f ` ), rather
a background study determined the appropriate maximum load level as 40 percent of the average
compressive strength based on age and curing regime of cylinders made during trial batches of
UHPC. This variation from breaking a companion specimen was permitted because of the high
predictability and low coefficient of variation (COV) of the compressive strength results based
on previous work conducted by Kollmorgen (2004) and Graybeal (2005).
46

A combined unbonded compressometer and extensometer equipped with two digital
indicators, which measured the transverse and tangential displacements. The set up is shown in
Figure 3-8. Load, transverse and tangential displacement, and time were recorded by DASYLab
Version 8.0 at a rate of 5 Hz. The specimens were loaded 150 psi per second until the
predetermined maximum load based on curing regime and age. The specimens were completely
unloaded at approximately the same rate and the gauges zeroed. This process occurred three
times for each specimen, following the ASTM procedure. The initial loading was used to seat
the gauges and the data disregarded. Data from the second and third loading was averaged and
reported as the results for the specimen.

Figure 3-8: Compressometer and Extensometer

3.5.3 Flexural Strength
Specimen Preparation – Flexure
Testing was conducted on 2 x 2 x 11.25 in. beam specimens, and not conducted on the
preferred size of 4 x 4 x 14 in. as suggested by ASTM C 1018. This smaller prism was used for
47

testing because, as Graybeal (2005) pointed out, UHPC elements subject to flexural forces are
not likely to be 4 in. thick, and it was desired to compare the results of this study with Graybeal’s
(2005) and Kollmorgen’s (2004) results. Additionally, the ASTM specifies that the cross section
need only be three times the fiber length which was satisfied with the 2 x 2 in. cross section. No
additional preparation was performed and all specimens remained in the laboratory at ambient
conditions after demolding until the time of testing, with the exception of the thermal treatment
duration.
Specimen Testing – Flexure
ASTM C 1018 Standard Test Method for Flexural Toughness and First-Crack Strength
of Fiber-Reinforced Concrete (Using a Beam with Third Point Loading) was used to determine
the first-crack strength and flexural toughness of Ductal
®
. This test consists of loading a small
prism at the third points, to create a constant moment region, and recording the load and
deflection so the data can be analyzed to give the flexural cracking stress, toughness, and
approximate flexural strength of the fiber reinforced concrete. Testing of beam specimens was
conducted on a 55 kip MTS load frame with Test Star II controls and Test Ware data acquisition.
When conducting this test, special attention needs to be given to the requirements of
ASTM C 1018 as it specifically calls for a testing configuration which can maintain the net
midspan deflection using a closed-loop servo-controlled testing machine. Controlling the rate of
deflection is the key to this test because if the test was conducted under load control, the post-
crack portion of the curve would not be suitable for analysis. The loading apparatus must allow
the net midspan deflection to be recorded and to control the rate of deflection. The ASTM
outlines two possible configurations, and the method used in this research uses a rectangular jig
to hold LVDTs that are averaged to determine net deflections as shown in Figure 3-9.
48


Figure 3-9: ASTM 1018 Loading Configuration

This configuration loaded the specimens at the third points of the span and created a
simple support condition as outlined in ASTM C 78 Standard Test Method for Flexural Strength
of Concrete (Using Simple Beam with Third-Point Loading) where the specification for the
loading apparatus is given. The deflection measuring jig was secured to the prism at the neutral
axis of the prism, directly above the support points. This allowed the jig to remain stationary as
the prism deflected. Two LVDT’s with strokes of ± 0.20 in. were used to measure the midspan
deflection from each side of the prism relative to the bar which was epoxied to the top surface of
the prism. The two LVDT’s signals were electronically averaged before being read by the
testing software which controlled the closed-loop control system. The midspan deflection rate
was chosen to be 0.003 in. per minute because ASTM C 1018 specifies that the first crack
deflection be reached in 30 to 60 seconds.
3.5.4 Rapid Chloride Penetration
For the rapid chloride penetration testing, ASTM C 1202 – Electrical Indication of
Concrete’s Ability to Resist Chloride Ion Penetration was followed for both the specimen
49

preparation and testing. The testing standard requires that specimens be vacuum saturated with
water and tested for electrical conductance. The electrical conductance is measured by applying
a 60-volt potential across a specimen that is mounted to a test cell with sodium chloride and
sodium hydroxide solutions for 6-hours. During that time the current passing is recorded, and
used for calculating the total charge in coulombs passing.
Specimen Preparation – RCPT
Preparation consisted of vacuum treating and impregnating the UHPC specimens with
water prior to the actual testing for chloride ion penetration. Despite UHPC’s highly
impermeable nature and the likelihood that it was never fully saturated through the specimen
depth, ASTM C 1202 was still followed to allow for a qualitative comparison between UHPC
and other concretes.
The specimens were initially cast in a 4 in. diameter by 3 in. high cylinder and, therefore,
two-days prior to testing the specimens were cut down to the required 2 in. height using a
kerosene cooled saw. The kerosene remaining on the surface of the UHPC specimens was then
evaporated off in an oven at 122°F for approximately two hours and then cooled to room
temperature over another two hours. A water cooled saw was specified in the ASTM method,
however, due to UHPC’s low permeability most of the kerosene was believed to be on the
surface and therefore evaporated in the oven drying process. After the specimens had cooled,
they were sealed on their side surface using Enviro-Tex self leveling epoxy mix and allowed to
cure for at least 12 hours (Figure 3-10). Specimens were then placed in a plastic container
inside a vacuum desiccator, and a vacuum was applied over the specimen for a total of four hours
(Figure 3-11). At three hours into the vacuuming process, de-airated water was introduced to
cover the specimen while the vacuum was maintained for the final hour. After this time, the
50

specimens were removed from the chamber and allowed to sit with the water covering them for
18 hours (± 2 hours). Water was then drained and the container sealed to maintain an
environment at 95% humidity until testing. To provide proper time for preparation, all of the
specimens were cut and epoxy coated on day 26 and vacuum treated on day 27, such that testing
could begin on 28-day old specimens.
Figure 3-10: Epoxy-coated UHPC Specimens
for RCPT
Figure 3-11: ASTM C 1202
Specimen Preparation Setup
Specimen Testing – RCPT
After the specimen preparation was completed, the prepared UHPC samples were then
tested following the ASTM C1202 procedures. A 60-volt potential was induced across the test
cells and specimen by a Kepco power supply (Figure 3-12). For 6-hours the current passing
through the specimen was monitored and recorded by an Agilent 34970A Data Acquisition unit,
and later would be used to calculate the total charge passing in coulombs. To replicate field
situations where the exposed surface of a UHPC beam would be smooth and finished, the bottom
cylinder surface (un-cut and finished) of the sample was placed facing the 3% sodium chloride
(NaCl) solution, while the saw cut side of the sample was placed facing the 0.3 N (0.3 normality)
sodium hydroxide (NaOH) solution.
51


Figure 3-12: UHPC Specimen Undergoing ASTM C 1202 Testing

Care was taken to ensure that the test cells did not leak and that no short circuit was
created due to the steel fiber reinforcement. If a short circuit was created, the specimen could
overheat leading to erroneous results and the boiling off of the solutions. While initially the
creation of a “short” circuit was a concern, no short circuiting occurred because of the random
distribution and the short nature of the steel fibers. Occasionally, when more than one specimen
was tested in a day, two cells were clamped together and a 60-volt potential was applied across
both specimens. The data acquisition unit recorded the current passing for each individual
specimen.
At the conclusion of the test, the data was compiled and a total charge passing was
calculated by integrating the area under the current (amperes) versus time (seconds) curve. The
integration was performed by using the trapezoidal rule (as suggested by the ASTM standard)
and the current passing at 30-minute intervals (Equation 3.1). However, the total charge passed
was based on a 3.75” diameter specimen according to ASTM standards and, therefore, Equation
52

3.2 was used to adjust the value obtained from Equation 3.1. Using this adjusted total charge
passed, Table 3.4 was then used to evaluate the chloride ion penetrability of the UHPC specimen.

0 30 60 300 330 360
900( 2 2 .... 2 2 ) Q I I I I I I = + + + + + + Equation 3.1
0
where:
charge passed (coulombs)
current (amperes) immediately after voltage is applied
current (amperes) at min after voltage is applied
t
Q
I
I t
=
=
=


2
3.75
s x
Q Q
x
⎛ ⎞
= ×
⎜ ⎟
⎝ ⎠
Equation 3.2
where:
charge passed (coulombs) through a 3.75-in.
diameter specimen
charge passed (coulombs) through a in. diameter specimen
diameter (in.) of the nonstandard specimen
s
x
Q
Q x
x
=
=
=



Table 3.4: Chloride Ion Penetrability Based on Charge Passed (ASTM C 1202)
Charge Passed
(coulombs)
Chloride Ion Penetrability
> 4,000 High
2,000 - 4,000 Moderate
1,000 - 2,000 Low
100 - 1,000 Very Low
< 100 Negligible

53

3.5.5 Freeze-Thaw Cyclic Testing
Freezing and thawing testing was performed in accordance with ASTM C 666 –
Resistance of Concrete to Rapid Freezing and Thawing standards. Eight cycles were completed
per day by means of freezing in air and thawing in water. Approximately every 32 cycles the
fundamental transverse frequency, length change, and mass were observed and recorded.
Specimens were tested until failure or 300 freeze-thaw cycles, whichever came first. Only one
minor variation was adapted for the testing of UHPC specimens as noted below.
Specimen Preparation – Freeze/Thaw
The freeze-thaw specimens were cast in 3 x 4 x 15.5 in. beam molds with 1-1/4 in.
stainless steel gauge studs embedded 1 in. into the ends of the specimen, leaving 1/4 in. of the
gauge studs exposed on each end. This produced specimens with a nominal length of 16 in.
(length between the two exposed ends of the gauge studs) and a gauge length of 13.5 in. (length
between the embedded ends of the two gauge studs). After demolding and curing for 27-days
and prior to testing, the UHPC specimens were cooled to the thaw temperature of the freeze-thaw
test machine by placing them in a 41°F water bath for at least 16 hours. Soaking of the
specimens according to ASTM C 666 in a lime bath for 48-hours was not performed to avoid
impacting the curing regimes. This may have allowed some UHPC freeze-thaw specimens to be
inadequately saturated. However, due to the high impermeability of UHPC, it is unlikely that
any of the specimens would have become fully saturated in 48-hours.
Specimen Testing – Freeze/Thaw
There are two methods specified in ASTM C 666 for determining the freeze-thaw
durability of a concrete – Procedure A and Procedure B. In Procedure A, the specimens are
54

frozen and thawed in water, while in Procedure B the specimens are frozen in air and thawed in
water. The current freeze-thaw testing machine at Michigan Tech is designed for Procedure B.
Testing for freeze-thaw durability was performed in an 80-specimen Scientemp freeze-
thaw chamber (Figure 3-13). As a chamber of this size contained too many slots to be
completely occupied by UHPC specimens during testing, a large number of NSC “dummy”
specimens were made to maintain a full-load and proper heating/cooling rates in the machine.
Furthermore, three UHPC control specimens were cast with type-T thermal couples embedded to
monitor and control the temperature in the chamber. Prior to using any specimen slot, these
control specimens were rotated to ensure each specimen slot’s freeze-thaw temperature cycle
conformity to the ASTM C 666 requirements.

Figure 3-13: MTU 80-specimen Freeze-Thaw Chamber (Procedure B)

During testing, the specimens’ fundamental transverse frequency, mass, and length
change were evaluated at regular intervals. The fundamental transverse frequency was measured
according to ASTM C 215 – Fundamental Transverse, Longitudinal, and Torsional Frequencies
of Concrete Specimens and involved striking the end of the UHPC specimen with a precision
55

weighted steel impact hammer and measuring the dynamic response of the specimen with an
accelerometer. A 2 in. thick Styrofoam pad was used to dampen any outside frequency
interference, and a fast-Fourier transform method in DASYLab (ver. 8.0 2004) calculated the
specimen’s fundamental transverse frequency (Figure 3-14). An average fundamental transverse
frequency of the specimen was determined from three strikes with the impact hammer. Using
this average fundamental transverse frequency, the relative dynamic modulus (RDM) of the
UHPC specimen was calculated as the ratio between the fundamental transverse frequency of a
specimen after n cycles and the fundamental transverse frequency of the specimen immediately
prior to testing. The specimen’s mass was also recorded to the nearest 0.01 lbs using a calibrated
scale.

Figure 3-14: Testing the Fundamental Transverse Frequency of an UHPC Specimen

While defining failure of a freeze-thaw specimen based on a length increase of 0.10% is
one option according to ASTM C 666, the 2001 Michigan Test Methods (MTM 115) specifically
Impact hammer
Styrofoam
Accelerometer
56

use this length change parameter as the test for specimen failure. Therefore, the length change of
each specimen was measured using equipment specified in ASTM C 490 – Use of Apparatus for
the Determination of Length Change of Hardened Cement Paste, Mortar and Concrete.
Specimens were measured horizontally by holding the specimen against the base of the length
comparator measuring the relative length on the dial gauge (Figure 3-15). The horizontal
measuring method was employed due to Note 1 in ASTM C 490, which indicates the need to use
a horizontal comparator to measure specimens with cross sections larger than 9 in.
2
. However,
because the length measurement was only used for a relative comparison between specimens a
vertical measurement may also be sufficient. A standard rod was used to zero the length
comparator before each test to guarantee that each measurement was made relative to a standard
length.

Figure 3-15: Length Change Measurement of an UHPC Freeze-Thaw Specimen

One freeze-thaw cycle was completed every 3-hours and the first measurements of initial
RDM, length, and mass were taken at 24-hours (8 cycles) after the specimens began cycling.
Repeat measurements were then taken at 96 hour intervals (32-cycles) until testing was complete
(300 cycles or specimen failure, which ever occurred first). Specimens were removed during the
thaw cycles at approximately 40°F, so that accurate length comparisons could be made.
57

Graybeal (2006a) observed that untreated UHPC specimens actually experienced as much
as a 10% increase in their relative dynamic modulus and a 0.2% mass gain after 300 cycles.
Therefore, two additional beams were cast in each batch (six in total) to investigate the impacts
of water absorption and additional hydration of unhydrated cement particles on the RDM and
mass of the UHPC specimens in the freeze-thaw chamber. To simulate the wetting and drying of
the specimens undergoing Procedure B freezing and thawing cycles without temperature change,
these extra “side-study” specimens were cycled every 24 hours between air and water in a water
bath kept at ambient temperature.
3.5.6 Coefficient of Thermal Expansion
The coefficient of thermal expansion for concrete was determined following a modified
version of AASHTO TP-60-00 – Coefficient of Thermal Expansion of Hydraulic Cement
Concrete (2004). Modifications to both specimen preparation and testing procedures were made
due to equipment specifications and in the interest of maintaining the integrity of UHPC curing
regimes. The standard process according to AASHTO TP-60-00 requires that specimens be
saturated with water prior to testing, placed into a test frame and submerged in a thermally
regulated water bath, and then subjected to heating and cooling cycles until an accurate CTE
measurement is obtained. Detailed changes to specimen preparation and testing for UHPC are
described below.
Specimen Preparation – CTE
Typically, specimen preparation for CTE testing involves saturating the concrete
specimen in a limewater bath for a minimum of 48-hours prior or until a mass change of less
than 0.5% is observed (whichever comes last). However, due to the extremely low water-cement
ratio in UHPC, some cement particles do not become fully hydrated even after the curing regime
58

is applied. Because the curing regime is one of the variables being investigated through this
research, it is important that the UHPC specimens do not absorb water that can react with
unhydrated cementitious particles. Therefore, the standard preparation of submerging specimens
in a limewater bath for 48-hours prior to testing was not employed. Instead, specimens were
preserved in the unsaturated condition by coating them with an epoxy sealant.
After curing for 26 days under the specified curing regime, the 4 x 8 in. cylinder
specimens were cut to size (4 x 7 in. cylinder) by removing the top 1 in. of the specimen using a
kerosene saw. The specimens were oven-dried at 122°F and then measured for perpendicularity.
To measure perpendicularity, 8 marks were made around the circumference of the cut side of the
cylinder and a 12 in. (±0.0005 in.) micrometer was used to measure the specimen length at these
8 points. The bottoms of the specimens were not cut due to the already smoothly formed surface
from the base of the mold. Nevertheless, some problems with perpendicularity arose on some
specimens. However, because the bottom of the cylinder rested on three steel support buttons in
the frame and was not in contact with the linear variable differential transformer (LVDT), it was
not considered harmful to the length change measurements.
Following the perpendicularity measurements, the UHPC specimens were completely
coated with epoxy (Figure 3-16) except for the locations on each end where the steel support
buttons of the test frame and the LVDT tip came into contact with the specimen. This allowed
for direct contact between the specimen and LVDT frame support buttons as well as the LVDT
transducer. After the epoxy application was complete, the epoxy was allowed to cure for at least
12 hours.
59


Figure 3-16: Epoxy-coating CTE Specimen

Specimen Testing – CTE
UHPC CTE testing utilized Pine Instrument Company’s AFCT1A Coefficient of Thermal
Expansion of Portland Cement Concrete Measurement System. The components of this system
included a 25-gallon water bath with a temperature stability of ±0.2°F over a temperature range
of 41°F to 160°F (Figure 3-17). The specimen test frame was calibrated in accordance with
manufacturer’s instructions and utilized a 1 in. stroke linear variable differential transformer with
a resolution of 0.0002 in. All of the equipment and programming was designed in agreement
with the FHWA Procurement Specification (Appendix B). While the majority of the FHWA
Specification follows AASHTO TP 60-00, some differences exist. The main variation was the
measurement tolerance of the specimens that was stipulated by the FHWA specification to be
0.00005 in. rather than the 0.00001 in. as specified by the AASHTO TP 60-00 specification.
Also, when determining if a specimen had reached thermal equilibrium, measurements were
taken every 5 minutes over a half-hour period for the FHWA procedure rather than every 10
minutes as noted by AASHTO. Additionally, the specimen measuring frame used a correction
factor procedure different from that outlined in AASHTO TP 60-00 Appendix X.2. The
60

correction factor procedure in the procurement specification provides for a two-point calibration
that allows a greater range of specimen lengths to be tested in a given frame (Pine 2006).
Overall, these changes were necessary for testing using the equipment available.

Figure 3-17: Pine CTE Specimen Test Frame and Water Bath

After a specimen had been prepared, mass and length were recorded prior to placing it in
the water bath (Figure 3-18) for testing. The CTE values of UHPC were obtained by taking the
average CTE of the specimen during a heating segment, 50°F to 122°F, and a cooling segment,
122°F to 50°F. One test cycle (one heating segment and one cooling segment) lasted
approximately 24 hours. However, several specimens required longer than one test cycle to meet
the AASHTO specification for successive CTE values (difference of no more than 0.5 micro
strain/°F between successive CTE values). These tests were continued until the AASHTO
specification was attained.
61


Figure 3-18: UHPC Specimen in Water Bath Undergoing CTE Testing


62











(This page intentionally left blank)

63

4.0 Results and Discussion
Several properties of UHPC were investigated during this research project. Compressive
strength, modulus of elasticity, Poisson’s ratio, flexural characteristics, rapid chloride
permeability (RCPT), freeze-thaw resistance (F/T), and coefficient of thermal expansion (CTE)
of an ultra-high performance concrete were tested with differing curing conditions and at
different ages to examine how these factors influence each of the properties. Table 4.1 lists the
total number of specimens tested for each UHPC property studied under several curing regimes.
Table 4.1: Experimental Test Matrix - Specimens Tested per Curing Regime
Curing
Regime


Specimen
Age at
Testing
(days)
Number of Specimens Tested
Compressive
Strength
Modulus
of
Elasticity
Poisson’s
Ratio
1
st

Flexure
Cracking RCPT F/T CTE
Air-cured 3 6 4 4 - - - 3
7 6 6 6 - - - 4
14 6 6 6 - - - 3
28 6 6 6 12 4 4 3
28 – NF

6 6 5 3 - - -
TT (Thermal
Treatment)
7 6 6 6 - 3 - 3
14 6 6 6 - - - -
28 6 6 6 12 4 4 3
28 - NF 5 5 5 3 - - -
DTT (delayed
TT)
14 6 6 6 - - - -
28 6 6 6 12 - - -
28 - NF 5 5 5 3 - - -
DDTT 28 5 5 5 - - - -

Curing Regime summary:
Air-cured = ambient lab conditions 72°F, 30-50% relative humidity
TT = Thermal treatment of 194°F, 100%RH for 48 hours beginning at age 3-days, ambient cure otherwise
DTT = Delayed thermal treatment beginning at age 10-days
DDTT = Doubly-delayed thermal treatment beginning at age 24-days

N.F. = refers to UHPC specimens cast without fibers

The resulting data from these tests was recorded, synthesized, and analyzed and the
summary data presented herein (additional data can be found in Appendix A). Calculating
sample mean values revealed the central tendencies of the data, while the standard deviation and
64

coefficient of variation (COV) were used to assess dispersion tendencies. The COV is the
standard deviation divided by the mean which shows the relative variability of a sample set.
A statistical analysis was conducted to compare the effects thermal treatment on the
properties of UHPC. Many options exist for testing several sample sets of data at the same time,
for example an analysis of variance (ANOVA). However, this test can only tell if one or more
sample group is statistically different, and does not locate which sample is different. Because the
focus of the study was to determine how the different curing regimes performed compared to
each other, the ANOVA was not an acceptable approach. When comparing data from two
populations (e.g. – comparing TT-cured UHPC freeze-thaw specimens to Air-cured freeze-thaw
specimens) a series of statistical F-tests and t-tests were performed to test the null and alternate
hypotheses on the sample sets from each curing regime and age, as well as each property
conducted. Complete details of the hypothesis testing and results can be found in Peuse (2008)
and Misson (2008). A summary of those results are provided here. Test results are also compared
to results in currently published literature.
4.1 Compression Strength
A total of 75 3x6 in. cylindrical specimens were tested for compressive strength. Table
4.2 shows the mean and COV for the compressive testing based on curing regime and age. Test
data recorded for each specimen is tabulated in Appendix A (Tables A.1, A.2, A.3).
4.1.1 Results
Table 4.2 summarizes the compressive stress test results for various curing regimes and
ages tested.
65

Table 4.2: Compressive Stress Test Results
Curing
Regime
Specimen
Age at
Testing
(days)
Number of
Specimens
(Sample
Size)
Sample
Mean (ksi)
Sample
COV (%)
Air-cured 3 6 14.4 3.8
7 6 19.9 1.8
14 6 22.3 3.2
28 6 23.9 2.2
28 - NF 6 24.9 3.0
TT (Thermal
Treatment)
7 6 30.3 2.9
14 6 30.1 4.6
28 6 31.1 1.3
28 - NF 5 31.9 6.2
DTT (delayed
TT)
14 6 29.7 3.5
28 6 29.9 2.2
28 - NF 3 31.6 3.7
DDTT 28 5 29.4 3.2

The data shows that the COV for all compressive specimens was very low and is
consistent with the COV shown in ASTM C 39. ASTM C 39 reports an expected within-test
COV of 2.4 percent for samples prepared from the same sample of concrete and tested at the
same age. These results were based on over 1200 test reports on 6 x 12 in. cylinders with
compressive strengths of 2000 psi to 8000 psi. Obviously the strength of UHPC is outside the
tested range and specimens tested came from different batches, but in general the COV for the
UHPC is similar to the expected value in ASTM C 39. The referenced standard also gives an
acceptance range of 7.8 percent for 3 individual cylinders which is based on two times the
standard deviation and thereby correlates to a COV of 3.9 percent in which all but two curing
regimes and ages fall under. In addition, Peuse (2008) compared intra-batch results as well as
inter-batch results to find no significant error was introduced by using one specimen per batch.
Figure 4-1 is a graphical representation of the mean compressive stress results for each
age and curing regime. The error bars plotted above and below the mean values show ± one
standard deviation for the respective specimen ages and curing regimes.
66


Figure 4-1: Mean Compressive Results for All Ages and Curing Regimes

It should be noted that a typical specimen failure was characterized as a shear failure by
ASTM C 39. The failure plane extended from the top corner to the opposite bottom corner and
the two pieces of the specimen experienced fiber pullout and fiber breakage. Specimens tested
without fibers were very explosive, leaving most of the specimens in small pieces.
4.1.2 Statistical Analysis and Discussion
The main focus of this research was determining the impact that age at thermal treatment
had on several properties of an ultra-high performance concrete, including after thermal
treatment was complete. For example, is the compressive strength of a cylinder which
underwent thermal treatment from 3-6 days the same on day 7 as it is on days 14 and 28? The
compressive stress of 3 x 6 in. cylinders under the four curing regimes of air-curing, thermal
treatment, delayed thermal treatment, and double delayed thermal treatment at various ages were
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
3 Day 7 Day 14 Day 28 Day 28 Day
No Fibers
M
e
a
n

C
o
m
p
r
e
s
s
i
v
e

S
t
r
e
s
s

(
k
s
i
)
Age of Specimens
Air
TT
DTT
DDTT
67

compared first within each curing regime and then later to each other to analyze the effect each
curing regime has on compressive strength.
Hypothesis testing for pairings within curing regimes were initially run through an F-test
to determine if the population variance for the two curing regimes were equal. Pairings were
then compared using a t-test with the null hypothesis (Ho) equating the population means and the
alternate hypothesis suggesting that they were unequal. Details of the statistical testing can be
found in Peuse (2008). Table 4.3 summarizes the statistical results from hypothesis testing for
comparison within each curing regime. Air-cured specimens continued to cure over the 28 day
duration and all pairings were found to be not equal. Section 4.1.3 discusses strength growth
with time. Air-cured specimens represent all of the thermally treated specimens up to the time of
thermal curing, e.g. TT specimens were not tested at an age of 3 days because thermal treatment
had not yet begun. The DDTT thermal curing occurred during days 24-27 which meant that only
28 day specimens were tested for this curing regime and no hypothesis testing was possible.
68

Table 4.3: Statistical Results for Compressive Strength Testing
Age of Specimen
Pairings (days)
Reject or Failed to Reject the Null
Hypothesis (t-test) Population Mean

Air-cured Specimens
3 vs 7 Reject Ho Not Equal
3 vs 14 Reject Ho Not Equal
3 vs 28 Reject Ho Not Equal
7 vs 14 Reject Ho Not Equal
7 vs 28 Reject Ho Not Equal
14 vs 28 Reject Ho Not Equal

TT – Thermal Treatment
7 vs 14 Failed to Reject Ho Equal
7 vs 28 Failed to Reject Ho Equal
14 vs 28 Failed to Reject Ho Equal

DTT – Delayed Thermal Treatment
14 vs 28 Failed to Reject Ho Equal

The purpose of conducting the hypothesis testing within the curing regimes was to
determine if the data had equal population variance and the same population mean. If all the age
groups (i.e. 7, 14, and 28 day) from a particular curing regime had the same population mean,
then all the data could be put together to make a larger sample set for each curing regime. These
larger sample sets could then be used to compare the curing regimes to one another. The results
showed that the Air-cured specimens did not have the same population mean, and therefore
could not be combined to form a larger sample set. Specimens tested at 28 days were chosen to
represent the Air curing regime because the sample set had the highest compressive strength.
Specimens undergoing TT and DTT did have the same population mean in their respective data
set. Therefore, data from each age group of TT and DTT were combined into two data sets, one
for TT and one for DTT. Table 4.4 shows the organization for the combined hypothesis testing.
69

Table 4.4: Combined Compressive Stress Results
Curing
Regime
Specimen
Age at
Testing
(days) Sample Size
Combined Sample
Mean Compressive
Stress (ksi)
Combined Sample
COV (%)
Air 28 6 23.9 2.2
TT 7, 14, 28 18 30.5 3.3
DTT 14, 28 11 29.8 2.6
DDTT 28 5 29.4 3.2

As with the individual curing regimes, hypothesis testing all combinations were initially
run through an F-test to determine if the population variance for the two curing regimes were
equal. The same combinations underwent hypothesis testing using a t-test with the null
hypothesis equating the population means and the alternate hypothesis suggesting that they were
unequal. For tests which include the Air curing regime, the null hypothesis was rejected and the
population means were not equal. However, the remaining pairings showed that the null
hypothesis was not to be rejected and that all three thermal treatment curing regimes shared the
same population mean. Table 4.5 is a summary of the results of the hypothesis testing based on
curing regime.
Table 4.5: Statistical Results for Combined Compressive Strength Testing
Curing Regime
Pairings
Reject or Failed to Reject the Null
Hypothesis (t-test) Population Mean
Air vs TT Reject Ho Not Equal
Air vs DTT Reject Ho Not Equal
Air vs DDTT Reject Ho Not Equal
TT vs DTT Failed to Reject Ho Equal
TT vs DDTT Failed to Reject Ho Equal
DTT vs DDTT Failed to Reject Ho Equal

Hypothesis testing indicated that specimens which had undergone thermal treatment,
regardless of when, had the same population mean. Therefore, the mean compressive stress of
70

specimens undergoing thermal treatment independent of age is 30.1 ksi, which is an increase of
25 percent over the mean of Air-cured specimens at 28 days which is 23.9 ksi. The valuable
conclusion of this statistical analysis is that for compressive strength measured after thermal
treatment has been applied, there is not a great difference in when thermal treatment occurs.
This would have a large benefit to the precasting industry if elements could be cast individually
on different days but cured together at some time in the future.
Previous work completed by Kollmorgen (2004) and Graybeal (2005) showed similar
compressive strengths and COV on 3 x 6 in. cylinders. Kollmorgen (2004) reported an average
compressive strength prior to thermal treatment (3 day air-cured) which was 40 percent lower
than results seen in this research and had a COV of 15 percent. However, the average
compressive strength of thermally treated specimens (7 day, 14 day, and 28 day) ranged from
only 3.3 to 4.6 percent lower than results reported by this research. The COV varied from 3.9 to
4.0 percent. Graybeal (2005) reported 28 day compressive strengths for air, steam, and delayed
steam cured specimens which were 23.6, 10.0, and 17.1 percent lower than what observed in this
research program, respectively. COV ranged from approximately 3 to 6 percent for Graybeal’s
reported data.
4.1.3 Air-Cured Compressive Strength Growth over Time
Previous sections have shown that once UHPC receives a thermal treatment its
compressive strength properties vary little. However, it was also shown that until UHPC
receives a thermal treatment its strength increases with time much like traditional NSC or HPC.
Determining how quickly and to what ultimate capacity UHPC will increase in strength under
ambient conditions is an important factor for the design and use of UHPC.
71

Graybeal (2005) tested specimens with four curing treatments for compressive strength,
modulus of elasticity, and linearity of the response over time. The mean results of the
compressive strength over time for specimens cured under ambient conditions can be seen along
with the mean results of Air-cured specimens for this research in Figure 4-2.

Figure 4-2: Compressive Stress Gain over Time for Air-Cured Specimens

Both sets of data present the classic asymptotic shape of the compressive strength
approaching some ultimate limit. In fact both data sets seem to have very similar curves, but the
data from this research has a greater compressive strength. Graybeal’s (2005) data appears to be
approaching 20 to 22 ksi while this research data appears to approach 25 to 27 ksi. There are
several possibilities which could cause this difference in results. The first is the age at which
specimens were demolded. Recall that specimens were demolded after 3 days for this research
program and 24 hours for work conducted by Graybeal (2005). As part of the material study on
0.0
5.0
10.0
15.0
20.0
25.0
30.0
0 10 20 30 40 50 60
C
o
m
p
r
e
s
s
i
v
e

S
t
r
e
s
s

(
k
s
i
)
Specimen Age (days)
Graybeal Data
Air Data
72

UHPC, Graybeal (2005) investigated the impact of age at demolding had on compressive
strength. The results showed that by leaving the specimens in the molds to 48 hours, an increase
in 28 day compressive strength of 5 ksi was seen in specimens receiving ambient curing. A
second possibility is the age of premix at the time of mixing. Both Kollmorgen (2004) and
Graybeal (2005) noted that the older the premix was the longer it took to mix each batch, and
Graybeal (2005) saw that older premix took longer to set. If the premix was old and specimens
had just enough strength to survive demolding and then were exposed to ambient conditions, the
specimens would likely dry out before using all available water for hydration, and would result
in lower strength (Graybeal 2005). The age of premix used for specimens in the compressive
strength specimens was approximately 2 to 4 months. A third possibility would be different
constituents or their proportions in the premix. The exact constituents and their proportions is
not know as this is a proprietary UHPC and the premix comes blended. Finally, the ambient
curing conditions could be different in each laboratory. Mixing, casting, and curing of
specimens for this research were conducted in the basement of Benedict Laboratory at Michigan
Tech during June, July, and August. Although the temperature and relative humidity did not
vary excessively during these months (~70°F, 30-50% RH), they were not controlled. It is
possible that UHPC curing in higher relative humidity could use the more readily available water
vapor continue hydration. None the less, data clearly shows that UHPC continues to gain
strength while curing at ambient conditions.
4.2 Modulus of Elasticity and Poisson’s Ratio
A total of 73 3x6 in. cylindrical specimens were tested to investigate the effects of age at
thermal treatment on modulus of elasticity and Poisson’s ratio. Both properties could be
measured simultaneously on any given specimen, and specimen preparation and testing
73

procedures were discussed in Chapter 3. In general, the process followed ASTM C 469 –
Standard Test Method for Static Modulus of Elasticity and Poisson’s Ratio of Concrete in
Compression, except the load rate was increased to 150 psi per second. Test data recorded for
each specimen is tabulated in Appendix A (Tables A.1, A.2, A.3).
4.2.1 Results
Table 4.6 summarizes the mean and COV for the modulus of elasticity and Poisson’s
Ratio based on curing regime and age. Two specimens which were Air-cured and compression
tested at 3 days were not tested for modulus of elasticity and Poisson’s ratio because of
difficulties with the data acquisition system.
Table 4.6: Modulus of Elasticity and Poisson’s Ratio Test Results
Curing
Regime
Specimen
Age at
Testing
(days)
Number of
Specimens
(Sample
Size)
Mod. of
Elasticity
Sample
Mean (ksi)
Mod. of
Elasticity
Sample
COV (%)
Poisson’s
Ratio
Sample
Mean
Poisson’s
Ratio
Sample
COV (%)
Air-cured 3 4 6,910 0.6% 0.198 2.7
7 6 7,520 1.9 0.205 2.0
14 6 7,865 0.8 0.206 1.3
28 6 7,863 1.8 0.205 4.2
28 - NF

6 7,696 1.2 0.200 1.7
TT (Thermal
Treatment)
7 6 8,056 1.6 0.206 2.2
14 6 8,215 1.3 0.206 2.4
28 6 8,114 0.7 0.205 2.1
28 - NF 5 7,889 1.1 0.203 2.5
DTT (delayed
TT)
14 6 8,177 1.5 0.205 2.6
28 6 8,161 0.6 0.207 4.4
28 - NF 3 7,741 2.2 0.200 1.1
DDTT 28 5 8,098 1.0 0.203 1.0

This table shows data with a very low COV and is very similar to previous results with
this type of UHPC (Kollmorgen 2004; Graybeal 2005). It is because of this low variability that
statistically significant results can be determined with low sample sizes.
74

Figure 4-3 shows a typical stress-strain graph for the calculation of modulus of elasticity.
This particular specimen was thermally treated from Batch 1 and tested at 14 days.

Figure 4-3: Typical Stress-Strain Curve for Calculating the Modulus of Elasticity

Figures 4-4 and 4-5 show the mean modulus of elasticity and Poisson’s ratio results,
respectively, for each age and curing regime. The error bars indicate ± one standard deviation
from the mean.
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
0.0000 0.0005 0.0010 0.0015 0.0020
A
x
i
a
l

S
t
e
s
s

(
k
s
i
)
Axial Strain (in/in)
Limits of Stress-Strain
Curve for Modulus of
Elasticity Calculation
(Gage reading 0.00005
in. to 90 kips)
75


Figure 4-4: Mean Modulus of Elasticity Results for All Ages and Curing Regimes


Figure 4-5: Mean Poisson’s Ratio Results for All Ages and Curing Regimes

The obvious trend which can be seen in the Table 4.6 and Figure 4-4 is that the modulus
value increases with time for the Air-cured specimens. The modulus value increases over 950
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
3 Day 7 Day 14 Day 28 Day 28 Day
No Fibers
M
e
a
n

M
o
d
u
l
u
s

o
f

E
l
a
s
t
i
c
i
t
y

(
k
s
i
)
Age of Specimens
Air
TT
DTT
DDTT
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
0.22
3 Day 7 Day 14 Day 28 Day 28 Day
No
Fibers
M
e
a
n

P
o
i
s
s
o
n
'
s

R
a
t
i
o
Age of Specimens
Air
TT
DTT
DDTT
76

ksi or 13.8 percent from 3 to 28 days. This is to be expected because of the increase in
compressive stress that was noted in the previous section. TT results only vary 159 ksi or about
2 percent independent of age. DTT mean modulus of elasticity values showed a difference of
only 20 ksi from 14 to 28 days. In fact, when comparing all means of specimens undergoing
thermal curing at any age the range in modulus values was from 8056 to 8215 ksi. A less
noticeable trend in the data is that the modulus of elasticity decreased when the fibers were
removed from the mix. Poisson’s Ratio for all curing regimes and ages were consistently
between 0.20 and 0.21, which is in line with typical concrete and the value given by AFGC
(2002).
4.2.2 Statistical Analysis and Discussion
Statistical analysis, similar to compressive stress testing, was conducted for modulus of
elasticity and Poisson’s Ratio comparing first within each curing regime and then later to each
other for each property to analyze the effect of each curing regime. Table 4.7 shows the results
of the hypothesis testing on all specimens for modulus of elasticity.

77

Table 4.7: Statistical Results for Modulus of Elasticity Testing
Age of Specimen
Pairings (days)
Reject or Failed to Reject the Null
Hypothesis (t-test)
Mod. of Elasticity
Population Mean

Air-cured Specimens
3 vs 7 Reject Ho Not Equal
3 vs 14 Reject Ho Not Equal
3 vs 28 Reject Ho Not Equal
7 vs 14 Reject Ho Not Equal
7 vs 28 Reject Ho Not Equal
14 vs 28 Failed to Reject Ho Equal

TT – Thermal Treatment
7 vs 14 Failed to Reject Ho Equal
7 vs 28 Failed to Reject Ho Equal
14 vs 28 Failed to Reject Ho Equal

DTT – Delayed Thermal Treatment
14 vs 28 Failed to Reject Ho Equal

Air-cured specimens continued to gain stiffness over the first 14 day duration and all
pairings were found to be not equal. However, once thermally treated, the Modulus was
achieved and remained constant independent of when thermal treatment was applied. Air-cured
specimens represent all of the thermally treated specimens up to the time of thermal curing, e.g.
TT specimens were not tested at an age of 3 days because thermal treatment had not yet begun.
The DDTT thermal curing occurred during days 24-27 which meant that only 28 day specimens
were tested for this curing regime and no hypothesis testing was possible. Section 4.2.3
discusses the relationship between concrete strength and stiffness (modulus of elasticity).
The purpose of conducting hypothesis testing on all curing regimes was to determine if
the specimens tested at different ages had the same population mean. The t-tests have shown
that the 14 and 28 day Air-cured sample sets had the same population mean, and that all TT
specimens shared the same population mean, and all DTT specimens shared the same population
78

mean. The data sharing the same population mean was further combined and hypothesis testing
was completed. Table 4.8 shows the combined values.
Table 4.8: Combined Modulus of Elasticity Results
Curing
Regime
Specimen
Age at
Testing
(days) Sample Size
Combined Sample
Mean Mod. of
Elasticity (ksi)
Combined Sample
COV (%)
Air 14, 28 12 7,864 1.3
TT 7, 14, 28 18 8,129 1.5
DTT 14, 28 11 8,168 1.1
DDTT 28 5 8,098 1.0

The F-test showed that all the combinations of curing regimes had an equal population
variance for the modulus of elasticity results. Conducting a t-test was the next step which
produced results showing that the population mean of the Air-cured specimens was different than
TT, DTT, and DDTT. Table 4.9 shows the results of the t-test.
Table 4.9: Statistical Results for Combined Modulus of Elasticity Testing
Curing Regime
Pairings
Reject or Failed to Reject the Null
Hypothesis (t-test) Population Mean
Air vs TT Reject Ho Not Equal
Air vs DTT Reject Ho Not Equal
Air vs DDTT Reject Ho Not Equal
TT vs DTT Failed to Reject Ho Equal
TT vs DDTT Failed to Reject Ho Equal
DTT vs DDTT Failed to Reject Ho Equal

When the specimens that received thermal curing were compared, all three combinations
showed that the population mean was equal for all curing regimes. This is the same result
obtained from the compressive hypothesis testing. For specimens receiving thermal treatment at
any age the mean modulus of elasticity is 8150 ksi, which is an increase of 3.8 percent over the
combined Air-cured value of 7850 ksi. It is interesting to note that the compressive strength of
79

Air-cured specimens continued to increase from day 14 to 28 but the modulus value appears to
have reached a plateau by 14 days. A similar result was noted by Graybeal (2005) as the data
showed the compressive strength increasing to 8 weeks after casting but the stiffness and peak
strain at failure curtailed at 4 weeks after casting.
All curing regimes were compared to determine the effects that the different curing
regimes had on the Poisson’s ratio. Table 4.10 shows the results of the hypothesis testing on all
specimens for Poisson’s ratio tests.
Table 4.10: Statistical Results for Poisson’s Ratio Testing
Age of Specimen
Pairings (days)
Reject or Failed to Reject the Null
Hypothesis (t-test)
Poisson’s Ratio
Population Mean

Air-cured Specimens
3 vs 7 Failed to Reject Ho Equal
3 vs 14 Failed to Reject Ho Equal
3 vs 28 Failed to Reject Ho Equal
7 vs 14 Failed to Reject Ho Equal
7 vs 28 Failed to Reject Ho Equal
14 vs 28 Failed to Reject Ho Equal

TT – Thermal Treatment
7 vs 14 Failed to Reject Ho Equal
7 vs 28 Failed to Reject Ho Equal
14 vs 28 Failed to Reject Ho Equal

DTT – Delayed Thermal Treatment
14 vs 28 Failed to Reject Ho Equal

Similar to compression stress and modulus of elasticity results and based on the above
analysis, Poisson’s ratio hypothesis testing results were combined for curing regimes and
compared for statistical differences. Results are summarized in Tables 4.11 and 4.12.
80

Table 4.11: Combined Poisson’s Ratio Results
Curing
Regime
Specimen
Age at
Testing
(days) Sample Size
Combined Sample
Mean Poisson’s
Ratio
Combined Sample
COV (%)
Air 3, 7, 14, 28 22 0.204 3.0
TT 7, 14, 28 18 0.206 2.1
DTT 14, 28 11 0.206 3.6
DDTT 28 5 0.203 1.0

Table 4.12: Statistical Results Poisson’s Ratio Testing
Curing Regime
Pairings
Reject or Failed to Reject the Null
Hypothesis (t-test) Population Mean
Air vs TT Failed to Reject Ho Equal
Air vs DTT Failed to Reject Ho Equal
Air vs DDTT Failed to Reject Ho Equal
TT vs DTT Failed to Reject Ho Equal
TT vs DDTT Failed to Reject Ho Equal
DTT vs DDTT Failed to Reject Ho Equal

Results show none of the null hypotheses were rejected and all six pairings show that all
curing regimes come from the same population mean. Even if there was a slight difference in the
population mean (a few thousandths), caution would have to be exercised in stating that a
difference was evident because Poisson’s ratio is reported and used in equations to two decimal
places. So if a difference shows up in the third or fourth decimal place, it has no practical
application. Based on this analysis, data for the UHPC specimens (Air-cured and all thermally
cured specimens independent of age at curing) have a mean Poisson’s ratio of 0.21.
Kollmorgen (2004) reported specimens tested before thermal treatment having modulus
values 5.1 percent less than those presented in this research. These specimens had a high COV
at 20 percent. Following thermal treatment and independent of age at testing a mean value of
9210 ksi was presented which is over 13 percent greater than the results for TT at 28 days and
81

the COV was 7.1 percent. These results were based on the utilization of three cylinder sizes and
LDTs as presented in Chapter 3. Poisson’s ratio results, for specimens tested before and after
thermal treatment, ranged from 0.16 to 0.22 with a mean of 0.19 and COV of 13.3 percent.
Graybeal (2005) presents modulus of elasticity results for air, steam, and delayed steam
cured specimens as 6200 ksi, 7650 ksi, and 7300 ksi respectively. When comparing these results
to those presented herein based on 28 day means, Graybeal’s (2005) results are 21.1 percent, 5.7
percent, and 10.6 percent lower than Air-cured, TT, and DTT results. COV for the three curing
treatments were very low ranging from 2 to 3 percent. The modulus of elasticity values were
calculated over the ranges of 10 to 40 percent of the ultimate capacity of each specimen. It was
noted that cylinders which underwent thermal curing reached 80 to 90 percent of their ultimate
capacity before deviating from a linear elastic behavior by 5 percent. This shows UHPC as
exhibiting nearly linear behavior to high stress levels.
While an in depth investigation on Poisson’s ratio was not conducted by Graybeal (2005),
a background study completed for determining an acceptable load rate indicated consistent
Poisson’s ratio results of 0.19 for thermally cured specimens and compares well with research
results presented herein.
4.2.3 Compressive Stress and Modulus of Elasticity Relationship
The modulus of elasticity of concrete is an important parameter and, therefore, a simple
yet acceptably accurate way to predict this property is necessary. As discussed in Chapter 2,
there are several different models which relate the compressive strength of UHPC to its modulus
of elasticity. It is not the intent of this research to propose a relationship, but rather to determine
if published equations predict the modulus of elasticity given the compressive strength for the
data collected herein. All specimens which were determined to have the same population mean
82

were used in the relationship of modulus of elasticity and compressive strength. For example,
TT specimens at 7, 14, and 28 days had the same population mean so all specimens were
averaged and included in the following figure. However, Air-cured specimens at 3, 7, 14, and 28
days had different population means and means are shown separately. Figure 4-6 shows the
experimental data plotted with the different prediction equations for modulus of elasticity based
on compressive strength.

Figure 4-6: Regression Model for Modulus of Elasticity vs. Compressive Strength

The predictive relationships vary significantly and none are a great match for all the data
gathered during this research. The sum of squares of the residuals was calculated for each
equation to determine the best fit for the data. The results showed that the sum of squares of the
residuals was minimized when the AFGC (2002) model was used. The value of the sum of
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
M
o
d
u
l
u
s

o
f

E
l
a
s
t
i
c
i
t
y

(
k
s
i
)
Compressive Stress (ksi)
ACI 318 (Eqn 2.1)
ACI 363 (Eqn 2.2)
AGFC (Eqn 2.3)
Iowa (Eqn 2.4)
Kollmorgen (Eqn 2.5)
Graybeal (Eqn 2.6)
Air Data
TT Data
DTT Data
DDTT Data
83

squares of the residuals was less than half of the next smallest value which was the Iowa model.
Just minimizing the sum of squares of the residuals does not guarantee a good fit. The coefficient
of determination which shows the goodness of fit and was calculated to be 0.34, which is not a
good fit. However, when the Air-cured specimens were removed from the analysis, the
coefficient of determination increased to 0.98 which is an excellent fit. Recall that the AFGC
(2002) equation requires the specimens to undergo a thermal treatment, so limiting the analysis
to the thermally treated specimen is warranted. However, this does not address the problem of
determining the modulus of elasticity for UHPC air-cured under ambient conditions.
Figure 4-7 presents the mean results (based on age at testing) of specimens cured under
ambient conditions and tested for compressive stress and modulus of elasticity. The results are
summarized with data from Kollmorgen (2004) and Graybeal (2005).

Figure 4-7: Mean Values of Compressive Stress and Modulus of Elasticity for Air-Cured
Specimens

0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
M
o
d
u
l
u
s

o
f

E
l
a
s
t
i
c
i
t
y

(
k
s
i
)
Compressive Stress (ksi)
Kollmorgen Data
Graybeal Data
Air Data
84

The mean Air-cured results for this research program fall in between the average results
presented by Kollmorgen (2004) and Graybeal (2005). This is the same trend observed in Figure
4-6 where Kollmorgen’s equation over predicts the modulus of elasticity, and Graybeal’s
equation under predicts the results over their appropriate ranges of compressive stress. At the
time of this report Graybeal’s (2005) raw data was not available to compile all the results from
the individual specimens and complete a regression model to determine an appropriate equation
for Air-Cured specimens. Further testing and/or analysis should be completed to determine the
relationship of compressive stress to modulus of elasticity for specimens cured under ambient
conditions. Therefore, data collected from this study shows the AGFC (2002) model to most
accurately predict the modulus of elasticity based on compressive strength of thermally treated
specimens, independent of age at thermal treatment application.
4.3 Flexural Strength Testing for First Cracking
Flexural testing was conducted at 28 days on 2 x 2 x 11.25 in. prisms that underwent
three curing regimes; Air-cured, thermal treatment, and delayed thermal treatment. ASTM C
1018 Standard Test Method for Flexural Toughness and First-Crack Strength of Fiber-
Reinforced Concrete (Using a Beam with Third Point Loading) was used to determine the first-
crack strength and flexural toughness of specimens. A total of 36 prisms were cast with fibers
and 9 were cast without fibers (Table 4.1) to see the effect that fibers had on the first crack
flexural stress. One specimen, Air-cured, was rejected because of a break outside of the
acceptable region. Additionally, three specimens had breaks within 5% of the constant moment
region and their values were reduced as outlined in ASTM 1018.
The first-crack flexural stress is used as an indicator of the maximum tensile stress for
UHPC. ASTM C 78 Standard Test Method for Flexural Strength of Concrete (Using Simple
85

Beam with Third-Point Loading) presents the equation for modulus of rupture ( R ), or cracking
stress, for beams with third-point loading.
2
*
*
d b
L P
R = (psi) Equation 4.1
Where: P = Load at first crack;
L = Span length;
b = Average specimen width;
d = Average specimen depth.
However, Graybeal (2005) points out that it has been widely observed that cracking stress
is an over estimate of the actual tensile capacity of UHPC. This discrepancy has been attributed
to depth and gradient effect set up by bending. To obtain the actual tensile cracking stress (
ct
f ),
AFGC (2002) recommends the flexural cracking stress be corrected with the use of Equation 4.2
which correlates the values back to a 100 x 100 mm (approximately 4 x 4 in.) prism.
7 . 0
7 . 0
* 0 . 2 1
* 0 . 2
*








+








=
o
o
ct
d
d
d
d
R f Equation 4.2
Where: f
ct
= corrected first-crack flexural stress;
o
d = reference depth of 100 mm.
4.3.1 Results
Table 4.13 lists the results of the flexural testing for the curing regimes of Air, TT, and
DTT for specimens with and without fibers. Test data recorded for each specimen is tabulated in
Appendix A (Tables A.4, A.5, A.6). Equation 4.2 was applied to the first-crack flexural stress to
calculate the corrected first-crack flexural stress or cracking tensile stress. The tensile strength
86

listed in the table for TT and DTT are slightly below the range of values for tensile strengths
reported by manufacturers.

Table 4.13: Flexural Stress, Deflection and Maximum Load Results
Curing
Regime
Specimen
Age and
fibers
(days)
Sample
Size
Sample Mean
First-Crack
Flexural Stress
Sample
COV First-
Crack
Sample
Mean
Corrected
First-Crack
Sample
COV (%)
Corrected
First-Crack
Air 28 12 1.34 10.9 0.74 10.9
28 –NF 3 1.50 9.5 0.83 9.5
TT 28 12 1.91 8.0 1.06 8.0
28 – NF 3 2.03 6.7 1.13 6.7
DTT 28 12 2.12 9.1 1.18 9.1
28 – NF 3 2.18 6.4 1.21 6.4

Curing
Regime
Specimen
Age and
fibers
(days)
Sample
Size
Sample Mean
First-Crack
Deflection (in.)
Sample
COV
(%)First-
Crack
Deflection
Sample
Mean First-
Crack Load
(kips)
Sample
COV (%)
First-Crack
Load
Air 28 12 0.00179 8.9 1.23 8.1
28 –NF 3 0.00205 4.2 1.36 7.4
TT 28 12 0.00240 7.4 1.70 7.3
28 – NF 3 0.00275 6.8 1.82 5.5
DTT 28 12 0.00273 6.2 1.93 7.0
28 – NF 3 0.00307 7.0 1.98 6.4

Curing
Regime
Specimen
Age and
fibers
(days)
Sample
Size
Sample Mean
Max. Load
Deflection (in.)
Sample
COV (%)
Max. load
Deflection
Sample
Mean Max.
Load (kips)
Sample
COV (%)
Max. Load
Air 28 12 0.0452 14.8 4.21 9.1
28 -NF 3 - - - -
TT 28 12 0.0498 14.8 4.83 5.9
28 - NF 3 - - - -
DTT 28 12 0.0430 27.3 4.55 8.6
28 - NF 3 - - - -

The flexural strength for fiber reinforced concretes is often difficult to calculate because
of the extensively cracked section and load carrying fibers. However, for convenience ASTM C
1018 allows the flexural strength to be calculated using Equation 4.1 by replacing the cracking
load with the ultimate load. As Graybeal (2005) points out this result does not have any physical
87

meaning, but is used for comparison purposes. The mean flexural strength for Air-cured, TT,
and DTT are 4.58 ksi, 5.44 ksi, and 4.99 ksi respectively. The results for flexural strength, like
those for tensile strength, are at the lower end of their respective range of values as reported by
the supplier. The results of the flexural testing show the largest COV for the mechanical
properties tested for this research. However, the COV presented is only slightly higher than the
one-sigma limit of 7 percent presented in ASTM C 1018 for first crack flexural stress.
Figure 4-8 shows the mean values and ± one standard deviation from the mean for the
corrected first-crack flexural stress for specimens with and without fibers.

Figure 4-8: Mean First-Crack Flexural Stress for All Curing Regimes

4.3.2 Statistical Analysis and Discussion
One of the more interesting observations, made from Table 4.13 and Figure 4-8, is that
the samples without fibers seem to have a larger first-crack stress than the specimen cast with
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
28 Day 28 Day No Fibers
C
o
r
r
e
c
t
e
d

M
e
a
n

F
i
r
s
t

C
r
a
c
k

F
l
e
x
u
r
a
l

S
t
r
e
s
s

(
k
s
i
)
Age of Specimens
Air
TT
DTT
88

fibers. This can be explained looking at the nature of the material from a homogeneous
standpoint. The fibers are geometrically the largest material in the mix and create irregularities
in the mix along the interface of the fibers and the paste. When looking at the small scale, fibers
create a non-homogeneous mixture and cause disruptions in the matrix and allow microcracking
to propagate more readily through the specimen than through the homogenous matrix without
fibers. Hence the lower first-cracking stress in UHPC with fibers. However, the difference in
first-cracking stress is very small and, as Mindess et al. (2003) points out, the purpose of the
fibers is to bridge the crack and provide post-crack ductility.
Flexural testing only had the three curing regimes of Air-cured, Thermal Treatment, and
Delayed Thermal Treatment. Additionally, flexural specimens were only tested at 28 days, and
are not compared within each curing regime. Unlike traditional unreinforced concrete, UHPC’s
first crack was not its last crack because of the fiber reinforcement. Therefore, the corrected first
crack strength and the maximum load could be analyzed to see the effects of the three curing
regimes. Recall that the corrected first crack strength is based on the SETRA procedure for
adjusting data to a normalized size of 100 mm x 100 mm (nominally 4 x 4in.) (AFGC 2002).
Similar to previous analyses, Table 4.14 displays the pairings and results of the hypothesis
testing for first-crack flexural strength.
Table 4.14: Corrected First-Crack Flexural Strength Hypothesis Testing
Curing Regime
Pairings
Reject or Failed to Reject the Null
Hypothesis (t-test) Population Mean
Air vs TT Reject Ho Not Equal
Air vs DTT Reject Ho Not Equal
TT vs DTT Reject Ho Not Equal

89

Hypothesis testing results for flexural strength are different than the compression and
modulus of elasticity testing where only the Air-cured specimens had a different population
mean, and Poisson’s ratio results where all population means were equal. UHPC’s ability to
continue to carry load after the element has cracked is one of the properties which make it
desirable for structural elements. However, flexural first cracking strength appears to be
influenced by the curing regime.
As previously noted Kollmorgen (2004) conducted flexural testing utilizing ASTM C 78
and did not correct the first-crack flexural stress. The mean first crack stress for thermally
treated 2 x 2 x 11.25 in. prisms at 28 days was 3.1 percent greater than results of this research
with a COV of 17 percent. The mean maximum load for the same specimens was 5.4 percent
lower and with COV of 8.0 percent.
When comparing the corrected first-crack data to results presented by Graybeal (2005),
using the same specimen size and testing configuration data, results were 73 percent, 35 percent,
and 10 percent greater than the Air-cured, TT, and DTT mean values, respectively, for this
research. No COV or standard deviation was reported for the mean values. Average peak load
values were 170 to 200 percent greater than the first crack load. Mean peak load values for this
research program were 340 percent, 280 percent, and 240 percent greater than the first-crack load
for Air-cured, TT and DTT, respectively. Large differences in testing results indicate that
ASTM procedures are not necessarily applicable to UHPC for flexural strength and further study
is warranted.
4.3.3 Flexural Toughness
Post-crack ductility is measured by the toughness that a fiber reinforced concrete
exhibits. Figure 4-9 is a graphical representation of a flexural elastic-plastic material, and
90

displays the areas under the load deflection curve used to calculate the reference toughness
indices. The toughness indices for plain concrete, elastic-plastic material, and observed ranges
for fibrous concrete are presented in Table 4.15. Figure 4-10 is a typical load deflection curve
for all specimens.

Figure 4-9: Load Deflection Curve for Elastic-Plastic Material (ASTM C 1018 Figure X1.1)

Table 4.15: Typical Toughness Values (ASTM C 1018 Figure X1.1)
Area Basis
A

Index
Designation
Deflection
Criteria


Values of Toughness Indices
Plain
Concrete
Elastic-Plastic
Material
Observed Range for
Fibrous Concrete
OACD I
5
3δ 1.0 5.0 1 to 6
OAEF I
10
5.5δ 1.0 10.0 1 to 12
OAGH I
20
10.5δ 1.0 20.0 1 to 25
A
Indices calculated by dividing this area by the area to the first crack OAB.

δ is the deflection at first-crack.

91


Figure 4-10: Typical Load Deflection Curve for Flexural Specimens

Toughness is an indication of the energy absorption capabilities of a material. Toughness
indices are calculated at set intervals of the first-crack deflection and provide a means to
compare the toughness of different materials. However, it is important to compare identical
specimen sizes and loading configurations as the toughness indices are not independent of
specimen dimension (Mindess et al. 2003). The toughness indices are a measure of how the
material responds versus the standard, a linear elastic material up to first crack followed by
perfectly plastic material behavior. For example, the index designation I
5
is the area under the
load deflection curve, up to 3 times the deflection at first-crack, normalized by dividing the
aforementioned area by the area under the curve up to the first-crack. More specifically, it is the
area of the polygon OACD divided by the area of the triangle OAB, as seen in Figure 4-9.
Additional toughness indices beyond those listed in Table 4.15 can be calculated in the same
0.0
1.0
2.0
3.0
4.0
5.0
6.0
0.00 0.01 0.02 0.03 0.04 0.05 0.06
L
o
a
d

(
k
i
p
)
Deflection (in)
1
st
Crack (δ)
5.5δ 15.5δ 20.5δ 10.5δ 3.5δ
92

manner. Toughness Indices I
30
and I
40
were calculated at a deflection of 15.5δ and 20.5δ,
respectively, for this research to compare to data published by Graybeal (2005). Also, in
research conducted by Chen et al. (1995) the toughness indices I
5
, I
10
, and to a lesser extent I
20

were not particularly sensitive to fiber addition rate or fiber type.
A second relative parameter evaluated from ASTM C 1018 is the residual strength factor,
which is has a standard value of 100 over two consecutive toughness indices values, for an
elastic-plastic material. For example, R
5,10
is calculated as 20(I
10
- I
5
) and R
20,30
is 10(I
30
– I
20
).
The residual strength factors represent the average level of strength retained after first crack as a
percentage of the first-crack strength. Plain concrete has a residual strength factor of zero.
The experimental results for flexural toughness indices and residual strength factors for
specimens with and without fibers are shown in Table 4.16.
Table 4.16: Experimental Toughness Indices and Residual Strength Factors
Curing
Regime
Test
Age
Sample
Size
Corrected Mean
First-Crack Flexural
Stress (ksi)

Toughness Indices
I
5
I
10
I
20
I
30
I
40

Air 28 12 0.74 6.8 17.6 44.7 74.9 107.0
28 –NF 3 0.83 1.0 1.0 1.0 1.0 1.0
TT 28 12 1.06 6.5 16.1 39.5 65.1 92.8
28 – NF 3 1.13 1.0 1.0 1.0 1.0 1.0
DTT 28 12 1.18 6.3 15.4 36.6 58.7 85.0
28 – NF 3 1.21 1.0 1.0 1.0 1.0 1.0
Curing
Regime
Test
Age
Sample
Size
Corrected Mean First-Crack
Flexural Stress (ksi)
Residual Strength Factors
R
5,10
R
10,20
R
20,30
R
30,40
Air 28 12 0.74 217 271 302 320
28 – NF 3 0.83 0 0 0 0
TT 28 12 1.06 193 234 256 273
28 – NF 3 1.13 0 0 0 0
DTT 28 12 1.18 182 212 224 241
28 – NF 3 1.21 0 0 0 0

The toughness indices I
5
and I
10
for Air-cured, TT, and DTT show results at the upper
end or slightly above the observed values listed in Table 4.15 for fiber reinforced concrete.
93

However, the remaining toughness indices, I
20,
I
30
and I
40
show substantially higher values than
those presented by ASTM C 1018. The residual strength factors show average levels of strength
retained from 240 to 320 percent of the first crack strength. They continue to increase to R
30,40

which captures the load deflection curve behavior over the deflection range of 15.5 to 20.5 times
the cracking deflection. This is consistent with the specimen’s load carrying capabilities
increasing from first crack to the point of maximum load. The high toughness indices and
residual strength factors indicate that this UHPC is highly ductile, a desirable property in almost
all concrete applications.
The toughness indices presented by Graybeal (2005) are approximately 12 to 40 percent
lower than those presented in Table 4.16. Likewise Graybeal’s (2005) residual strength factors
are 16 to 46 percent lower than the observed results. The differences in the results may be
partially explained by the following. Recall that the toughness indices are normalized by
dividing the area under the load deflection curve at a given deflection by the area under the curve
up to the first crack. Results provided by Graybeal (2005), with a 2 x 2 in. cross section with a 9
inch span, for the first-crack flexural stress were 10 to 73 percent higher than the observed values
presented in this research. Because the area under the load deflection curve is triangular to first-
crack, a greater flexural stress will result in a larger area and ultimately smaller toughness
indices.
Kollmorgen (2004) calculated toughness using a procedure outlined by AFGC (2002)
which calculates toughness at the first crack, ultimate load, and a deflection limit of 0.030 in.
based on a 2.0 x 2.0 x 11.25 in. specimen. The toughness values have units of in-kips, and are
not directly comparable to results from this study. Kollmorgen (2004) suggests that the values
94

be normalized by dividing by the toughness at first-crack or the toughness of specimens without
fibers.
Specimens cast without fibers were included in this study to provide guidance as to the
impact the fibers had on the first-crack stress, and as previously noted the fibers have a tendency
to reduce the first-crack stress. Testing specimens without fibers would allow toughness values,
which were the summation of the area under the load deflection curve, to be normalized for
comparison between different fiber addition rates and fiber types. Jamet et al. (1995) calculates
the effective toughness of fibers at different addition rates by subtracting the toughness of the
unreinforced high strength concrete from the toughness of the fiber reinforced high strength
concrete. By removing the toughness of the unreinforced matrix the effects of fiber addition
rates, fiber types, and interactions with different mix designs could be examined. Because only
one fiber addition rate, one fiber type, one mix design, and one specimen geometry were utilized
in the presented research, analyzing the effect of the different fiber addition rates or fiber types
was not possible.
4.4 Rapid Chloride Penetration Test
UHPC specimens of three unique age/curing regime applications were tested for chloride
penetration according to ASTM C 1202: 7-day thermally treated, 28-day air-treated, and 28-day
thermally treated. These curing regimes were applied as described in Section 3.4 and the
specimen age refers to the age of the specimen when tested. Following curing and specimen
preparation, the chloride ion penetrability was measured by the total charged passed in coulombs
over a 6-hour period.
95

4.4.1 Results
Initially, three specimens for each age/curing regime were tested. However, additional
specimens were available for 28-day Air and 28-day TT cured specimens and were included in
the analysis. Test data recorded for each specimen is tabulated in Appendix A (Tables A.7, A.8).
Additionally, one of the 7-day TT specimens was tested on day 8 and yielded a total charge
passing of 10 coulombs. As this was within the standard deviation for 7-day TT specimens, it
was also included in the analysis. Results from these tests are summarized in Table 4.17. While
the standard deviation and COV values in Table 4.17 seemed high upon initial observation, the
ASTM C 1202 standard specifies a 42% COV value for a single operator on concrete samples
from one batch.
Table 4.17: Michigan Tech Rapid Chloride Penetration Summary Data
Curing
regime
Age at
testing
(days)
No.
Specimens
Charge Passed
(coulombs)
Chloride Ion
Penetrability Average
Standard
Deviation
COV
(%)
Air 28 4 75 15 20 Negligible
TT 7 3* 10 1.5 15 Negligible
TT 28 4 15 3.5 24 Negligible
*one specimen tested at 8 days
4.4.2 Discussion
All of the UHPC specimens tested exhibited chloride ion penetration values in the
negligible range (< 100 coulombs passed). Nonetheless, the values for the TT-cured specimens
were lower than the Air-cured specimens. Correspondingly, a t-test statistical analysis
demonstrated that the amount of charge passed for the thermally treated specimens was
statistically lower than the air cured specimen results (Misson 2008). In addition, this data was
congruent with other research data (Bonneau et al. 1997) that reported very high resistance
96

(negligible penetration) to ionic transport in steam treated UHPC specimens, and somewhat
higher post-thermal treatment resistances (Graybeal 2006a). Comparatively, the 28-day TT-
cured UHPC specimens tested herein had an average total charge passing equal to Graybeal’s
(2006a) 28-day steam treated specimens with a 95% confidence interval (Table 4.18). Another
statistical analysis on thermally treated specimens tested herein revealed that ionic movement in
thermally treated UHPC was independent of whether the specimen is 7-day or 28-day within a
95% confidence interval.
Table 4.18: Graybeal (2006a) Rapid Chloride Penetration Summary Data
Curing regime
Age at
testing
(days)
No.
Specimens
Tested
Charge Passed
(coulombs)
Chloride Ion
Penetrability Average
St.
Dev.
COV
(%)
Steam 28 3 18 1 6 Negligible
Untreated 28 2 360 2 1 Very Low
Untreated 56 3 76 18 24 Negligible
Tempered Steam 28 3 39 1 3 Negligible
Tempered Steam 56 3 26 4 15 Negligible
Delayed Steam 28 3 18 5 28 Negligible

However, the charge passed by the 28-day Air-cured specimens were lower than the
results reported by Graybeal (2006a), who reported an average of 360 coulombs passing for 28-
day untreated UHPC specimens. One possible reason for the difference may be the different
preparation methods used (kerosene saw vs. water cooled saw) where the kerosene may inhibit
ion migration. Graybeal also used an accelerator in mixing. Nevertheless, a statistical
comparison confirmed that Graybeal’s (2006a) 56-day untreated specimens and the 28-day Air
(untreated) specimens tested herein were equivalent, having an average of 76 coulombs passing
and a standard deviation of 18 compared to 75 coulombs passing and a standard deviation of 15
for the 28-day old specimens from this study, respectively. Further testing may be needed to
97

evaluate the differences and similarities between the Air-cured UHPC specimens at different
ages.
Visual observation of the UHPC specimens after testing revealed that the specimens
experienced some corrosion of the steel fibers (Figure 4-11). These stains appeared to be limited
to the surface of the UHPC directly in contact with the sodium chloride solution and no other
distress was visible. All of the specimens tested exhibited similar staining patterns. This
observation was similar to that observed in previous research (Graybeal 2006a).

Figure 4-11: Surface Staining of UHPC Specimen after ASTM C 1202 Test

It should be noted that the ASTM C 1202 test does not specifically measure any one type
of ion movement and instead measures the bulk flow of ions through the specimen (Stanish et al.
2000). Also, the method does not measure permeability as is sometimes understood, but again
measures the ionic movement through the specimen. Finally, in materials containing high
amounts of silica fume (like UHPC), the test method tends to indicate a lower chloride
movement rate than would normally be expected (Perenchio 1994; Mindess et al. 2003), but no
98

correlation between movement rate and silica fume quantity has been investigated to date for
applicability to UHPC.
4.5 Freeze-Thaw Cyclic Testing
Freezing and thawing testing was performed in accordance with ASTM C 666 –
Resistance of Concrete to Rapid Freezing and Thawing standards. This test procedure involves
the rapid freezing and thawing of concrete by means of freezing in air, and thawing in water.
Eight cycles were completed per day, and approximately every 32 cycles the fundamental
transverse frequency, length change, and mass were observed. Specimens were tested until
failure or 300 freeze-thaw cycles, whichever came first. Only one minor variation was adapted
for the testing of UHPC specimens and as outlined in Section 3.5.5.
4.5.1 Results
Freeze-thaw testing following ASTM C 666 Procedure B (freezing in air, thawing in
water) was performed on four 28-day Air cured and four 28-day TT-cured UHPC specimens to
monitor UHPC’s resistance to freeze-thaw damage. Deterioration due to freeze-thaw (cracking,
spalling, or disintegration) was observed mechanically through the monitoring of a specimen’s
relative dynamic modulus (RDM), length change, and mass change (Table 4.19). Decreases in a
specimen’s RDM indicated disruptions to the transfer of vibrations through the material due to
microcrack formation, while increases in length was a sign of cracks and microcracks creating
void space in the specimen, and decreases in mass signified spalling or disintegration of material.
These three tests were performed every 32 freeze-thaw cycles (96 hours) and the RDM test was
performed on all eight specimens, while the length and mass change was recorded for six of the
specimens. Occasionally, due to equipment malfunction, these parameters were recorded at
99

periods greater than 32 cycles but never at an interval of more than 36 cycles. Also, on some
occasions the specimens were stored in the frozen condition until the equipment was functional
again. The effects of water absorption on the RDM and mass of the six additional side-study
UHPC specimens (three air cured and three thermally treated) were also analyzed (specimens
denoted by “SS”). Length change was not documented on these side study specimens.
Table 4.19: Effects of Freeze-Thaw Cycles on UHPC
Curing and
(Testing)
Regime
+

No.
Spec.
Average
RDM at end
of cycling
(%)
COV
(%)
Avg.
Length
Change
(%)
COV
(%)
Avg.
Mass
Change
(%)
COV
(%)
Air (F-T) 4 101.57 0.32 0.0004* 20.4 0.54* 5.5
TT (F-T) 4 100.27 0.12 0.00014* 14.4 0.08* 44.0
Air (SS) 3 101.91 0.50 N.A. N.A. 0.22 28.66
TT (SS) 3 100.10 0.50 N.A. N.A. 0.06 0.54
*Note: Only three of the F-T specimens were tested for length change and mass change
+
F-T for freeze-thaw cycles, SS for side-study wet-dry cycles

4.5.2 Discussion
Failure of a specimen undergoing freeze-thaw cycles as stated by ASTM C 666 has been
reached when the specimen’s relative dynamic modulus of elasticity reaches 60% of its initial
modulus, or if a 0.10% expansion in length of the specimen is attained. However, all of the
UHPC freeze-thaw specimens maintained their integrity and exhibited an increase in RDM (<
2%) as testing continued. Table 4.19 shows the summary data for the specimens tested (data is
listed in Appendix A, Figures A.9, A.10, A.11). UHPC freeze-thaw specimens showed only a
small increase (< 1%) in mass and negligible (< 0.01%) change in length. The side-study
specimens (which underwent wet-dry cycles without temperature changes) demonstrated similar
increases in RDM and mass. Lastly, the increases for the Air cured F-T specimens (RDM - 1.57
100

%; Mass – 0.54%) were significantly higher than the increases for the thermally treated F-T
specimens (RDM - 0.27%; Mass – 0.08%).
Figure 4-12 displays the correlation between the increase in RDM for the freeze-thaw and
side-study specimens. After 300 cycles, all eight freeze-thaw UHPC specimens had higher
RDM’s than at the beginning of testing, suggesting that the specimens did not deteriorate at all,
but rather continued to hydrate. The similar increases also indicated that the increase in RDM
from both testing regimes was due to the effects from cycling the specimens in and out of water.
The Air-cured side-study specimens’ RDM increased with a similar trend as the Air-cured
freeze-thaw specimens. However, the increase of the TT-cured specimens was small in
comparison to the Air-cured specimens. This can be primarily attributed to the greater amounts
of unhydrated cement particles in the Air-cured specimens that can become hydrated in the
presence of water.
101


Figure 4-12: Effects of Freeze-Thaw Cycling on the Average Relative Dynamic Modulus of
UHPC Samples

Testing by others (Graybeal 2006a; Lee et al. 2005) also reported a similar phenomenon
for freeze-thaw specimens. In both studies, specimens undergoing freeze-thaw cycles increased
in RDM, and Graybeal also reported that the specimens also showed a mass increase. Lee
demonstrated that reactive powder concrete (a precursor to UHPC) cubes increased in
compressive strength after 300 cycles of freeze-thaw testing, and further testing by Graybeal
(2006a) revealed that untreated UHPC specimens immersed in a water bath (without wet/dry
cycles) also increased in compressive strength. These studies suggest that the submersing of
UHPC in water can increase both compressive strength and RDM, even when being exposed to a
harsh freeze-thaw environment. The data in Figure 4-12 supports this as the side-study
99.5
100.0
100.5
101.0
101.5
102.0
102.5
0 50 100 150 200 250 300 350
Freeze-Thaw Cycles
A
v
e
r
a
g
e

R
e
l
a
t
i
v
e

D
y
n
a
m
i
c

M
o
d
u
l
u
s

(
%
)
Air cured F-T specimens
Air cured side study specimens
Thermally treated F-T specimens
Thermally treated side study specimens
102

specimens exhibited similar changes in RDM and mass as the specimens undergoing freeze-thaw
cycling. The overall increase in RDM (1.57 % for Air-cured specimens) in this study was
different than in other research (approximately 10% for untreated specimens) (Graybeal 2006a).
However, Procedure B (freezing in air, thawing in water) was used in this study rather than
Procedure A (freezing in water, thawing in water) which was used in Graybeal’s study.
Lengthier exposure of the UHPC specimens to water during Procedure A may allow for a greater
probability of water reacting with unhydrated cement particles.
Although little visual damage was noted, small microcracks were observed on the
surfaces of the freeze-thaw specimens upon removal from the freeze-thaw chamber, especially
the air cured specimens (Figure 4-13). These cracks were only evident while the surfaces of the
specimens were wet, and quickly disappeared from view once the specimens dried. Despite the
visual observation of cracking in the specimens, the RDM values for all of the specimens were
greater than 100 upon the completion of testing.

Figure 4-13: Cracks in Air Cured UHPC Specimens Following Freeze-Thaw Testing

103

An analysis of the fundamental transverse frequencies recorded provided further evidence
of the low variability of UHPC. Figure 4-14 shows the change in mean resonant frequencies of
UHPC freeze-thaw and side study specimen as testing was performed. The tight range of the
initial resonant frequencies revealed that the air-cured and thermally treated specimens
undergoing freeze-thaw cycles were within 8 Hz of their identically treated side-study
specimens. These resonant frequencies (2720-2770 Hz) were also very close to the frequencies
observed by Graybeal (2400 – 2600 Hz). The somewhat higher frequencies (5% - 15%) in this
testing were likely due, in part, to the shorter length dimension (3% shorter) of the specimens
due to the length change studs (3 x 4 x 15 ½ in. with two ¼ in. of exposed studs as opposed to 3
x 4 x 16 in.). Also note that the initial fundamental resonant frequency of the Air-cured
specimens is lower than that of the TT-cured specimens, an observation also noted by Graybeal
(2006a). However, precise absolute length measurements were not recorded prior testing, so
further comparison of the resonant frequencies is beyond the scope of this project.
104


Figure 4-14: Average Resonant Frequencies of UHPC Freeze-Thaw and Side-Study
Specimens

The UHPC specimens were stored in ambient conditions in the lab after freeze-thaw
testing was completed. Approximately six months after testing was completed, additional
resonant frequency testing revealed that the UHPC specimens displayed frequency responses not
initially observed upon the completion of the testing. First, a noticeable decrease in the resonant
frequencies of the Air-cured freeze-thaw specimens, -1.29%, was observed (Table 4.20). The
thermally treated specimens subjected to freeze-thaw cycling also experienced a decrease in their
resonant frequencies over time, though not as substantial (-0.34%) (Figure 4-15). However, a
similar decrease was not observed in the side study specimens for either curing regimes (Figure
2700
2710
2720
2730
2740
2750
2760
2770
0 50 100 150 200 250 300 350
Freeze-Thaw Cycles
A
v
e
r
a
g
e

R
e
s
o
n
a
n
t

F
r
e
q
u
e
n
c
y

(
H
z
)

Air cured F-T specimens Air cured side study specimens
Thermally treated F-T specimens Thermally treated side study specimens
105

4-16). The decrease in resonant frequency suggests that damage to the UHPC specimens from
freeze-thaw cycling may be greater than that observed immediately after testing was completed.
Table 4.20: Change in Resonant Frequency of UHPC Specimens after Testing Completed
Curing
Regime
Testing
Regime Specimen ID
Resonant
Frequency at
end of testing
(Hz)
Resonant
Frequency 6
months after
testing (Hz)
%
change
Average
change
(%)
Air F-T M-FT-A-28 2729 2715 0.51
-1.29
Air F-T P-FT-A-28 2756 2693 2.29
Air F-T R-FT-A-28 2744 2725 -0.69
Air F-T S-FT-A-28 2749 2703 -1.67
TT F-T M-FT-TT-28 2762 2754 -0.29
0.34
TT F-T P-FT-TT-28 2766 2757 -0.33
TT F-T R-FT-TT-28 2769 2756 -0.47
TT F-T S-FT-TT-28 2767 2759 -0.29
Air SS P-FT-SSA-28 2771 2781 0.36
0.34 Air SS R-FT-SSA-28 2742 2751 0.33
Air SS S-FT-SSA-28 2761 2770 0.33
TT SS P-FT-SSTT-28 2754 2759 0.18
0.23 TT SS R-FT-SSTT-28 2748 2756 0.29
TT SS S-FT-SSTT-28 2772 2778 0.22

Figure 4-15: Resonant Frequencies of UHPC Specimens after Freeze-Thaw Testing
106


Figure 4-16: Resonant Frequencies of UHPC Specimens after Side-Study Testing

Further evidence that frequency responses were different than those initially observed
after testing was the change in the shape of the frequency distribution curves for the air cured
freeze-thaw specimens. During freeze-thaw testing and immediately after testing, the curves
used to determine the transverse resonant frequency of the beams were bell shaped for all
specimens (freeze-thaw and side-study) as shown by the curve in Figure 4-17 of a side-study
specimen. However, after testing had been completed and the specimens were stored for several
months at ambient lab conditions, the curves for the Air-cured freeze-thaw specimens exhibited
the more skewed shape depicted in Figure 4-18. This change in shape indicates a change in the
distribution of frequencies acquired from the impact resonance test and may indicate a change in
material behavior.
107


Figure 4-17: Typical Bell Shaped Resonant Frequency Output of Air-cured UHPC
Specimen (Frequency in Hz)


Figure 4-18: Typical Skewed Resonant Frequency Output of an Air-cured UHPC Specimen
Six Months after Freeze-Thaw Testing (Frequency in Hz)

Overall, these test results indicate that the UHPC specimens underwent some form of
autogenous healing similar to what Jacobsen and Sellevold (1996) observed in HPC freeze-thaw
specimens. Increases in RDM and mass during freeze-thaw cycling were both indicators that
additional hydration may be taking place within the specimens. Additionally, UHPC specimens
outside of freeze-thaw cycling had improved RDM values similar to those in the freeze-thaw
108

chamber. However, this healing appears to have limited long term impact as RDM values
decreased just months after testing was completed.
4.6 Coefficient of Thermal Expansion
The coefficient of thermal expansion for concrete was determined following a modified
version of AASHTO TP-60-00 – Coefficient of Thermal Expansion of Hydraulic Cement
Concrete. Modifications to both specimen preparation and testing procedures were made due to
equipment specifications and in the interest of maintaining the integrity of UHPC curing
regimes. The standard process according to AASHTO TP-60-00 requires that specimens be
saturated with water prior to testing, placed into a test frame and submerged in a thermally
regulated water bath, and then subjected to heating and cooling cycles until an accurate CTE
measurement is obtained.
4.6.1 Results
Coefficient of thermal expansion testing was performed on Air-cured UHPC specimens
ranging from 3-days in age to 28-days and on 7-day and 28-day TT-cured specimens. These
curing regimes were applied upon specimen demolding, and the specimen age refers to the age of
the specimen when tested (Chapter 3). A total of 22 specimens were tested. Typically, CTE
tests began on the day of the stated specimen age (e.g – a specimen tested for 7-day CTE values
began testing on day 7) and lasted 24-36 hours. However, due to one operable test frame and
tests that ran longer than 1-day, three of the specimens were not tested on their appropriate test
day. Therefore, additional specimens were cast for each of the missed testing times such that
sample size of at least three test specimens was available for each curing regimes. The data for
the three specimens not tested on their appropriate test day is not included in this section, but is
109

included in Appendix A for reference. Data for individual CTE tests are listed in Appendix A
(Tables A.12, A.13).
All of the UHPC specimens were kept unsaturated during testing through the use of an
epoxy coating to avoid the potential hydration effects of water on unhydrated cement particles in
the cement matrix. Further motivating this decision was UHPC’s low permeability that would
create an increased likelihood of unequal degrees of saturation in each specimen. This reasoning
follows previous CTE research conducted on UHPC specimens (Graybeal 2006a). However,
moisture content does affect the CTE values of normal strength concrete specimens (Mindess et
al. 2003), and therefore this data does not represent the CTE values of saturated or partially
saturated UHPC specimens.
A summary of the CTE values for unsaturated UHPC specimens is presented in Table
4.21. For unsaturated 28-day TT-cured specimens, the average CTE was 8.16 x 10
-6
/°F, and the
CTE value for unsaturated 28-day Air-cured specimens had an average CTE value of 7.74 x 10
-
6
/°F. Values for the thermally-treated 7-day specimens (8.20 x 10
-6
/°F) were similar to the TT-
cured 28-day specimens and comparatively higher than the Air-cured 7-day specimens (7.62 x
10
-6
/°F). Also, a statistical t-test confirmed that regardless of specimen age, TT-cured UHPC
specimens had a statistically higher CTE value than the Air-cured specimens.

110

Table 4.21: Coefficient of Thermal Expansion (CTE) Test Summary
Curing
regime
Specimen
age at
testing
(days)
No.
samples
tested
Average
CTE
value
(/°F)
St. Dev
(/°F) COV (%)
Air 3 3 7.53E-06 0.08E-06 1.1
Air 7 4 7.62E-06 0.11E-06 1.0
Air 14 3 7.69E-06 0.13E-06 1.8
Air 28 3 7.74E-06 0.15E-06 1.9
TT 7 3 8.20E-06 0.06E-06 0.7
TT 28 3 8.18E-06 0.15E-06 2.1

4.6.2 Discussion
The results demonstrate that the age of a specimen at testing plays a more significant role
in Air-cured UHPC specimens than in TT-cured specimens. The averages of the data in Table
4.21 indicate an increasing CTE value as the Air-cured specimens age (see Figure 4-19). Figure
4-19 shows the CTE values for the Air-cured UHPC specimens and their respective standard
deviations. Yet after performing a statistical t-test on the Air-cured data, only the 3-day and the
7-day specimens were statistically smaller (88 percent and 78 percent confidence, respectively)
than the next testing age (7-day and 14-day, respectively). That is, the 3-day air specimen
exhibited a statistically lower CTE (88% confidence) than the 7-day Air-cured specimen, and the
7-day Air-cured specimen exhibited a statistically lower CTE (78% confidence) than the 14-day
Air-cured specimen. However, the 14-day Air-cured specimen’s CTE value was not statistically
lower than the CTE value for the 28-day Air-cured specimen.
111


Figure 4-19: Average CTE Values for Air-cured UHPC Specimens

Conversely, the 28-day TT-cured specimens showed little change from the 7-day
specimens (Table 4.21). In fact, a two-sample t-test (95% confidence interval) determined that
TT-cured UHPC specimens maintained the same CTE value whether tested at 7-days or at 28-
days. Again, this supports the assertion that thermally steam treating UHPC “locks” in
properties so that specimen properties change little post-treatment.
The UHPC CTE values are slightly higher than the values typically reported for normal
and high strength concretes which tend to have CTE values of approximately 4.1-7.3 x 10
-6
/°F
(FHWA 2006). However, a closer look at the factors that influence the CTE of a concrete
reveals that CTE values obtained for UHPC are reasonable. By and large the CTE of a concrete
is most greatly influenced by the CTE of its coarse and fine aggregates, but UHPC has no coarse
aggregate. Instead, UHPC consists mostly of fine sand (41%) and portland cement (29%).
Therefore, the CTE value for UHPC should fall between the range of a 1:6 cement/natural silica
sand mortar (6.7 x 10
-6
/°F ) (Mehta and Monteiro 2006) and saturated portland cement pastes (10
7.40E-06
7.50E-06
7.60E-06
7.70E-06
7.80E-06
7.90E-06
8.00E-06
0 5 10 15 20 25 30
C
T
E

(
i
n
/
i
n
/
°
F
)
Specimen Age at Testing (days)
112

to 11 x 10
-6
/°F). Equation 2.1 reveals an expected CTE value for UHPC of approximately 8.5 x
10
-6
/°F. Therefore, the CTE value of UHPC is not unexpectedly high.
The CTE values measured in this testing vary from published data (Table 4.22). The
Japanese recommendations (JSCE 2006) suggest a CTE value of 7.5 x 10
-6
/°F for steam treated
UHPC samples while Graybeal’s thermally treated specimens (Graybeal 2006a) had a CTE value
of 8.7 x 10
-6
/°F. Measured CTE values (8.16 x 10
-6
/°F) appear larger than Japan’s data and
smaller than Graybeal’s data. No statistical tests were performed to compare the data because of
differing specimen ages at the time of testing. However, the COV values for Graybeal’s data and
the data presented in this report are similar, showing that both test sets have tight fitting data and
little variation of CTE values between specimens. Some explanations for the variations in data
include slightly different batching and curing procedures, the influence of admixtures used by
others, and the different ages of the specimens at the time of testing.
113

Table 4.22: Comparison of Some Published UHPC CTE Data
Curing
regime
Specimen
Age
(days)
Average
CTE Value
(in/in/°F)
St. Dev
(in/in/°F) COV (%)
Michigan Tech CTE Data
Air 28 7.74E-06 1.5E-07 1.9
TT 28 8.16E-06 1.7E-07 2.1
Graybeal CTE Data (Graybeal 2006a)
Air > 60 days 8.17E-06 2.2E-07 2.7
TT > 60 days 8.67E-06 1.7E-07 1.9
Japan CTE Data (JSCE 2006)
TT* N.A. 7.50E-06 N.A. N.A.
*TT – Definition of exact procedure unknown at the time of publishing
4.6.3 Study of water absorption
To provide alternative methods for sealing UHPC specimens prior to CTE testing, a study
was performed to determine the sealing properties of two types of sealant – concrete driveway
sealant and epoxy sealant. The objective of the study was to determine whether concrete
driveway sealant would provide a simpler and more effective method of preparing UHPC
specimens prior to CTE testing. For that reason, a driveway concrete sealant was used to coat
the UHPC specimens that were used to shakedown the CTE equipment. The shakedown
involved testing the repeatability of the equipment following the same testing procedures
previously outlined. However, the UHPC specimens coated in driveway concrete sealant
exhibited a mass increase of approximately 0.4% throughout the shakedown. Therefore, epoxy
sealant was employed during testing and a side study to examine the effect of water absorption
on the weight and length of specimens coated in epoxy was developed. During testing it was
found that specimens coated in epoxy only increased 0.02% in mass during a typical CTE test
(Table A.13 – Appendix A) or approximately a gram of water absorbed by each specimen. For
that reason, epoxy was determined to be an adequate sealant. Application of concrete driveway
114

sealer was a simpler process than epoxy application, but it provided less adequate protection
against water absorption. A more detailed study to account for minute length changes due to
water absorption in UHPC was beyond the scope of this project.
115

5.0 Conclusions of the Experimental Studies
The purpose of this research was to determine the impact that age of thermal treatment
had on the mechanical properties of compressive strength, modulus of elasticity, Poisson’s Ratio,
and flexural strength and toughness of an ultra-high performance concrete. Four different curing
regimes were used to measure the impact on the mechanical properties, Air-cured, Thermal
Treatment, Delayed Thermal Treatment, and Double Delayed Thermal Treatment (Air, TT, DTT,
and DDTT, respectively). Additionally, a UHPC using two curing regimes was compared for
resistance to rapid chloride penetration and freeze-thaw, and to determine the coefficient of
thermal expansion. A summary of the tests conducted was listed previously in Table 4.1.
Results and discussion were also provided in Chapter 4. This chapter summarizes the major
conclusions of the experimental studies.
In general, the following conclusions can be made. Specific conclusions for each test
type are listed below in separate sections.
1. UHPC durability properties researched herein exceed those of normal strength concretes
and high performance concretes.
2. Mixing time increases as the age of the premix of UHPC increases, however material
properties do not show significant changes.
3. UHPC test results are repeatable between different laboratories when comparing to
Graybeal’s work at the FHWA Turner-Fairbank Laboratory (2005), although the research
reported herein is much expanded over the preliminary studies conducted at FHWA.

116

5.1 Compression Strength
1. Compressive strength testing showed that there was no difference in the compressive
stress after thermal treatment was applied. The mean compressive stress for all TT, DTT,
and DDTT cylinders was 30.1 ksi. The compressive stress was independent of age at
which thermal treatment was applied as well as the age at which the specimen was tested
following thermal treatment. This could have a large impact on how UHPC is used in
industry, by allowing a precaster to cast several elements over a period of time and then
thermally treat them simultaneously, allowing more flexibility in the casting and curing
sequence.
2. Air-cured specimens showed an increase of strength with age and at 28 days had a
compressive stress of 23.9 ksi. The Air-cured specimens appear to be asymptotically
approaching a maximum compressive stress of 25 to 27 ksi.
5.2 Modulus of Elasticity and Poisson’s Ratio
1. Modulus of elasticity was scarcely impacted by the four curing regimes. The Air-cured
specimens at 14 and 28 days, which were shown to have the same population mean, had a
mean modulus of elasticity of 7850 ksi. Like the compressive stress samples, the three
curing regimes of TT, DTT, and DDTT had the same population mean and a combined
modulus of elasticity of 8150 ksi. By conducting thermal curing, the modulus value was
only increased by 3.8 percent where the compressive stress increased by 25.9 percent
over air curing based on 28 day information.
2. Modulus of elasticity data for this research is best predicted by the model proposed by
AFGC (2002) which is:
( )
3
` * 000 , 262
ATT c
f E = (psi) Equation 2.3
117

Where: f`
ATT
= Compressive stress of UHPC after thermal treatment.
Like the compressive stress, having the same modulus value independent of when the
thermal cure was applied or when the specimen was tested would be a measurable benefit
not only to the precast industry but the design industry as well because the casting and
curing of the material is flexible.
3. The four curing regimes had no impact on Poisson’s ratio as all specimens, independent
of age or curing regime, had the same population mean. The mean value for all samples
was 0.21 which is slightly greater than 0.20, the commonly accepted value for normal
strength concrete. Again, by having Poisson’s ratio independent of when thermal curing
is applied makes the manufacturing process much more flexible.
5.3 Flexural Strength Testing for First Cracking and Toughness
1. One age (28 days) and three curing regimes were considered for flexural testing, Air, TT,
and DTT. Data was corrected for specimen size as discussed in Section 4.3. Unlike
compression and modulus, all curing regimes had different population means for
corrected first crack stress. The sample means for corrected first crack stress are as
follows: Air 0.76 ksi, TT 1.06 ksi, and DTT 1.18 ksi. The difference between TT and
DTT does not coincide with the compression and modulus trends were specimens which
received thermal curing had the same population means, which would slightly complicate
the production process; however this could be easily overcome by using the conservative
value for calculations.
2. Flexural toughness and residual strength factors are significantly enhanced by the use of
fibers. Fibers provide post-crack ductility.

118

5.4 Rapid Chloride Penetration Test
1. Rapid chloride penetration resistance of UHPC is superior to NSC and HSC regardless of
curing regime. All UHPC specimens, whether Air-cured or TT-cured had negligible
chloride ion penetrability.
2. Thermally treating UHPC specimens enhances its chloride penetration resistance by
limiting ionic movement to even lower levels than that of Air-cured specimens.
3. 7-day and 28-day thermally treated specimens exhibited statistically similar total charge
passing results, indicating that specimen age does not play a major role in thermally
treated UHPC chloride ion resistance.
5.5 Freeze-Thaw Cyclic Testing
1. UHPC demonstrated a high resistance to freeze-thaw damage (100+ durability factor and
less than 0.01% length change) with no large cracking or spalling of the material
regardless of curing regime.
2. Air-cured and TT-cured UHPC specimens increased in both relative dynamic modulus
and mass during freeze-thaw testing at rates similar to companion UHPC specimens
undergoing wet-dry cycles. The greatest increases documented were in the Air-cured
specimens, but were still less than 2.0%.
3. UHPC exhibits signs of autogenous healing that leads to increased RDM and mass gain
when submerged in water, even while undergoing freeze-thaw testing. However, the
long term impact of healing decreased just months after testing was completed as noted
by the resonant frequency of UHPC specimens declining.

119

5.6 Coefficient of Thermal Expansion
1. A coefficient of thermal expansion of 8.2x10
-6
/°F is recommended for UHPC once it has
been thermally treated regardless of age.
2. Coefficient of thermal expansion values were tested on unsaturated UHPC specimens and
increased with age in Air-cured UHPC specimens, although the only statistically
significant changes occurred before specimens aged 14-days. TT-cured specimens
maintained CTE values regardless of age.
3. 28-day TT-cured UHPC specimens had a statistically higher CTE value than 28-day Air-
cured specimens (8.18 x 10
-6
/°F and 7.74 x 10
-6
/°F, respectively).
4. UHPC has a coefficient of thermal expansion value slightly higher than NSC. However,
this value can be estimated based on the volumetric proportions and CTE values of
UHPC constituent materials.
5. Epoxy coating performed better than concrete driveway sealer when sealing test cylinders
for maintaining water saturation levels in UHPC specimens.


120











(This page intentionally left blank)

121

6.0 Preliminary Life Cycle Costs of a UHPC Superstructure
A preliminary life cycle cost analysis comparing the initial and long term costs of UHPC
and normal strength concrete (NSC) bridges has been performed. Two scenarios were evaluated
for the UHPC superstructure; scenario one involving a UHPC bridge of only UHPC girders and
an NSC cast-in-place deck and scenario two involving a UHPC bridge of UHPC girders and
deck panels. The Mars Hill Bridge in Wapello County, Iowa served as model for these scenarios
and is composed of UHPC girders and NSC deck and sub-structure. It should be noted that the
Mars Hill Bridge did not use UHPC in an optimized section, which could lead to excessively
higher costs. However, it is currently the only bridge in the U.S. that is built using UHPC. A
cost estimate was also performed for the control bridge, a bridge using a NSC superstructure
with correspondingly larger beams for the constant bridge span and width.
6.1 Bridge Components
The control and similar UHPC bridges were divided into three components:
superstructure, deck, and sub-structure. The superstructure consisted of the three I-beams and
the sub-structure included the abutments and footings. All three model bridges were a 110 ft.
single-span with a width of 24.5 ft. (Moore 2006) to accommodate two lanes of traffic. Precast,
prestressed concrete I-beams and a NSC deck, typical of current MDOT construction practices,
were assumed for the control bridge. The NSC deck was a nine inch, cast-in-place (CIP)
concrete slab. The UHPC model bridges used a modified 45-inch Bulb Tee girders.
Modifications were a two inch reduction in the lower flange and web width and a one inch
reduction in the upper flange (Moore 2006). The abutments were assumed to be equal to the
bridge width, 12.5 ft. high and have a width of 3.0 ft. while footings were assumed to be 3.5 ft.
122

high and have a width of 8.5 ft. The NSC deck of UHPC scenario one was the same as the
control bridge. While the original Wapello County Bridge used an 8 in. deck, costs were
adjusted to a 9 in. deck for comparison to MDOT standard practice. For scenario two, the deck
was assumed to be four inch thick, precast UHPC deck panels, with a similar width and span of
the control bridge, topped with a water-proofing membrane and asphalt wearing surface.
6.2 Construction
RS Means data were used to find the unit cost of each bridge component and activity,
which are described in Tables 6.1 and 6.2. Reference numbers are provided for further
information (RS Means 2005).
Table 6.1: Bridge Component Unit Costs
Item Unit Cost
Reference Number
(RS Means 2005)
Abutment Cubic Yard $345 02800-02850-205-1050
Approach Railing Linear Foot $119 02800-02850-205-4000
Concrete Deck Cubic Yard $298 02800-02850-205-1000
Prefabricated I-beam
Each (100-120 ft
span)
$16,000 02800-02850-205-1620
Footing Cubic Yard $298 02800-02850-205-1000
Parapet Cubic Yard $585 02800-02850-205-1150
Reinforcing,( Epoxy
Coated)
Ton $3,550 02800-02850-205-2100
Sidewalk
Square Foot $17.55 02800-02850-205-1230

Table 6.2: Construction Activities Unit Costs
Item
Unit Cost Reference Number
Machine Excavation for Abutments Cubic Yard $10.80 02300-02315-462-6050
Mobilization and Demobilization
Up to 25 miles (Dozer, loader, backhoe)
Each $305 02300-02305-250-0100
Additional 5 mile haul distance + 5 mi. Each 10% 02300-02305-250-2500
Mob. & Demob. Truck-mounted Crane Each $120 02300-02305-250-2000
123


The total cost of each component and activity for the control bridge was then calculated
using assumed quantities and the above unit costs. Total costs for the approach railings,
parapets, and sidewalks were added to the cost of the deck. The information obtained from RS
Means was cross checked against information from recent Michigan bids available through
MERL (Michigan Engineers’ Resource Library, a database of Michigan bid items that allows
project managers and engineers to make road and bridge project estimates). An estimated cost of
$432,000, taken from the Mars Hill Bridge, was used as a guide for the UHPC structures
(Endicott 2006). Given that Scenario 1 represented the Mars Hill Bridge, the application of a
3.0% inflation rate per year was the only modification to the cost of the UHPC structure so as to
compare in 2007 dollars. The added cost of a UHPC deck, as opposed to a NSC deck, was
included in the cost of the UHPC structure in scenario two, along with the 3.0% inflation rate per
year. Table 6.3 contains the estimated construction costs for the control and UHPC bridges.
Table 6.3: Estimated Construction Costs
Bridge Component
Control Bridge
(NSC Girders and Deck)
Scenario 1
(UHPC girders, NSC Deck)
Scenario 2
(UHPC Girders, UHPC deck
panels)
Deck, railing, parapet $33,000 $33,000 $82,000

Sub-structure $40,000 Not Available Not Available
Superstructure $48,000 Not Available Not Available
Reinforcing, Epoxy Coated $107,000 Not Available Not Available
Activities $4,000 Not Available Not Available
Design (10% of costs) $23,100 Not Available Not Available
Total
(Control in 2005 $, Scenario 1 & 2 in 2006 $) $255,000 $434,000* $483,000
Total (2007 $)
Control-3.4% inflation average 2005-2007
UHPC-3% inflation 2006-2007 $273,000 $447,000 $497,000
*Reported cost adjusted to 9 in. deck from 8 in. deck, all other costs estimated.


Includes waterproofing membrane and asphalt wearing surface.
124

6.3 Maintenance
Bridge maintenance was assumed according to the MDOT Bridge Preservation Timeline
for a concrete deck with epoxy coated rebar and prestressed concrete beams (MDOT 2007b).
Unit costs were obtained from the Ohio Department of Transportation’s Preventative
Maintenance page, as this information was readily available (Ohio LTAP 2007). Maintenance
unit costs can be seen in Table 6.4.
Table 6.4: Unit Costs of Maintenance Activities
Activity
Unit Cost
Beam End Rehab CF $50.00
BIT Overlay SF $15.00
Deck Patch SF $100.00
Joint Replace LF $40.00
Overlay (Deep and Shallow) SF $35.00
Sub Repair: Diagonal Cracking LF $20.00
Sub Repair: Deteriorated Concrete SF $45.00

The MDOT Bridge Preservation Timeline stipulates regular Capital Scheduled
Maintenance (CSM) throughout the life of the bridge to prevent deterioration of the structure.
Capital Scheduled Maintenance, as defined in the CSM Manual, includes superstructure
washing, vegetation control, spot painting, joint repair/replacement, concrete coating/sealing,
minor concrete patching and repair, concrete crack sealing, approach pavement relief joints, and
slope paving repair (MDOT 2007a). Given that it is difficult to assume an average value for
these tasks, an annual deck maintenance cost was utilized in place of the CSM. The annual deck
maintenance cost includes washing, flushing, patching, and sealing. An annual cost of $0.20 per
square foot for Michigan bridges was obtained from a questionnaire sent to several DOT’s,
consultants, and contractors pertaining to procedures and costs of concrete bridge decks (Lopez-
125

Anido 1998). Michigan was one of twelve DOT’s that participated in the survey, and the price
stated above reflects the individual response of MDOT.
The Bridge Preservation Timeline obtained from MDOT suggests a 90 year design life
for the control bridge, at the end of which an evaluation of the bridge is made and replacement of
the superstructure or bridge is decided, along with additional rehabilitation needs (MDOT
2007b). It was assumed that the deck and I-beams of the control bridge would need replacement
at the end of the 90 year design life. Additional rehabilitation needs included sub-structure repair
of diagonal cracking and deteriorated concrete.
A design life of 180 years was chosen for the UHPC structures because UHPC is
expected to outperform traditional structural concrete by at least twice as much. This assumption
is based on results of mechanical and durability tests of the material performed herein. The
Bridge Preservation Timeline was followed for all NSC components of the UHPC structures and
a modified preservation timeline was used for the UHPC components. Tables 6.5 and 6.6
compare the NSC and UHPC maintenance of the bridge girders and deck.
Table 6.5: Bridge Girder Maintenance
Control
UHPC
Year Maintenance Year Maintenance
40 Beam end rehab
90 Beam replace 90 Beam end rehab
130 Beam end rehab
126


Table 6.6: Bridge Deck Maintenance
Control Deck UHPC Deck
Year Maintenance† Year Maintenance†
12 DP & JR
25 DP & JR
40 DO & JR 45 JR
52 DP & JR
65 SO & JR
80 BIT Overlay
90 Deck Replace 90 JR
102 DP & JR
115 DP & JR
130 DO & JR 130 JR
142 DP & JR
155 SO & JR 155 JR
170 BIT Overlay
†DP = Deck Patch, JR = Joint Replace, DO = Deep Overlay, and SO = Shallow Overlay

The UHPC structure in scenario two was also expected to not require annual deck
maintenance. It was assumed that little maintenance would be needed for a UHPC deck, in
comparison to a NSC deck, and the maintenance schedule was reduced to regular joint
replacements.
6.4 Preliminary Life Cycle Costs
The cost-benefit analysis for each bridge included the construction and maintenance costs
over a 180 year period. An analysis period of 180 years was chosen because it is the first
incidence point of the control and UHPC structures, or the first point at which both structures are
at similar maintenance needs. Maintenance costs and repair costs over the 180 year design life
were reduced to a net present value (NPV) based on a seven percent discount rate and a three
percent inflation rate. Equation 6.1 details the method of reduction to NPV where r is the real
discount rate and n is the difference in years from the present to future date.

127

NPV = (present $) / (1+r)
n
Equation 6.1

Table 6.7 summarizes the construction and maintenance costs for the control and UHPC
bridges for the 180 year common time increment.
Table 6.7: Costs of Control and UHPC Bridges, 2007 $
Activity
Control Bridge
(NSC Girders and Deck)
Scenario 1
(UHPC girders, NSC
Deck)
Scenario 2
(UHPC Girders, UHPC deck
panels)
Construction $273,000 $447,000 $497,000
Maintenance $71,000 $69,000 $4,000
Total
(3% inflation/year & 7%
discount rate)
$344,000 $516,000 $501,000
Incremental increase
(%)
(baseline) 50.0% 45.6%

It is interesting to note that the control bridge still provides the lowest cost bridge over
the 180 year period, which includes a total deck replacement. This is easily explained by the
cost of the UHPC material. The cost per cubic yard of the UHPC used in the experimental study
for this research project was around 20 times more expensive than the cost per cubic yard of
typical concrete ($2000/yd versus $100/yd). However, if this cost were to drop, UHPC usage
may become more advantageous in life cycle cost assessments. Figure 6-1 depicts the change in
the cost of scenario two, given a change in the cost per yard of UHPC. The changing cost for
scenario two was found by taking the current unit cost of UHPC and assuming quantities for the
bridge deck and beams, based on their dimensions. When calculating the price for the beams a
10% design cost was added, along with 5% for additional charge that may be unforeseen. These
unforeseen charges were not assumed to include plant modifications needed for UHPC or
transportation costs for the bridge components. Based on Figure 6-1, for it to be advantageous in
128

scenario two the target cost would be about $1,750 per yard, or about a 12.5% decrease over
current prices.


Figure 6-1: Target Cost of UHPC

6.5 Conclusion of the Preliminary Life Cycle Cost Analysis
The preliminary life cycle cost analysis shows that the control bridge will cost $172,000
less than a UHPC bridge constructed of UHPC girders and a NSC deck, and $157,000 less than a
UHPC bridge constructed of UHPC girders and UHPC deck panels, over a 180 year design life.
However, the analysis performed only accounted for the construction and maintenance costs and
does not incorporate additional user costs or benefits of either bridge. Also, the model bridge
chosen incurred increased costs due to transportation of the bridge components. Furthermore,
the UHPC bridge design and actual costs considered did not take full advantage of UHPC by
using an optimized cross-section. After finding the target cost of UHPC it was apparent that if
the unit cost of the UHPC material were to drop about 12.5%, life cycle costs may be less over
the 180 year design life. This would entail that the transportation cost, cost of modifying the
$0.00
$50,000.00
$100,000.00
$150,000.00
$200,000.00
$250,000.00
$300,000.00
$350,000.00
$0 $1,000 $2,000 $3,000
S
c
e
n
a
r
i
o

2
,

2
0
0
7

$
Cost per yd, 2007 $
Target Price of Ductal
UHPC
NSC
129

current plant, and any additional charges would not raise the cost of UHPC above the stated
target cost. It is acknowledged that the initial costs of a current NSC control bridge is much
lower, but the reduced maintenance of the UHPC bridges, in particular for the bridge model
using the UHPC deck panels, is significantly less over time.
Additional costs and benefits to note relating to the user are the delays due to
construction or maintenance, bridge safety, and the effects of deterioration on the performance of
the bridge. The assumed maintenance of the UHPC bridges in this report can only be
considered a rough estimate; further observation of UHPC structures will be needed for a better
assumption.
Also, UHPC is a new material to the construction market, and little is known about the
potential optimization of UHPC components and the future cost of the material after product
familiarization has occurred. Component optimizations in this report are conservative estimates
and may not truly reflect the capabilities of UHPC. As more UHPC structures are implemented,
the limits of the material will be better understood and design of the components will improve.
The future cost of the product is also hard to forecast, however, current prices may be higher due
to research costs, low material production, and the cost of precast plants to modify current
technologies to fit the needs of UHPC.
6.6 Future Work
A thorough cost-benefit analysis is needed to gain more insight on the feasibility of
UHPC bridges in Michigan. This analysis would include the user costs and benefits listed above,
as well as other costs or benefits that may be pertinent, and a risk analysis to assess the
sensitivity of the outcome to input variation. The risk analysis could be as simple as fluctuating
the discount or inflation rates or as complicated as applying computer simulation of weighted
130

input variables. Monte Carlo Simulation is one method of computer simulation that could be
utilized (Walls and Smith 1998).
Eventually a detailed life cycle analysis is needed for UHPC structures. This analysis
would include the environmental impact of the structure, costs and benefits of the structure, and
sustainability of the materials used. As more UHPC structures are built, it is important that the
proper data and observations are made throughout their lives, so that the life cycle analysis can
be simplified.
131

7.0 Recommendations, Implementation and Future Work
Recommendations are made for adapting current test procedures to properly evaluate
UHPC properties. Implementation activities and suggestions for future work are also included in
this chapter.

7.1 Recommendations for UHPC Testing Procedures
While ASTM and AASHTO Standards are accepted for normal strength concretes,
usually up to about 10,000 psi, these procedures may not always be applicable for assessing
UHPC performance. As such, the spirit of ASTM and AASHTO standards were followed for
testing UHPC but some modifications were necessary. This chapter summarizes information on
modifications to the testing procedures used.
7.1.1 Compression Testing
Base Procedure: ASTM C 39
Procedure Modifications for UHPC:
• The load rate was increased from 35 psi per second to 150 psi per second.
• 3 x 6 in. cylinders were used instead of 4 x 8 in. or 6 x 12 in. cylinders
7.1.2 Modulus of Elasticity and Poisson’s Ratio
Base Procedure: ASTM C 469 using Compressometer-Extensometer
Procedure Modifications for UHPC:
• The load rate was increased from 35 psi per second to 150 psi per second.
132

• Instead of breaking a companion cylinder and determining 0.40
c
f ` from that test,
specimens were loaded to a predetermined load based on curing regime and age at time
of testing.
7.1.3 Flexural Strength Testing for First Cracking and Flexural Toughness
Base Procedure: ASTM C 1018
Procedure Modifications for UHPC:
• The midspan deflection rate was chosen as 0.003 in. per minute independent of curing
regime tested.
• The test was carried out to 20.5 times the first crack deflection so a greater number of
toughness indices and residual strength factors could be calculated.
7.1.4 Rapid Chloride Penetration Test
Base procedure: ASTM C 1202
Procedure Modifications for UHPC: None
Preparation Notes:
• Specimens were cast in a 4 in. diameter x 3 in. high cylinder and one day before treating
were cut down from the top to 4 in. diameter x 2 in. high a day before treating. The
bottom surface of the cylinder should not be cut due to the fact that the finished surface
of UHPC will normally be exposed.
• Specimens may never be fully saturated during preparation. Further research is needed to
discover whether or not vacuum preparation actually saturates specimens.
133

Testing Notes:
• Finished bottom of specimens shall be exposed to the NaCl solution to simulate finished
surface of UHPC. Both ends may be cut, but the significance must be checked.
7.1.5 Freeze-Thaw Cyclic Testing
Base procedure: ASTM C 666, Procedure B
Procedure Modifications for UHPC: None
Preparation Notes:
• The time that UHPC specimens should be soaked in water prior to testing should be
determined based on a balanced evaluation of the need to maintain curing regime
integrity versus the need to have a saturated specimen for freeze-thaw cycling.
Testing Notes:
• The optional length change test should not replace the RDM test due to the potential for
autogenous healing of UHPC specimens while undergoing freeze-thaw cycling.
• Length change tests should be performed vertically to allow for easier handling of the
specimens while maintaining accuracy.
• Visual inspection should be carefully performed to monitor for minute micro-cracking
and erosion of the outer surface.
• Future freeze-thaw testing of UHPC should test the specimens for several months after
testing is complete to monitor for changes in RDM.
7.1.6 Coefficient of Thermal Expansion
Base procedure: AASHTO TP 66-00
Procedure Modifications for UHPC:
134

• Do not soak the UHPC samples in a lime bath prior to testing. This may hydrate
additional cement particles and change the properties of the UHPC sample. Instead,
samples must be coated in an epoxy resin to allow for testing in the unsaturated state.
Samples must be completely coated in epoxy except for the points where the LVDT
will be in contact with the sample and where the support frame buttons are in contact
with the sample.
Preparation Notes:
• (Not performed during this research, but suggested for future testing) Specimens shall
be cast in a 4 x 8 in. cylinder mold and then cut to 4 x 7 in. Both ends of the sample
should be cut 1/2 in. to ensure a plane surface of contact for both the LVDT and the
support frame buttons.
Testing Notes: None

7.2 Draft U.S. Design Recommendations for UHPC
Ultra-High Performance Concrete (UHPC) is one of the latest advances in concrete
technology and it addresses the shortcomings of many concretes today. UHPC is typically
defined as a concrete achieving high compressive strength in excess of 25,000 psi (sometimes
greater than 30,000 psi) with exceptional durability performance. As such, innovative solutions
can be created for long-term applications.
The purpose of these draft design recommendations is to begin the development of a
design code for using the material in the U.S. bridges. Recommendations are based on testing
conducted at Michigan Technological University and reported herein. Tests were conducted on
135

Ductal
®
, the only UHPC currently available in the U.S. market at the time of testing. Steel fiber
reinforcement (8x10
-3
in. φ by 0.50 in. length) was provided at a rate of 2% by volume.
Typical recommendations would include three parts: (I) material characterization, (II)
design and analysis of UHPC structures, and (III) durability. While several sections of the
recommendations are outlined below, only those sections with applicable results from the
research study herein are drafted. Other sections are noted as “future sections”.

Part I – Material Characteristics 
Characterization of material behavior is necessary to design systems using UHPC.
Recommended testing procedures for UHPC material and durability characterization were listed
previously in Section 7.1.
Part I.1 – Mixing and Placing 
The concrete mix is batched in a laboratory within manufacturer recommended procedures to
attain design characteristics for the proposed application. Trial batching at a plant should be
conducted to ensure that all parties are familiar with the material prior to casting final elements.
UHPC must be covered throughout the casting process and immediately after casting to avoid
moisture evaporation that can lead to reduced hydration and excessive surface shrinkage.
Part I.2 – Effects of Thermal Treatment 
Some UHPC applications benefit from a thermal curing, while other applications may not
be able to allow for curing other than in ambient conditions. The main benefits of thermal curing
are:
• Increased compressive and tensile strengths
• Improved durability
• Apparent reduced creep and shrinkage after curing
136

• Increased time to maturity so as to eliminate waiting for a 28-day or longer cure time

Thermal treatment is defined as a 100% humidity steam treatment at 194°F for 48 hours.
Thermal treatment begins with a 6 hour ramp up period, 48 hours at the specified humidity and
temperature, followed by a 6 hour ramp down period. Ambient air-curing is typically considered
as 72°F at 30-50% humidity.
Part I.2.x – UHPC Maturity (future section) 
Comments: UHPC maturity to be defined by the set time, such that thermal treatment is
applied only after the set to avoid the potential of DEF (delayed ettringite formation).

Part I.3 – Compressive Strength 
Compressive behavior is characterized most often by ultimate strength. Test specimens
are typically 3x6 in. cylinders cast vertically or horizontally. Horizontal casting allows for
parallel ends and eliminates the need for end grinding as neoprene pads and high-strength sulfer
capping are not appropriate for UHPC’s high strengths.
The ultimate compressive strength of thermally treated UHPC can be expected to achieve
25-30 ksi independent of when thermal curing is applied. The ultimate compressive strength of
ambient cured UHPC is time dependent and shall be measured for specific applications. If no
other information exists at the design stage, a preliminary design compressive strength of 24 ksi
at 28-days can be assumed.
Part I.3.x – Compressive Strength Gain with Time (future section) 

137

Part I.4 – Tensile Strength 
Tensile behavior has been characterized in two stages: an elastic stage and a post-
cracking stage. The fiber strength and disbursement can strongly influence the tensile capacity
of a UHPC member. A direct tensile test is considered the most accurate measure of tensile
strength, however no such test has proved successful for UHPC in the U.S. to date. The flexural
tensile strength at first cracking can be compared to current U.S. design standards for mild steel
reinforced concrete sections. As additional test results become available, tensile behavior of thin
slabs may be different than beam elements.
Part I.4.x – Direct Tensile Strength (future section) 
Part I.4.x – Splitting Tensile Strength (future section) 
 
Part I.4.x – Flexural Tensile Strength 
The flexural tensile strength of UHPC at first cracking is dependent on the curing method
applied. Testing should be conducted to verify results for the intended curing method. Full size
specimen testing may be needed for thin slab applications. If no other information is known in
the early design stages, a first crack flexural strength of 0.75 ksi can be assumed for ambient
curing at 28-days, and 1.05 ksi for thermally cured specimens.
Part I.4.x – Post­cracking Behavior (future section) 

Part I.5 – Modulus of Elasticity 
Part I.5.1 – Static Modulus of Elasticity 
The following relationship can be used to estimate the modulus of elasticity for UHPC
elements (AFGC 2002):
E
c
= 262,uuu × (V¡'
C
) (psi)
where: f’
c
= compressive strength of UHPC (psi).
138

If in the initial design stages and information is unknown, the modulus of elasticity for
thermally cured fiber reinforced specimens can be estimated as 8150 ksi. For ambient cured
specimens, an estimate of 7800 ksi after 14-days can be assumed.
Part I.5.2 – Dynamic Modulus of Elasticity (future section) 

Part I.6 – Poisson’s Ratio 
Poisson’s ratio for UHPC can be generally assumed to be 0.21 within the elastic range.

Part I.7 – Thermal Characteristics 
Part I.7.1 – Coefficient of Thermal Expansion 
A value of 8.2x10
-6
/°F can be assumed as the coefficient of thermal expansion for UHPC
that has been thermally cured. This value should be reduced to 7.7x10
-6
/°F for 28-day ambient
cured UHPC elements.
Part I.7.x – Thermal Conductivity, Thermal Diffusivity, Specific Heat (future section) 

Part I.8 – Shrinkage Behavior (future section) 
Part I.8.x – Early Age Behavior  
Part I.8.x – Post Thermal Treatment Behavior 
Part I.9 –Creep Behavior (future section) 
Part I.9.x – Early Age Behavior  
Part I.9.x – Post Thermal Treatment Behavior 
Part I.10 – Fatigue Strength (future section) 

Part I.11 – Impact Strength (future section) 
139


Part II – Structural Analysis and Design  

Part II.1 – Loads 
Loading for UHPC elements does not vary from loadings applied to normal strength
concrete members. Applied loadings and induced stresses due to loadings shall be calculated in
accordance with accepted methods and the current edition of the AASHTO Bridge Design
Specifications.
The unit weight of UHPC shall be assumed to be 155 pcf.
Part II.2 – Serviceability and Ultimate Limit States (future sections) 
This section should include discussions for serviceability and ultimate limit states,
capacities, and general detailing. The following outline is proposed and can be adjusted as
research becomes available to support findings:
• General detailing – cover, spacing, beveling
• Serviceability – deflections, cracking, vibration
• Moment capacity and strain compatibility
• Shear capacity – one-way beam shear, punching shear, interface (horizontal) shear
• Torsional capacity
• Fatigue resistance
• Buckling of slender members
• Anchorage – bond strength, and confinement steel for bursting zones
• Connection details

Part III – Durability 
Part III.1 Chloride Ion Ingress 
UHPC has demonstrated superior resistance to chloride ion ingress (negligible
penetrability) regardless of curing method.
140

Part III.2 – Freeze/Thaw Resistance 
UHPC has demonstrated a high resistance to freeze-thaw damage (100+ durability factor
and less than 0.01% length change) with no large cracking or spalling of the material regardless
of curing method.
Part III.3 – Carbonation (future section) 
Comment: Carbonation has been found to be negligible (Japan 2006)
Part III.4 – Chemical Attack (future section) 
Part III.5 – Alkali­Silica Reactivity (future section) 
Part III.6 – Fire Resistance (future section) 
Comment: No fire resistance testing has been performed on UHPC to date in the U.S.
End of DRAFT U.S. Design Recommendations

7.3 Implementation Activities
UHPC has been shown to have extreme durability through high resistance to freeze-thaw
cycling and chloride penetration, and advanced mechanical performance through increased
compressive and flexural strengths, making it a prime candidate for structural highway systems,
especially those exposed to aggressive environments like those found in northern regions of the
U.S. and coastal areas. However, because of the enhanced properties of UHPC (such as higher
compressive strengths), the direct implementation of UHPC into highway systems without a
proper design code could result in an inefficiency of the material use. Wasting material is not
only expensive; it is irresponsible, particularly in a decade in which engineers are well aware of
the adverse effects that CO
2
emissions (from cement production) can have on global climate
change. As such, it is imperative to consider efficient designs through optimization of bridge
girders sections and deck systems.
141

UHPC appears to lock in some properties through the use of thermal treatment.
Compressive strength, modulus of elasticity, Poisson’s ratio, chloride penetration, freeze-thaw
resistance and the coefficient of thermal expansion exhibited no change in properties once
thermally cured. However, several properties and their impact on design and performance are
still not well understood. Suggestions for further study address many of these items.
7.4 Suggested Future Work
As a result of this work, several items should be considered in future research. In this
research Air-cured specimens were tested at a maximum of 28 days. It is of interest to see what
happens over the period of months or years. Do the specimens continue to increase in strength or
change modulus and approach an asymptotic ceiling, and if they do what is the limit of an Air-
cured specimen? The point at which a thermal cure does not have an impact needs to be located.
Also, finding a relationship between the compressive stress and modulus of elasticity of non-
thermally cured specimens is of interest.
UHPC exhibits increasing load carrying capacity beyond first-crack because of the fiber
reinforcement. More research is needed to better understand, and account for, the increase in
flexural capacity of UHPC as a function of crack growth/development. Furthermore, differences
in first-crack flexural strength for specimens tested under several curing regimes need to be
identified for proper implementation into design codes. Also, the practical limits of allowable
cracking for use in the design of structural elements need to be determined.

Thermally curing the specimens while under load is an unanswered question which needs
work. Creep-testing research to date has been performed on small thermally cured specimens.
The reality is that all prestressed elements would be creep loaded prior to production curing. The
142

strands would be released and the element would have the compressive force applied, then the
element may be cured immediately or possibly stockpiled for some length of time before curing.
There are two issues which need addressing. The first is the creep loading on a non-thermally
cured specimen, and the second, what happens to the specimen when it is under load and
exposed to the high temperatures of a thermal cure.
The durability properties of UHPC may show great improvements and it is foreseeable
that it could be used as a sacrificial or wearing course over normal strength concrete. The bond
characteristics of these two surfaces would need to be investigated.
Supplemental work is necessary to develop a broad understanding of all types of
concretes classified as ultra-high performance concretes. Currently only research on Ductal
®
has
been performed in the United States, and other UHPC materials should be investigated to
develop a comprehensive understanding of UHPC performance.
The effects of self-healing in UHPC should be investigated further to determine whether
long term benefits exist. Performing an ESEM or petrographic analysis on freeze-thaw
specimens post testing may shed more light on this interesting effect. Additional petrographic
analysis on the continued hydration of Air-cured versus thermally treated UHPC specimens will
also help in properly describing the self-healing phenomena of UHPC.
A closer look at the dynamic response of UHPC undergoing freeze-thaw testing may
provide information about the nature of the “skewed” effect on the frequency curves used to
determine relative dynamic modulus. Furthermore, UHPC’s resistance to freeze-thaw cycling in
a saline environment should also be investigated, along with UHPC’s resistance to deterioration
if cracked prior to freeze-thaw cycling.
143

Research into the coefficient of thermal expansion of saturated UHPC specimens should
also be investigated. The majority of the research to date has focused on the unsaturated CTE
values of UHPC. Additionally, the interaction between UHPC elements and NSC or HPC
elements due to thermal gradients should be researched to provide practical recommendations
when using UHPC with other concretes. Additional studies on the effects of specimen age on
UHPC’s CTE value are valuable to prestressed concrete manufacturers to accurately estimating
strand stress. Determining UHPC’s saturated CTE value, thermal interaction with non-UHPC
materials, and age effects are crucial to the implementation of UHPC as a viable structural
material.
Besides superplasticizer, no admixtures were used in this research. The effects on long
term durability by including accelerators during the mixing process may be important for those
looking for rapid strength gain and long term durability. Alternate curing regimes, such as water
baths have also not been examined thoroughly in the U.S. These studies can provide further
flexibility when designing structures using UHPC.
And lastly, while UHPC shows promise as a material of choice for transportation
infrastructure, design code development is an integral part of introducing any new material for
application. Because of the increased ductility afforded by UHPC, design codes may need to
consider a crack-width based approach to design instead of the current stress-based limit states.
A rigorous study of section optimization (such as stout double-tee shapes for girders or waffle-
slabs systems for slab bridges or deck panels) is warranted. Incorporating results from tests
reported herein, in conjunction with other testing results by others as well as those suggested for
further study, will provide a comprehensive document for designing UHPC structures.
144










(This page intentionally left blank)

145

References
American Concrete Institute-ACI (2005), Building Code Requirements for Structural Concrete,
ACI Committee 318, Farmington Hills, Michigan.

American Concrete Institute - ACI (2002). “Guide to Durable Concrete.” ACI 201.2R-92, ACI
Committee 201, Farmington Hills, Michigan.

American Concrete Institute-ACI (1997), “State-of-the Art Report on High-Strength Concrete,”
Manual of Concrete Practice-Part I, ACI Committee 363, Farmington Hills, Michigan.

Ahlborn, T.M., Steinberg, E.P., Hartmann, J.L., Graybeal, B.A., Potter, J.L., and Volygi, J.
(2003). “Ultra-High Performance Concrete-Study Tour 2002.” 3
rd
International
Symposium on High Performance Concrete, PCI, Orlando, Florida.

Aitcin, P. (1998) “High-Performance Concrete,” Routledge, New York, pp. 549, 551-552.

American Association of State Highway and Transportation Officials – AASHTO (2005)

AASHTO Designation: TP 60-00, Standard Method of Test for Coefficient of Thermal
Expansion of Hydraulic Cement Concrete, 2004.

American Society for Testing and Materials. Annual Book of ASTM Standards. (2006)

ASTM C 39, Standard Test Method for Compressive Strength of Cylindrical Concrete
Specimens, 2004.

ASTM C 78, Standard Test Method for Flexural Strength of Concrete (Using Simple
Beam with Third-Point Loading), 2002.

ASTM C 192, Standard Practice for Making and Curing Concrete Test Specimens in the
Laboratory, 2000.

ASTM C 215, Standard Test Method for Fundamental Transverse, Longitudinal, and
Torsional Frequencies of Concrete Specimens, 2002.

ASTM C 490, Standard Practice for Use of Apparatus for the Determination of Length
Change of Hardened Cement Paste, Mortar, and Concrete, 2004.

ASTM C 469, Standard Test Method for Static Modulus of Elasticity and Poisson’s Ratio
of Concrete in Compression, 2002.

ASTM C 617, Standard Practice for Capping Cylindrical Concrete Specimens, 1998.

ASTM C 666, Standard Test Method for Resistance of Concrete to Rapid Freezing and
Thawing, 2003.

146

ASTM C 1018, Standard Test Method for Flexural Toughness and First-Crack Strength
of Fiber-Reinforced Concrete (Using Beam With Third-Point Loading), 1997,
(Withdrawn 2006).

ASTM C 1202, Standard Test Method for Electrical Indication of Concrete’s Ability to
Resist Chloride Ion Penetration, 2005.

ASTM C 1231, Standard Practice for Use of Unbonded Caps in Determination of
Compressive Strength of Hardened Concrete Cylinders, 2000.

ASTM C 1437, Standard Test Method for Flow of Hydraulic Cement Mortar, 2001.

AFGV (2002) Association Française de Génie Civil Interim Recommendations for Ultra High
Performance Fibre-Reinforced Concretes. SETRA (Service d’études techniques des
routes et autoroutes). (Bétons Fibrés à Ultra-Hautes Performacnes – Recommandations
Provisoires), France.

Banta, T.E. (2005). “Horizontal Shear Transfer Between Ultra High Performance Concrete and
Lightweight Concrete.” MS Thesis, Virginia Polytechnic Institute and State University,
Blacksburg, VA.

Behloul, M., and Cheyrezy, M. (2002a). “Ductal Reference Projects.” Presentation at Bouygues
Headquarters – Challenger, St Quentin en Yveslines, France.

Behloul, M., and Cheyrezy, M. (2002b). “Seonyu Foot Bridge.” Presentation at Bouygues
Headquarters – Challenger, St Quentin en Yveslines, France.

Blais, P., and Couture, M. (1999). “Precast, Prestressed Pedestrian Bridge-World’s First Reactive
Powder Concrete Structure.” PCI Journal, Vol. 44, No. 5, September-October 1999, pp.
60-71.

Bonneau, O., Lachemi, M., Dallaire, E., Dugat, J., and Aitcin, P. (1997). “Mechanical Properties
and Durability of Two Industrial Reactive Powder Concretes.” ACI Materials Journal,
Vol. 94, No. 4, July/August 1997, pp. 286-290.

Bonneau, O., Vernet, C., Moranville, M., and Aitcin, P. (2000). “Characterization of the
Granular Packing and Percolation Threshold of Reactive Powder Concrete.” Cement and
Concrete Research, 30(12), pp. 1861-1867.

Chen, L., Mindess, S., Morgan, D.R., Shah, S.P., Johnston, C.D., and Pigeon, M. (1995).
“Comparative Toughness Testing of Fiber Reinforced Concrete.” Testing of Fiber
Reinforced Concrete, eds. Stevens, D.J., Banthia, N., Gopalaratnam, V.S., and Tatnall
P.C. American Concrete Institute, Detroit, Michigan, pp. 41-75, 1995.

147

Cheyrezy, M., Behloul, M., Dowd, W., and Dauriac, C. (1998) “Reactive Powder Concrete
(RPC) Application for Seismic Design.” American Concrete Institute Convention,
October 29, 1998, pp. 1-15.

Collins, A.R. (1944). “The destruction of concrete by frost.” Journal of the Institution of Civil
Engineers, 23(1), pp. 29-41.

DASYLab (2004). “DASYLab 8.0.” ver. 8.00.04, DasyTech USA, Inc.

Dugat, J., Roux, N., and Rernier, G. (1996). “Mechanical Properties of Reactive Powder
Concretes.” Materials and Structures/Matériaux et Constuctions, Vol. 29, No. 188, May
1996, pp. 233-240.

Ductal
®
Reference, Lafarge North American, Calgary, Canada.

Ductal
®
Reference T 001, Operating Procedure Compressive Test.

Ductal
®
Reference T 002, Operating Procedure Cylinder & Prism Preparation.

Ductal
®
Reference T 006, Operating Procedure Flow Test.

Ductal
®
Reference T 009, Operating Procedure Cylinder End Preparation.

Edvardsen, C. (1999). “Water permeability and autogenous healing of cracks in concrete.” ACI
Materials Journal, Title No. 96-M56, pp. 448-454.

Emanuel, J. and Hulsey, J. (1977). “Prediction of the thermal coefficient of expansion of
concrete.” ACI Journal, 149-155.

Endicott, W.A. (2006) “Iowa Bridge gives Glimpse into the Future.” Ascent, Summer 2006, pp.
41-46.

Federal Highway Administration (FHWA) (2006). “Achieving the Promise of Ultra-High
Performance Concrete.” FOCUS. Nov. Turner-Fairbank Highway Research Center
Online. Online. 26 Jan. 2007.

Fehling, E., Schmidt, M., Bunje, K., Schreiber, W. (2008) “The “Gartnerplatzbrucke” Design of
first hybrid UHPC-steel bridge across the river Fulda in Kassel, Germany.” Ultra-high
performance concrete (UHPC), Second International Symposium on Ultra High
Performance Concrete, Kassel University Press, Kassel, Germany, pp. 581-588.

Gatty, L., Bonnamy, S., Feylessoufi, A., Clinard, C., Richard, P., and VanDamme, H. (1998).
“Silica Fume Distribution and Reactivity in Reactive Powder Concretes.” Fly ash, silica
fume, slag & natural pozzolans in concrete : proceedings : Sixth CANMET/ACI
International Conference : Bangkok, Thailand, SP 178, American Concrete Institute,
Farmington Hills, MI, pp.931-935.
148


Granger, S., Loukili, A., Pijaudier-Cabot, G., and Chanvillard, G. (2007). “Experimental
characterization of the self-healing of cracks in an ultra high performance cementitious
material: Mechanical tests and acoustic emission analysis.” Cement and Concrete
Research, 37, 519-527.

Graybeal, B. (2006a). “Material property characterization of ultra-high performance concrete.”
Rep. No. FHWA-HRT-06-103, Federal Highway Administration, Washington, D.C.

Graybeal, B. (2006b). “Nation’s First Ultra-High Performance Concrete Bridge Opens in Iowa.”
Research & Technology Transporter, FHWA, June 2006.

Graybeal, B.A. (2005). “Characterization of the Behavior of Ultra-High Performance Concrete.”
PhD Dissertation, University of Maryland, College Park, Maryland.

Graybeal, B.A., and Hartmann, J.L. (2003) “Strength and Durability of Ultra-High Performance
Concrete,” 3
rd
International Symposium on High Performance Concrete, PCI, Orlando,
Florida.

Graybeal, B., Hartmann, J., and Perry, V. (2004). “Ultra-High Performance Concrete for
Highway Bridges.” AFGC: Concrete Structures: The Challenge of Creativity, Fib
Symposium 2004, Avignon, France.

Graybeal, B. and Tanesi, J. (2007). “Durability of an Ultrahigh-Performance Concrete.” Journal
of Materials in Civil Engineering, ASCE, 19(10), pp. 848-854.

Hanle, L., Jayaraman, K., and Smith, J. (2004). “CO2 Emissions Profile of the U.S. Cement
Industry.” 13th Annual Emission Inventory Conference, U.S. EPA, Clearwater, FL.

Harris, D.K. (2004). “Characterization of Punching Shear Capacity of Thin UHPC Plates.” MS
Thesis, Virginia Polytechnic Institute and State University, Blacksburg, Virginia.

Harris, D.K. and Roberts-Wollmann, C.L. (2005). “Characterization of the Punching Shear
Capacity of Thin Ultra-High Performance Concrete Slabs.” Virginia Transportation
Research Council and Virginia Department of Transportation, Charlottesville, Virginia.

IDOT (2007). Email to C.G. Gilbertson. 9 August. “Wapello County UHPC Project.”

Hartmann, J. and Graybeal, B. (2001). “U-HPC Testing: Shear,” FHWA Structural Testing
Laboratory.

Hartmann, J. and Graybeal, B. (2002). “Testing of Ultra-High-Performance Concrete Girders.”
PCI Journal, 47(2), January-February 2002, pp. 148-149.

Jacobsen, S. and Sellevold, E.J. (1996). “Self healing of high strength concrete after deterioration
by freeze/thaw.” Cement and Concrete Research, 26(1), 55-62.
149


JSCE (2006) Japan Society of Civil Engineers. “Recommendations for Design and Construction
of Ultra High Strength Fiber Reinforced Concrete Structures (Draft).” Niwa, J., ed.

Jamet, D., Gettu, R., Gopalaratnam, V.S., and Aguado, A. (1995). “Toughness of Fiber-
Reinforced High-Strength Concrete from Notched Beam Tests.” Testing of Fiber
Reinforced Concrete, eds. Stevens, D.J., Banthia, N., Gopalaratnam, V.S., and Tatnall
P.C. American Concrete Institute, Detroit, Michigan, pp. 23-39, 1995.

Jungwirth, J., and Muttoni, A. (2004). “Structural Behavior of Tension Members in Ultra High
Performance Concrete.” International Symposium on UHPC, Kassel, Germany.

Keierleber, B., Bierwagen, D., Fanous, F., Phares, B., and Couture, I. (2007) “Design of
Buchanan County, Iowa, Bridge Using Ultra High Performance Concrete and PI
Girders.” Proceedings of the 2007 Mid-Continent Transportation Research Symposium,
Ames, Iowa, August 2007.

Kollmorgen, G.A. (2004). “Impact of Age and Size on the Mechanical Behavior of and Ultra-
High Performance Concrete.” MS Thesis in Civil Engineering, Michigan Technological
University, Houghton, Michigan.

Kowald, T. (2004). “Influence of surface modified Carbon Nanotubes on Ultra-High
Performance Concrete.” Proceedings of the International Symposium on Ultra High
Performance Concrete, Kassel University Press, Kassel, Germany, pp 195-202.

Lafarge North America (2006a). “Detroit-Columns.”
<http://www.imagineductal.com/imagineductal/public/eng/pdf/Detroit-Columns.pdf>
(March 16, 2008).

Lafarge North America (2006b). “Wapello Bridge.”
<http://www.imagineductal.com/imagineductal/public/eng/pdf/WapelloBridge.pdf>
(March 16, 2008).

Lee, M.-G., Chiu, C.-T., and Wang, Y.-C. (2005). “The study of bond strength and bond
durability of reactive powder concrete.” Journal of ASTM International, 2(7).

Litvan, G.G. (1972). “Phase transitions of adsorbates IV – Mechanism of frost action in hardened
cement paste.” Journal of the American Ceramic Society, 55(1), 38-42.

Lopez-Anido, R. (1998). “The Life-Cycle Cost Evaluation Report of Fiber Reinforced Polymer
(FRP) Composite Bridge Decks.” University of Maine. 2 June.

Loukili, A., Richard, P., and Lamirault, J. (1998). “A Study on Delayed Deformations of an Ultra
High Strength Cementitious Material.” Recent Advances in Concrete Technology :
proceedings : Fourth CANMET/ACI/JCI International Conference : 136 Tokushima,
Japan, SP-179, American Concrete Institute, Farmington Hills, MI, pp. 929-949.
150


Mamlouk, M., and Zaniewski, J. (1999) "Materials for Civil and Construction Engineers."
Addison Wesley Longman, Inc., Menlo Park, CA, pp. 31-192.

Mehta, P.K., and Monteiro, P.J.M. (2006). “Concrete Microstructure, Properties, and Materials.”
The McGraw-Hill Companies, Inc., New York, NY.

MDOT (2007a). “Capital Scheduled Maintenance Bridge Manual.” Michigan.gov. Online. 23
July.

MDOT (2007b). Email to Dr. Theresa Ahlborn. 25 April. “UHPC progress report - real discount
rate and deck timeline.”

MERL (2006). “Michigan Engineers’ Resource Library.” (Database of Michigan bid items that
allows project managers and engineers make road and bridge project estimates).

Mindess, S., Young, J.F., and Darwin, D. (2003). “Concrete 2
nd
Edition.” Pearson Education,
Inc., Upper Saddle River, New Jersey.

Misson, Donald Li (2008) “Influence of Curing Regimes on the Durability of an Ultra-High
Performance Concrete.” MS Thesis in Civil Engineering, Michigan Technological
University, Houghton, Michigan.

Mitchell, L.J. (1953). “Thermal expansion tests on aggregates, neat cement, and concretes.”
Proceedings, ASTM, 53, 963-977.

Moore, B., and Bierwagen, D. (2006), “Ultra High Performance Concrete Highway Bridge.”
CTRE Online. Online. Iowa State University. 26 Feb. 2007.

Office of Management and Budget (2007). “Circular No. A-94 Revised.” Whitehouse.gov.
Online. OMB. 15 July.

Ohio LTAP Center (2007). “Preventative Maintenance/Repair Guidelines for Bridges and
Culverts.” ODOT On-line Bridge Maintenance Manual. Online. 15 July.

Nemegeer, D.E., and Tatnall, P.C. (1995). “Measuring Toughness Characteristics of SFRC-A
Critical Reviwe of ASTM C 1018.” Testing of Fiber Reinforced Concrete, eds. Stevens,
D.J., Banthia, N., Gopalaratnam, V.S., and Tatnall P.C. American Concrete Institute,
Detroit, Michigan, pp. 41-75, 1995.

O’Neil, E.F., Dauriac, C.E, and Gililand, S.K. (1997). “Development of Reactive Powder
Concrete (RPC) Products in the United States Construction Market.” High- Strength
Concrete : An International Perspective : papers presented at three halfday sessions at
the ACI Convention in Montreal, SP-167, American Concrete Institute, Farmington Hills,
MI, pp. 249-261.

151

Perenchio, W. F. (1994). “Corrosion of Reinforcing Steel.” Significance of Tests and Properties
of Concrete and Concrete-Making Materials. STP 169C. Publication Code Number 04-
169030-07. American Society for Testing and Materials, Philadelphia, PA. pp. 164-172.

Perry, V.H. (2003) “The World’s First Long-Span Roof Constructed in Ductal
®
.” Nova Award,
2003, http://www.cif.org/Nom2003/Nom22_03.pdf Viewed 04/17/2007.

Perry, V.H., and Zakariasen, D. (2004). “First Use of Ultra-High Performance Concrete for an
Innovative Train Station Canopy.” Concrete Technology Today, PCA, 25(2).

Perry, V.H., and Zakariasen, D. (2003) “Overview of UHPC Technology, Materials, Properties,
Markets & Manufacturing.” 3
rd
International Symposium on High Performance
Concrete, PCI, Orlando, Florida.

Peuse, Erron J., (2008) “Ultra-High Performance Concrete for Michigan Bridges – Material
Behavior.” MS Thesis in Civil Engineering, Michigan Technological University,
Houghton, Michigan.

Pigeon, M. and Pleau, R. (1995). “Durability of Concrete in Cold Climates.” E & FN Spon,
London, UK, pp. 11-23.

Pine Instrument Company (2006). “AFCT1A / AFCT1C Coefficient of Thermal Expansion of
Portland Cement Concrete Measurement System Operation Manual.” Version 071206,
Grove City, PA.

Powers, T.C. (1945). “A working hypothesis for further studies of frost resistance.” Journal of
the American Concrete Institute, 16(4), 245-272.

Powers, T.C. and Helmuth, R.A. (1953). “Theory of volume changes in hardened portland
cement pastes during freezing.” Proceedings of the Highway Research Board, 29, pp.
184-211.

Rebentrost, M. and Cavill, B. (2006). “Reactive Powder Concrete Bridges.” AustRoads
Conference, Perth, Australia.

Resplendino, J., and Petitjean, J. (2003). “Ultra-High Performance Concrete: First
Recommendations and Examples of Application.” 3
rd
International Symposium on High
Performance Concrete, PCI, Orlando, Florida.

Richard, P., and Cheyrezy, M. (1995). “Composition of reactive Powder Concretes.” Cement and
Concrete Research, Vol. 25, No. 7 October 1995, pp. 1501-1511.

Richard, P. and Cheyrezy, M. (1996). “Reactive Powder Concretes with High Ductility and 200-
800 MPa Compressive Strength,” Concrete Technology: Past, Present, and Future, SP-
144, American Concrete Institute, Farmington Hills, MI, pp. 507-518.

152

RS Means (2005). Heavy Construction Cost Data 19
th
edition. Kingston: Reed, 2004.

Russell, H. (1999). "ACI Defines High-Performance Concrete," Concrete International, 21(2),
February, pp. 56-57.

Schmidt, M., Fehling, E., Teichmann, T., Kai, B., and Roland, B. (2003). “Ultra-High
Performance Concrete: Perspective for the precast concrete industry.” Concrete
Precasting Plant and Technology, 69(3), pp. 16-29.

Schmidt, M. and Fehling, E. (2005). “Ultra-High-Performance Concrete: Research, Development
and Application in Europe.” Seventh International Symposium on the Utilization of High-
Strength/High-Performance Concrete, ACI Special Publication, Washington, D.C., USA,
228, pp 51-78

Searls, D. (2007). “Ductal Concrete by Lafarge.” Concrete Decor,
<http://www.concretedecor.net/All_Access/703/CD703-Technology.cfm> (March 16,
2008).

Semioli, W.J. (2001). “The New Concrete Technology.” Concrete International, 2(11),
November, pp. 75-79.

Sritharan, S., Bristow, B.J., and Perry, V.H. (2003) “Characterizing an Ultra High Performance
Material for Bridge Applications Under Extreme Loads.” 3
rd
International Symposium on
High Performance Concrete, PCI, Orlando, Florida.

Stanish, K.D., Hooton, R.D., and Thomas, M.D.A. (2000). “Testing the Chloride Penetration
Resistance of Concrete: A Literature Review.” FHWA Contract DTFH61-97-R-00022,
University of Toronto, Toronto, Ontario, Canada.

Torres, S., and Eggers, J. (2005). “Capping Systems for High-Strength Concrete.”
Transportation Research Board Annual Meeting CD-ROM, 2006

Toutanji, H., McNeil, S., and Bayasi, Z. (1998). “Chloride Permeability and Impact Resistance
of Polypropylene-Fiber-Reinforced Silica Fume Concrete.” Cement and Concrete
Research, 28(7), pp. 961-968

U.S. Department of Labor 2007. “Bureau of Labor Statistics.” 30 Nov. 2006. U.S. Department of
Labor Online. Online. 24 July. 2007.

Walker, S., Bloem, D.L., and Mullen, W.G. (1952). “Effects of temperature changes on concrete
as influenced by aggregates.” ACI Journal Proceedings, 48(8), 661-680.

Walls, J. III, and Smith, M.R. (1998), “Life-Cycle Cost Analysis in Pavement Design.” FHWA-
SA-98-079. FHWA, Pavement Division: Washington, DC. Sep.

A-1
APPENDIX A: EXPERIMENTAL TEST DATA
Specimen naming scheme ............................................................................................. 2
Compression Strength, Modulus of Elasticity, Poisson’s Ratio, and Flexural
Tests ...................................................................................................................... 2
Rapid Chloride Penetration, Freeze-Thaw Resistance, Coefficient of Thermal
Expansion .............................................................................................................. 3
Table A.1: Data for Air-Cured Cylindrical Specimens ............................................ 5
Table A.2: Data for TT Cylindrical Specimens ........................................................ 6
Table A.3: Data for DTT and DTT Cylindrical Specimens ...................................... 7
Table A.4: Data for Air-Cured Flexural Specimens ................................................. 8
Table A.5: Data for TT Flexural Specimens ............................................................. 9
Table A.6: Data for DTT Flexural Specimens ........................................................ 10
Table A.7: RCPT specimen data sorted by batch .................................................. 11
Table A.8: RCPT specimen data sorted by curing regime ..................................... 12
Figure A.9: Sample freeze-thaw cycle temperature – Position A17 ...................... 13
Figure A.10: Average mass change of UHPC freeze-thaw and side study
specimens ................................................................................................................ 14
Figure A.11: Average length change of UHPC specimens undergoing freeze-thaw
cycling ..................................................................................................................... 15
Table A.12: Michigan Tech CTE Summary Report on UHPC .............................. 16
Table A.13: Mass change study on epoxy coated UHPC CTE specimens ............ 18

A- 2
Specimen naming scheme
This appendix includes data from all specimens tested herein. A naming system was used to
identify each specimen.

Compression Strength, Modulus of Elasticity, Poisson’s Ratio, and Flexural
Tests
The naming system followed the format of Batch number-Specimen geometry-Curing
regime-Testing age-(optional character). The batch number followed the form of
B(number) where the number was 1 through 7 was used with 7 being the batch without
fibers. Specimen geometry was designated by C3 for 3 in. diameter cylinders and B2 for 2
in. x 2 in. x 11.25 in. beams. The curing regime followed the same designation as is laid out
in Table 1.3.1 where A was Air-cured, TT was thermal treatment, DTT was delayed thermal
treatment, and DDTT was double delayed thermal treatment. The testing age designation
was the age in days at which the specimen was to be tested. The optional character was used
only for the 2 in. x 2 in. x 11.25 in. beams because they were the only specimens in which
more than one particular “geometry-curing regime-testing age” came from one batch. In this
case, a letter A or B was included to facilitate record keeping.

An example of the specimen nomenclature would be B3-B2-A-28-B. This specimen would
have been cast out of batch 3, was a 2 in. x 2 in. x 11.25 in. beam, air treated, tested at an age
of 28 days and been the second of the 2 in. x 2 in. x 11.25 in. beams made from the batch.
Another example would be B2-C3-TT-14. This specimen was cast from the second batch, a
A-3
3 in. x 6 in. cylinder, thermally treated, and tested at 14 days. Because no other 3 in. x 6 in.
cylinders thermally treated and being tested at 14 days came from batch 2, the optional A or
B tag was left off the specimen identifier.
Rapid Chloride Penetration, Freeze-Thaw Resistance, Coefficient of Thermal
Expansion
A total of nine batches were cast for testing (with several more cast for specimen shakedown)
and 79 samples were tested to observe the durability characteristics of UHPC (Error!
Reference source not found.). A specimen nomenclature was used to identify batch
number, test procedure, curing regime and testing age. All specimens were marked with the
following nomenclature after demolding: Batch number-Test procedure-(optional side study
label)Curing regime-Testing age(optional letter within batch). Batch numbers began at A
and ended at D for the batches incorporating coefficient of thermal expansion and rapid
chloride penetration test specimens, and began again at M and ended at S (skipping the letter
O label to avoid confusion) for batches comprising of primarily freeze-thaw test specimens.
Supplementary CTE and RCPT specimens were also cast in batches M through S with excess
material. The test procedures were designated on each specimen as CTE for the coefficient
of thermal expansion test, RCP for the rapid chloride penetration test, FT for the freeze-thaw
test, and C for the compression test. Curing regime designations followed the format of TT
for thermal treatment curing and A for ambient air curing (Note: In this report, when
referring to the type of curing regime, the abbreviation “Air” is used. When referring to a
specific specimen’s nomenclature, the entire nomenclature will be used). Additionally, when
a specimen was used for a side study test rather than actual testing, SS was added before the
curing regime notation. So a side study air cured specimen would have the label of SSA
A- 4
rather than A. The final number in the nomenclature specifies the number of days after
casting that a specimen was tested. For example, Specimen X marked N-FT-A-28 would
imply that Specimen X was cast from batch N for the purpose of freeze-thaw testing and was
ambient air cured for 28-days prior to testing. Similarly, Specimen Y marked N-FT-SSA-28
means that Specimen Y was cast from batch N for the purpose of freeze-thaw side study
testing and was ambient cured for 28-days prior to testing. However, when two or more
specimens from a particular batch were cured the same, for the same amount of time, and
tested under the same test (e.g. – companion compression test cylinders), an additional letter
(A – D) was added at the end of the specimen nomenclature. For example, Batch S included
three specimens designated as S-C-TT-28A, S-C-TT-28B, and S-C-TT-28C to distinguish
between the individual specimens. In this study, only companion compression test cylinders
employed this additional nomenclature scheme.


A-5
Table A.1: Data for Air-Cured Cylindrical Specimens
Specimen
Compressive
Stress (ksi)
Modulus of
Elasticity (ksi)
Poisson's
Ratio
Degrees out of
Perpendicularity
B1-C3-A-3 13.517 NR NR 0.327
B2-C3-A-3 14.656 NR NR 0.418
B3-C3-A-3 14.162 6859.481 0.2003 0.155
B4-C3-A-3 14.578 6912.080 0.1900 0.189
B5-C3-A-3 15.098 6954.998 0.2020 0.132
B6-C3-A-3 14.667 6911.809 0.1995 0.195
B1-C3-A-7 20.378 7688.747 0.2068 0.332
B2-C3-A-7 20.141 7493.304 0.2044 0.327
B3-C3-A-7 19.948 7666.874 0.2031 0.218
B4-C3-A-7 19.939 7537.427 0.2064 0.074
B5-C3-A-7 19.281 7356.481 0.2109 0.103
B6-C3-A-7 19.828 7376.915 0.1989 0.120
B1-C3-A-14 22.049 7777.159 0.2085 0.172
B2-C3-A-14 22.727 7899.073 0.2095 0.149
B3-C3-A-14 21.710 7906.809 0.2050 0.149
B4-C3-A-14 21.972 7813.602 0.2020 0.126
B5-C3-A-14 21.796 7850.511 0.2065 0.218
B6-C3-A-14 23.542 7943.388 0.2062 0.160
B1-C3-A-28 24.606 7905.299 0.2114 0.115
B2-C3-A-28 23.378 7727.810 0.1961 0.149
B3-C3-A-28 23.316 8056.596 0.2190 0.223
B4-C3-A-28 24.037 7993.116 0.1999 0.235
B5-C3-A-28 23.934 7736.127 0.2035 0.149
B6-C3-A-28 24.354 7756.123 0.1995 0.212
B7-C3-A-28A 25.585 7795.399 0.1952 0.109
B7-C3-A-28B 24.501 7792.813 0.2022 0.080
B7-C3-A-28C 25.716 7540.652 0.2012 0.109
B7-C3-A-28D 24.742 7668.635 0.2048 0.132
B7-C3-A-28E 25.160 7675.370 0.1980 0.109
B7-C3-A-28F 23.713 7701.819 0.1995 0.092




A- 6
Table A.2: Data for TT Cylindrical Specimens
Specimen
Compressive
Stress (ksi)
Modulus of
Elasticity (ksi)
Poisson's
Ratio
Degrees out of
Perpendicularity
B1-C3-TT-7 30.503 8047.005 0.2012 0.338
B2-C3-TT-7 31.180 8079.182 0.2090 0.418
B3-C3-TT-7 30.152 7879.787 0.2060 0.080
B4-C3-TT-7 28.591 7948.310 0.2135 0.126
B5-C3-TT-7 30.572 8151.638 0.2024 0.120
B6-C3-TT-7 30.676 8231.884 0.2047 0.092
B1-C3-TT-14 30.215 8175.464 0.1968 0.138
B2-C3-TT-14 32.014 8359.577 0.2064 0.080
B3-C3-TT-14 30.081 8155.241 0.2063 0.172
B4-C3-TT-14 28.153 8127.066 0.2124 0.103
B5-C3-TT-14 29.046 8127.497 0.2063 0.115
B6-C3-TT-14 31.000 8348.119 0.2064 0.120
B1-C3-TT-28 31.095 8166.650 0.2068 0.138
B2-C3-TT-28 30.845 8080.693 0.2050 0.115
B3-C3-TT-28 30.808 8124.252 0.2119 0.115
B4-C3-TT-28 30.994 8048.575 0.2046 0.143
B5-C3-TT-28 31.884 8198.986 0.2025 0.138
B6-C3-TT-28 30.944 8066.015 0.1991 0.126
B7-C3-TT-28A 33.025 7793.664 0.1968 0.052
B7-C3-TT-28B 29.456 7844.217 0.1992 0.074
B7-C3-TT-28C 30.154 7858.390 0.2096 0.097
B7-C3-TT-28D 32.667 7929.705 0.2054 0.103
B7-C3-TT-28E 34.063 8020.766 0.2029 0.103








A-7
Table A.3: Data for DTT and DDTT Cylindrical Specimens
Specimen
Compressive
Stress (ksi)
Modulus of
Elasticity (ksi)
Poisson's
Ratio
Degrees out of
Perpendicularity
B1-C3-DTT-14 29.859 9003.986 0.2233 0.097
B2-C3-DTT-14 29.712 8098.644 0.2115 0.166
B3-C3-DTT-14 28.572 8082.995 0.1987 0.097
B4-C3-DTT-14 28.869 8393.630 0.2078 0.092
B5-C3-DTT-14 30.795 8180.276 0.2005 0.103
B6-C3-DTT-14 30.740 8128.955 0.2063 0.166
B1-C3-DTT-28 30.424 8181.310 0.2085 0.080
B2-C3-DTT-28 29.375 8172.989 0.1977 0.057
B3-C3-DTT-28 29.210 8096.720 0.2226 0.086
B4-C3-DTT-28 29.420 8144.191 0.2091 0.086
B5-C3-DTT-28 30.378 8129.613 0.1980 0.120
B6-C3-DTT-28 30.761 8239.005 0.2085 0.115
B7-DTT-28A 24.785 7786.891 0.1995 0.109
B7-DTT-28B 32.340 7807.098 0.2025 0.126
B7-DTT-28C 32.278 7872.662 0.1988 0.103
B7-DTT-28D 30.282 7544.063 0.1984 0.155
B7-DTT-28E 30.015 7979.124 0.1994 0.086
B1-C3-DDTT-28 28.098 8178.926 0.2025 0.298
B2-C3-DDTT-28 30.706 8121.966 0.2031 0.109
B3-C3-DDTT-28 29.154 8164.106 0.2012 0.103
B4-C3-DDTT-28 29.726 8001.333 0.2025 0.097
B5-C3-DDTT-28 29.431 8026.161 0.2065 0.149








A- 8
Table A.4: Data for Air-Cured Flexural Specimens
Specimen
Corrected
First Crack
Strength (ksi)
First Crack
Strength
(ksi)
Deflection at
First Crack
(in.)
Equivalent
Flexural
Strength (ksi)
Ultimate
Load
(kip)
I
5
I
10
I
20
I
30
I
40
R
5,10
R
10,20
R
20,30
R
30,40
B1-B2-A-28A 0.778 1.402 0.00190 4.862 4.322 6.8 17.4 44.2 74.2 107.4 212 268 300 332
B1-B2-A-28B 0.757 1.365 0.00183 4.637 4.225 6.9 18.5 46.8 77.1 109.5 232 284 302 325
B2-B2-A-28A 0.772 1.391 0.00176 5.482 4.994 6.9 18.0 46.7 78.5 113.4 221 288 318 348
B2-B2-A-28B 0.772 1.392 0.00182 4.603 3.990 6.7 17.7 45.2 76.2 108.8 221 275 310 326
B3-B2-A-28A 0.716 1.291 0.00168 4.643 4.127 6.9 18.0 46.4 78.8 112.1 222 285 323 333
B3-B2-A-28B 0.566 1.021 0.00146 3.575 3.724 6.9 17.6 44.9 76.6 109.9 214 273 317 334
B4-B2-A-28A 0.700 1.262 0.00163 4.878 4.445 6.8 18.1 46.9 79.5 114.2 225 288 327 347
B4-B2-A-28B 0.642 1.157 0.00168 4.862 4.770 7.3 19.3 49.4 84.2 121.1 238 302 348 369
B5-B2-A-28A 0.747 1.348 0.00198 4.213 4.069 6.9 17.4 42.7 70.3 100.4 211 253 277 300
B5-B2-A-28B 0.823 1.484 0.00185 4.308 3.829 6.1 16.3 40.8 67.5 95.3 204 245 267 278
B6-B2-A-28A 0.792 1.429 0.00186 4.308 3.829 6.6 16.7 41.3 67.4 93.6 202 246 261 262
B6-B2-A-28B 0.873 1.574 0.00203 4.626 4.215 6.6 16.5 41.2 68.7 97.7 198 247 275 290
B7-B2-A-28A 0.888 1.601 0.00215
B7-B2-A-28B 0.741 1.336 0.00198
B7-B2-A-28C 0.864 1.559 0.00203
Not Applicable
Not Applicable
Not Applicable






A-9
Table A.5: Data for TT Flexural Specimens

Specimen
Corrected
First Crack
Strength (ksi)
First Crack
Strength
(ksi)
Deflection at
First Crack
(in.)
Equivalent
Flexural
Strength (ksi)
Ultimate
Load
(kip)
I
5
I
10
I
20
I
30
I
40
R
5,10
R
10,20
R
20,30
R
30,40
B1-B2-TT-28A 1.078 1.943 0.00248 5.764 5.124 6.6 16.0 39.3 65.2 92.9 188 233 258 277
B1-B2-TT-28B 1.101 1.986 0.00263 5.367 4.770 6.8 16.5 40.1 66.0 91.9 194 236 259 259
B2-B2-TT-28A 1.038 1.873 0.00240 5.644 5.017 6.5 16.4 40.6 67.4 NR 199 242 269 NR
B2-B2-TT-28B 0.967 1.744 0.00247 5.197 4.619 7.0 17.5 42.6 70.4 99.4 211 250 278 291
B3-B2-TT-28A 0.967 1.743 0.00230 5.230 4.649 6.5 16.5 40.4 67.1 95.7 200 238 268 286
B3-B2-TT-28B 1.042 1.878 0.00233 4.987 4.433 6.0 14.8 35.9 59.7 85.3 177 211 238 257
B4-B2-TT-28A 0.965 1.739 0.00203 5.622 4.997 6.7 16.8 42.1 69.6 98.9 202 253 275 293
B4-B2-TT-28B 0.989 1.783 0.00229 5.444 4.960 6.7 16.9 42.6 70.5 99.3 204 257 279 288
B5-B2-TT-28A 1.187 2.140 0.00264 4.978 4.314 6.5 15.7 37.2 59.2 NR 185 215 220 NR
B5-B2-TT-28B 1.209 2.180 0.00258 6.124 5.307 6.2 15.4 38.1 62.6 88.8 184 227 246 262
B6-B2-TT-28A 1.134 2.045 0.00238 5.348 4.873 6.1 15.0 36.0 59.4 84.3 178 210 234 248
B6-B2-TT-28B 1.047 1.889 0.00226 5.539 4.924 6.2 15.9 38.8 63.8 91.1 194 229 250 272
B7-B2-TT-28A 1.051 1.896 0.00253
B7-B2-TT-28B 1.123 2.026 0.00284
B7-B2-TT-28C 1.201 2.166 0.00288
Not Applicable
Not Applicable
Not Applicable




A- 10
Table A.6: Data for DTT Flexural Specimens

Specimen
Corrected
First Crack
Strength (ksi)
First Crack
Strength
(ksi)
Deflection at
First Crack
(in.)
Equivalent
Flexural
Strength (ksi)
Ultimate
Load
(kip)
I
5
I
10
I
20
I
30
I
40
R
5,10
R
10,20
R
20,30
R
30,40
B1-B2-DTT-28A 1.219 2.197 0.00277 NR NR 6.5 16.1 39.2 NR NR 192 231 NR NR
B1-B2-DTT-28B 1.184 2.136 0.00276 5.773 5.132 6.6 16.1 39.0 64.0 90.7 192 229 249 267
B2-B2-DTT-28A 1.161 2.094 0.00256 5.185 4.609 6.5 15.5 37.5 61.4 NR 181 220 239 NR
B2-B2-DTT-28B 1.133 2.043 0.00249 4.575 4.067 6.0 14.7 35.3 56.9 NR 175 205 217 NR
B3-B2-DTT-28A 1.013 1.827 0.00247 4.534 4.365 6.3 15.2 37.1 60.7 84.7 178 219 236 239
B3-B2-DTT-28B 0.961 1.733 0.00267 4.832 4.768 6.7 16.7 40.2 66.2 93.5 200 235 259 273
B4-B2-DTT-28A 1.136 2.049 0.00271 4.481 4.083 6.1 14.8 35.2 56.2 76.7 174 204 209 205
B4-B2-DTT-28B 1.168 2.106 0.00267 4.732 4.311 6.2 15.3 35.7 57.3 79.3 183 204 215 220
B5-B2-DTT-28A 1.284 2.315 0.00285 4.475 4.077 6.2 14.7 33.4 51.4 NR 171 187 180 NR
B5-B2-DTT-28B 1.250 2.255 0.00287 5.227 4.646 6.1 14.8 34.7 55.9 NR 175 199 212 NR
B6-B2-DTT-28A 1.305 2.354 0.00295 5.405 4.925 6.1 14.8 34.5 55.7 NR 175 197 212 NR
B6-B2-DTT-28B 1.288 2.322 0.00299 5.642 5.015 6.4 15.6 36.8 59.9 NR 183 213 231 NR
B7-B2-DTT-28A 1.142 2.059 0.00282
B7-B2-DTT-28B 1.187 2.141 0.00314
B7-B2-DTT-28C 1.293 2.331 0.00324
Not Applicable
Not Applicable
Not Applicable





A-11
Table A.7: RCPT Specimen Data Sorted by Batch
Specimen ID Age
Charge Passed
(coulombs)
Chloride Ion
Penetrability
1A-RCP-TT-7 7 12 Negligible
1A-RCP-A-28 28 93 Negligible
1A-RCP-TT-28 28 16 Negligible

1B-RCP-TT-7 7 9 Negligible
1B-RCP-A-28 28 78 Negligible
1B-RCP-TT-28 28 13 Negligible

1C-RCP-A-28 28 57 Negligible
1C-RCP-TT-28 28 11 Negligible

D-RCP-A-28 28 73 Negligible
D-RCP-TT-28 28 19 Negligible

S-RCP-TT-7 8 10 Negligible


A- 12
Table A.8: RCPT Specimen Data Sorted by Curing Regime
Specimen ID
Specimen age at time
of testing (days)
Charge Passed
(coulombs)
Chloride Ion
Penetrability
1A-RCP-TT-7 7 12 Negligible
1B-RCP-TT-7 7 9 Negligible
S-RCP-TT-7 8 10 Negligible
Average 10
St. Dev. 1.5
COV (%) 15

1A-RCP-TT-28 28 16 Negligible
1B-RCP-TT-28 28 13 Negligible
1C-RCP-TT-28 28 11 Negligible
D-RCP-TT-28 28 19 Negligible
Average 15
St. Dev. 3.5
COV (%) 24

1A-RCP-A-28 28 93 Negligible
1B-RCP-A-28 28 78 Negligible
1C-RCP-A-28 28 57 Negligible
D-RCP-TT-28 28 73 Negligible
Average 75
St. Dev. 15
COV (%) 20

A-13
Figure A.9: Sample Freeze-Thaw Cycle Temperature – Position A17

A- 14
Figure A.10: Average Mass Change of UHPC Freeze-Thaw and Side Study Specimens






A-15

Figure A.11: Average Length Change of UHPC Specimens Undergoing Freeze-Thaw Cycling

A- 16
Table A.12: Michigan Tech CTE Summary Report on UHPC
Date Core
Frame
S/N L
o
ΔL
m1
ΔL
m2
CTE
1
CTE
2

CTE
avg

(
o
C)
CTE
avg

(
o
F)
Not tested
on correct
day
5/22/2007 1A-CTE-A-3 1135 175.49
-
0.1050 0.1067 13.3E-06 13.5E-06 13.4E-06 7.44E-06
5/27/2007 1A-CTE-A-7 1135 177.70 0.1067 -0.1056 14.0E-06 13.8E-06 13.9E-06 7.73E-06 x
6/3/2007 1A-CTE-A-14 1135 177.20 0.1077 -0.1077 14.0E-06 14.0E-06 14.0E-06 7.79E-06
6/15/2007 1A-CTE-A-28 1135 177.99 0.1040 -0.1040 13.7E-06 13.7E-06 13.7E-06 7.59E-06
6/16/2007 1A-CTE-TT-28 1135 177.80 0.1088 -0.1093 14.3E-06 14.4E-06 14.3E-06 7.97E-06

5/26/2007 1B-CTE-A-3 1135 178.12
-
0.1029 0.1040 13.5E-06 13.7E-06 13.6E-06 7.56E-06
5/30/2007 1B-CTE-A-7 1135 177.75 0.1023 -0.1023 13.4E-06 13.4E-06 13.4E-06 7.44E-06 x
6/5/2007 1B-CTE-A-14 1135 178.76 0.1018 -0.1023 13.5E-06 13.6E-06 13.6E-06 7.53E-06
6/19/2007 1B-CTE-A-28 1135 177.82
-
0.1072 0.1083 14.1E-06 14.2E-06 14.2E-06 7.89E-06

5/29/2007 1C-CTE-A-3 1135 176.73 0.1066 -0.1050 13.8E-06 13.5E-06 13.7E-06 7.59E-06
6/1/2007 1C-CTE-A-7 1135 177.25
-
0.1061 0.1067 13.8E-06 13.9E-06 13.8E-06 7.69E-06
6/8/2007 1C-CTE-A-14 1135 177.03 0.1067 -0.1077 13.8E-06 14.0E-06 13.9E-06 7.73E-06
6/22/2007 1C-CTE-A-28 1135 178.07 0.1061 -0.1050 14.0E-06 13.8E-06 13.9E-06 7.73E-06
6/24/2007 1C-CTE-TT-28 1135 177.64 0.1120 -0.1137 14.7E-06 15.0E-06 14.8E-06 8.25E-06 x

6/6/2007 1D-CTE-A-7 1135 177.05 0.1061 -0.1050 13.8E-06 13.6E-06 13.7E-06 7.61E-06

7/3/2007 M-CTE-A-7 1135 176.09 0.1066 -0.1056 13.6E-06 13.5E-06 13.6E-06 7.53E-06
7/23/2007 M-CTE-TT-28 1135 177.80 0.1142 -0.1126 15.1E-06 14.8E-06 15.0E-06 8.31E-06

A-17

Table A.12 (continued): Michigan Tech CTE Summary Report on UHPC
Date Core
Frame
S/N L
o
ΔL
m1
ΔL
m2
CTE
1
CTE
2

CTE
avg

(
o
C)
CTE
avg

(
o
F)
Not tested
on correct
day
7/6/2007 N-CTE-A-7 1135 176.36 -0.1083 0.1072 13.9E-06 13.8E-06 13.9E-06 7.69E-06
7/26/2007 N-CTE-TT-28 1135 177.09 0.1137 -0.1126 14.8E-06 14.7E-06 14.8E-06 8.21E-06

7/11/2007 P-CTE-TT-7 1135 178.41 0.1110 -0.1110 14.7E-06 14.7E-06 14.7E-06 8.19E-06

7/14/2007 R-CTE-TT-7 1135 177.85 0.1110 -0.1115 14.6E-06 14.7E-06 14.7E-06 8.14E-06

7/17/2007 S-CTE-TT-7 1135 177.58 0.1131 -0.1131 14.9E-06 14.9E-06 14.9E-06 8.26E-06




A- 18
Table A.13: Mass change study on epoxy coated UHPC CTE specimens
Air-cured UHPC Specimens Thermally-treated UHPC Specimens
Specimen
Age of
specimen at
testing (days)
Mass Change
(%) Specimen
Age of specimen
at testing (days)
Mass Change
(%)
1A-CTE-A-3 3 0.00 1A-CTE-TT-28 28 0.03
1A-CTE-A-7 7 0.00 1C-CTE-TT-28 28 0.03
1A-CTE-A-14 14 0.05 M-CTE-TT-28 28 N/A
1A-CTE-A-28 28 0.03 N-CTE-TT-28 28 N/A
1B-CTE-A-3 3 N/A P-CTE-TT-7 7 0.05
1B-CTE-A-7 7 0.05 R-CTE-TT-7 7 0.00
1B-CTE-A-14 14 0.00 S-CTE-TT-7 7 0.01
1B-CTE-A-28 28 0.03 Mean 0.02
1C-CTE-A-3 3 N/A Standard Dev. 0.02
1C-CTE-A-7 7 0.03 COV (%) 84.94
1C-CTE-A-14 14 0.03
1C-CTE-A-28 28 0.03
D-CTE-A-7 7 0.03
D-CTE-A-28 28 0.03
M-CTE-A-7 7 0.03
N-CTE-A-7 7 0.03
Mean 0.03
Standard Dev. 0.02
COV (%) 66.44

B-1

APPENDIX B – CTE Test Procedure Modifications




B-2






B-3






B-4



















(This page intentionally left blank)

(This page intentionally left blank)

Technical Report Documentation Page 1. Report No. 2. Government Accession No. Research Report RC-1525 4. Title and Subtitle Ultra-High-Performance-Concrete for Michigan Bridges Material Performance – Phase I 7. Author(s) Dr. Theresa M. Ahlborn, Mr. Erron J. Peuse, Mr. Donald Li Misson 9. Performing Organization Name and Address Center for Structural Durability Michigan Technological University 1400 Townsend Drive Houghton MI 49931-1295 12. Sponsoring Agency Name and Address Michigan Department of Transportation Construction and Technology Division PO Box 30049 Lansing MI 48909 15. Supplementary Notes

3. MDOT Project Manager Roger Till, P.E. 5. Report Date November 13, 2008 6. Performing Organization Code MTU 8. Performing Org Report No. CSD-2008-11

10. Work Unit No. (TRAIS) 11. Contract Number 2003-0063 11(a). Authorization Number Auth 21 Rev.1 13. Type of Report and Period Covered Final Report 14. Sponsoring Agency Code

16. Abstract One of the latest advancements in concrete technology is Ultra-High Performance Concrete (UHPC). UHPC is defined as concretes attaining compressive strengths exceeding 25 ksi (175 MPa). It is a fiber-reinforced, denselypacked concrete material which exhibits increased mechanical performance and superior durability to normal and high strength concretes. UHPC has great potential to be used in the bridge market in the United States. However, to gain acceptance by designers, contractors, and owners this material needs to be tested according to American Society for Testing and Materials (ASTM) International and American Association of State Highway Transportation Officials (AASHTO) standards, and new practices must be developed. The focus of this research was to investigate how the age at which UHPC undergoes a steam (thermal) treatment affects some mechanical and durability properties. Four mechanical properties (compressive strength, modulus of elasticity, Poisson’s ratio, and flexural characteristics) and properties related to durability (chloride ion penetration resistance, freeze-thaw durability, and coefficient of thermal expansion) were investigated. The testing was conducted with differing curing conditions and at different ages to examine how these factors influence each of the measured properties. Specimens, independent of age at thermal treatment, yielded compressive strengths of over 30 ksi, modulus of elasticity values in excess of 8000 ksi, and a Poisson’s ratio of 0.21. Flexural characteristics were dependent on curing regime. Testing consistently validated that UHPC had negligible chloride ion penetration, a high resistance to freeze-thaw cycling (durability factor of 100), and coefficient of thermal expansion values similar to that of normal strength concretes for both ambient cured and thermally treated specimens. Additional results revealed UHPC’s autogenous healing properties while undergoing freeze-thaw cycling, low variability between batches, and the reproducibility of results between different U.S. laboratories. Lastly, recommendations were developed for future testing of UHPC durability properties and a preliminary lifecycle cost comparison showed that the low life-maintenance costs of UHPC can offset higher initial costs, especially as the use of UHPC in the U.S. increases and the initial cost of the material decreases. 17. Key Words: 18. Distribution Statement Ultra High Performance Concrete, UHPC, Bridge Materials, No restrictions. This document is Compressive Strength, Modulus, Poisson’s Ratio, Flexure, Rapid available to the public through the Chloride Penetration, Freeze-Thaw, Coefficient of Thermal Michigan Department of Expansion, Life Cycle Cost Transportation. 19. Security Classification (report) 20. Security Classification (Page) 21. No of Pages 22. Price Unclassified Unclassified 181 Report RC-1525

(This page intentionally left blank) .

Associate Professor and CSD Director 906/487-2625 tess@mtu. Puese and Mr.E. Donald Li Misson Former Graduate Research Assistants .edu Mr. Ahlborn. Michigan 49931 Fax: 906/487-1620 Dr. 1400 Townsend Dr. Dept.Ultra-High-Performance-Concrete for Michigan Bridges Material Performance – Phase I Submitted by the Michigan Tech CENTER FOR STRUCTURAL DURABILITY A Michigan DOT Center of Excellence Submitted to: Final Report – November 2008 Submitted by: Michigan Technological University Civil & Environmental Eng. Houghton. Erron J. Theresa M. P.

(This page intentionally left blank) .

undergraduate research assistant for assistance with the preliminary cost-benefit study. for oversight of the experimental studies.ACKNOWLEDGEMENTS This project was financially supported by the Michigan Department of Transportation in cooperation with the Federal Highway Administration. This document is disseminated under the sponsorship of the Michigan Department of Transportation in the interest of information exchange. MTTI Operations Manager. suggestions. for technical editing of the final report. Kari Klaboe.E. The authors would also like to acknowledge the contributions of Mr. P.. and Mr. DISCLAIMER The content of this report reflects the views of the authors. . Research Engineer. Chris Gilbertson. P. Charles Mott. The authors would like to thank the members of the Michigan Department of Transportation (MDOT) Research Advisory Panel (RAP).. The Michigan Department of Transportation assumes no liability for the content of this report of its use thereof.E. who are responsible for the facts and accuracy of the information presented herein. Roger Till. including Project Manager Mr. and patience throughout the course of the project. Ms. for their guidance.

(This page intentionally left blank) .

The testing was conducted with differing curing conditions and at different ages to examine how these factors influence each of the measured properties.21. contractors. recommendations were developed for future testing of UHPC durability properties and for a future design code. Flexural characteristics were dependent on curing regime. modulus of elasticity. Specimens. The focus of this research was to investigate how the age at which UHPC undergoes a steam (thermal) treatment affects some mechanical and durability properties. low variability between batches. and new practices must be developed. UHPC has great potential to be used in the bridge market in the United States. Testing consistently validated that UHPC had negligible chloride ion penetration.S. and a preliminary life-cycle cost comparison showed that the low life-maintenance costs of UHPC can offset higher initial costs. and coefficient of thermal expansion) were investigated. Poisson’s ratio. a high resistance to freeze-thaw cycling (durability factor of 100).S. independent of age at thermal treatment. increases and the initial cost of the material decreases. It is a fiber-reinforced. and the reproducibility of results between different U. and coefficient of thermal expansion values similar to that of normal strength concretes for both ambient cured and thermally treated specimens. and flexural characteristics) and properties related to durability (chloride ion penetration resistance. Additional results revealed UHPC’s autogenous healing properties while undergoing freeze-thaw cycling. Four mechanical properties (compressive strength.Abstract One of the latest advancements in concrete technology is Ultra-High Performance Concrete (UHPC). densely-packed concrete material which exhibits increased mechanical performance and superior durability to normal and high strength concretes. UHPC is defined as concretes attaining compressive strengths exceeding 25 ksi. i . especially as the use of UHPC in the U. However. and owners this material needs to be tested according to American Society for Testing and Materials (ASTM) International and American Association of State Highway Transportation Officials (AASHTO) standards. laboratories. freezethaw durability. yielded compressive strengths of over 30 ksi. Lastly. and a Poisson’s ratio of 0. modulus of elasticity values in excess of 8000 ksi. to gain acceptance by designers.

(This page intentionally left blank) ii .

............................................................................. 23  Freeze-Thaw Testing ...................................................... 3  1.... 8  2...2  2........ 33  3....... 40  3.........................................................0  Review of UHPC ........................................ viii  1. 20  Thermal Treatment ............................... 15  2...................................................................... 25  Coefficient of Thermal Expansion ......................................... iii  List of Figures ............................................................................3  Casting Specimens ........................................................................................5.....................................................4.............................................................................................................................................................................................................................................................1  2................................................................................................................4  Chloride Ion Penetration ............................ 21  2.....................5......................2  Types of UHPC ....................................................................................5  Durability Improvements .......................... 33  3..3  Scope ..........................................6  Other UHPC research ....................................................................................................................................................................................................................................... 29  3....................................4  Curing Regimes ............ i  Contents ............................................. vi  List of Tables ......... 4  2.........5... 39  3.......................................................2  2....................... 10  2............................................................................................................................... 22  2.....4.....1  UHPC Composition.................................. 5  2...............................................1  Introduction ................................................ 17  First-Crack Flexural Strength and Flexural Toughness .............................3  2................2  UHPC Mixing Procedure ..........................................................3  Applications of UHPC ......Contents Abstract ...................................................5  Specimen Preparation and Test Procedures ................................................................. 27  Additional Durability Research .......4..................................... 34  3........0  Introduction to Ultra-High Performance Concrete (UHPC) ..................4  Compressive Strength ...4  Mechanical Properties ..3  2.............................................................................................5.......................0  Methodology ..................................................................................................................... 29  2........................................ 1  1........................................................... 15  Modulus of Elasticity and Poisson’s Ratio .......... 41  iii ... 10  2.4......2  Objectives ............................................................1  2.

......................................4.............................................. 116  iv .................................................................3 Flexural Strength ......6..... 85  4........................................ 81  4............... 53  3......................................... 42  3......3  Air-Cured Compressive Strength Growth over Time .........................5 Freeze-Thaw Cyclic Testing ......1  Results ................... 98  4.........................................................................0  Results and Discussion ..................3................................................... 87  4............. 64  4.. 70  4........................................2.......... 108  4.....................................1.. 46  3.........2  Discussion ... 94  4...........................................................2..3  Flexural Strength Testing for First Cracking .................................. 98  4.......................................................................................................6 Coefficient of Thermal Expansion ............................................................................................ 116  5.........................................................................................................1  Compression Strength .......5.......................................4 Rapid Chloride Penetration ........1  Results ..........2  Statistical Analysis and Discussion .............1  Compression Strength ............... 95  4. 48  3..............5.....................5............................6................0  Conclusions of the Experimental Studies ...................5................................................. 115  5.................................. 95  4.......................................................................................................................................5.........1 Compressive Strength ..............2 Modulus of Elasticity and Poisson’s Ratio ...................... 63  4..........2  Discussion ..................................................................6  Coefficient of Thermal Expansion ...... 76  4............................................ 64  4.....1  Results ...................................... 99  4............................................................................................................................................................5  Freeze-Thaw Cyclic Testing .............................................. 72  4.....................3  Compressive Stress and Modulus of Elasticity Relationship ..............1.... 84  4...5.............3................ 113  5.................. 89  4......................................................................................... 57  4....................... 108  4...........2  Statistical Analysis and Discussion ........ 110  4......................................... 45  3.3................2  Statistical Analysis and Discussion ...........5...........4  Rapid Chloride Penetration Test ...........................................4..6.... 73  4....................................................................................................1  Results .............................................1  Results ..........................................................3  Study of water absorption ...1  Results ...............................2  Discussion .................3....2.............................. 66  4..................................................................................5....2  Modulus of Elasticity and Poisson’s Ratio .....1.............2  Modulus of Elasticity and Poisson’s Ratio .............3  Flexural Toughness ...................................................

...................................................................3  Maintenance ................5  7...........2  Draft U........ 134  7..................6  Future Work ....................................................... 129  7.......... B-1  v ....................................................................................4  Suggested Future Work .....2  7....1.......................................................................................................................................................... 131  7.............................. 118  5............6  Compression Testing ... 122  6.....................................................................................1......... Design Recommendations for UHPC ......................3  7..........1... 141  References ............................................1.............. A-1  Appendix B – CTE Test Procedure Modifications ....................S..0  Preliminary Life Cycle Costs of a UHPC Superstructure ................................................................................................................3  Implementation Activities ................. 121  6.......................5  Freeze-Thaw Cyclic Testing .. 133  Coefficient of Thermal Expansion ................................ 132  Freeze-Thaw Cyclic Testing ............... 133  7...........................4  Preliminary Life Cycle Costs .............5  Conclusion of the Preliminary Life Cycle Cost Analysis ...........0  Recommendations.....1  7............................................................ 132  Rapid Chloride Penetration Test ....................................................... 118  5............................................... 131  Modulus of Elasticity and Poisson’s Ratio .4  7...................................................... 121  6.................................................................1  Bridge Components ....................... 124  6.......................6  Coefficient of Thermal Expansion ...... 131  7.... 128  6.........1...................... 117  5..3  Flexural Strength Testing for First Cracking and Toughness .........................................................1  Recommendations for UHPC Testing Procedures ..................................................................................... 145  A  B  Appendix A – Experimental Test Data ........................................ 131  Flexural Strength Testing for First Cracking and Flexural Toughness ................................................. 119  6.... 126  6..........................5................................................................................................................ Implementation and Future Work .......... 140  7...1........4  Rapid Chloride Penetration Test ......................2  Construction .....

.List of Figures Figure 2-1: UHPC Example: Sherbrooke Footbridge (Resplendino and Petitjean 2003) ............ 102  Figure 4-14: Average Resonant Frequencies of UHPC Freeze-Thaw and Side-Study Specimens ................................................................................ 87  Figure 4-9: Load Deflection Curve for Elastic-Plastic Material (ASTM C 1018 Figure X1....................................................................1) . 55  Figure 3-15: Length Change Measurement of an UHPC Freeze-Thaw Specimen .................... 56  Figure 3-16: Epoxy-coating CTE Specimen .............................................................................................. 66  Figure 4-2: Compressive Stress Gain over Time for Air-Cured Specimens ............................... 44  Figure 3-8: Compressometer and Extensometer ..................................................... 83  Figure 4-8: Mean First-Crack Flexural Stress for All Curing Regimes......................................................................................................... 97  Figure 4-12: Effects of Freeze-Thaw Cycling on the Average Relative Dynamic Modulus of UHPC Samples .......................................... 106  vi ............................... 74  Figure 4-4: Mean Modulus of Elasticity Results for All Ages and Curing Regimes ..................... 46  Figure 3-9: ASTM 1018 Loading Configuration .................................................................................................................................................................... Iowa (Lafarge 2006b) ....................................... 50  Figure 3-12: UHPC Specimen Undergoing ASTM C 1202 Testing ......................... 91  Figure 4-11: Surface Staining of UHPC Specimen after ASTM C 1202 Test ................................... 35  Figure 3-2: Turning Point of UHPC ............................................... 14  Figure 2-4: UHPC Girder Testing........................ 71  Figure 4-3: Typical Stress-Strain Curve for Calculating the Modulus of Elasticity............. 51  Figure 3-13: MTU 80-specimen Freeze-Thaw Chamber (Procedure B) ..... 13  Figure 2-3: UHPC Construction Examples.. 104  Figure 4-15: Resonant Frequencies of UHPC Specimens after Freeze-Thaw Testing ......... 90  Figure 4-10: Typical Load Deflection Curve for Flexural Specimens ................................................................... 54  Figure 3-14: Testing the Fundamental Transverse Frequency of an UHPC Specimen .................................... 101  Figure 4-13: Cracks in Air Cured UHPC Specimens Following Freeze-Thaw Testing ........................ 30  Figure 2-5: Mars Hill Bridge in Wapello County...................................................................... 75  Figure 4-6: Regression Model for Modulus of Elasticity vs...................................................................................................................................................................................................................................................................................................................................... 39  Figure 3-4: Michigan Tech’s UHPC Thermal Treatment Cure Chamber .............................................. 42  Figure 3-6: End Perpendicularity Set-up ............................. 11  Figure 2-2: UHPC Footbridges ..................... 75  Figure 4-5: Mean Poisson’s Ratio Results for All Ages and Curing Regimes ................................................ 82  Figure 4-7: Mean Values of Compressive Stress and Modulus of Elasticity for Air-Cured Specimens ......................................... 37  Figure 3-3: Impact Table Measurement of UHPC’s Flow....................................................................... 60  Figure 3-18: UHPC Specimen in Water Bath Undergoing CTE Testing ....................................... Compressive Strength ......................................................... 41  Figure 3-5: Reid Surface Grinder ................ 31  Figure 3-1: Doyon Mixer .............................................................. 105  Figure 4-16: Resonant Frequencies of UHPC Specimens after Side-Study Testing ................................. 43  Figure 3-7: Baldwin CT 300 Compression Testing Machine .............................................................. 61  Figure 4-1: Mean Compressive Results for All Ages and Curing Regimes . 48  Figure 3-10: Epoxy-coated UHPC Specimens for RCPT ....................................................................................................................... 50  Figure 3-11: ASTM C 1202 Specimen Preparation Setup.... 59  Figure 3-17: Pine CTE Specimen Test Frame and Water Bath .................

.........................................................................................Figure 4-17: Typical Bell Shaped Resonant Frequency Output of Air-cured UHPC Specimen (Frequency in Hz) ............................................................................. 107  Figure 4-19: Average CTE Values for Air-cured UHPC Specimens ...... 128  vii ...................................... 107  Figure 4-18: Typical Skewed Resonant Frequency Output of an Air-cured UHPC Specimen Six Months after Freeze-Thaw Testing (Frequency in Hz)................. 111  Figure 6-1: Target Cost of UHPC ....................................................................

...............................................1) ............. 52  Table 4.... 63  Table 4........................................... 125  Table 6.13: Flexural Stress............. 2007 $ ..................................................................3: Flow Domain Classifications of Freshly Mixed UHPC ........ 65  Table 4............Specimens Tested per Curing Regime ...........................List of Tables Table 2...12: Statistical Results Poisson’s Ratio Testing .............9: Statistical Results for Combined Modulus of Elasticity Testing ..... 8  Table 3.......5: Bridge Girder Maintenance ................................... 35  Table 3....................10: Statistical Results for Poisson’s Ratio Testing .......................................................... 69  Table 4............................ 110  Table 4.....21: Coefficient of Thermal Expansion (CTE) Test Summary ..........8: Combined Modulus of Elasticity Results .....................4: Combined Compressive Stress Results ...... 69  Table 4.6: Modulus of Elasticity and Poisson’s Ratio Test Results ............5: Statistical Results for Combined Compressive Strength Testing ..................... 96  Table 4..... 78  Table 4.............6: Bridge Deck Maintenance . Deflection and Maximum Load Results .....................20: Change in Resonant Frequency of UHPC Specimens after Testing Completed .............................. 73  Table 4....................................... 126  Table 6....................................................................................................................................... 123  Table 6.................. 95  Table 4...........................................................................16: Experimental Toughness Indices and Residual Strength Factors ................................1: Comparison of UHPC Material Properties to Other Concrete Classifications .................. 68  Table 4.... 124  Table 6...... 105  Table 4..... 99  Table 4......................................4: Unit Costs of Maintenance Activities .................................... 122  Table 6.....................65 ft3 batch .............19: Effects of Freeze-Thaw Cycles on UHPC .............................................2: Composition of a Typical UHPC Mix ................................ 78  Table 4............................4: Chloride Ion Penetrability Based on Charge Passed (ASTM C 1202) ...........3: Statistical Results for Compressive Strength Testing ....................................1: Bridge Component Unit Costs ..................................................................................................... 127  viii .......................... 86  Table 4...........7: Statistical Results for Modulus of Elasticity Testing .... 39  Table 3........................18: Graybeal (2006a) Rapid Chloride Penetration Summary Data ....................22: Comparison of Some Published UHPC CTE Data ....2: Construction Activities Unit Costs ....... 90  Table 4...........................1: Experimental Test Matrix .... 113  Table 6.......... 88  Table 4....................14: Corrected First-Crack Flexural Strength Hypothesis Testing ...............11: Combined Poisson’s Ratio Results ......................... 79  Table 4......................... 7  Table 2.......2: Typical and Adjusted Mixing Procedures ........................................... 77  Table 4............................................................ 36  Table 3.............................................. 80  Table 4...2: Compressive Stress Test Results ....3: Estimated Construction Costs ............. 122  Table 6....................................... 80  Table 4.................................. 92  Table 4.............................................................7: Costs of Control and UHPC Bridges..................................17: Michigan Tech Rapid Chloride Penetration Summary Data ..15: Typical Toughness Values (ASTM C 1018 Figure X1...........................1: Ductal® Mix Proportions for 0..........................................................................

with embedded steel reinforcement replaces normal strength concrete in many structural applications. While normal strength concrete. This very low permeability allows UHPC to withstand many distresses normally associated with NSC and HPC such as freeze-thaw deterioration. and chemical ingress. high-performance concrete. or HPC (10.000 psi . resistance to water. has long been able to achieve compressive strengths of 3. and its ability to be easily formed and placed according to need. Now labeled and classified as ultra-high performance concretes (UHPC). UHPC is also nearly impermeable. low ductility. Today.000 psi).1.000 psi compressive strengths).000 – 12. these materials address many of the durability performance deficiencies associated with both NSC and HPC. 1 . low tensile strength. Ultra-High Performance Concrete (UHPC) is one of the latest advances in concrete technology and it addresses the shortcomings of many concretes today: low strength to weight ratio. called reactive powder concrete (RPC).000 psi (sometimes greater than 30. durability in addition to strength must be considered as a principal design concern. and volume instability.0 Introduction to Ultra-High Performance Concrete (UHPC) Concrete has been one of the most widely used building materials because of its compressive strength. with compressive strengths comparable to that of some steels. However. NSC. Research over the past decade has yielded a new classification of highly resilient concrete. corrosion of embedded steel. issues with deterioration and an increasing desire to build larger and more robust structures with smaller members has driven researchers to explore ever stronger and more durable concrete materials. as concrete structures begin to be constructed in ever more aggressive environments. In addition to achieving high compressive strengths in excess of 25.000 – 5.

bridges and buildings that were all but thought impossible may now be realized. Additionally. bridge industry. the amount of cement used in the lifetime of a UHPC structure may be far less than the amount used for several lifetimes of a NSC or HSC structure. Additionally. Also. or a bridge support pier enduring the constant harsh freezing and thawing of the Straits of Mackinac. Overall. While UHPC requires higher cement quantities than normal concretes. Increased span lengths mean fewer support structures such as piers which can lead to improved safety when traveling under overpasses and lower environmental impact in water crossings. and increased material efficiency. 2 . UHPC requires much less maintenance than its concrete counterparts and in turn fewer materials are required for repair or rehabilitation. Similarly. Cement production is a leading contributor to industrial process-related emission sources (Hanle et al. The improved durability of UHPC may lead to lower bridge repair costs and less downtime due to repair construction. 2004).S. lower transportation costs. longer lasting structures minimize the impact on the environment. The need for a structural material to perform in harsh environments is a reality whether the structure is a local bridge subjected to the constant winter salting. the greatest impact of UHPC materials may lie in the improved durability of concrete structures. The higher strengths afforded to UHPC could allow increased girder spans while maintaining similar or smaller cross-sectional areas. UHPC bridges or structures constructed in aggressive environments may remain structurally safe for generations. Costs may be reduced as the lower span to depth ratio of UHPC bridges require less embankment fill while providing more aesthetically pleasing profiles.The implementation of UHPC in bridge construction around the world has sparked new research investigating the potential utilization of UHPC in the U. beam spacing can be increased allowing for faster construction times.

S. A UHPC research initiative by the Federal Highway Administration’s (FHWA) produced the first substantial UHPC research in the U.S. material properties of UHPC need verification testing to substantiate proprietary claims on strength and durability. and included a baseline case of ambient-cure. • Conduct a preliminary life cycle cost comparison between a typical prestressed concrete bridge built using standard building materials and the same bridge built using UHPC. and coefficient of thermal expansion. The age at thermal treatment for curing varied from 3.S. in late 2006 (Graybeal 2006a). the effects of curing regimes and specimen age on the mechanical and durability properties of UHPC require a more thorough investigation. 3 . flexural first-crack strength.Despite these apparent benefits. The goals necessary to accomplish this objective are outlined below: • Characterize some UHPC material properties and build upon previous research at Michigan Tech and throughout the U. but there is need for additional testing and inter-laboratory confirmation of some of the tests. and tested according to American Society for Testing and Materials (ASTM) and American Association of State Highway Transportation Officials (AASHTO) standards where appropriate.2 Objectives The primary objective of this research is to present the history of ultra-high performance concrete and to evaluate some material properties for potential use in durable highway structures. Poisson’s ratio. 10. chloride permeability. Properties include compressive strength. 1. and 24 days. Moreover. freeze/thaw behavior. Similarly. results from UHPC research abroad must be validated here in the U. • Identify the impacts of UHPC material behavior on bridge design and construction. modulus of elasticity. • Consider the impact that different curing regimes had on the above mentioned properties.

and ionic transport. testing of properties was performed to analyze the effects of curing regime and cure time. Research is continuing at many universities and a summary of past and current research related to selected UHPC properties is presented herein. also began investigating this new material and in late 2006 FHWA published the first large scale report on UHPC (Graybeal 2006a).S. Europe has led the way and produced substantial research about its material properties and durability. a product of Lafarge. In the area of ultra-high performance concretes. Ductal® BS1000 was used throughout this research program. However.S. 1. More specifically. physical distress. specifications for UHPC material and durability testing. 4 . Moreover. These suggested procedures can be used for further research on UHPC or as a foundation for developing U. suggested UHPC test procedures and methods were also developed for the various tests.3 Scope To better understand UHPC and its potential impact on the transportation industry. is Ductal®. Inc. due to the unique nature of the material. Additionally.• Develop recommendations for testing mechanical and durability properties of UHPC evaluated in this research project. previous research and testing data regarding mechanical and durability properties of UHPC was compiled and synthesized. The results were then compared to previous research to further characterize the material behavior.S. It was for this reason that Ductal® was the only UHPC considered. recently the U. Currently there is a large amount of research being pursued across the globe involving many different types of UHPC materials. age of specimen. However. the only UHPC that is commercially available in the U.

with the help of Sika Corporation and Lafarge Corporation. Eiffage Group with Sika Corporation created BSI® which is noted as being coarser than other UHPCs (Jungwirth and Muttoni 2004). and Virginia Tech. Many of these research projects are funded through the state transportation departments (DOT) including Virginia. Federal Highway Administration (FHWA) Turner Fairbank Laboratory recently completed an extensive material property characterization study on a proprietary UHPC material. several universities are conducting research on UHPC behavior including Georgia Tech. but several other codes have been developed in Europe (AFGC 2002) and Japan (JSCE 2006). there is no design code for UHPC in the U. Ohio University. and the partnership between Boygues and Lafarge produced Ductal®. developed two different UHPC’s which exhibit similar properties (Harris 2004). Iowa State. While this literature review will cover many of the known properties of UHPC.S. and Iowa DOT’s. This review incorporates and condenses the current body of information related to UHPC material behavior and current applications. the U.. procedures and standards. Eiffage Group and Boygues Construction. The majority of the information on UHPC comes from sources outside the United States that have been published since the mid-1990’s.0 Review of UHPC UHPC is a new family of concretes which exhibits superior mechanical and durability properties over traditional normal strength concrete (NSC) and high performance concrete (HPC). Georgia. 5 . respectively. In addition to the research completed at FHWA. In the early 1990’s two separate French contractors. However. and serves to provide a basis for understanding UHPC durability.S.2. Currently.S. only a few studies have been conducted using accepted U.

1 compares some of UHPC’s properties to HPC and NSC. the development of UHPC materials have benefited from both improved aggregate gradations and the use of a high-range water reducer. and improving microstructure of cement paste by heat treatment application (Richard and Cheyrezy 1995). Although UHPC use of non-continuous steel fibers does not aid in increasing compressive strength. fibers do aid in improving UHPC’s ductility and tensile strength. or superplasticizer.Coming on the heels of continued developments in high performance concrete (HPC). UHPC was first developed as a reactive powder concrete (RPC) with compressive strengths ranging from 29 to 116 ksi. 6 . Table 2. These high strengths were the products of improving homogeneity by eliminating coarse aggregates. optimizing the granular mixture.

(in2) Water Absorption at 225 hours. (lb/ft2) Removal Removal 0. Hartmann and Graybeal (2001).00 0. (in) w/c Ratio (water/cement ratio) Mechanical Properties Compression Strength.9 10-15% 4500-8000 0.0-33.8-1.205 0.024 0. (ksi) Split Cylinder Tensile Strength.1: Comparison of UHPC Material Properties to Other Concrete Classifications Material Characteristics Maximum Aggregate Size. (lb/in2) Chloride ion diffusion coefficient (by steady state diffusion).ultimate.000570. (k-in/in2) Young's Modulus.0 1.057-0.19-0.0-14.27 UHPC 25. (ksi) Flexure Strength .55x10-9 HPC 0.4-3.6 x10-6 250 Times > NSC UHPC 100% Durable < 100 0 7.4-0. No Autogenous Shrinkage After Cure 7.6-1.35 HPC 6. (2003).1x10-11 None Mass Mass Mass Removal Scaling Resistance.Table 2.0-6.40-0.002 >0.6 4.21 2.38-0.5 0.8 2-6% 0.11-0.24 0.00086 2000-6000 0.2 3. (1997).50 0.2-0. Cu Porosity Fracture Energy.75x10-10 UHPC 0.228 8000-9000 2.0 1.0-9.75-1. (ksi) Modulus of Rupture 1st crack. Mindess et al. Mamlouk and Zaniewski (1999).016 Note: Table and information adapted from Kollmorgen (2004).14-0.016-0. (in2/s) Penetration of Carbon / Sulfates NSC 0.36-0. O’Neil et al.0-3.3x10-6 NSC 10% Durable > 2000 4. (ksi) Shrinkage Coefficient of Thermal Expansion (per °F) Ductility Durability Characteristics Freeze/Thaw Resistance Chloride Penetration (coulombs passing) Air Permeability (k) at 24 hrs and 40oC.0 Post Cure <1x10-5.65x10-14 4x10-3 1.45 0.1x10-5 3.5-8. Russell (1999).0 0.1-7.2 Post Cure 40-80x10-5 HPC 90% Durable 500-2000 0 5x10-4 7. Mehta and Monteiro (2006).35 20-25% 0. (ksi) Poisson's Ratio Creep Coefficient.24-0. Aitcin (1998) 7 .70 NSC 3.

320 5-8 Superplasticizer 20 .770 39 . x 0.500 in.530 8-9 Ground Quartz 0 .2. unhydrated cement grains exist in the matrix and act as particle packing material.180 . Blais and Couture (1999). The very low water to cementitious materials ratio prevents all the cement from hydrating. and also reduces permeability (Schmidt and Fehling 2005). but at a much higher proportion rate than in NSC or even HPC.1.710 27 .41 -3 Metallic Fibers (8. long). including carbon nanotubes (Kowald 2004). Table 2. water. After thermal treatment. UHPC consists mostly of the same constituents as normal strength concrete such Portland cement.) 245 . and Richard and Cheyrezy (1996).00 x10 in.5 in.0.1 UHPC Composition While considered a relatively new material. it also includes finely ground quartz.27 -material and included in this ratio) Note: Information condensed from the following references: Hartmann and Graybeal (2001). by 0.1.2. (1996). Cement with high proportions of tricalcium aluminate (CA3) and tricalcium silicate (C3S). steel fibers (0.293 . most UHPC mixes consist of these basic elements.0 Water 260 .2: Composition of a Typical UHPC Mix Constituent lb/yd3 % by Weight Premix Portland Cement 1.38 Silica Fume 385 . Portland cement is the primary binder used in UHPC. The combination of these components creates a dense packing matrix that improves rheological and mechanical properties. and a lower Blaine fineness are desirable for 8 .008 in. and quartz sand. silica fume. However.1. A breakdown of the basic constituents of a typical UHPC is shown in Table 2. Dugat et al.390 0-8 Fine Sand 1. While other constituents have also been investigated.5 . and superplasticizer. dia. dia.14 .30 0.350 5-8 Water/Cementitious Material Ratio (silica fume content is considered a cementitious 0.

2003). Finally.20 would reach discontinuous capillary porosity when 26% hydration of cement has occurred (Bonneau et al. Both the ground quartz (4. the elimination of coarse aggregates aids in improving the durability of UHPC. the polycarboxylate superplasticizer also contributes to improving workability and durability. For this particular application of 9 . the addition of steel fibers aids in preventing the propagation of microcracks and macrocracks and thereby limits crack width and permeability. and the lower Blaine fineness reduces the water demand (Mindess et al. 2003). This zone is the area around any inclusion in the cementitious matrix. Additionally. Despite the large amount of particles left unhydrated. 2000).024 in. and because it is highly pozzolanic. and therefore. and pozzalonic reactivity (reaction with the weaker hydration product calcium-hydroxide) leading to the production of additional calcium-silicates (Richard and Cheyrezy 1995). increasing flowability due to spherical nature. is the largest constituent aside from the steel fibers. the most permeable portion of a concrete tends to be the interfacial transition zone (ITZ) between coarse aggregates and the cement matrix (Mehta and Monteiro 2006). as the CA3 and C3S contribute to high early strength.0 x 10-4 in. an RPC with a water-to-cementitious material ratio of 0. and therefore permeability.) and quartz sand contribute to the optimized packing. By reducing the amount of water necessary to produce a fluid mix. aids in increased strength and reduced permeability. Reduction of the ITZ zone increases the tensile strength and decreases the porosity of the cementitious matrix (Mindess et al.UHPC. and is where the cement grains have difficulty growing because of the presence of a large surface which impedes crystal growth. Silica fume (the smallest component in UHPC with a diameter of 0.2 μm) helps fill this region. The addition of silica fume fulfills several roles including particle packing. Quartz sand with a maximum diameter of 0.

The choice and quantity of this fiber was chosen because of its availability. precast elements. and likelihood that it will be used in the structures industry. it reinforces the concrete on the micro level and eliminates the need for secondary reinforcement in prestressed bridge girders (Graybeal 2005). 2. there is a heavy push to develop many new and innovative types of UHPC materials. Utilization of UHPC in the U. BSI®. etc) and geometries (crimped. but its international roots have led to many different applications in Canada. This is the largest particle in the mix and is added at 2 percent by volume to the mix. durability. the proper market has yet to be discovered to utilize its increased strength. To date this versatile material has been used in artwork. Europe. Eiffage Group.S. and many new UHPC materials are in the process of being developed. use in previous research. respectively. Asia. pedestrian bridges. straight high carbon steel fibers with a diameter of 0. organic. 2. has been limited. While the various UHPC materials differ slightly in composition. are used. etc) are available. 2003) which are marketed by Lafarge.5 in.3 Applications of UHPC As UHPC is being developed. and flexural capacity. While many of the applications 10 . a basic understanding of UHPC material behavior and its potential implementation remains a priority for the U. Ductal® has been promoted in North America by the Lafarge North America group and is the brand of UHPC studied in this report.UHPC. and a few highway bridges.2 Types of UHPC In Europe. acoustical panels. specifically bridges. Other fiber types (polymers.S. and Australia.008 in. and Laboratoire of Central des Ponts et Chausses of France. Because of its size relative to the other constituents. and length of 0. but were not investigated herein. and CEMENTEC (Ahlborn et al. Several that have already been developed include Ductal®. hooked.

the Sherbrooke pedestrian bridge. however. Kollmorgen 2004.2 in. Among many other benefits.have been related to the transportation industry. Six match cast segments compose the main span. The environment is extremely corrosive and UHPC was chosen because of its durability properties with the expectation of reduced or eliminated maintenance. in France. Three years later an AFGC-SETRA working group visited the site and 11 . Figure 2-1: UHPC Example: Sherbrooke Footbridge (Resplendino and Petitjean 2003) In 1997 UHPC’s durability received a test when it was used to replace steel beams in the cooling towers of the Cattenom power plant. Only a brief overview of UHPC functions in the world are presented here. a long term monitoring program was also implemented on the bridge to monitor deflections and forces in the prestressing tendons. Development of UHPC began in the early 1990’s. was constructed in Quebec. the enhanced mechanical properties of UHPC allowed for the use of a deck top of only 1. but also UHPC’s durability. Canada. thick (Semioli 2001). To develop an understanding of how UHPC works in actual applications. and in 1997 the first structure made of UHPC. Schmidt and Fehling 2005). more and more uses for this innovative material are being discovered to not only reap the benefits of its strength. The 197 foot long structure is a post tension open space truss (Figure 2-1). more detailed investigations of these uses can be found in other sources (Behloul and Cheyrezy 2002a and 2002b.

Most recently. the Gärtnerplatz Bridge was completed in Kassel. Japan. arch height of only 49 ft . (Brouwer 2001).under a normal layer of sediment no deterioration of the UHPC was noted (Resplendino and Petitjean 2003). The Footbridge of Peace in Seoul. the Sakata-Mirai footbridge (Figure 2-2c) was completed in 2002 and demonstrated how a perforated webs in a UHPC superstructure can both reduce weight and be aesthetically pleasing (Tanaka et al. highly fire resistant footbridge (Figure 2-2d) at a Chryso Plant in Rhodia (Behloul and Cheyrezy 2002a). and 4 in. France utilized UHPC’s fire resistant capabilities and high load carrying properties to construct an aesthetically pleasing yet. 12 . 2008). South Korea (Figure 2-2a and Figure 2-2b). France. Germany (Figure 2-2e) (Fehling et al. In Japan. and a deck thickness varying anywhere between 1. and Germany.2 in. is an arch-bridge with a span of 394 ft. Other transit applications include footbridges constructed in South Korea. 2002).

2002). (b) and during the day (Lafarge in Searls 2007). (c) perforated hollow UHPC bridge girder (Tanaka et al.Kassel. (d) fire resistant UHPC footbridge in Rhodia. South Korea at night (Behloul and Cheyrezy 2002a). Germany 13 . France.(b) (a) (d) (c) (e) Figure 2-2: UHPC Footbridges (a) Footbridge of Peace in Seoul. (e) Gärtnerplatz Bridge .

Utilizing UHPC saved time and labor as the roof was constructed faster and with fewer workers than the two companion metal roofed silos. The 24 wedge-shaped precast panels with a thickness of 0. while reducing the dead weight by over half (Rebentrost and Cavill 2006). It spans approximately 145 feet with two equal spans consisting of 5 π-shaped prestressed elements. diameter silo. (a) (b) Figure 2-3: UHPC Construction Examples (a) UHPC panels on Joppa clinker silo (Behloul and Cheyrezy 2002a). Australia used UHPC to carry four lanes of traffic over a skewed (16°) single span of 49 ft. the Shepards Creek Bridge in New South Wales. Continuing in the cement industry.5 in. (b) 54 ft UHPC columns in Detroit 14 . UHPC has since been used to create columns with a smaller cross section in a cement terminal in Detroit. UHPC made the transition to the United States in 2001 with the construction of the roof of a clinker silo (Figure 2-3a) in Joppa.In 2001 the Bourg Les Valence Bridge in France was the 1st vehicle bridge constructed using UHPC. Michigan (Figure 2-3b) which allows for five more feet of truck width clearance for the three loading bays (Lafarge North America 2006a). The π-shaped elements were connected by casting UHPC in situ (Resplendino and Petitjean 2003). Additionally. covered the 58 ft. Illinois (Perry 2003).

Graybeal and 15 .. however.4. 2.. Graybeal (2005) showed an increase of 53 percent over non-thermally cured specimens of the same age. In the United States the standard size for a concrete cylinder is 6 x 12 in. the same cylinder would take over 13 minutes to bring to failure instead of the normal 2 to 6 minutes with NSC. and suggested that a 3 x 6 in. The following sections discuss the basic mechanical properties. In addition. a 28 ksi cylinder of this size would require almost 800 kips to break. If the load rate of 35 psi per second specified by ASTM C 39 Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens was followed. When undergoing a 48 hour thermal treatment of 194°F at 95 percent relative humidity.1 Compressive Strength One of the most noticeable assets of UHPC is its high compressive strength. Testing UHPC with traditional standards is difficult because of its high compressive strength. and thermal curing of UHPC. can be attributed to the particle packing and selection of specific constituents. The size of compression machine and the length of the test may prove to be a barrier for production use in the U.2. This increase in compressive strength may allow UHPC to get a foot hold in the long span and low span-to-depth ratio market segments which have been dominated by steel. Perry and Zakariasen (2003) demonstrated that UHPC is capable of reaching compressive strengths of 2533 ksi.S. Kollmorgen (2004) showed that there was no size effect for UHPC for cylinders as small as 3 x 6 in. over NSC or HPC. The increase in compressive strength. This was supported by Kollmorgen (2004) with research showing a compressive strength of over 28 ksi.4 Mechanical Properties Characterization of the mechanical properties is imperative to the efficient design and use of UHPC. creating choices for designers and owners. cylinder be used for standard testing of UHPC.

Specimens were cured under ambient conditions for three days before being demolded. and 2 x 4 in. The average compressive stress exceeded 8. an accelerator was added to the mix and the specimens were demolded at 16 .. 28..) and two different cube (3. and two prismatic geometries were used to complete the testing. Thermal treatment included a six hour ramp up to 194°F at 100 percent relative humidity. To reduce the time before the specimens could be demolded. and after thermal treatment (7.Hartmann (2003) showed that increasing the load rate to 150 psi per second does not affect the results and greatly reduces the time required to complete a compression test. a 48 hour hold period. as well as the long term stability under various loading and environmental conditions. Kollmorgen (2004) investigated the mechanical behavior of thermally treated UHPC at different ages and with different sized specimens. For a 3 x 6 in. two cube. the total load required to break the cylinder is decreased to approximately 200 kips and the test duration reduced to just over three minutes. and 2 x 2 x 2 in. and 56 days after mixing). Over 240 compressive specimens were tested at various ages and with different geometries.94 x 3.94 x 3. cylinder was recommended for use for compression testing.94 in. Three different cylindrical (4 x 8 in. Graybeal (2005) conducted a material characterization study prior to performing full scale tests on AASHTO Type II girders made of UHPC. It was shown that the age and size effects were minimal on compression specimens. This characterization study included defining mechanical and durability properties. Specimens were tested before thermal treatment (3 days after mixing). 3 x 6 in. Three cylindrical. then tested or thermally treated. 14.) geometries were tested in compression. and a 3 x 6 in. cylinder. and a ramp down over night.5 ksi for specimens tested 3 days after casting before thermal treatment and over 28 ksi for all specimens undergoing thermal treatment regardless of age.

Because testing the modulus of 17 . Specimens were then subject to one of the four curing treatments of air. 28. 2. 24. delayed steam.2 Modulus of Elasticity and Poisson’s Ratio The modulus of elasticity is a material dependent property which is often described as a mathematical relationship between stress and strain. Several recommendations were made based on the research. 3 x 6 in. and tempered steam.8 ksi for air. and tempered steam treatments. Graybeal (2005) tested nearly 1000 cylindrical and cubic compression specimens which underwent one of the four curing treatments and yielded an average 28 day compressive stress of 18. Specimens undergoing air curing treatment remained at standard laboratory atmospheric conditions until the time of testing. Delayed steam treatment was similar to steam treatment except it commenced on the 15th day after mixing. respectively.8. delayed steam.4. steam. cylinders can be utilized for compression testing and the load rate for compression testing can be increased to 150 psi per second. steam. followed by a 44 hour hold. Typically when the value is given for concrete. it is referencing the elastic portion of the compressive stress-strain curve up to 40 percent of the ultimate compressive strength (0. Steam treatment began within 4 hours of demolding and consisted of a 2 hour ramp up to 194°F at 95 percent relative humidity. Tempered steam treatment was similar to steam treatment except the temperature was limited to 140°F. and a 2 hour ramp down to atmospheric conditions.0.approximately 24 hours after casting.3. The modulus of elasticity is used in design calculations to predict deflection behavior of the element so the design can often satisfy the specified limit states. and 24. The slope of the elastic portion of the stress-strain curve is the modulus of elasticity.40 f `c ) as specified in ASTM C 469 Standard Test Method for Static Modulus of Elasticity and Poisson’s Ratio of Concrete in Compression.

ACI Committee 318 (ACI 2005) presents an equation which relates the 28 day compressive strength (f`c) of normal strength concrete to the modulus of elasticity (Ec). Additionally.1 1 E c = wc . The document states that the material property should be determined in the later design stages by conducting a test. E c = 262.000 psi.000 * f `c + 1. and requires additional testing jigs and software to determine. The ACI 318-05 equation. does not have an equation relating compressive strength to modulus of elasticity (AFGC 2002). efforts have been undertaken to develop a relationship between modulus of elasticity and compressive strength. which was requested by the Association Française de Génie Civil (AFGC) and published by Service d’études Techniques des Routes et Autoroutes (SETRA). However.elasticity is time consuming.3 18 . the Interim Recommendations for UHPC. is shown as Equation 2.0 * 10 6 (psi) Equation 2. for concrete with a unit weight ( wc ) of 90 to 155 pcf. a relationship is given from the work conducted at the Cattenom nuclear power plant on 196 cylindrical specimens with diameters of 7 cm (2. Ec = 40.).76 in. ACI Committee 363 (ACI 1997) produced a relationship for higher strength concrete with compressive strengths from 3. The equation is shown below after it was converted to English units.000 * ( 3 f `ATT (psi) ) Equation 2.2 The previous two equations do not apply as the high compressive strength of UHPC lies above the range of applicable compressive strengths.1 However.000 to 12.5 * 33 * f `c (psi) Equation 2. The equation is shown below. HPC has a much greater compressive strength than normal strength concrete.

delayed steam.000 * ( 3. Ec = 50. steam. E c = 351. (2003) at Iowa State University on five 3 x 6 in. The modulus of elasticity was determined using local deformation transducers (LDT) made out of strips of phosphorus bronze with strain gauges attached to each side of the metallic strip. and tempered steam. and 4 x 8 in. cylinders produced an equation which took the following form. were solidly attached to the specimens. Hinges were glued to the specimens at a set gage distance and the LDT installed. Twenty-four cylindrical specimens of 2 x 4 in. and any of the aforementioned curing treatments.200 * f `c (psi) Equation 2. the curing treatments were air. Graybeal (2005) developed yet another relationship using a total of 148 specimens undergoing one of four different curing regimes. The relationship was shown to apply to UHPC with compressive strengths between 4 and 28 ksi.000 * f `ATT (psi) Equation 2.4 Where: f`ATT = Compressive stress of UHPC after thermal treatment.Where: f`ATT = Compressive stress of UHPC after thermal treatment. As previously noted. Two parallel solid rings with a gage distance of 2 in.5 Where: f`ATT = Compressive stress of UHPC after thermal treatment.14 f `ATT (psi) ) Equation 2. Research conducted by Sritharan et al.6 19 .. 3 x 6 in. Work completed by Kollmorgen (2004) at Michigan Tech resulted in an equation relating compressive strength of thermally treated specimens to modulus of elasticity. E c = 46. were used to determine the following relationship with an applicable range of 5 to 30 ksi.. The upper ring held three LVDTs which end bears on the lower ring.

G.0 ksi and had a toughness of 250 times that of normal strength 20 . (1998) showed that UHPC was capable of reaching a flexural strength as high as 7. no standards are available for UHPC but ASTM C 1018 can be adapted. however it can overestimate the tensile strength when small scale prisms are utilized (Graybeal 2005). E = 2 * G * (1 + ν ) Equation 2. Small prismatic specimens are loaded at the third point to create a region of constant moment in the specimen. The applied load and resulting deflection are recorded to be used in determining the first-crack strength and post crack flexural toughness. This value is also the widely accepted value for Portland cement concrete of normal strength.8 SETRA (AFGC 2002) gives a Poisson’s ratio of 0. 2.3 First-Crack Flexural Strength and Flexural Toughness ASTM C 1018 Standard Test Method of Flexural Toughness and First-Crack Strength of Fiber-Reinforced Concrete (Using Beam with Third-Point Loading) is used to evaluate the first crack strength and flexural toughness of portland cement concrete.2 if it is not determined by direct testing. and modulus of elasticity. Flexural toughness is calculated as the area under the load deflection curve and is an indication of the energy absorption capabilities.7 This ratio is also used to relate the shear modulus.Poisson’s ratio (ν) is defined as the relationship of the transverse strain ( ε trans ) divided by the longitudinal strain ( ε longitudinal ) as shown in the equation below. E. ν= ε trans ε longitudinal Equation 2.4. Research by Cheyrezy et al. The first-crack strength is a useful indicator of the tensile strength of UHPC. However.

concrete.2 ksi and an ultimate flexural strength of 4. and 3 x 3 x 11. Kollmorgen (2004) conducted flexural testing on 58 specimens. Graybeal and Hartmann (2003) attributed the increase in the flexural behavior of UHPC to the particle packing and the addition of fibers which hold the cement matrix together after cracking has occurred. spans and loading applied at the third points..0 – 7. 9 in. The specimens had two different geometries to determine if a size effect existed on small scale prisms. and utilized five different geometries/loading configurations. Dugat et al. Specimens had span lengths of 6 in. less damaging cracks to form. A constant displacement rate of 1. 2. Average values of first-crack strengths. which controls the rate of deflection of the prism.25 in. was used to test both specimen geometries. Ultimate load and toughness values based on the procedure outlined in ASTM C 1018 were reported. based on AGFC (2002) were reported. the full hydration potentials of the cement and silica fume are never reached. at the testing machine head. Testing was conducted on 2 x 2 x 11.. with a cross section of 2 x 2 in. and toughness values. improved performance has 21 .50 x 10-4 in/sec. span with a 3 x 4 in.4 Thermal Treatment Due to the very low water-to-cementitious material ratio in UHPC. and a 12 in. cross section.25 in. maximum loads. (1996) reported average modulus of rupture values of 3.6 ksi. UHPC exhibits ductility because as the specimen begins to microcrack the small scale fibers reinforce the matrix causing smaller. Graybeal (2005) tested 71 flexural specimens utilizing the procedure outlined in ASTM C 1018.0 ksi which confirmed Cheyrezy’s findings. Corrections were applied to calculate a more representative tensile strength from the first-crack strength. As previously noted the specimens underwent one of four curing treatments.4. and 15 in. However. Perry and Zakariasen (2003) showed that UHPC had flexural strengths ranging from 5. 12 in. prismatic specimens with 9 in..

further details the durability of concrete as that which is able to resist weathering. fire. Graybeal 2006). The thermal treatment appears to allow continued hydration of the portland cement and pozzolanic reaction of the silica fume (Gatty et al. or delayed ettringite formation). Generally. The American Concrete Institute. In addition to improved mechanical properties. and pressure treatments (Loukili et al. steam. 1995). or salt crystallization) or chemical agents (alkali-silica reaction. or other processes of deterioration (ACI 2002). In general. chemical attack. Loukili et al. 1998. but also due to the method of curing. Concrete can experience deterioration from either physical attack (abrasion.been observed after thermally treating UHPC using combinations of heat. 1998. Graybeal (2006) observed improved durability characteristics including increased resistance to chloride penetration and abrasion. (1998) noted that after treating UHPC in 194°F water. Kollmorgen 2004. These findings indicate that the full promises of UHPC’s benefits are not only realized because of particle packing. 2. the durability of a concrete can be summarized as the capability of a concrete to continue performing its designed functions while maintaining its dimensional stability in a given environment. All of these issues can lead to additional durability problems or build upon already existing problems. abrasion. chloride ingress and corrosion of embedded steel. up to 65% of the cement is hydrated (compared to 48% before treatment). outside of 22 . sulfate attack.5 Durability Improvements Concrete durability has become an ever more important aspect in the design of structural concrete. more and more research is focused on investigating the durability aspects of concretes. freezing and thawing. Cheyrezy et al. While compressive strength has long been the standard for determining the quality of a concrete. Aitcin (1998) defines durability of concrete as “the resistance of concrete to the attack of physical or chemical aggressive agents”. or ACI.

5. Cheyrezy et al.1 x 10-5 lb/in. Mehta and Monteiro 2005).2 (O’Neil et al. Permeability testing demonstrated that UHPC has an oxygen permeability of less than 1. Once chloride ions reach embedded steel. However. The very low water/cement ratio and densely packed matrix of UHPC contribute to permeability results even lower than HPC. 2000) is one of the most problematic durability issues associated with low permeability concretes.7 x 10-5 lb/in.poor material selection leading to internal attack (high chloride content in cement paste or alkaliaggregate reaction) or poor construction practices. Mehta and Monteiro (2005) define permeability as the ease with which a fluid under pressure flows through a solid. corrosion can take place through an electro-chemical reaction that expands the steel up to 600%.6 x 10-15 in. HPC on the other hand had an air permeability of 1900 x 10-15 in. or diffusion (Stanish et al.2 which is on the extremely low end of testing (AFGC 2002). much more susceptible to chloride ingress which eventually leads to corrosion of reinforcing steel. 1997). hydrostatic pressure.2 and water absorption of 49. (1997) reported water absorption of 7. (1995) used mercury 23 . A concrete with high permeability is. Steel corrosion is such a large problem that a 1991 FHWA report on the status of reinforced concrete bridges linked corrosion as a cause of distress for a majority of cases (Mehta and Monteiro 2005). previous research demonstrated that UHPC exhibited almost no permeability and was not susceptible to chloride ingress.1 Chloride Ion Penetration Chloride ion migration through a concrete by means of capillary absorption. UHPC has an extremely low water/cement ratio and a densely packed matrix that may contribute to a very low permeability. therefore. 2. 2003. high permeability in a concrete is the main cause of durability failures (Mindess et al. On the other hand.2. while O’Neil et al.

Therefore. reduced the permeability of the concrete.74 x 10-3 in.. In the U.39 μin/yr. Similarly. The Japan Society of Civil Engineers 24 . (1998) revealed that adding 0. thus preventing deterioration. the addition of silica fume greatly reduced the conductivity of the specimen. carbonization. Toutanji et al.48 x 10-7 in. Another method to determine whether a concrete is susceptible to chloride ingress uses an applied electric potential across a specimen load cell to determine concrete’s conductance (ASTM C 1202). (2003) supported claims of high resistance to aggressive agents such as de-icing salts. (1997) reported less than 10 Coulombs passing (over a six hour period) through UHPC specimens (negligible chloride ion penetrability) that were water cured at varying times and temperatures. However.50 in. and only surface corrosion of steel fibers when exposed to corrosive chemical conditions (AFGC 2002). The highly dense structure of the composite and the reduction in pore volume restricts aggressive chemicals and water from entering UHPC’s cementitious matrix.S. While Graybeal (2006a) demonstrated that the steel fibers did not contribute to a short circuit effect during UHPC testing.75 in. the reduction was not proportional to the amount of silica fume added (Toutanji et al. results from rapid chloride penetration testing of UHPC should reflect these claims and demonstrate UHPC’s high resistance to chloride penetration. 1998). the French recommendations report that UHPC has an electrical resistivity of 2878 kW/in. polypropylene fibers increased the permeability of concrete and adding shorter fibers 0. Bonneau et al. and chloride ion attack.intrusion to demonstrate that the porosity of an RPC is less than 9% in volume for the pore diameter range of 1. additional research by Graybeal (2006a) demonstrated that UHPC had negligible chloride ion penetration when thermally treated and only very low penetration when not thermally treated. to 3. Work by Schmidt et al. a rate of reinforcement corrosion less than 0. Furthermore.

acknowledges this characteristic by mandating only a 0.79 in. cover for prestressed strands used in UHPC (JSCE 2006). Australian publications also reported chloride diffusion on the order of 31 x 10-12 in.2/s to have replaced collapsed weir covers exposed to salt-water spray at the Eraring Power Station with UHPC panels having a life expectancy greater than 100 years (Rebentrost and Cavill 2006). 2.5.2 Freeze-Thaw Testing Another mechanism of concrete distress that hinges on a concrete’s permeability is freeze-thaw durability. Typically, concrete specimens exhibit distress, and eventually deterioration, in the form of cracking, spalling, and disintegration as freeze-thaw cycling persists. Freezing and thawing of saturated concrete occurs regularly in northern climates and over a period of time can disintegrate both cement matrix and aggregates. Several mechanisms are believed to be at work in this process including hydraulic pressures developed due to expanding ice (water expands 9% when frozen) (Powers 1945), osmotic pressure (Powers and Helmuth 1953), water expulsion (Litvan 1972), or the movement of water towards frozen water to form ice lenses (Collins 1944). While some or all of these theories may be at work in a specimen, the methods to avoid freeze-thaw damage are more regularly accepted. The two methods of producing freeze-thaw resistant concrete are: entraining air voids that allow pressure to dissipate or using a sufficiently low w/c ratio to reduce capillary porosity which reduces the amount of freezable water (Pigeon and Pleau 1995). The freeze-thaw durability of UHPC comes from its highly impermeable matrix that effectively eliminates capillary porosity (Bonneau et al. 2000). Most research to date has revealed that UHPC has a durability factor of 100 or greater (no deterioration of specimens after 300 freeze-thaw cycles) (Bonneau et al. 1997). In fact, research by Lee et al. (2005) and Graybeal (2006a) demonstrated that UHPC actually increased its 25

relative dynamic modulus, or RDM (ratio of squared resonant frequency at the end of freezethaw testing to squared resonant frequency before testing – revealing amount of deterioration in specimen), and gained mass as the freeze-thaw cycles continued. A mass increase of 0.2% after 125 cycles was documented by Graybeal and Tanesi (2007) and led to a further study of mass change in UHPC specimens submerged in water. Normally, concretes lose mass due to material spalling and experience a decrease in relative dynamic modulus as micro-cracking occurs. In spite of this, compiled research on autogenous healing (sometimes referred to as self healing) revealed that concrete can heal when exposed to water after or during deterioration (Jacobsen and Sellevold 1996). Jacobsen and Sellevold also demonstrated that high strength concretes can recover from damage due to freeze-thaw if submerged in water after testing. Concretes that lost up to 50% of their RDM values, gained nearly all of it back after being submerged in water for three months. However, the submerged specimens regained only 4-5% of their compressive strengths after losing nearly 22-29% of it due to freezing and thawing. Granger et al. (2007) investigated self-healing cracks in a UHPC material and determined that precracked specimens stored in water regained stiffness upon reloading while precracked specimens stored in air did not regain stiffness. The methods for self healing may come from the gathering of debris in the cracks, hydration of unhydrated cement particles in low w/c concretes, or precipitation of calcium carbonate (CaCO3) (Edvardsen 1999; Granger et al. 2007). This phenomenon was further investigated by Graybeal (2006a) on air treated and steam treated UHPC specimens not subjected to distress. Untreated UHPC specimens that were submerged in water, but not subjected to freeze-thaw cycling, gained 0.25 – 0.35% mass compared to 0.09 – 0.18% mass gain for specimens remaining in an ambient air environment. Additionally, the untreated UHPC specimens submerged in water increased in compressive strength by 12%

26

relative to the compressive strengths of specimens remaining in air. On the other hand, the compressive strengths of the steam treated specimens submerged in water increased by only 3% relative to their counterparts that remained in an ambient air environment (Graybeal and Tanesi 2007). 2.5.3 Coefficient of Thermal Expansion Temperature fluctuations that are not in the freezing range can also play an important role in structural design. The coefficient of thermal expansion (CTE) of a material is the strain in the material per a given change in temperature. The coefficient of thermal expansion needs to be considered when more than one type of material is used in construction or when member expansion could lead to overstressing under large strains. Thermal strains of concrete structural elements can lead to cracking and, therefore, susceptibility to chloride ingress or freeze-thaw damage. Given UHPC’s unique material composition, establishing an appropriate CTE value for UHPC is extremely important before utilizing it with other structural materials. Because concrete is a composite material, its coefficient of thermal expansion value correlates to the proportions (volume) of its constituents (Walker et al. 1952). As concretes generally contain a majority coarse aggregate by volume, the CTE value largely depends on the CTE of the aggregate (Mindess et al. 2003). Other important factors when considering a material’s CTE are the w/c ratio, specimen age, and moisture content (Mindess et al. 2003). Some of these factors can greatly affect the recorded CTE value of a material. Partially saturated specimens can have CTE values as much as 1.8 times higher than fully saturated specimens (Emanuel and Hulsey 1977), and the CTE of cement pastes can increase as much as 25% as the cement fineness increases from 590 ft2/lb to 1300 ft2/lb (Mitchell 1953). However, dry specimens exhibit only slightly higher CTE values on the order of 1.17 times larger than fully 27

saturated specimens (Emanuel and Hulsey 1977). the French Recommendations for UHPC (AFGC 2002) use CTE values from European UHPC suppliers of 6. To date. However. Values of approximately 8.3 x 10-6/°F (Mindess et al.86 .0 . Emanuel and Hulsey used these factors and developed the following equation to estimate the CTE of a concrete: α C = fT [ f M f A β Pα S + β FAα FA + β CAα CA ] where: Equation 2.3 x 10-6/°F have been document by Graybeal (2006a).9 αC αS α CA α FA = CTE of concrete = CTE of cement paste = CTE of coarse aggregate = CTE of fine aggregate fT = correction factor for temperature alternations (1. However. 28 . These CTE values are slightly higher than those of most normal strength concretes which are on the range of 4.controlled. The Japanese specification for UHPC design states that the coefficient of thermal expansion can change substantially and depends on the moisture content.1 x 10-6/°F to 7.8 x 10-6/°F for BSI® which are more in line with typical concretes.outside) f M = correction factor for moisture f A = correction factor for age β P = proportion by volume of paste β FA = proportion by volume of fine aggregate β CA = proportion by volume of coarse aggregate The CTE value can be very important where differential heating occurs or where a variety of materials are used. 2003). CTE values of 7. little research is published about the coefficient of thermal expansion of UHPC.6 x 10-6/°F for Ductal® and 5. the coefficient of thermal expansion of UHPC has not been tested in the United States. with most coming from manufacturers. 0.5 x 10-6/°F are given by the specification for specimens after thermal treatment (JSCE 2006).

Additionally. and inter-laboratory repeatability.S. with a more comprehensive and thorough understanding of UHPC and its distinctive material properties.5. The research reported here will address several of these issues including the age of thermal treatment. however. the body of research on UHPC durability. Variables such as specimen age at time of thermal treatment and procedure variations (ASTM C 666 . Relatively few statistical analyses of the test results have been reported due to the small sample sizes tested. While these girders did not efficiently take 29 . and aesthetically pleasing bridges and concrete structures. procedure variations. areas of potential future research will be addressed so that continual progress can be made towards providing stronger. full scale testing of two 80 ft prestressed AASHTO Type II UHPC girders (Figure 2-4a) was completed in 2001 (Hartmann and Graybeal 2002). especially in the United States. but has seen little done in the way of showing that these results can be replicated in other laboratories or that variability between batches is low. more sustainable. 2. Both mechanical and durability properties of a UHPC were investigated. the main material property tests have been conducted by FHWA at their Turner-Fairbank Highway Research Center (TFHRC). is limited and incomplete. The FHWA material property characterization research was completed in 2006 (Graybeal 2006a) and provided a broad. Also.4 Additional Durability Research Despite these findings. Procedure B) have not been considered. These findings will help provide the body of UHPC research in the U.Procedure A vs. understanding of how UHPC performs under ASTM and AASHTO testing methods. yet basic. Current research has provided information about the durability properties of UHPC.2.6 Other UHPC research Several state Department’s of Transportation and a few universities have been investigating this relatively new material.

and a new design has been proposed (Keierleber et al. including transverse loading and fiber distribution. was comprised of three 110 ft Modified Iowa Bulb-Tee prestressed UHPC beams (Graybeal 2006b). replacing the 73-year-old dilapidated truss bridge. the Iowa Department of Transportation and Lafarge North America constructed the first UHPC bridge in the United States in Wapello County. The 110 ft simple span Mars Hill bridge. deep UHPC deck were tested for both long-term and loading effects (Figure 2-4b).advantage of the UHPC properties by minimizing the cross-sectional area of the beams. Iowa in 2006 (Figure 2-5). deep. Initially. 2007) Using the research from FHWA and collaborating with Iowa State University. (a) (b) Figure 2-4: UHPC Girder Testing (a) Testing of a UHPC I-girder (Hartmann and Graybeal 2002) (b) UHPC π-girder at the TFHRC (Keierleber et al.S. precast facilities. 70 ft span girders with a double-tee (or π-shaped) cross-section with an integrated 8 ft wide. 2007). testing revealed problems with the girder shape. Results from the follow-up investigation are pending. 2004). Current research at the TFHRC is investigating an optimized girder/deck configuration for use on short span road bridges (Graybeal et al. they provided valuable information about how UHPC can be cast in U. However. 3 in. no mild steel 30 . In addition. four 33 in.

reinforcement was used (steel use was limited to the steel fibers in the UHPC matrix and to the prestressing tendons). To monitor the behavior of the beams and bridge, a 2-year performance monitoring program was implemented.

Figure 2-5: Mars Hill Bridge in Wapello County, Iowa (Lafarge 2006b)

Additional UHPC research at the Virginia DOT and Virginia Polytechnic Institute & State University characterized some of the structural aspects of UHPC including punching shear of thin UHPC plates (Harris and Roberts-Wollmann 2005) and horizontal shear between bridge decks and beams (Banta 2005). Both studies demonstrated that current AASHTO and ACI design standards could be used for design of UHPC members for punching shear and horizontal shear.

31

(This page intentionally left blank)

32

3.0
3.1

Methodology
Introduction The objectives of this chapter are to outline the materials and equipment used to mix, cast

and cure UHPC. The areas investigated were compressive strength (ASTM C39), modulus of elasticity and Poisson’s ratio (ASTM C 469), and flexural properties (ASTM C 1018), rapid chloride penetration testing (ASTM C 1202), freeze-thaw cycling (ASTM C 666), and coefficient of thermal expansion (AASHTO TP 60-00). Some ASTM standards were modified slightly to facilitate use for UHPC. Modifications are discussed herein. Additionally, studies were performed for the freeze-thaw and coefficient of thermal expansion tests specimens, to determine the impact of water absorption on UHPC specimens under testing. While previous research by Graybeal (2006a) showed a strong intra-batch correlation in the durability data (i.e. – test results correlated well when specimens were from the same batch), an inter-batch correlation is desirable to demonstrate repeatability of the batching and curing process. Therefore, one sample for each test and corresponding test age was often cast in each batch. For example, only one thermal treatment cured specimen for the 28-day coefficient of thermal expansion test was cast in Batch A, and Batch B similarly had only one thermally cured specimen for the 28-day coefficient of thermal expansion test. While this may have introduced some variability due the fact that there was only one sample of a test age for each test, the homogeneous nature of UHPC materials and batching processes allowed for accurate comparisons between specimens from different batches (as shown by low coefficient of variation results in upcoming sections). In addition, to ensure that any variability present did not greatly affect the results, three or four compression specimens for various age testing were cast in each batch to check the variability between each batch. If the compression cylinders did not achieve 33

the batch would be considered to have had an error during the mixing or curing process. Mixing UHPC requires special equipment and procedures to develop consistency in batching. Graybeal 2005). five 40 lbs. but high carbon/high strength steel fibers were chosen as these were consistent with other research being conducted using Ductal® for precast bridge elements (Kollmorgen 2004. bags of premix.2 UHPC Mixing Procedure The Ductal® product used in this project was purchased from Lafarge North. casting. 34 . Enough material to make about three fourths of a cubic yard was supplied in one shipment.proper strength. The straight steel fibers measured 0. The shipment included thirty-three 80 lbs. and curing in a timely fashion. long. in diameter x 0. A high shear capacity mixer along with vibratory table and steam cure chamber capable of maintaining 100% humidity at 194°F is required (Kollmorgen 2004).5 in. 3. The UHPC was batched according to procedures given by a Lafarge North America Technician and more detailed description of procedures can be found in Kollmorgen (2004). Mixing UHPC in Benedict Laboratory at Michigan Technological University used a Doyon BTF-060 planetary mixer (Figure 3-1). Three bags were damaged in transport and used for preliminary testing to ensure previous procedures outlined at Michigan Tech were still valid.008 in. boxes of steel fibers. and five gallons of Chryso® Fluid Prema 150 superplasticizer. UHPC can be mixed with several different fiber types.

1.65 ft3 batch Constituents Premix Water Superplasticizer Steel Fibers w/c ratio 1. 0. except for one batch which was cast without fibers to determine unreinforced mechanical properties. 51 lb. 5.84 lb.65 cubic feet (ft3) in volume were mixed using the amounts shown in Table 3.20 0.Figure 3-1: Doyon Mixer Batches of up to 0.20 UHPC was packaged as a blended premix that included the proprietary proportions of Portland cement. 0. 263 lb. This mix design was followed for all batches. 6.1. Table 3. silica fume.3 lb. the batching weights were kept consistent but the steel fibers were eliminated.0 yd3 3700 lb. This premix was added to the mixer first and then disturbed with dry mixing to break up any clumps in the 35 . 219 lb. and quartz sand already combined.1: Ductal® Mix Proportions for 0.7 lb. ground quartz.32 lb.65 ft3 96. In that case.

half blended with water and added near the beginning of the mix. After the turning point was reached. the remainder of the superplasticizer was added followed by the slow addition of steel fibers.2: Typical and Adjusted Mixing Procedures Typical Mix Time Start Mixing Addition of Water + ½ Superplasticizer First Speed Increase Second Speed Increase Turning Point Addition of ½ Superplasticizer Addition of Fibers Reduce Speed Stop Mix Mixer Speed 1 1 4 5 5 5 3 1 1 0:00 2:00 4:15 6:00 11:00 12:00 14:00 16:00 18:00 Mix Time for Premix age 2 mo. The turning point of UHPC was defined as the time at which all of premix and water.2). The turning point was determined by the mixer attaining peak amperage. The other constituents (water. Misson 2008). superplasticizer. Table 3. The superplasticizer was added in two stages. and half of the superplasticizer are completely mixed so that the UHPC begins clumping together and falling from the sides of the mixing bowl.material. Figure 3-2 (Kollmorgen 2004). and the other half added after the turning point (Table 3. 0:00 4:00 6:15 8:00 20:00 22:00 24:00 27:00 30:00 36 . and steel fibers) were added at the appropriate times during the mixing sequence (Peuse 2008. 0:00 2:00 4:15 6:00 11:00 12:00 14:00 16:00 18:00 Mix Time for Premix age 11 mo.

As the premix used for this testing aged from two to eleven months. it is suggested by Lafarge North-America that the material’s internal temperature remain below 86°F to ensure the longest pot life of the material. There is no known way of controlling this internal temperature. Table 3. However. The typical mix procedure called for a total mixing time of 18 minutes and for the time from the 37 . the corresponding mixing time was also longer. several of the batches exceeded this temperature during mixing and attained mix temperatures up to 95°F. During mixing. rather it is an implicit measure.Figure 3-2: Turning Point of UHPC Several issues with the prescribed procedures developed during the mixing process including internal temperature and mixing times. This was not viewed as a significant problem however. Another concern was that mixing time increased for batches made at later times in the research program.2 indicates typical and adjusted timing. as consistency of the material rheology and compressive strengths were still achieved. Previous research (Graybeal 2006a) indicated that mix times increased as premix age increased.

all of the older premix bags used for specimen batching had 1 in. twenty times. for the older premix the total mix time increased to nearly 30 minutes and the average time from the addition of water to the turning point averaged 15-16 minutes. as the beginning of the mix procedure was adjusted to break apart these clumps. However. The same four diameter dimensions of the UHPC flow were measured again and a domain classification of the mix assigned according to Table 3. Also. The premix clumping may have inhibited the ability of the water and superplasticizer to adequately wet all of premix in the shorter mixing time used previously. and then the impact table was dropped 0. or larger sized clumps of the premix material present before mixing.3. compressive strengths. UHPC was first placed on the impact table in a short steel cone which was then lifted off slowly to allow the concrete to flow evenly about the table. in spite of these changes.. the rheology of the UHPC mix was tested for consistency. All of the batches mixed were 38 . An adjusted ASTM C 1437 Standard Test Method for Flow of Hydraulic Cement method was used so that the recommendations outlined in Ductal® reference T006 (Operating Procedure – Flow Test) were followed (20 impacts compared to the ASTM specified 25 impacts). and durability properties comparable to that of other research.addition of water to the turning point to be approximately 8-9 minutes. the UHPC material obtained mix properties.5 in. While specific reasons for the delay in mix time were not investigated. However. Four measurements of the diameter of the flow were taken at equally spaced locations (Figure 3-3). the time between the addition of water and the turning point is considered a better comparison than contrasting total mix times. The increase in mix time may lead to higher water evaporation and therefore poor rheological properties. After mixing was completed.

These sealed tops prevented moisture loss 39 .classified under Domain B.3: Flow Domain Classifications of Freshly Mixed UHPC Domain A Stiff Average Flow Measurements after 20 blows < 200 mm Domain B Fluid 200 mm . and an acrylic plastic top sealed with weather-strip covered the beam molds. Beam molds were placed on the table and filled from one end. Upon removal from the table. Figure 3-3: Impact Table Measurement of UHPC’s Flow Table 3. so all specimens were cast according to Ductal® reference T002 Cylinder and Prism Preparation for Domain B mixes. amplitude set by Ductal® reference T006. After the molds were filled. they remained on the vibrating table for an additional 30 seconds to consolidate the mix. and the mix was allowed to flow to the opposite end to fill the mold.020 in. the cylinders were sealed with a fitting cap. Cylinder molds were held on the table and filled in two equal lifts.250 mm Domain C Highly Fluid > 250 mm 3.3 Casting Specimens Cylinders and beams were cast on a vibrating table capable of reproducing the specified 0.

3. Specimens were either air cured at ambient laboratory conditions (Air). 100% humidity.from the fresh specimens. overuse of mineral oil resulted in more air bubbles appearing on the sides of specimens. Air-cured specimens for testing at 3-days were demolded two to three hours before testing to allow ample time to prepare and test the three day specimens in the allotted time by ASTM C 39 which was 3 days ± 2 hours. almost glassy. The TT-cured specimens were subjected to a 48-hour. The specimens were placed on a bench in the casting room and allowed to remain at ambient conditions for three days until being demolded. however. Specimens destine for delayed thermal treatment (DTT) or doubly40 . All specimens were removed after the 60 hour cure process and allowed to return to ambient temperature before end grinding or testing occurred.4 Curing Regimes The curing regime was a primary variable in this study. Air-cured specimens tested at later ages were also demolded at this time and placed on a laboratory shelf in ambient conditions until being tested. At the end of this time. finish. the environment was allowed to ramp down to lab conditions over 6 hours by opening the outer lid of the cure chamber (Figure 3-4). Specimens with both types of surface conditions were tested. the sides of demolded cylinders had a smooth. During batching. The thermal treatment began with a 6-hour ramp up period to 194°F and 100% humidity followed by a 48-hour hold at the elevated temperature and relative humidity. In the plastic cylinder molds. steam treatment at 194°F upon demolding at 3-days. Moreover. yet the phenomenon appeared to have had no effect on the material properties analyzed. or thermally treated (TT). mineral oil was used to prevent the concrete specimens from sticking to the steel prism molds as well as the plastic cylinder molds. when no mineral oil was used to lubricate the forms.

Therefore. and testing procedures are outlined so that these tests may be reproduced and properly understood. However. Also. 41 . a clear understanding of how the specimens were handled and tested is necessary to properly interpret and apply the results. specimen preparation. equipment. it is essential that the materials. several sections of the procedures were adjusted in an effort to maintain the integrity of the curing practices. Recommended specimen preparation and testing procedures for UHPC are summarized below for each experimental test conducted. due to the unique nature of UHPC.5 Specimen Preparation and Test Procedures ASTM and AASHTO standards were used as a baseline for investigating the properties of UHPC.delayed thermal treatment (DDTT) were demolded at 3-days and placed on a laboratory shelf in ambient conditions until starting the thermal curing process at days 10 and 24. Figure 3-4: Michigan Tech’s UHPC Thermal Treatment Cure Chamber 3. respectively.

5. Specimens were prepared one day before their scheduled testing time with the exception of the three day Air-cured cylinders as they were demolded and end-ground a few hours before testing.0 ksi stated ASTM C 1231 and C 617. After several passes of the wheel. Currently. and therefore could not be used for testing UHPC. the best alternative is to grind the ends of the cylinders using a surface grinder. Figure 3-5 shows the Reid surface grinder used to prepare specimens for this research.1 Compressive Strength Specimen Preparation – Compressive Strength The traditional methods of preparing the ends of cylinders for compressive testing by using unbonded caps (neoprene pads) or sulfur caps greatly exceeded the allowable stress of 12. Figure 3-5: Reid Surface Grinder Cylinders were held in place using a v-shaped jig which kept the specimens perpendicular to the grinding wheel. the ends were checked for perpendicularity using the procedure set forth by Ductal® reference T009 Operating Procedure Cylinder End Preparation.3. The cylinder end planeness limit of 1° corresponds to the limit set by ASTM C 39 which states that neither end of the cylinder can depart from 42 . respectively.

4 x 8 or 6 x 12 in. Figure 3-6: End Perpendicularity Set-up Specimen Testing – Compressive Strength ASTM C 39 – Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens was used as the baseline for compression testing. A machine of this size is not typically used in the precast industry or testing laboratories.perpendicularity by more than 0. the smaller 3 x 6 in. testing a 6 x 12 in.5°. However. Research by Kollmorgen (2004) and Graybeal (2005) showed that the size effect of 3 x 6. cylindrical specimens was not statistically significant regarding the compressive strength. UHPC cylinder after thermal treatment would potentially require a compression machine to have a capacity of approximately 800 kips. Hence. End perpendicularity was measured using a digital dial gage as seen in Figure 3-6 utilizing the referenced procedure (Peuse 2008). All cylinders were ground within the 1° tolerance with the initial grinding. cylinder was 43 .

chosen.0 (2004). Tests performed on 3 x 6 in. Additionally. This lengthy time period would be unacceptable for the time required to break specimens for production use. Figure 3-7: Baldwin CT 300 Compression Testing Machine A slight modification to ASTM C 39 was made to make the testing of UHPC more practical. Graybeal and Hartmann (2003) conducted research showing there was no 44 . Ductal® reference T001-Operating Procedure Compressive Test suggests a load rate of 150 psi per second which reduces the length of a test to approximately three minutes and also falls within the allotted time requirement of ASTM C 39. This hydraulic load frame has a capacity of 300 kips and was operated by manual controls. affording a test machine capacity of about 250 kips which is readily available in testing labs and at precast plants in the United States. cylinders were conducted on a Baldwin CT 300 (Figure 3-7). The current standard sets the load rate at 35 ± 7 psi per second which would dictate that a specimen of UHPC could take up to 15 minutes to break. namely the increase of the load rate applied to the specimen. Data was externally collected by DASYLab Version 8.

All specimens remained in the laboratory at ambient conditions after demolding until the time of testing. except the load rate was increased to 150 psi per second as mentioned for the compression testing. This variation from breaking a companion specimen was permitted because of the high predictability and low coefficient of variation (COV) of the compressive strength results based on previous work conducted by Kollmorgen (2004) and Graybeal (2005). and the fact that previous research completed at Michigan Tech used the manufacturer prescribed load rate.significant difference in UHPC compressive strength when changing the load rate from the ASTM standard to the recommended UHPC load rate. companion compression cylinders were not tested to determine the maximum applied load (0.5. 45 . Specimen Testing – Ec and ν The testing process followed ASTM C 469 – Standard Test Method for Static Modulus of Elasticity and Poisson’s Ratio of Concrete in Compression. For this reason. rather a background study determined the appropriate maximum load level as 40 percent of the average compressive strength based on age and curing regime of cylinders made during trial batches of UHPC. a load rate for compression testing of 150 psi per second was used. with the exception of the time that thermal treatment was applied. Load data was recorded at 5 Hz until ultimate failure of the specimen. Additionally. Specimen ends were prepared as described for compression testing one day before their scheduled testing time with the exception of the three day Air-cured cylinders as they were demolded and end-ground a few hours before testing. 3.2 Modulus of Elasticity and Poisson’s Ratio Specimen Preparation – Ec and ν The modulus of elasticity and Poisson’s ratio were conducted on 3 x 6 in.40 f `c ). cylindrical specimens.

This process occurred three times for each specimen. The specimens were loaded 150 psi per second until the predetermined maximum load based on curing regime and age. following the ASTM procedure. transverse and tangential displacement. Figure 3-8: Compressometer and Extensometer 3.A combined unbonded compressometer and extensometer equipped with two digital indicators. Load. as suggested by ASTM C 1018.0 at a rate of 5 Hz. Data from the second and third loading was averaged and reported as the results for the specimen.3 Flexural Strength Specimen Preparation – Flexure Testing was conducted on 2 x 2 x 11. and time were recorded by DASYLab Version 8.5. This smaller prism was used for 46 . The set up is shown in Figure 3-8. which measured the transverse and tangential displacements. The initial loading was used to seat the gauges and the data disregarded. and not conducted on the preferred size of 4 x 4 x 14 in.25 in. The specimens were completely unloaded at approximately the same rate and the gauges zeroed. beam specimens.

The ASTM outlines two possible configurations. Controlling the rate of deflection is the key to this test because if the test was conducted under load control. and recording the load and deflection so the data can be analyzed to give the flexural cracking stress. thick. to create a constant moment region. This test consists of loading a small prism at the third points. Additionally. the postcrack portion of the curve would not be suitable for analysis. and the method used in this research uses a rectangular jig to hold LVDTs that are averaged to determine net deflections as shown in Figure 3-9. The loading apparatus must allow the net midspan deflection to be recorded and to control the rate of deflection. with the exception of the thermal treatment duration. as Graybeal (2005) pointed out. the ASTM specifies that the cross section need only be three times the fiber length which was satisfied with the 2 x 2 in. 47 . toughness. Testing of beam specimens was conducted on a 55 kip MTS load frame with Test Star II controls and Test Ware data acquisition. and approximate flexural strength of the fiber reinforced concrete. cross section. and it was desired to compare the results of this study with Graybeal’s (2005) and Kollmorgen’s (2004) results. UHPC elements subject to flexural forces are not likely to be 4 in. No additional preparation was performed and all specimens remained in the laboratory at ambient conditions after demolding until the time of testing. When conducting this test.testing because. Specimen Testing – Flexure ASTM C 1018 Standard Test Method for Flexural Toughness and First-Crack Strength of Fiber-Reinforced Concrete (Using a Beam with Third Point Loading) was used to determine the first-crack strength and flexural toughness of Ductal®. special attention needs to be given to the requirements of ASTM C 1018 as it specifically calls for a testing configuration which can maintain the net midspan deflection using a closed-loop servo-controlled testing machine.

Two LVDT’s with strokes of ± 0.20 in.5. were used to measure the midspan deflection from each side of the prism relative to the bar which was epoxied to the top surface of the prism. The two LVDT’s signals were electronically averaged before being read by the testing software which controlled the closed-loop control system.4 Rapid Chloride Penetration For the rapid chloride penetration testing. The deflection measuring jig was secured to the prism at the neutral axis of the prism.Figure 3-9: ASTM 1018 Loading Configuration This configuration loaded the specimens at the third points of the span and created a simple support condition as outlined in ASTM C 78 Standard Test Method for Flexural Strength of Concrete (Using Simple Beam with Third-Point Loading) where the specification for the loading apparatus is given. This allowed the jig to remain stationary as the prism deflected. ASTM C 1202 – Electrical Indication of Concrete’s Ability to Resist Chloride Ion Penetration was followed for both the specimen 48 . per minute because ASTM C 1018 specifies that the first crack deflection be reached in 30 to 60 seconds.003 in. The midspan deflection rate was chosen to be 0. directly above the support points. 3.

preparation and testing. height using a kerosene cooled saw. The specimens were initially cast in a 4 in. During that time the current passing is recorded. The testing standard requires that specimens be vacuum saturated with water and tested for electrical conductance. ASTM C 1202 was still followed to allow for a qualitative comparison between UHPC and other concretes. however. the 49 . The kerosene remaining on the surface of the UHPC specimens was then evaporated off in an oven at 122°F for approximately two hours and then cooled to room temperature over another two hours. A water cooled saw was specified in the ASTM method. After the specimens had cooled. due to UHPC’s low permeability most of the kerosene was believed to be on the surface and therefore evaporated in the oven drying process. two-days prior to testing the specimens were cut down to the required 2 in. and used for calculating the total charge in coulombs passing. high cylinder and. Despite UHPC’s highly impermeable nature and the likelihood that it was never fully saturated through the specimen depth. The electrical conductance is measured by applying a 60-volt potential across a specimen that is mounted to a test cell with sodium chloride and sodium hydroxide solutions for 6-hours. Specimens were then placed in a plastic container inside a vacuum desiccator. Specimen Preparation – RCPT Preparation consisted of vacuum treating and impregnating the UHPC specimens with water prior to the actual testing for chloride ion penetration. de-airated water was introduced to cover the specimen while the vacuum was maintained for the final hour. After this time. At three hours into the vacuuming process. and a vacuum was applied over the specimen for a total of four hours (Figure 3-11). they were sealed on their side surface using Enviro-Tex self leveling epoxy mix and allowed to cure for at least 12 hours (Figure 3-10). diameter by 3 in. therefore.

For 6-hours the current passing through the specimen was monitored and recorded by an Agilent 34970A Data Acquisition unit. all of the specimens were cut and epoxy coated on day 26 and vacuum treated on day 27. To replicate field situations where the exposed surface of a UHPC beam would be smooth and finished. A 60-volt potential was induced across the test cells and specimen by a Kepco power supply (Figure 3-12). 50 . such that testing could begin on 28-day old specimens. and later would be used to calculate the total charge passing in coulombs. Figure 3-10: Epoxy-coated UHPC Specimens for RCPT Specimen Testing – RCPT Figure 3-11: ASTM C 1202 Specimen Preparation Setup After the specimen preparation was completed. Water was then drained and the container sealed to maintain an environment at 95% humidity until testing. the prepared UHPC samples were then tested following the ASTM C1202 procedures. the bottom cylinder surface (un-cut and finished) of the sample was placed facing the 3% sodium chloride (NaCl) solution.specimens were removed from the chamber and allowed to sit with the water covering them for 18 hours (± 2 hours).3 N (0. To provide proper time for preparation.3 normality) sodium hydroxide (NaOH) solution. while the saw cut side of the sample was placed facing the 0.

While initially the creation of a “short” circuit was a concern. At the conclusion of the test.1). Occasionally. therefore. the specimen could overheat leading to erroneous results and the boiling off of the solutions. However. the total charge passed was based on a 3. two cells were clamped together and a 60-volt potential was applied across both specimens.Figure 3-12: UHPC Specimen Undergoing ASTM C 1202 Testing Care was taken to ensure that the test cells did not leak and that no short circuit was created due to the steel fiber reinforcement. when more than one specimen was tested in a day. The data acquisition unit recorded the current passing for each individual specimen. no short circuiting occurred because of the random distribution and the short nature of the steel fibers. the data was compiled and a total charge passing was calculated by integrating the area under the current (amperes) versus time (seconds) curve.75” diameter specimen according to ASTM standards and. Equation 51 . The integration was performed by using the trapezoidal rule (as suggested by the ASTM standard) and the current passing at 30-minute intervals (Equation 3. If a short circuit was created.

75 ⎞ Qs = Qx × ⎜ ⎟ ⎝ x ⎠ 2 Equation 3.) of the nonstandard specimen Table 3. Q = 900( I 0 + 2 I 30 + 2 I 60 + .3.4: Chloride Ion Penetrability Based on Charge Passed (ASTM C 1202) Charge Passed (coulombs) Chloride Ion Penetrability High Moderate Low Very Low Negligible > 4.000 < 100 52 .4 was then used to evaluate the chloride ion penetrability of the UHPC specimen.1.2 was used to adjust the value obtained from Equation 3.2 where: Qs = charge passed (coulombs) through a 3. diameter specimen Qx = charge passed (coulombs) through a x in. diameter specimen x = diameter (in.000 2..000 100 . Using this adjusted total charge passed.1 Equation 3..000 .2..000 1.000 .1.75-in.4. Table 3. + 2 I 300 + 2 I 330 + I 360 ) where: Q = charge passed (coulombs) I 0 = current (amperes) immediately after voltage is applied I t = current (amperes) at t min after voltage is applied ⎛ 3.

After demolding and curing for 27-days and prior to testing. into the ends of the specimen. Only one minor variation was adapted for the testing of UHPC specimens as noted below. Specimen Preparation – Freeze/Thaw The freeze-thaw specimens were cast in 3 x 4 x 15. Soaking of the specimens according to ASTM C 666 in a lime bath for 48-hours was not performed to avoid impacting the curing regimes. leaving 1/4 in. due to the high impermeability of UHPC.3. (length between the embedded ends of the two gauge studs).5 Freeze-Thaw Cyclic Testing Freezing and thawing testing was performed in accordance with ASTM C 666 – Resistance of Concrete to Rapid Freezing and Thawing standards. (length between the two exposed ends of the gauge studs) and a gauge length of 13. stainless steel gauge studs embedded 1 in.5 in. In Procedure A. Approximately every 32 cycles the fundamental transverse frequency. whichever came first.5.5 in. This may have allowed some UHPC freeze-thaw specimens to be inadequately saturated. beam molds with 1-1/4 in. However. This produced specimens with a nominal length of 16 in. it is unlikely that any of the specimens would have become fully saturated in 48-hours. the UHPC specimens were cooled to the thaw temperature of the freeze-thaw test machine by placing them in a 41°F water bath for at least 16 hours. of the gauge studs exposed on each end. Specimens were tested until failure or 300 freeze-thaw cycles. and mass were observed and recorded. Specimen Testing – Freeze/Thaw There are two methods specified in ASTM C 666 for determining the freeze-thaw durability of a concrete – Procedure A and Procedure B. the specimens are 53 . Eight cycles were completed per day by means of freezing in air and thawing in water. length change.

The fundamental transverse frequency was measured according to ASTM C 215 – Fundamental Transverse. and Torsional Frequencies of Concrete Specimens and involved striking the end of the UHPC specimen with a precision 54 . and length change were evaluated at regular intervals. As a chamber of this size contained too many slots to be completely occupied by UHPC specimens during testing. Figure 3-13: MTU 80-specimen Freeze-Thaw Chamber (Procedure B) During testing. Longitudinal. these control specimens were rotated to ensure each specimen slot’s freeze-thaw temperature cycle conformity to the ASTM C 666 requirements. Prior to using any specimen slot. three UHPC control specimens were cast with type-T thermal couples embedded to monitor and control the temperature in the chamber. Furthermore. Testing for freeze-thaw durability was performed in an 80-specimen Scientemp freezethaw chamber (Figure 3-13). The current freeze-thaw testing machine at Michigan Tech is designed for Procedure B. the specimens’ fundamental transverse frequency. a large number of NSC “dummy” specimens were made to maintain a full-load and proper heating/cooling rates in the machine. while in Procedure B the specimens are frozen in air and thawed in water. mass.frozen and thawed in water.

weighted steel impact hammer and measuring the dynamic response of the specimen with an accelerometer. The specimen’s mass was also recorded to the nearest 0. Using this average fundamental transverse frequency. An average fundamental transverse frequency of the specimen was determined from three strikes with the impact hammer. 8. the relative dynamic modulus (RDM) of the UHPC specimen was calculated as the ratio between the fundamental transverse frequency of a specimen after n cycles and the fundamental transverse frequency of the specimen immediately prior to testing.01 lbs using a calibrated scale. Accelerometer Impact hammer Styrofoam Figure 3-14: Testing the Fundamental Transverse Frequency of an UHPC Specimen While defining failure of a freeze-thaw specimen based on a length increase of 0.10% is one option according to ASTM C 666. thick Styrofoam pad was used to dampen any outside frequency interference.0 2004) calculated the specimen’s fundamental transverse frequency (Figure 3-14). A 2 in. the 2001 Michigan Test Methods (MTM 115) specifically 55 . and a fast-Fourier transform method in DASYLab (ver.

However. length. Figure 3-15: Length Change Measurement of an UHPC Freeze-Thaw Specimen One freeze-thaw cycle was completed every 3-hours and the first measurements of initial RDM. and mass were taken at 24-hours (8 cycles) after the specimens began cycling. the length change of each specimen was measured using equipment specified in ASTM C 490 – Use of Apparatus for the Determination of Length Change of Hardened Cement Paste. Specimens were removed during the thaw cycles at approximately 40°F. Specimens were measured horizontally by holding the specimen against the base of the length comparator measuring the relative length on the dial gauge (Figure 3-15). so that accurate length comparisons could be made. which indicates the need to use a horizontal comparator to measure specimens with cross sections larger than 9 in. Therefore.2. The horizontal measuring method was employed due to Note 1 in ASTM C 490. because the length measurement was only used for a relative comparison between specimens a vertical measurement may also be sufficient. Repeat measurements were then taken at 96 hour intervals (32-cycles) until testing was complete (300 cycles or specimen failure. Mortar and Concrete.use this length change parameter as the test for specimen failure. which ever occurred first). A standard rod was used to zero the length comparator before each test to guarantee that each measurement was made relative to a standard length. 56 .

Therefore. two additional beams were cast in each batch (six in total) to investigate the impacts of water absorption and additional hydration of unhydrated cement particles on the RDM and mass of the UHPC specimens in the freeze-thaw chamber. 3. and then subjected to heating and cooling cycles until an accurate CTE measurement is obtained. Modifications to both specimen preparation and testing procedures were made due to equipment specifications and in the interest of maintaining the integrity of UHPC curing regimes. Detailed changes to specimen preparation and testing for UHPC are described below. some cement particles do not become fully hydrated even after the curing regime 57 . The standard process according to AASHTO TP-60-00 requires that specimens be saturated with water prior to testing. Specimen Preparation – CTE Typically.5% is observed (whichever comes last). due to the extremely low water-cement ratio in UHPC.6 Coefficient of Thermal Expansion The coefficient of thermal expansion for concrete was determined following a modified version of AASHTO TP-60-00 – Coefficient of Thermal Expansion of Hydraulic Cement Concrete (2004).5.2% mass gain after 300 cycles. However. To simulate the wetting and drying of the specimens undergoing Procedure B freezing and thawing cycles without temperature change. these extra “side-study” specimens were cycled every 24 hours between air and water in a water bath kept at ambient temperature. placed into a test frame and submerged in a thermally regulated water bath.Graybeal (2006a) observed that untreated UHPC specimens actually experienced as much as a 10% increase in their relative dynamic modulus and a 0. specimen preparation for CTE testing involves saturating the concrete specimen in a limewater bath for a minimum of 48-hours prior or until a mass change of less than 0.

58 .) micrometer was used to measure the specimen length at these 8 points. Therefore. the standard preparation of submerging specimens in a limewater bath for 48-hours prior to testing was not employed. 8 marks were made around the circumference of the cut side of the cylinder and a 12 in. After the epoxy application was complete. The bottoms of the specimens were not cut due to the already smoothly formed surface from the base of the mold. cylinder) by removing the top 1 in. Instead. it is important that the UHPC specimens do not absorb water that can react with unhydrated cementitious particles. because the bottom of the cylinder rested on three steel support buttons in the frame and was not in contact with the linear variable differential transformer (LVDT). Because the curing regime is one of the variables being investigated through this research. To measure perpendicularity. the 4 x 8 in. specimens were preserved in the unsaturated condition by coating them with an epoxy sealant. Following the perpendicularity measurements. Nevertheless. After curing for 26 days under the specified curing regime. However. cylinder specimens were cut to size (4 x 7 in. the epoxy was allowed to cure for at least 12 hours. The specimens were oven-dried at 122°F and then measured for perpendicularity. (±0. some problems with perpendicularity arose on some specimens. it was not considered harmful to the length change measurements. of the specimen using a kerosene saw. This allowed for direct contact between the specimen and LVDT frame support buttons as well as the LVDT transducer. the UHPC specimens were completely coated with epoxy (Figure 3-16) except for the locations on each end where the steel support buttons of the test frame and the LVDT tip came into contact with the specimen.is applied.0005 in.

some differences exist. The components of this system included a 25-gallon water bath with a temperature stability of ±0. measurements were taken every 5 minutes over a half-hour period for the FHWA procedure rather than every 10 minutes as noted by AASHTO. All of the equipment and programming was designed in agreement with the FHWA Procurement Specification (Appendix B). the specimen measuring frame used a correction factor procedure different from that outlined in AASHTO TP 60-00 Appendix X. rather than the 0.2.00005 in. The specimen test frame was calibrated in accordance with manufacturer’s instructions and utilized a 1 in. as specified by the AASHTO TP 60-00 specification. Additionally. The main variation was the measurement tolerance of the specimens that was stipulated by the FHWA specification to be 0. Also. stroke linear variable differential transformer with a resolution of 0.00001 in.Figure 3-16: Epoxy-coating CTE Specimen Specimen Testing – CTE UHPC CTE testing utilized Pine Instrument Company’s AFCT1A Coefficient of Thermal Expansion of Portland Cement Concrete Measurement System. The 59 .2°F over a temperature range of 41°F to 160°F (Figure 3-17).0002 in. when determining if a specimen had reached thermal equilibrium. While the majority of the FHWA Specification follows AASHTO TP 60-00.

Overall. However. 60 . These tests were continued until the AASHTO specification was attained. several specimens required longer than one test cycle to meet the AASHTO specification for successive CTE values (difference of no more than 0. 50°F to 122°F. and a cooling segment. 122°F to 50°F.5 micro strain/°F between successive CTE values). these changes were necessary for testing using the equipment available. Figure 3-17: Pine CTE Specimen Test Frame and Water Bath After a specimen had been prepared. The CTE values of UHPC were obtained by taking the average CTE of the specimen during a heating segment. mass and length were recorded prior to placing it in the water bath (Figure 3-18) for testing.correction factor procedure in the procurement specification provides for a two-point calibration that allows a greater range of specimen lengths to be tested in a given frame (Pine 2006). One test cycle (one heating segment and one cooling segment) lasted approximately 24 hours.

Figure 3-18: UHPC Specimen in Water Bath Undergoing CTE Testing 61 .

(This page intentionally left blank) 62 .

NF 28 Number of Specimens Tested Compressive Strength Modulus of Elasticity Poisson’s Ratio 1st Flexure Cracking Curing Regime† Air-cured RCPT F/T CTE TT (Thermal Treatment) DTT (delayed TT) DDTT † 6 6 6 6 6 6 6 6 5 6 6 5 5 4 6 6 6 6 6 6 6 5 6 6 5 5 4 6 6 6 5 6 6 6 5 6 6 5 5 12 3 12 3 12 3 - 4 3 4 - 4 4 - 3 4 3 3 3 3 - Curing Regime summary: Air-cured = ambient lab conditions 72°F. modulus of elasticity. and analyzed and the summary data presented herein (additional data can be found in Appendix A). and coefficient of thermal expansion (CTE) of an ultra-high performance concrete were tested with differing curing conditions and at different ages to examine how these factors influence each of the properties. rapid chloride permeability (RCPT). Compressive strength.NF 14 28 28 .Specimens Tested per Curing Regime Specimen Age at Testing (days) 3 7 14 28 28 – NF‡ 7 14 28 28 .0 Results and Discussion Several properties of UHPC were investigated during this research project. Calculating sample mean values revealed the central tendencies of the data. synthesized. 30-50% relative humidity TT = Thermal treatment of 194°F. ambient cure otherwise DTT = Delayed thermal treatment beginning at age 10-days DDTT = Doubly-delayed thermal treatment beginning at age 24-days ‡ N.1 lists the total number of specimens tested for each UHPC property studied under several curing regimes. freeze-thaw resistance (F/T). 100%RH for 48 hours beginning at age 3-days. = refers to UHPC specimens cast without fibers The resulting data from these tests was recorded. while the standard deviation and 63 . Table 4.F. flexural characteristics.4.1: Experimental Test Matrix . Table 4. Poisson’s ratio.

2. 4.1. 4. cylindrical specimens were tested for compressive strength. A. However. for example an analysis of variance (ANOVA). The COV is the standard deviation divided by the mean which shows the relative variability of a sample set. Complete details of the hypothesis testing and results can be found in Peuse (2008) and Misson (2008). – comparing TT-cured UHPC freeze-thaw specimens to Air-cured freeze-thaw specimens) a series of statistical F-tests and t-tests were performed to test the null and alternate hypotheses on the sample sets from each curing regime and age.coefficient of variation (COV) were used to assess dispersion tendencies.2 shows the mean and COV for the compressive testing based on curing regime and age. the ANOVA was not an acceptable approach.2 summarizes the compressive stress test results for various curing regimes and ages tested. Test results are also compared to results in currently published literature. this test can only tell if one or more sample group is statistically different. as well as each property conducted. A summary of those results are provided here. Table 4.1.3). A statistical analysis was conducted to compare the effects thermal treatment on the properties of UHPC. Test data recorded for each specimen is tabulated in Appendix A (Tables A. Many options exist for testing several sample sets of data at the same time. When comparing data from two populations (e. and does not locate which sample is different. Because the focus of the study was to determine how the different curing regimes performed compared to each other.1 Compression Strength A total of 75 3x6 in.1 Results Table 4. A.g. 64 .

5 2.NF 14 28 28 .4 percent for samples prepared from the same sample of concrete and tested at the same age.9 22.8 1. but in general the COV for the UHPC is similar to the expected value in ASTM C 39.0 2.9 24. Peuse (2008) compared intra-batch results as well as inter-batch results to find no significant error was introduced by using one specimen per batch.2: Compressive Stress Test Results Specimen Age at Testing (days) 3 7 14 28 28 . 65 . Figure 4-1 is a graphical representation of the mean compressive stress results for each age and curing regime.NF 7 14 28 28 .2 2.2 3.9 4.4 19.NF 28 Curing Regime Air-cured Number of Specimens (Sample Size) 6 6 6 6 6 6 6 6 5 6 6 3 5 Sample Mean (ksi) 14.3 23.9 percent in which all but two curing regimes and ages fall under.9 31.3 30.9 30.9 29. ASTM C 39 reports an expected within-test COV of 2.6 29.Table 4.8 3. The referenced standard also gives an acceptance range of 7.8 percent for 3 individual cylinders which is based on two times the standard deviation and thereby correlates to a COV of 3.4 Sample COV (%) 3. cylinders with compressive strengths of 2000 psi to 8000 psi.7 29.2 3.1 31. The error bars plotted above and below the mean values show ± one standard deviation for the respective specimen ages and curing regimes.2 TT (Thermal Treatment) DTT (delayed TT) DDTT The data shows that the COV for all compressive specimens was very low and is consistent with the COV shown in ASTM C 39.3 6. In addition.1 31.7 3.2 3. These results were based on over 1200 test reports on 6 x 12 in. Obviously the strength of UHPC is outside the tested range and specimens tested came from different batches.6 1.

35.0
Mean Compressive Stress (ksi)

30.0 25.0
Air

20.0 15.0 10.0 5.0 0.0 3 Day 7 Day 14 Day 28 Day 28 Day No Fibers

TT DTT DDTT

Age of Specimens

Figure 4-1: Mean Compressive Results for All Ages and Curing Regimes

It should be noted that a typical specimen failure was characterized as a shear failure by ASTM C 39. The failure plane extended from the top corner to the opposite bottom corner and the two pieces of the specimen experienced fiber pullout and fiber breakage. Specimens tested without fibers were very explosive, leaving most of the specimens in small pieces. 4.1.2 Statistical Analysis and Discussion The main focus of this research was determining the impact that age at thermal treatment had on several properties of an ultra-high performance concrete, including after thermal treatment was complete. For example, is the compressive strength of a cylinder which underwent thermal treatment from 3-6 days the same on day 7 as it is on days 14 and 28? The compressive stress of 3 x 6 in. cylinders under the four curing regimes of air-curing, thermal treatment, delayed thermal treatment, and double delayed thermal treatment at various ages were 66

compared first within each curing regime and then later to each other to analyze the effect each curing regime has on compressive strength. Hypothesis testing for pairings within curing regimes were initially run through an F-test to determine if the population variance for the two curing regimes were equal. Pairings were then compared using a t-test with the null hypothesis (Ho) equating the population means and the alternate hypothesis suggesting that they were unequal. Details of the statistical testing can be found in Peuse (2008). Table 4.3 summarizes the statistical results from hypothesis testing for comparison within each curing regime. Air-cured specimens continued to cure over the 28 day duration and all pairings were found to be not equal. Section 4.1.3 discusses strength growth with time. Air-cured specimens represent all of the thermally treated specimens up to the time of thermal curing, e.g. TT specimens were not tested at an age of 3 days because thermal treatment had not yet begun. The DDTT thermal curing occurred during days 24-27 which meant that only 28 day specimens were tested for this curing regime and no hypothesis testing was possible.

67

Table 4.3: Statistical Results for Compressive Strength Testing

Age of Specimen Pairings (days)

Reject or Failed to Reject the Null Hypothesis (t-test) Air-cured Specimens Reject Ho Reject Ho Reject Ho Reject Ho Reject Ho Reject Ho TT – Thermal Treatment Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho DTT – Delayed Thermal Treatment Failed to Reject Ho

Population Mean

3 vs 7 3 vs 14 3 vs 28 7 vs 14 7 vs 28 14 vs 28

Not Equal Not Equal Not Equal Not Equal Not Equal Not Equal

7 vs 14 7 vs 28 14 vs 28

Equal Equal Equal

14 vs 28

Equal

The purpose of conducting the hypothesis testing within the curing regimes was to determine if the data had equal population variance and the same population mean. If all the age groups (i.e. 7, 14, and 28 day) from a particular curing regime had the same population mean, then all the data could be put together to make a larger sample set for each curing regime. These larger sample sets could then be used to compare the curing regimes to one another. The results showed that the Air-cured specimens did not have the same population mean, and therefore could not be combined to form a larger sample set. Specimens tested at 28 days were chosen to represent the Air curing regime because the sample set had the highest compressive strength. Specimens undergoing TT and DTT did have the same population mean in their respective data set. Therefore, data from each age group of TT and DTT were combined into two data sets, one for TT and one for DTT. Table 4.4 shows the organization for the combined hypothesis testing.

68

Table 4. 14. the mean compressive stress of 69 .3 2. Table 4.4 Combined Sample COV (%) 2.2 3. Table 4. Therefore.5 is a summary of the results of the hypothesis testing based on curing regime.5 29. regardless of when. However.6 3. 28 28 Curing Regime Air TT DTT DDTT Sample Size 6 18 11 5 Combined Sample Mean Compressive Stress (ksi) 23.2 As with the individual curing regimes. had the same population mean. For tests which include the Air curing regime. hypothesis testing all combinations were initially run through an F-test to determine if the population variance for the two curing regimes were equal. The same combinations underwent hypothesis testing using a t-test with the null hypothesis equating the population means and the alternate hypothesis suggesting that they were unequal. the null hypothesis was rejected and the population means were not equal.5: Statistical Results for Combined Compressive Strength Testing Curing Regime Pairings Air vs TT Air vs DTT Air vs DDTT TT vs DTT TT vs DDTT DTT vs DDTT Reject or Failed to Reject the Null Hypothesis (t-test) Reject Ho Reject Ho Reject Ho Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho Population Mean Not Equal Not Equal Not Equal Equal Equal Equal Hypothesis testing indicated that specimens which had undergone thermal treatment.9 30. the remaining pairings showed that the null hypothesis was not to be rejected and that all three thermal treatment curing regimes shared the same population mean.8 29. 28 14.4: Combined Compressive Stress Results Specimen Age at Testing (days) 28 7.

cylinders. respectively. there is not a great difference in when thermal treatment occurs. Determining how quickly and to what ultimate capacity UHPC will increase in strength under ambient conditions is an important factor for the design and use of UHPC.1 percent lower than what observed in this research program. the average compressive strength of thermally treated specimens (7 day. Kollmorgen (2004) reported an average compressive strength prior to thermal treatment (3 day air-cured) which was 40 percent lower than results seen in this research and had a COV of 15 percent.3 Air-Cured Compressive Strength Growth over Time Previous sections have shown that once UHPC receives a thermal treatment its compressive strength properties vary little.9 to 4.0. 14 day. and 28 day) ranged from only 3. The valuable conclusion of this statistical analysis is that for compressive strength measured after thermal treatment has been applied.specimens undergoing thermal treatment independent of age is 30. steam.6. 70 . Previous work completed by Kollmorgen (2004) and Graybeal (2005) showed similar compressive strengths and COV on 3 x 6 in. The COV varied from 3. COV ranged from approximately 3 to 6 percent for Graybeal’s reported data. 4.6 percent lower than results reported by this research.1. However.0 percent. and delayed steam cured specimens which were 23. and 17. However. 10. This would have a large benefit to the precasting industry if elements could be cast individually on different days but cured together at some time in the future. Graybeal (2005) reported 28 day compressive strengths for air.9 ksi. it was also shown that until UHPC receives a thermal treatment its strength increases with time much like traditional NSC or HPC.3 to 4.1 ksi. which is an increase of 25 percent over the mean of Air-cured specimens at 28 days which is 23.

As part of the material study on 71 . The mean results of the compressive strength over time for specimens cured under ambient conditions can be seen along with the mean results of Air-cured specimens for this research in Figure 4-2.0 0 10 20 30 40 Specimen Age (days) 50 60 Figure 4-2: Compressive Stress Gain over Time for Air-Cured Specimens Both sets of data present the classic asymptotic shape of the compressive strength approaching some ultimate limit.0 25. There are several possibilities which could cause this difference in results.Graybeal (2005) tested specimens with four curing treatments for compressive strength. In fact both data sets seem to have very similar curves. 30. Recall that specimens were demolded after 3 days for this research program and 24 hours for work conducted by Graybeal (2005).0 5.0 10. but the data from this research has a greater compressive strength. and linearity of the response over time.0 15. Graybeal’s (2005) data appears to be approaching 20 to 22 ksi while this research data appears to approach 25 to 27 ksi. modulus of elasticity. The first is the age at which specimens were demolded.0 Compressive Stress (ksi) 20.0 Graybeal Data Air Data 0.

casting. and would result in lower strength (Graybeal 2005).2 Modulus of Elasticity and Poisson’s Ratio A total of 73 3x6 in. The age of premix used for specimens in the compressive strength specimens was approximately 2 to 4 months. A second possibility is the age of premix at the time of mixing. and specimen preparation and testing 72 . the ambient curing conditions could be different in each laboratory. The exact constituents and their proportions is not know as this is a proprietary UHPC and the premix comes blended.UHPC. Both properties could be measured simultaneously on any given specimen. None the less. A third possibility would be different constituents or their proportions in the premix. It is possible that UHPC curing in higher relative humidity could use the more readily available water vapor continue hydration. July. Mixing. Graybeal (2005) investigated the impact of age at demolding had on compressive strength. they were not controlled. Both Kollmorgen (2004) and Graybeal (2005) noted that the older the premix was the longer it took to mix each batch. and Graybeal (2005) saw that older premix took longer to set. Although the temperature and relative humidity did not vary excessively during these months (~70°F. and August. and curing of specimens for this research were conducted in the basement of Benedict Laboratory at Michigan Tech during June. data clearly shows that UHPC continues to gain strength while curing at ambient conditions. cylindrical specimens were tested to investigate the effects of age at thermal treatment on modulus of elasticity and Poisson’s ratio. an increase in 28 day compressive strength of 5 ksi was seen in specimens receiving ambient curing. 4. The results showed that by leaving the specimens in the molds to 48 hours. If the premix was old and specimens had just enough strength to survive demolding and then were exposed to ambient conditions. 30-50% RH). the specimens would likely dry out before using all available water for hydration. Finally.

696 8.177 8.7 1.741 8.1 Results Table 4.200 0.3).205 0.2.910 7.7 2.098 Mod.206 0.NF 14 28 28 .NF 28 Curing Regime Air-cured Number of Specimens (Sample Size) 4 6 6 6 6 6 6 6 5 6 6 3 5 Mod. A.0 TT (Thermal Treatment) DTT (delayed TT) DDTT This table shows data with a very low COV and is very similar to previous results with this type of UHPC (Kollmorgen 2004.205 0.865 7.6 summarizes the mean and COV for the modulus of elasticity and Poisson’s Ratio based on curing regime and age.2 1.8 1.0 1.procedures were discussed in Chapter 3.215 8.2 1.2. It is because of this low variability that statistically significant results can be determined with low sample sizes.200 0. 4.2 1. except the load rate was increased to 150 psi per second.1.6 1.161 7.1 1.203 0. Two specimens which were Air-cured and compression tested at 3 days were not tested for modulus of elasticity and Poisson’s ratio because of difficulties with the data acquisition system. Graybeal 2005). Test data recorded for each specimen is tabulated in Appendix A (Tables A.205 0.NF‡ 7 14 28 28 . 73 . In general.198 0. of Elasticity Sample Mean (ksi) 6. A. the process followed ASTM C 469 – Standard Test Method for Static Modulus of Elasticity and Poisson’s Ratio of Concrete in Compression.203 Poisson’s Ratio Sample COV (%) 2.6% 1.863 7.7 2.205 0.056 8.1 1.206 0.5 0.520 7.6 4.2 2.206 0.1 2.4 1.8 1.4 2.889 8.0 Poisson’s Ratio Sample Mean 0.114 7.9 0.6: Modulus of Elasticity and Poisson’s Ratio Test Results Specimen Age at Testing (days) 3 7 14 28 28 . of Elasticity Sample COV (%) 0.3 0. Table 4.6 2.3 4.5 2.207 0.

0 12. for each age and curing regime.0020 Limits of Stress-Strain Curve for Modulus of Elasticity Calculation (Gage reading 0.0005 0.0 0.0015 0. to 90 kips) Figure 4-3: Typical Stress-Strain Curve for Calculating the Modulus of Elasticity Figures 4-4 and 4-5 show the mean modulus of elasticity and Poisson’s ratio results.0 0.00005 in. 74 .0 6.Figure 4-3 shows a typical stress-strain graph for the calculation of modulus of elasticity. The error bars indicate ± one standard deviation from the mean.0 2.0 4.0 8. respectively.0010 Axial Strain (in/in) 0. 14.0000 0. This particular specimen was thermally treated from Batch 1 and tested at 14 days.0 Axial Stess (ksi) 10.

6 and Figure 4-4 is that the modulus value increases with time for the Air-cured specimens.000 6.00 3 Day 7 Day 14 Day 28 Day 28 Day No Fibers Age of Specimens Mean Poisson's Ratio Air TT DTT DDTT Figure 4-5: Mean Poisson’s Ratio Results for All Ages and Curing Regimes The obvious trend which can be seen in the Table 4.9.06 0.10 0.04 0.12 0.20 0.000 4.18 0.02 0.000 7. The modulus value increases over 950 75 .22 0.000 3.14 0.16 0.08 0.000 1.000 5.000 2.000 Mean Modulus of Elasticity (ksi) 8.000 0 3 Day 7 Day 14 Day 28 Day 28 Day No Fibers Air TT DTT DDTT Age of Specimens Figure 4-4: Mean Modulus of Elasticity Results for All Ages and Curing Regimes 0.

8 percent from 3 to 28 days. Table 4. DTT mean modulus of elasticity values showed a difference of only 20 ksi from 14 to 28 days. was conducted for modulus of elasticity and Poisson’s Ratio comparing first within each curing regime and then later to each other for each property to analyze the effect of each curing regime. similar to compressive stress testing. This is to be expected because of the increase in compressive stress that was noted in the previous section.2 Statistical Analysis and Discussion Statistical analysis. which is in line with typical concrete and the value given by AFGC (2002). Poisson’s Ratio for all curing regimes and ages were consistently between 0.2. 4. A less noticeable trend in the data is that the modulus of elasticity decreased when the fibers were removed from the mix.21. when comparing all means of specimens undergoing thermal curing at any age the range in modulus values was from 8056 to 8215 ksi. TT results only vary 159 ksi or about 2 percent independent of age. In fact.7 shows the results of the hypothesis testing on all specimens for modulus of elasticity. 76 .20 and 0.ksi or 13.

e.7: Statistical Results for Modulus of Elasticity Testing Age of Specimen Pairings (days) Reject or Failed to Reject the Null Hypothesis (t-test) Air-cured Specimens Reject Ho Reject Ho Reject Ho Reject Ho Reject Ho Failed to Reject Ho TT – Thermal Treatment Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho DTT – Delayed Thermal Treatment Failed to Reject Ho Mod.2.g. and all DTT specimens shared the same population 77 . However. the Modulus was achieved and remained constant independent of when thermal treatment was applied. and that all TT specimens shared the same population mean. The t-tests have shown that the 14 and 28 day Air-cured sample sets had the same population mean.3 discusses the relationship between concrete strength and stiffness (modulus of elasticity). TT specimens were not tested at an age of 3 days because thermal treatment had not yet begun. The purpose of conducting hypothesis testing on all curing regimes was to determine if the specimens tested at different ages had the same population mean. of Elasticity Population Mean 3 vs 7 3 vs 14 3 vs 28 7 vs 14 7 vs 28 14 vs 28 Not Equal Not Equal Not Equal Not Equal Not Equal Equal 7 vs 14 7 vs 28 14 vs 28 Equal Equal Equal 14 vs 28 Equal Air-cured specimens continued to gain stiffness over the first 14 day duration and all pairings were found to be not equal. once thermally treated. The DDTT thermal curing occurred during days 24-27 which meant that only 28 day specimens were tested for this curing regime and no hypothesis testing was possible. Air-cured specimens represent all of the thermally treated specimens up to the time of thermal curing.Table 4. Section 4.

8 shows the combined values. 14. 28 28 Curing Regime Air TT DTT DDTT Sample Size 12 18 11 5 Combined Sample Mean Mod.8: Combined Modulus of Elasticity Results Specimen Age at Testing (days) 14. of Elasticity (ksi) 7.168 8. all three combinations showed that the population mean was equal for all curing regimes.1 1. It is interesting to note that the compressive strength of 78 .0 The F-test showed that all the combinations of curing regimes had an equal population variance for the modulus of elasticity results.129 8.8 percent over the combined Air-cured value of 7850 ksi. 28 14. Conducting a t-test was the next step which produced results showing that the population mean of the Air-cured specimens was different than TT.5 1. DTT. Table 4. Table 4. The data sharing the same population mean was further combined and hypothesis testing was completed. This is the same result obtained from the compressive hypothesis testing.9: Statistical Results for Combined Modulus of Elasticity Testing Curing Regime Pairings Air vs TT Air vs DTT Air vs DDTT TT vs DTT TT vs DDTT DTT vs DDTT Reject or Failed to Reject the Null Hypothesis (t-test) Reject Ho Reject Ho Reject Ho Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho Population Mean Not Equal Not Equal Not Equal Equal Equal Equal When the specimens that received thermal curing were compared.098 Combined Sample COV (%) 1. which is an increase of 3.3 1. Table 4.mean. For specimens receiving thermal treatment at any age the mean modulus of elasticity is 8150 ksi. Table 4. 28 7. and DDTT.864 8.9 shows the results of the t-test.

10: Statistical Results for Poisson’s Ratio Testing Age of Specimen Pairings (days) Reject or Failed to Reject the Null Hypothesis (t-test) Air-cured Specimens Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho TT – Thermal Treatment Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho DTT – Delayed Thermal Treatment Failed to Reject Ho Poisson’s Ratio Population Mean 3 vs 7 3 vs 14 3 vs 28 7 vs 14 7 vs 28 14 vs 28 Equal Equal Equal Equal Equal Equal 7 vs 14 7 vs 28 14 vs 28 Equal Equal Equal 14 vs 28 Equal Similar to compression stress and modulus of elasticity results and based on the above analysis. Table 4. Results are summarized in Tables 4.11 and 4. Poisson’s ratio hypothesis testing results were combined for curing regimes and compared for statistical differences. All curing regimes were compared to determine the effects that the different curing regimes had on the Poisson’s ratio.12.10 shows the results of the hypothesis testing on all specimens for Poisson’s ratio tests. 79 . Table 4. A similar result was noted by Graybeal (2005) as the data showed the compressive strength increasing to 8 weeks after casting but the stiffness and peak strain at failure curtailed at 4 weeks after casting.Air-cured specimens continued to increase from day 14 to 28 but the modulus value appears to have reached a plateau by 14 days.

1 3.206 0.203 Combined Sample COV (%) 3. These specimens had a high COV at 20 percent. caution would have to be exercised in stating that a difference was evident because Poisson’s ratio is reported and used in equations to two decimal places. Kollmorgen (2004) reported specimens tested before thermal treatment having modulus values 5. 7.204 0. 14. So if a difference shows up in the third or fourth decimal place. 28 14.Table 4.1 percent less than those presented in this research.6 1.21. it has no practical application. Based on this analysis. 14.206 0.11: Combined Poisson’s Ratio Results Specimen Age at Testing (days) 3. 28 7.12: Statistical Results Poisson’s Ratio Testing Curing Regime Pairings Air vs TT Air vs DTT Air vs DDTT TT vs DTT TT vs DDTT DTT vs DDTT Reject or Failed to Reject the Null Hypothesis (t-test) Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho Failed to Reject Ho Population Mean Equal Equal Equal Equal Equal Equal Results show none of the null hypotheses were rejected and all six pairings show that all curing regimes come from the same population mean. Even if there was a slight difference in the population mean (a few thousandths). Following thermal treatment and independent of age at testing a mean value of 9210 ksi was presented which is over 13 percent greater than the results for TT at 28 days and 80 .0 2. data for the UHPC specimens (Air-cured and all thermally cured specimens independent of age at curing) have a mean Poisson’s ratio of 0. 28 28 Curing Regime Air TT DTT DDTT Sample Size 22 18 11 5 Combined Sample Mean Poisson’s Ratio 0.0 Table 4.

All specimens which were determined to have the same population mean 81 . The modulus of elasticity values were calculated over the ranges of 10 to 40 percent of the ultimate capacity of each specimen. It was noted that cylinders which underwent thermal curing reached 80 to 90 percent of their ultimate capacity before deviating from a linear elastic behavior by 5 percent. While an in depth investigation on Poisson’s ratio was not conducted by Graybeal (2005). steam. there are several different models which relate the compressive strength of UHPC to its modulus of elasticity. therefore.the COV was 7. for specimens tested before and after thermal treatment. It is not the intent of this research to propose a relationship. and DTT results. but rather to determine if published equations predict the modulus of elasticity given the compressive strength for the data collected herein. 7650 ksi.1 percent. These results were based on the utilization of three cylinder sizes and LDTs as presented in Chapter 3. This shows UHPC as exhibiting nearly linear behavior to high stress levels.16 to 0. 5. As discussed in Chapter 2. 4.19 and COV of 13.7 percent. Graybeal’s (2005) results are 21. COV for the three curing treatments were very low ranging from 2 to 3 percent.3 percent. ranged from 0. and 7300 ksi respectively. a simple yet acceptably accurate way to predict this property is necessary. a background study completed for determining an acceptable load rate indicated consistent Poisson’s ratio results of 0. and 10.22 with a mean of 0.2.3 Compressive Stress and Modulus of Elasticity Relationship The modulus of elasticity of concrete is an important parameter and. Graybeal (2005) presents modulus of elasticity results for air. and delayed steam cured specimens as 6200 ksi.6 percent lower than Air-cured.1 percent. TT. Poisson’s ratio results. When comparing these results to those presented herein based on 28 day means.19 for thermally cured specimens and compares well with research results presented herein.

However.3) Iowa (Eqn 2.4) Kollmorgen (Eqn 2. TT specimens at 7.5) Graybeal (Eqn 2. For example. 14.6) Air Data TT Data DTT Data DDTT Data 30. Figure 4-6 shows the experimental data plotted with the different prediction equations for modulus of elasticity based on compressive strength.0 25.0 15.0 20.0 Compressive Stress (ksi) ACI 363 (Eqn 2. 7.1) 5000 4000 3000 2000 1000 0 0. 10000 9000 8000 Modulus of Elasticity (ksi) 7000 6000 ACI 318 (Eqn 2. Air-cured specimens at 3. and 28 days had different population means and means are shown separately.0 35. The results showed that the sum of squares of the residuals was minimized when the AFGC (2002) model was used.0 5. The sum of squares of the residuals was calculated for each equation to determine the best fit for the data.2) AGFC (Eqn 2.0 10. Compressive Strength The predictive relationships vary significantly and none are a great match for all the data gathered during this research. and 28 days had the same population mean so all specimens were averaged and included in the following figure.0 Figure 4-6: Regression Model for Modulus of Elasticity vs. 14. The value of the sum of 82 .were used in the relationship of modulus of elasticity and compressive strength.

0 20.0 10.34.0 35.98 which is an excellent fit.squares of the residuals was less than half of the next smallest value which was the Iowa model. Just minimizing the sum of squares of the residuals does not guarantee a good fit. Recall that the AFGC (2002) equation requires the specimens to undergo a thermal treatment. so limiting the analysis to the thermally treated specimen is warranted.0 30. However. The coefficient of determination which shows the goodness of fit and was calculated to be 0. Figure 4-7 presents the mean results (based on age at testing) of specimens cured under ambient conditions and tested for compressive stress and modulus of elasticity. However. the coefficient of determination increased to 0. when the Air-cured specimens were removed from the analysis.0 Compressive Stress (ksi) Figure 4-7: Mean Values of Compressive Stress and Modulus of Elasticity for Air-Cured Specimens 83 . The results are summarized with data from Kollmorgen (2004) and Graybeal (2005).0 5. 10000 9000 Modulus of Elasticity (ksi) 8000 7000 6000 5000 4000 3000 2000 1000 0 0.0 15. which is not a good fit.0 Kollmorgen Data Graybeal Data Air Data 25. this does not address the problem of determining the modulus of elasticity for UHPC air-cured under ambient conditions.

The first-crack flexural stress is used as an indicator of the maximum tensile stress for UHPC. prisms that underwent three curing regimes.The mean Air-cured results for this research program fall in between the average results presented by Kollmorgen (2004) and Graybeal (2005).3 Flexural Strength Testing for First Cracking Flexural testing was conducted at 28 days on 2 x 2 x 11. thermal treatment. independent of age at thermal treatment application. One specimen. A total of 36 prisms were cast with fibers and 9 were cast without fibers (Table 4. ASTM C 78 Standard Test Method for Flexural Strength of Concrete (Using Simple 84 . ASTM C 1018 Standard Test Method for Flexural Toughness and First-Crack Strength of FiberReinforced Concrete (Using a Beam with Third Point Loading) was used to determine the firstcrack strength and flexural toughness of specimens. Air-cured. three specimens had breaks within 5% of the constant moment region and their values were reduced as outlined in ASTM 1018.25 in. and Graybeal’s equation under predicts the results over their appropriate ranges of compressive stress. This is the same trend observed in Figure 4-6 where Kollmorgen’s equation over predicts the modulus of elasticity. was rejected because of a break outside of the acceptable region. Further testing and/or analysis should be completed to determine the relationship of compressive stress to modulus of elasticity for specimens cured under ambient conditions. Additionally. data collected from this study shows the AGFC (2002) model to most accurately predict the modulus of elasticity based on compressive strength of thermally treated specimens. Air-cured. At the time of this report Graybeal’s (2005) raw data was not available to compile all the results from the individual specimens and complete a regression model to determine an appropriate equation for Air-Cured specimens.1) to see the effect that fibers had on the first crack flexural stress. Therefore. and delayed thermal treatment. 4.

Beam with Third-Point Loading) presents the equation for modulus of rupture ( R ). The tensile strength 85 .2 was applied to the first-crack flexural stress to calculate the corrected first-crack flexural stress or cracking tensile stress.2 which correlates the values back to a 100 x 100 mm (approximately 4 x 4 in. A.7 ⎛ d ⎞ 1 + 2. or cracking stress. AFGC (2002) recommends the flexural cracking stress be corrected with the use of Equation 4. Test data recorded for each specimen is tabulated in Appendix A (Tables A. ⎛ d ⎞ 2. Graybeal (2005) points out that it has been widely observed that cracking stress is an over estimate of the actual tensile capacity of UHPC.1 Where: P = Load at first crack. TT.5.1 Results 0. b = Average specimen width.) prism.2 Table 4. d = Average specimen depth. and DTT for specimens with and without fibers. To obtain the actual tensile cracking stress ( f ct ). However.6).0 * ⎜ ⎟ ⎜d ⎟ ⎝ o⎠ Where: fct = corrected first-crack flexural stress. d o = reference depth of 100 mm. R= P*L b*d 2 (psi) Equation 4. This discrepancy has been attributed to depth and gradient effect set up by bending. Equation 4.0 * ⎜ ⎟ ⎜d ⎟ ⎝ o⎠ f ct = R * 0. for beams with third-point loading.13 lists the results of the flexural testing for the curing regimes of Air.4.7 Equation 4. A. 4. L = Span length.3.

50 1. Load 9.0 Sample Mean FirstCrack Load (kips) 1.) 0. for convenience ASTM C 1018 allows the flexural strength to be calculated using Equation 4. Table 4.82 1.00275 0.23 1.0 6. However.9 9.listed in the table for TT and DTT are slightly below the range of values for tensile strengths reported by manufacturers.1 7.0 6.7 9.00205 0.NF 28 28 .18 1.0430 - Sample COV (%) Max.83 1.06 1.21 Sample COV (%) Corrected First-Crack 10.03 2.36 1. As Graybeal (2005) points out this result does not have any physical 86 .8 27.1 by replacing the cracking load with the ultimate load. load Deflection 14.NF Sample Size 12 3 12 3 12 3 Sample Mean Max.1 5.00307 Sample COV (%)FirstCrack Deflection 8.8 14.8 6.0452 0.5 8. Load Deflection (in.91 2.3 - Sample Mean Max.7 9.93 1.34 1.6 - The flexural strength for fiber reinforced concretes is often difficult to calculate because of the extensively cracked section and load carrying fibers.70 1.00240 0.18 Sample COV FirstCrack 10.2 7.1 6.83 4.4 Sample Mean Corrected First-Crack 0.2 7.9 4.5 8.0 6.1 6. Deflection and Maximum Load Results Specimen Age and fibers (days) 28 28 –NF 28 28 – NF 28 28 – NF Curing Regime Air TT DTT Sample Size 12 3 12 3 12 3 Sample Mean First-Crack Flexural Stress 1.12 2.4 6.74 0.00273 0.4 7.9 8.) 0.4 Curing Regime Air TT DTT Specimen Age and fibers (days) 28 28 -NF 28 28 .0498 0.9 9.3 5.21 4.98 Sample COV (%) First-Crack Load 8.13: Flexural Stress.55 - Sample COV (%) Max.13 1.5 7. Load (kips) 4.00179 0.4 Curing Regime Air TT DTT Specimen Age and fibers (days) 28 28 –NF 28 28 – NF 28 28 – NF Sample Size 12 3 12 3 12 3 Sample Mean First-Crack Deflection (in.

0 0. The results for flexural strength.2 0. but is used for comparison purposes. The results of the flexural testing show the largest COV for the mechanical properties tested for this research.58 ksi. are at the lower end of their respective range of values as reported by the supplier. the COV presented is only slightly higher than the one-sigma limit of 7 percent presented in ASTM C 1018 for first crack flexural stress. 5.4 0.meaning.44 ksi.99 ksi respectively.3. like those for tensile strength. The mean flexural strength for Air-cured.6 0.4 1. Corrected Mean First Crack Flexural Stress (ksi) 1. is that the samples without fibers seem to have a larger first-crack stress than the specimen cast with 87 .2 Statistical Analysis and Discussion One of the more interesting observations.2 Air 1.0 28 Day 28 Day No Fibers TT DTT Age of Specimens Figure 4-8: Mean First-Crack Flexural Stress for All Curing Regimes 4. Figure 4-8 shows the mean values and ± one standard deviation from the mean for the corrected first-crack flexural stress for specimens with and without fibers. TT.13 and Figure 4-8.8 0. and DTT are 4. made from Table 4. and 4. However.

and Delayed Thermal Treatment.14: Corrected First-Crack Flexural Strength Hypothesis Testing Curing Regime Pairings Air vs TT Air vs DTT TT vs DTT Reject or Failed to Reject the Null Hypothesis (t-test) Reject Ho Reject Ho Reject Ho Population Mean Not Equal Not Equal Not Equal 88 . Therefore. The fibers are geometrically the largest material in the mix and create irregularities in the mix along the interface of the fibers and the paste.14 displays the pairings and results of the hypothesis testing for first-crack flexural strength. the purpose of the fibers is to bridge the crack and provide post-crack ductility. the difference in first-cracking stress is very small and. flexural specimens were only tested at 28 days. Additionally. Table 4. Table 4. However. Recall that the corrected first crack strength is based on the SETRA procedure for adjusting data to a normalized size of 100 mm x 100 mm (nominally 4 x 4in. When looking at the small scale.fibers. fibers create a non-homogeneous mixture and cause disruptions in the matrix and allow microcracking to propagate more readily through the specimen than through the homogenous matrix without fibers. the corrected first crack strength and the maximum load could be analyzed to see the effects of the three curing regimes. Similar to previous analyses. (2003) points out. Hence the lower first-cracking stress in UHPC with fibers. and are not compared within each curing regime. Flexural testing only had the three curing regimes of Air-cured. UHPC’s first crack was not its last crack because of the fiber reinforcement. Unlike traditional unreinforced concrete. This can be explained looking at the nature of the material from a homogeneous standpoint.) (AFGC 2002). as Mindess et al. Thermal Treatment.

As previously noted Kollmorgen (2004) conducted flexural testing utilizing ASTM C 78 and did not correct the first-crack flexural stress. for this research. results were 73 percent. using the same specimen size and testing configuration data. Mean peak load values for this research program were 340 percent.3. The mean maximum load for the same specimens was 5. and 240 percent greater than the first-crack load for Air-cured.25 in. UHPC’s ability to continue to carry load after the element has cracked is one of the properties which make it desirable for structural elements.3 Flexural Toughness Post-crack ductility is measured by the toughness that a fiber reinforced concrete exhibits. However.Hypothesis testing results for flexural strength are different than the compression and modulus of elasticity testing where only the Air-cured specimens had a different population mean. and 10 percent greater than the Air-cured. No COV or standard deviation was reported for the mean values. prisms at 28 days was 3.0 percent. The mean first crack stress for thermally treated 2 x 2 x 11. flexural first cracking strength appears to be influenced by the curing regime. and 89 . TT and DTT. respectively. Large differences in testing results indicate that ASTM procedures are not necessarily applicable to UHPC for flexural strength and further study is warranted. When comparing the corrected first-crack data to results presented by Graybeal (2005).4 percent lower and with COV of 8. Average peak load values were 170 to 200 percent greater than the first crack load. Figure 4-9 is a graphical representation of a flexural elastic-plastic material. TT. 35 percent. 280 percent. respectively. 4. and DTT mean values.1 percent greater than results of this research with a COV of 17 percent. and Poisson’s ratio results where all population means were equal.

and observed ranges for fibrous concrete are presented in Table 4.0 1 to 25 Indices calculated by dividing this area by the area to the first crack OAB.1) Area BasisA OACD OAEF OAGH A † Index Designation I5 I10 I20 Deflection Criteria† 3δ 5.15.0 1 to 12 20.displays the areas under the load deflection curve used to calculate the reference toughness indices. δ is the deflection at first-crack.15: Typical Toughness Values (ASTM C 1018 Figure X1.0 1 to 6 10.5δ Plain Concrete 1.1) Table 4.0 1. elastic-plastic material. 90 .5δ 10.0 1. Figure 4-9: Load Deflection Curve for Elastic-Plastic Material (ASTM C 1018 Figure X1.0 Values of Toughness Indices Elastic-Plastic Observed Range for Material Fibrous Concrete 5. The toughness indices for plain concrete. Figure 4-10 is a typical load deflection curve for all specimens.

normalized by dividing the aforementioned area by the area under the curve up to the first-crack. For example.0 2.5δ 5. it is important to compare identical specimen sizes and loading configurations as the toughness indices are not independent of specimen dimension (Mindess et al.5δ 20. The toughness indices are a measure of how the material responds versus the standard. a linear elastic material up to first crack followed by perfectly plastic material behavior.0 Load (kip) 3.05 0. Toughness indices are calculated at set intervals of the first-crack deflection and provide a means to compare the toughness of different materials.0 0. 2003). the index designation I5 is the area under the load deflection curve.06 Figure 4-10: Typical Load Deflection Curve for Flexural Specimens Toughness is an indication of the energy absorption capabilities of a material.04 0. Additional toughness indices beyond those listed in Table 4.0 3.00 0. However.01 0. as seen in Figure 4-9.5δ 10.15 can be calculated in the same 91 .5δ 0.02 0. it is the area of the polygon OACD divided by the area of the triangle OAB. up to 3 times the deflection at first-crack.0 1st Crack (δ) 4.0 5.0 1.6. More specifically.03 Deflection (in) 0.5δ 15.

30 R30.0 85.13 1.4 1.10 R10. I10. respectively.0 I10 17.8 1.74 0.40 217 0 193 0 182 0 271 0 234 0 212 0 302 0 256 0 224 0 320 0 273 0 241 0 The toughness indices I5 and I10 for Air-cured. Also. R5.83 1. TT.0 I20 44.20 R20.5 1.5δ and 20. A second relative parameter evaluated from ASTM C 1018 is the residual strength factor.6 1.0 65. (1995) the toughness indices I5.5 1. Table 4.7 1. Toughness Indices I30 and I 40 were calculated at a deflection of 15.18 1. 92 . The residual strength factors represent the average level of strength retained after first crack as a percentage of the first-crack strength.30 is 10(I30 – I20).0 6.0 36.21 Toughness Indices I5 6. and to a lesser extent I20 were not particularly sensitive to fiber addition rate or fiber type.06 1.10 is calculated as 20(I10 .21 R5.0 58.0 I40 107.8 1.16: Experimental Toughness Indices and Residual Strength Factors Corrected Mean First-Crack Flexural Stress (ksi) 0. The experimental results for flexural toughness indices and residual strength factors for specimens with and without fibers are shown in Table 4.3 1. Plain concrete has a residual strength factor of zero.1 1.I5) and R20.13 1.15 for fiber reinforced concrete.0 Curing Regime Air TT DTT Test Age 28 28 –NF 28 28 – NF 28 28 – NF Sample Size 12 3 12 3 12 3 Residual Strength Factors Curing Regime Air TT DTT Test Age 28 28 – NF 28 28 – NF 28 28 – NF Sample Size 12 3 12 3 12 3 Corrected Mean First-Crack Flexural Stress (ksi) 0.0 1.0 16. which is has a standard value of 100 over two consecutive toughness indices values. and DTT show results at the upper end or slightly above the observed values listed in Table 4.5δ.06 1.83 1.0 1. in research conducted by Chen et al.18 1.manner. for an elastic-plastic material.0 6. For example.9 1.74 0.7 1. for this research to compare to data published by Graybeal (2005).0 I30 74.0 92.1 1.16.6 1.0 15.0 39.

and are not directly comparable to results from this study. The toughness indices presented by Graybeal (2005) are approximately 12 to 40 percent lower than those presented in Table 4. Because the area under the load deflection curve is triangular to firstcrack.0 x 2. the remaining toughness indices. based on a 2. Kollmorgen (2004) calculated toughness using a procedure outlined by AFGC (2002) which calculates toughness at the first crack. with a 2 x 2 in.16. I30 and I40 show substantially higher values than those presented by ASTM C 1018.0 x 11. The residual strength factors show average levels of strength retained from 240 to 320 percent of the first crack strength. I20. for the first-crack flexural stress were 10 to 73 percent higher than the observed values presented in this research. Kollmorgen (2004) suggests that the values 93 . Recall that the toughness indices are normalized by dividing the area under the load deflection curve at a given deflection by the area under the curve up to the first crack. They continue to increase to R30.5 times the cracking deflection. a desirable property in almost all concrete applications. and a deflection limit of 0.40 which captures the load deflection curve behavior over the deflection range of 15. The toughness values have units of in-kips. cross section with a 9 inch span.However. specimen.25 in. ultimate load. The differences in the results may be partially explained by the following. This is consistent with the specimen’s load carrying capabilities increasing from first crack to the point of maximum load.5 to 20.030 in. Results provided by Graybeal (2005). a greater flexural stress will result in a larger area and ultimately smaller toughness indices. The high toughness indices and residual strength factors indicate that this UHPC is highly ductile. Likewise Graybeal’s (2005) residual strength factors are 16 to 46 percent lower than the observed results.

to be normalized for comparison between different fiber addition rates and fiber types. Testing specimens without fibers would allow toughness values. one mix design. which were the summation of the area under the load deflection curve. and 28-day thermally treated. and interactions with different mix designs could be examined. Because only one fiber addition rate. These curing regimes were applied as described in Section 3. By removing the toughness of the unreinforced matrix the effects of fiber addition rates. and one specimen geometry were utilized in the presented research. Specimens cast without fibers were included in this study to provide guidance as to the impact the fibers had on the first-crack stress. one fiber type. 28-day air-treated. 4. fiber types. and as previously noted the fibers have a tendency to reduce the first-crack stress. 94 .be normalized by dividing by the toughness at first-crack or the toughness of specimens without fibers. analyzing the effect of the different fiber addition rates or fiber types was not possible. the chloride ion penetrability was measured by the total charged passed in coulombs over a 6-hour period. Jamet et al.4 Rapid Chloride Penetration Test UHPC specimens of three unique age/curing regime applications were tested for chloride penetration according to ASTM C 1202: 7-day thermally treated.4 and the specimen age refers to the age of the specimen when tested. Following curing and specimen preparation. (1995) calculates the effective toughness of fibers at different addition rates by subtracting the toughness of the unreinforced high strength concrete from the toughness of the fiber reinforced high strength concrete.

4.5 COV (%) 20 15 24 Chloride Ion Penetrability Negligible Negligible Negligible *one specimen tested at 8 days 4. a t-test statistical analysis demonstrated that the amount of charge passed for the thermally treated specimens was statistically lower than the air cured specimen results (Misson 2008). Test data recorded for each specimen is tabulated in Appendix A (Tables A. this data was congruent with other research data (Bonneau et al.17.5 3. three specimens for each age/curing regime were tested.7.2 Discussion All of the UHPC specimens tested exhibited chloride ion penetration values in the negligible range (< 100 coulombs passed). it was also included in the analysis. A. Table 4.8).4. the values for the TT-cured specimens were lower than the Air-cured specimens. 1997) that reported very high resistance 95 . the ASTM C 1202 standard specifies a 42% COV value for a single operator on concrete samples from one batch. one of the 7-day TT specimens was tested on day 8 and yielded a total charge passing of 10 coulombs. While the standard deviation and COV values in Table 4. Results from these tests are summarized in Table 4.17 seemed high upon initial observation. Correspondingly. additional specimens were available for 28-day Air and 28-day TT cured specimens and were included in the analysis.17: Michigan Tech Rapid Chloride Penetration Summary Data Age at testing (days) 28 7 28 Charge Passed (coulombs) Curing regime Air TT TT No.1 Results Initially. Additionally. In addition.4. As this was within the standard deviation for 7-day TT specimens. Nonetheless. Specimens 4 3* 4 Average 75 10 15 Standard Deviation 15 1. However.

18: Graybeal (2006a) Rapid Chloride Penetration Summary Data Age at testing (days) 28 28 56 28 56 28 No. Further testing may be needed to 96 . Comparatively. the 28-day TTcured UHPC specimens tested herein had an average total charge passing equal to Graybeal’s (2006a) 28-day steam treated specimens with a 95% confidence interval (Table 4. Specimens Tested 3 2 3 3 3 3 Charge Passed (coulombs) Curing regime Steam Untreated Untreated Tempered Steam Tempered Steam Delayed Steam Average 18 360 76 39 26 18 St. Nevertheless.18). the charge passed by the 28-day Air-cured specimens were lower than the results reported by Graybeal (2006a). Graybeal also used an accelerator in mixing. water cooled saw) where the kerosene may inhibit ion migration. a statistical comparison confirmed that Graybeal’s (2006a) 56-day untreated specimens and the 28-day Air (untreated) specimens tested herein were equivalent. Dev. 1 2 18 1 4 5 COV (%) 6 1 24 3 15 28 Chloride Ion Penetrability Negligible Very Low Negligible Negligible Negligible Negligible However. respectively. One possible reason for the difference may be the different preparation methods used (kerosene saw vs. who reported an average of 360 coulombs passing for 28day untreated UHPC specimens. Another statistical analysis on thermally treated specimens tested herein revealed that ionic movement in thermally treated UHPC was independent of whether the specimen is 7-day or 28-day within a 95% confidence interval.(negligible penetration) to ionic transport in steam treated UHPC specimens. having an average of 76 coulombs passing and a standard deviation of 18 compared to 75 coulombs passing and a standard deviation of 15 for the 28-day old specimens from this study. and somewhat higher post-thermal treatment resistances (Graybeal 2006a). Table 4.

evaluate the differences and similarities between the Air-cured UHPC specimens at different ages. but again measures the ionic movement through the specimen. Figure 4-11: Surface Staining of UHPC Specimen after ASTM C 1202 Test It should be noted that the ASTM C 1202 test does not specifically measure any one type of ion movement and instead measures the bulk flow of ions through the specimen (Stanish et al. 2003). the method does not measure permeability as is sometimes understood. Visual observation of the UHPC specimens after testing revealed that the specimens experienced some corrosion of the steel fibers (Figure 4-11). the test method tends to indicate a lower chloride movement rate than would normally be expected (Perenchio 1994. but no 97 . All of the specimens tested exhibited similar staining patterns. 2000). Finally. These stains appeared to be limited to the surface of the UHPC directly in contact with the sodium chloride solution and no other distress was visible. Mindess et al. Also. This observation was similar to that observed in previous research (Graybeal 2006a). in materials containing high amounts of silica fume (like UHPC).

due to equipment malfunction. and mass change (Table 4. these parameters were recorded at 98 . and mass were observed. Occasionally. length change. while the length and mass change was recorded for six of the specimens. Only one minor variation was adapted for the testing of UHPC specimens and as outlined in Section 3. This test procedure involves the rapid freezing and thawing of concrete by means of freezing in air. Deterioration due to freeze-thaw (cracking. length change. 4. and thawing in water.correlation between movement rate and silica fume quantity has been investigated to date for applicability to UHPC.5. thawing in water) was performed on four 28-day Air cured and four 28-day TT-cured UHPC specimens to monitor UHPC’s resistance to freeze-thaw damage. Eight cycles were completed per day. 4. whichever came first. These three tests were performed every 32 freeze-thaw cycles (96 hours) and the RDM test was performed on all eight specimens.5. and approximately every 32 cycles the fundamental transverse frequency. Decreases in a specimen’s RDM indicated disruptions to the transfer of vibrations through the material due to microcrack formation. while increases in length was a sign of cracks and microcracks creating void space in the specimen.5. spalling.1 Results Freeze-thaw testing following ASTM C 666 Procedure B (freezing in air. and decreases in mass signified spalling or disintegration of material.19).5 Freeze-Thaw Cyclic Testing Freezing and thawing testing was performed in accordance with ASTM C 666 – Resistance of Concrete to Rapid Freezing and Thawing standards. or disintegration) was observed mechanically through the monitoring of a specimen’s relative dynamic modulus (RDM). Specimens were tested until failure or 300 freeze-thaw cycles.

Lastly.A. N. N.66 0.A.5. 4 4 3 3 Average RDM at end of cycling (%) 101. Mass Change (%) 0.00014* N. the increases for the Air cured F-T specimens (RDM . Avg.27 101.10 COV (%) 0.10% expansion in length of the specimen is attained. SS for side-study wet-dry cycles 4.08* 0. COV (%) 20.50 Avg.2 Discussion Failure of a specimen undergoing freeze-thaw cycles as stated by ASTM C 666 has been reached when the specimen’s relative dynamic modulus of elasticity reaches 60% of its initial modulus. Figures A. Spec.4 14.1.57 99 .12 0. on some occasions the specimens were stored in the frozen condition until the equipment was functional again.4 N. or if a 0.9. Length Change (%) 0.19: Effects of Freeze-Thaw Cycles on UHPC Curing and (Testing) Regime+ Air (F-T) TT (F-T) Air (SS) TT (SS) + No.06 COV (%) 5. A. The side-study specimens (which underwent wet-dry cycles without temperature changes) demonstrated similar increases in RDM and mass. The effects of water absorption on the RDM and mass of the six additional side-study UHPC specimens (three air cured and three thermally treated) were also analyzed (specimens denoted by “SS”).54 *Note: Only three of the F-T specimens were tested for length change and mass change F-T for freeze-thaw cycles.22 0. A.19 shows the summary data for the specimens tested (data is listed in Appendix A.01%) change in length. UHPC freeze-thaw specimens showed only a small increase (< 1%) in mass and negligible (< 0.5 44. Table 4.57 100. Length change was not documented on these side study specimens.0 28.50 0.A.54* 0. all of the UHPC freeze-thaw specimens maintained their integrity and exhibited an increase in RDM (< 2%) as testing continued.0004* 0. However.periods greater than 32 cycles but never at an interval of more than 36 cycles.10.11).91 100. Also.32 0.A. Table 4.

After 300 cycles. but rather continued to hydrate. suggesting that the specimens did not deteriorate at all.27%. the increase of the TT-cured specimens was small in comparison to the Air-cured specimens.%.0. This can be primarily attributed to the greater amounts of unhydrated cement particles in the Air-cured specimens that can become hydrated in the presence of water. Figure 4-12 displays the correlation between the increase in RDM for the freeze-thaw and side-study specimens. all eight freeze-thaw UHPC specimens had higher RDM’s than at the beginning of testing. 100 . However.54%) were significantly higher than the increases for the thermally treated F-T specimens (RDM . The Air-cured side-study specimens’ RDM increased with a similar trend as the Air-cured freeze-thaw specimens. Mass – 0. Mass – 0.08%). The similar increases also indicated that the increase in RDM from both testing regimes was due to the effects from cycling the specimens in and out of water.

and further testing by Graybeal (2006a) revealed that untreated UHPC specimens immersed in a water bath (without wet/dry cycles) also increased in compressive strength. and Graybeal also reported that the specimens also showed a mass increase.0 101.5 102. specimens undergoing freeze-thaw cycles increased in RDM.0 99.Average Relative Dynamic Modulus (%) 102.5 101. even when being exposed to a harsh freeze-thaw environment. The data in Figure 4-12 supports this as the side-study 101 . Lee demonstrated that reactive powder concrete (a precursor to UHPC) cubes increased in compressive strength after 300 cycles of freeze-thaw testing.0 100. Lee et al.5 0 50 100 150 200 250 300 350 Freeze-Thaw Cycles Air cured F-T specimens Air cured side study specimens Thermally treated F-T specimens Thermally treated side study specimens Figure 4-12: Effects of Freeze-Thaw Cycling on the Average Relative Dynamic Modulus of UHPC Samples Testing by others (Graybeal 2006a. These studies suggest that the submersing of UHPC in water can increase both compressive strength and RDM. 2005) also reported a similar phenomenon for freeze-thaw specimens.5 100. In both studies.

thawing in water) which was used in Graybeal’s study. small microcracks were observed on the surfaces of the freeze-thaw specimens upon removal from the freeze-thaw chamber. Lengthier exposure of the UHPC specimens to water during Procedure A may allow for a greater probability of water reacting with unhydrated cement particles.specimens exhibited similar changes in RDM and mass as the specimens undergoing freeze-thaw cycling. especially the air cured specimens (Figure 4-13). thawing in water) was used in this study rather than Procedure A (freezing in water. However. Procedure B (freezing in air. These cracks were only evident while the surfaces of the specimens were wet. Although little visual damage was noted. and quickly disappeared from view once the specimens dried. The overall increase in RDM (1. the RDM values for all of the specimens were greater than 100 upon the completion of testing.57 % for Air-cured specimens) in this study was different than in other research (approximately 10% for untreated specimens) (Graybeal 2006a). Despite the visual observation of cracking in the specimens. Figure 4-13: Cracks in Air Cured UHPC Specimens Following Freeze-Thaw Testing 102 .

with two ¼ in.). precise absolute length measurements were not recorded prior testing. However. Also note that the initial fundamental resonant frequency of the Air-cured specimens is lower than that of the TT-cured specimens.15%) in this testing were likely due. These resonant frequencies (2720-2770 Hz) were also very close to the frequencies observed by Graybeal (2400 – 2600 Hz). The somewhat higher frequencies (5% . in part. an observation also noted by Graybeal (2006a). to the shorter length dimension (3% shorter) of the specimens due to the length change studs (3 x 4 x 15 ½ in. 103 . Figure 4-14 shows the change in mean resonant frequencies of UHPC freeze-thaw and side study specimen as testing was performed. of exposed studs as opposed to 3 x 4 x 16 in.An analysis of the fundamental transverse frequencies recorded provided further evidence of the low variability of UHPC. The tight range of the initial resonant frequencies revealed that the air-cured and thermally treated specimens undergoing freeze-thaw cycles were within 8 Hz of their identically treated side-study specimens. so further comparison of the resonant frequencies is beyond the scope of this project.

a noticeable decrease in the resonant frequencies of the Air-cured freeze-thaw specimens. though not as substantial (-0. additional resonant frequency testing revealed that the UHPC specimens displayed frequency responses not initially observed upon the completion of the testing. However.29%. The thermally treated specimens subjected to freeze-thaw cycling also experienced a decrease in their resonant frequencies over time. a similar decrease was not observed in the side study specimens for either curing regimes (Figure 104 .20).2770 Average Resonant Frequency (Hz) 2760 2750 2740 2730 2720 2710 2700 0 50 100 150 200 250 300 350 Freeze-Thaw Cycles Air cured F-T specimens Thermally treated F-T specimens Air cured side study specimens Thermally treated side study specimens Figure 4-14: Average Resonant Frequencies of UHPC Freeze-Thaw and Side-Study Specimens The UHPC specimens were stored in ambient conditions in the lab after freeze-thaw testing was completed. First.34%) (Figure 4-15). was observed (Table 4. -1. Approximately six months after testing was completed.

34 0.51 2.34 0.33 0.23 Figure 4-15: Resonant Frequencies of UHPC Specimens after Freeze-Thaw Testing 105 .67 -0.29 -0.29 0.47 -0.22 0.29 0.29 Curing Regime Air Air Air Air TT TT TT TT Air Air Air TT TT TT Testing Regime F-T F-T F-T F-T F-T F-T F-T F-T SS SS SS SS SS SS Specimen ID M-FT-A-28 P-FT-A-28 R-FT-A-28 S-FT-A-28 M-FT-TT-28 P-FT-TT-28 R-FT-TT-28 S-FT-TT-28 P-FT-SSA-28 R-FT-SSA-28 S-FT-SSA-28 P-FT-SSTT-28 R-FT-SSTT-28 S-FT-SSTT-28 % change 0. Table 4.18 0.4-16).29 -0.20: Change in Resonant Frequency of UHPC Specimens after Testing Completed Resonant Frequency at end of testing (Hz) 2729 2756 2744 2749 2762 2766 2769 2767 2771 2742 2761 2754 2748 2772 Resonant Frequency 6 months after testing (Hz) 2715 2693 2725 2703 2754 2757 2756 2759 2781 2751 2770 2759 2756 2778 Average change (%) -1.36 0.33 0.69 -1.33 -0. The decrease in resonant frequency suggests that damage to the UHPC specimens from freeze-thaw cycling may be greater than that observed immediately after testing was completed.

the curves for the Air-cured freeze-thaw specimens exhibited the more skewed shape depicted in Figure 4-18. During freeze-thaw testing and immediately after testing. 106 . This change in shape indicates a change in the distribution of frequencies acquired from the impact resonance test and may indicate a change in material behavior. the curves used to determine the transverse resonant frequency of the beams were bell shaped for all specimens (freeze-thaw and side-study) as shown by the curve in Figure 4-17 of a side-study specimen.Figure 4-16: Resonant Frequencies of UHPC Specimens after Side-Study Testing Further evidence that frequency responses were different than those initially observed after testing was the change in the shape of the frequency distribution curves for the air cured freeze-thaw specimens. However. after testing had been completed and the specimens were stored for several months at ambient lab conditions.

Additionally.Figure 4-17: Typical Bell Shaped Resonant Frequency Output of Air-cured UHPC Specimen (Frequency in Hz) Figure 4-18: Typical Skewed Resonant Frequency Output of an Air-cured UHPC Specimen Six Months after Freeze-Thaw Testing (Frequency in Hz) Overall. Increases in RDM and mass during freeze-thaw cycling were both indicators that additional hydration may be taking place within the specimens. these test results indicate that the UHPC specimens underwent some form of autogenous healing similar to what Jacobsen and Sellevold (1996) observed in HPC freeze-thaw specimens. UHPC specimens outside of freeze-thaw cycling had improved RDM values similar to those in the freeze-thaw 107 .

4. additional specimens were cast for each of the missed testing times such that sample size of at least three test specimens was available for each curing regimes. but is 108 . These curing regimes were applied upon specimen demolding.6. and then subjected to heating and cooling cycles until an accurate CTE measurement is obtained. three of the specimens were not tested on their appropriate test day. Modifications to both specimen preparation and testing procedures were made due to equipment specifications and in the interest of maintaining the integrity of UHPC curing regimes. Typically. placed into a test frame and submerged in a thermally regulated water bath. CTE tests began on the day of the stated specimen age (e.g – a specimen tested for 7-day CTE values began testing on day 7) and lasted 24-36 hours. this healing appears to have limited long term impact as RDM values decreased just months after testing was completed. A total of 22 specimens were tested. due to one operable test frame and tests that ran longer than 1-day. and the specimen age refers to the age of the specimen when tested (Chapter 3).6 Coefficient of Thermal Expansion The coefficient of thermal expansion for concrete was determined following a modified version of AASHTO TP-60-00 – Coefficient of Thermal Expansion of Hydraulic Cement Concrete.1 Results Coefficient of thermal expansion testing was performed on Air-cured UHPC specimens ranging from 3-days in age to 28-days and on 7-day and 28-day TT-cured specimens. 4. However. The data for the three specimens not tested on their appropriate test day is not included in this section.chamber. However. The standard process according to AASHTO TP-60-00 requires that specimens be saturated with water prior to testing. Therefore.

For unsaturated 28-day TT-cured specimens. Further motivating this decision was UHPC’s low permeability that would create an increased likelihood of unequal degrees of saturation in each specimen. and therefore this data does not represent the CTE values of saturated or partially saturated UHPC specimens.62 x 10-6/°F).21.74 x 106 /°F. This reasoning follows previous CTE research conducted on UHPC specimens (Graybeal 2006a). Data for individual CTE tests are listed in Appendix A (Tables A. 2003). A. However.12. A summary of the CTE values for unsaturated UHPC specimens is presented in Table 4. and the CTE value for unsaturated 28-day Air-cured specimens had an average CTE value of 7. 109 . Also. All of the UHPC specimens were kept unsaturated during testing through the use of an epoxy coating to avoid the potential hydration effects of water on unhydrated cement particles in the cement matrix. a statistical t-test confirmed that regardless of specimen age.16 x 10-6/°F. moisture content does affect the CTE values of normal strength concrete specimens (Mindess et al. Values for the thermally-treated 7-day specimens (8. the average CTE was 8.included in Appendix A for reference.20 x 10-6/°F) were similar to the TT- cured 28-day specimens and comparatively higher than the Air-cured 7-day specimens (7. TT-cured UHPC specimens had a statistically higher CTE value than the Air-cured specimens.13).

7 2. Figure 4-19 shows the CTE values for the Air-cured UHPC specimens and their respective standard deviations.9 0.11E-06 0.74E-06 8.21: Coefficient of Thermal Expansion (CTE) Test Summary Specimen age at testing (days) 3 7 14 28 7 28 No. the 3-day air specimen exhibited a statistically lower CTE (88% confidence) than the 7-day Air-cured specimen. Dev (/°F) 0.53E-06 7. The averages of the data in Table 4.21 indicate an increasing CTE value as the Air-cured specimens age (see Figure 4-19).2 Discussion The results demonstrate that the age of a specimen at testing plays a more significant role in Air-cured UHPC specimens than in TT-cured specimens.1 4.08E-06 0. the 14-day Air-cured specimen’s CTE value was not statistically lower than the CTE value for the 28-day Air-cured specimen.06E-06 0.69E-06 7.0 1. 110 .18E-06 Curing regime Air Air Air Air TT TT St. However. Yet after performing a statistical t-test on the Air-cured data.15E-06 0.1 1.15E-06 COV (%) 1. respectively) than the next testing age (7-day and 14-day.20E-06 8. respectively).6.Table 4. samples tested 3 4 3 3 3 3 Average CTE value (/°F) 7. That is. only the 3-day and the 7-day specimens were statistically smaller (88 percent and 78 percent confidence.8 1.62E-06 7.13E-06 0. and the 7-day Air-cured specimen exhibited a statistically lower CTE (78% confidence) than the 14-day Air-cured specimen.

In fact. the CTE value for UHPC should fall between the range of a 1:6 cement/natural silica sand mortar (6. Again. The UHPC CTE values are slightly higher than the values typically reported for normal and high strength concretes which tend to have CTE values of approximately 4. Instead. UHPC consists mostly of fine sand (41%) and portland cement (29%).90E-06 7.80E-06 CTE (in/in/°F) 7. However. a two-sample t-test (95% confidence interval) determined that TT-cured UHPC specimens maintained the same CTE value whether tested at 7-days or at 28days.70E-06 7. a closer look at the factors that influence the CTE of a concrete reveals that CTE values obtained for UHPC are reasonable.3 x 10-6/°F (FHWA 2006). By and large the CTE of a concrete is most greatly influenced by the CTE of its coarse and fine aggregates.50E-06 7. this supports the assertion that thermally steam treating UHPC “locks” in properties so that specimen properties change little post-treatment.00E-06 7.40E-06 0 5 10 15 20 25 30 Specimen Age at Testing (days) Figure 4-19: Average CTE Values for Air-cured UHPC Specimens Conversely.8.60E-06 7.21). Therefore.7 x 10-6/°F ) (Mehta and Monteiro 2006) and saturated portland cement pastes (10 111 . the 28-day TT-cured specimens showed little change from the 7-day specimens (Table 4. but UHPC has no coarse aggregate.1-7.

112 . The Japanese recommendations (JSCE 2006) suggest a CTE value of 7.5 x 10-6/°F for steam treated UHPC samples while Graybeal’s thermally treated specimens (Graybeal 2006a) had a CTE value of 8. the influence of admixtures used by others. The CTE values measured in this testing vary from published data (Table 4. the COV values for Graybeal’s data and the data presented in this report are similar.to 11 x 10-6/°F). the CTE value of UHPC is not unexpectedly high. However. Equation 2. Some explanations for the variations in data include slightly different batching and curing procedures.22). Therefore.5 x 10-6/°F. Measured CTE values (8. and the different ages of the specimens at the time of testing.1 reveals an expected CTE value for UHPC of approximately 8.16 x 10-6/°F) appear larger than Japan’s data and smaller than Graybeal’s data. No statistical tests were performed to compare the data because of differing specimen ages at the time of testing. showing that both test sets have tight fitting data and little variation of CTE values between specimens.7 x 10-6/°F.

However. Dev (in/in/°F) COV (%) Michigan Tech CTE Data Air 28 7.02% in mass during a typical CTE test (Table A.2E-07 1. The objective of the study was to determine whether concrete driveway sealant would provide a simpler and more effective method of preparing UHPC specimens prior to CTE testing.4% throughout the shakedown. the UHPC specimens coated in driveway concrete sealant exhibited a mass increase of approximately 0.3 Study of water absorption To provide alternative methods for sealing UHPC specimens prior to CTE testing.50E-06 1. Application of concrete driveway 113 . *TT – Definition of exact procedure unknown at the time of publishing 4. The shakedown involved testing the repeatability of the equipment following the same testing procedures previously outlined.67E-06 Japan CTE Data (JSCE 2006) TT* N.5E-07 1.A.1 2.7E-07 2. For that reason. Therefore. epoxy sealant was employed during testing and a side study to examine the effect of water absorption on the weight and length of specimens coated in epoxy was developed.Table 4.17E-06 TT > 60 days 8. epoxy was determined to be an adequate sealant. N.9 N.22: Comparison of Some Published UHPC CTE Data Curing regime Specimen Age (days) Average CTE Value (in/in/°F) St. a driveway concrete sealant was used to coat the UHPC specimens that were used to shakedown the CTE equipment.7 1.9 2. During testing it was found that specimens coated in epoxy only increased 0.7E-07 1.74E-06 TT 28 8.A.13 – Appendix A) or approximately a gram of water absorbed by each specimen. For that reason.A. 7.16E-06 Graybeal CTE Data (Graybeal 2006a) Air > 60 days 8. a study was performed to determine the sealing properties of two types of sealant – concrete driveway sealant and epoxy sealant.6.

A more detailed study to account for minute length changes due to water absorption in UHPC was beyond the scope of this project. but it provided less adequate protection against water absorption.sealer was a simpler process than epoxy application. 114 .

and Double Delayed Thermal Treatment (Air. Four different curing regimes were used to measure the impact on the mechanical properties. UHPC test results are repeatable between different laboratories when comparing to Graybeal’s work at the FHWA Turner-Fairbank Laboratory (2005). In general. UHPC durability properties researched herein exceed those of normal strength concretes and high performance concretes. 2. and flexural strength and toughness of an ultra-high performance concrete. modulus of elasticity. 1. Air-cured. Mixing time increases as the age of the premix of UHPC increases. and DDTT. 3. A summary of the tests conducted was listed previously in Table 4. Poisson’s Ratio. DTT. Thermal Treatment.0 Conclusions of the Experimental Studies The purpose of this research was to determine the impact that age of thermal treatment had on the mechanical properties of compressive strength. Delayed Thermal Treatment. Additionally. Results and discussion were also provided in Chapter 4. 115 .1. Specific conclusions for each test type are listed below in separate sections. although the research reported herein is much expanded over the preliminary studies conducted at FHWA. respectively). a UHPC using two curing regimes was compared for resistance to rapid chloride penetration and freeze-thaw.5. however material properties do not show significant changes. the following conclusions can be made. This chapter summarizes the major conclusions of the experimental studies. and to determine the coefficient of thermal expansion. TT.

had a mean modulus of elasticity of 7850 ksi. 2. 5. and DDTT had the same population mean and a combined modulus of elasticity of 8150 ksi.1 Compression Strength 1. DTT.000 * 116 ( 3 f `ATT (psi) ) Equation 2. which were shown to have the same population mean.1 ksi.5. Air-cured specimens showed an increase of strength with age and at 28 days had a compressive stress of 23. 2.3 . The compressive stress was independent of age at which thermal treatment was applied as well as the age at which the specimen was tested following thermal treatment. The mean compressive stress for all TT. the three curing regimes of TT. Modulus of elasticity was scarcely impacted by the four curing regimes. by allowing a precaster to cast several elements over a period of time and then thermally treat them simultaneously.8 percent where the compressive stress increased by 25. By conducting thermal curing. The Air-cured specimens appear to be asymptotically approaching a maximum compressive stress of 25 to 27 ksi.2 Modulus of Elasticity and Poisson’s Ratio 1. Modulus of elasticity data for this research is best predicted by the model proposed by AFGC (2002) which is: Ec = 262. DTT. allowing more flexibility in the casting and curing sequence. The Air-cured specimens at 14 and 28 days. and DDTT cylinders was 30. Compressive strength testing showed that there was no difference in the compressive stress after thermal treatment was applied. This could have a large impact on how UHPC is used in industry. the modulus value was only increased by 3.9 ksi.9 percent over air curing based on 28 day information. Like the compressive stress samples.

Again. One age (28 days) and three curing regimes were considered for flexural testing. Fibers provide post-crack ductility. Unlike compression and modulus. TT 1.3 Flexural Strength Testing for First Cracking and Toughness 1. Like the compressive stress. independent of age or curing regime. 2. The difference between TT and DTT does not coincide with the compression and modulus trends were specimens which received thermal curing had the same population means. the commonly accepted value for normal strength concrete. The mean value for all samples was 0.18 ksi. and DTT. Data was corrected for specimen size as discussed in Section 4.3. had the same population mean. The four curing regimes had no impact on Poisson’s ratio as all specimens. 3.06 ksi.21 which is slightly greater than 0. having the same modulus value independent of when the thermal cure was applied or when the specimen was tested would be a measurable benefit not only to the precast industry but the design industry as well because the casting and curing of the material is flexible.76 ksi. and DTT 1. however this could be easily overcome by using the conservative value for calculations.20. 117 . all curing regimes had different population means for corrected first crack stress. 5. TT. Air. The sample means for corrected first crack stress are as follows: Air 0. by having Poisson’s ratio independent of when thermal curing is applied makes the manufacturing process much more flexible. which would slightly complicate the production process.Where: f`ATT = Compressive stress of UHPC after thermal treatment. Flexural toughness and residual strength factors are significantly enhanced by the use of fibers.

but were still less than 2. All UHPC specimens. Air-cured and TT-cured UHPC specimens increased in both relative dynamic modulus and mass during freeze-thaw testing at rates similar to companion UHPC specimens undergoing wet-dry cycles. UHPC demonstrated a high resistance to freeze-thaw damage (100+ durability factor and less than 0. 5. 118 .4 Rapid Chloride Penetration Test 1. 2. UHPC exhibits signs of autogenous healing that leads to increased RDM and mass gain when submerged in water. Rapid chloride penetration resistance of UHPC is superior to NSC and HSC regardless of curing regime. 2. indicating that specimen age does not play a major role in thermally treated UHPC chloride ion resistance. The greatest increases documented were in the Air-cured specimens. 7-day and 28-day thermally treated specimens exhibited statistically similar total charge passing results. Thermally treating UHPC specimens enhances its chloride penetration resistance by limiting ionic movement to even lower levels than that of Air-cured specimens. However. whether Air-cured or TT-cured had negligible chloride ion penetrability. the long term impact of healing decreased just months after testing was completed as noted by the resonant frequency of UHPC specimens declining. 3.5 Freeze-Thaw Cyclic Testing 1.0%. 3. even while undergoing freeze-thaw testing.01% length change) with no large cracking or spalling of the material regardless of curing regime.5.

Epoxy coating performed better than concrete driveway sealer when sealing test cylinders for maintaining water saturation levels in UHPC specimens. 3. A coefficient of thermal expansion of 8. 28-day TT-cured UHPC specimens had a statistically higher CTE value than 28-day Aircured specimens (8.5. However. TT-cured specimens maintained CTE values regardless of age. this value can be estimated based on the volumetric proportions and CTE values of UHPC constituent materials. 4. 119 .74 x 10-6/°F. 5.2x10-6/°F is recommended for UHPC once it has been thermally treated regardless of age.6 Coefficient of Thermal Expansion 1.18 x 10-6/°F and 7. although the only statistically significant changes occurred before specimens aged 14-days. respectively). UHPC has a coefficient of thermal expansion value slightly higher than NSC. 2. Coefficient of thermal expansion values were tested on unsaturated UHPC specimens and increased with age in Air-cured UHPC specimens.

(This page intentionally left blank) 120 .

1 Bridge Components The control and similar UHPC bridges were divided into three components: superstructure.S.5 ft. a bridge using a NSC superstructure with correspondingly larger beams for the constant bridge span and width. typical of current MDOT construction practices. were assumed for the control bridge. high and have a width of 3. 6.5 ft. deck. The Mars Hill Bridge in Wapello County.0 ft. Precast. single-span with a width of 24.0 Preliminary Life Cycle Costs of a UHPC Superstructure A preliminary life cycle cost analysis comparing the initial and long term costs of UHPC and normal strength concrete (NSC) bridges has been performed. scenario one involving a UHPC bridge of only UHPC girders and an NSC cast-in-place deck and scenario two involving a UHPC bridge of UHPC girders and deck panels. The superstructure consisted of the three I-beams and the sub-structure included the abutments and footings. prestressed concrete I-beams and a NSC deck. and sub-structure. Iowa served as model for these scenarios and is composed of UHPC girders and NSC deck and sub-structure. However. while footings were assumed to be 3. It should be noted that the Mars Hill Bridge did not use UHPC in an optimized section. which could lead to excessively higher costs. Modifications were a two inch reduction in the lower flange and web width and a one inch reduction in the upper flange (Moore 2006).6. (Moore 2006) to accommodate two lanes of traffic. The NSC deck was a nine inch. The UHPC model bridges used a modified 45-inch Bulb Tee girders. it is currently the only bridge in the U.5 ft. Two scenarios were evaluated for the UHPC superstructure. The abutments were assumed to be equal to the bridge width. A cost estimate was also performed for the control bridge. cast-in-place (CIP) concrete slab. 12. that is built using UHPC. All three model bridges were a 110 ft. 121 .

80 $305 10% $120 Reference Number 02300-02315-462-6050 02300-02305-250-0100 02300-02305-250-2500 02300-02305-250-2000 Additional 5 mile haul distance Mob.550 $17. with a similar width and span of the control bridge. Reference numbers are provided for further information (RS Means 2005). which are described in Tables 6. While the original Wapello County Bridge used an 8 in.5 ft. Truck-mounted Crane 122 .2: Construction Activities Unit Costs Item Machine Excavation for Abutments Mobilization and Demobilization Up to 25 miles (Dozer. costs were adjusted to a 9 in.55 Reference Number (RS Means 2005) 02800-02850-205-1050 02800-02850-205-4000 02800-02850-205-1000 02800-02850-205-1620 02800-02850-205-1000 02800-02850-205-1150 02800-02850-205-2100 02800-02850-205-1230 Sidewalk Table 6.2 Construction RS Means data were used to find the unit cost of each bridge component and activity. Table 6.1: Bridge Component Unit Costs Item Abutment Approach Railing Concrete Deck Prefabricated I-beam Footing Parapet Reinforcing. For scenario two.000 $298 $585 $3. the deck was assumed to be four inch thick. topped with a water-proofing membrane and asphalt wearing surface. deck. The NSC deck of UHPC scenario one was the same as the control bridge. precast UHPC deck panels. Each Each Cost $10.( Epoxy Coated) Unit Cubic Yard Linear Foot Cubic Yard Each (100-120 ft span) Cubic Yard Cubic Yard Ton Square Foot Cost $345 $119 $298 $16. 6. backhoe) Unit Cubic Yard Each + 5 mi.2. & Demob. deck for comparison to MDOT standard practice.1 and 6.high and have a width of 8. loader.

Total costs for the approach railings. Table 6. † Includes waterproofing membrane and asphalt wearing surface.000* $447.000 $48. parapets.The total cost of each component and activity for the control bridge was then calculated using assumed quantities and the above unit costs.100 $255. 123 . deck from 8 in. was included in the cost of the UHPC structure in scenario two. UHPC deck panels) $33. An estimated cost of $432.000† Not Available Not Available Not Available Not Available Not Available $483. Table 6.3 contains the estimated construction costs for the control and UHPC bridges.000.000 Total (2007 $) Control-3.000 $33. parapet Sub-structure Superstructure Reinforcing. a database of Michigan bid items that allows project managers and engineers to make road and bridge project estimates). and sidewalks were added to the cost of the deck.000 $82. railing.000 $273. The added cost of a UHPC deck.000 $107.000 $497.000 $23.4% inflation average 2005-2007 UHPC-3% inflation 2006-2007 *Reported cost adjusted to 9 in. Scenario 1 & 2 in 2006 $) Control Bridge (NSC Girders and Deck) Scenario 1 (UHPC girders. Epoxy Coated Activities Design (10% of costs) Total (Control in 2005 $.0% inflation rate per year. as opposed to a NSC deck.000 $4. Given that Scenario 1 represented the Mars Hill Bridge. taken from the Mars Hill Bridge. all other costs estimated.0% inflation rate per year was the only modification to the cost of the UHPC structure so as to compare in 2007 dollars. NSC Deck) Scenario 2 (UHPC Girders. the application of a 3. The information obtained from RS Means was cross checked against information from recent Michigan bids available through MERL (Michigan Engineers’ Resource Library. deck. was used as a guide for the UHPC structures (Endicott 2006).3: Estimated Construction Costs Bridge Component Deck.000 Not Available Not Available Not Available Not Available Not Available $434.000 $40. along with the 3.

as this information was readily available (Ohio LTAP 2007). approach pavement relief joints. and contractors pertaining to procedures and costs of concrete bridge decks (Lopez- 124 . Unit costs were obtained from the Ohio Department of Transportation’s Preventative Maintenance page. as defined in the CSM Manual. concrete coating/sealing. and sealing. Maintenance unit costs can be seen in Table 6. The annual deck maintenance cost includes washing.00 $15.3 Maintenance Bridge maintenance was assumed according to the MDOT Bridge Preservation Timeline for a concrete deck with epoxy coated rebar and prestressed concrete beams (MDOT 2007b).00 The MDOT Bridge Preservation Timeline stipulates regular Capital Scheduled Maintenance (CSM) throughout the life of the bridge to prevent deterioration of the structure.4: Unit Costs of Maintenance Activities Activity Beam End Rehab BIT Overlay Deck Patch Joint Replace Overlay (Deep and Shallow) Sub Repair: Diagonal Cracking Sub Repair: Deteriorated Concrete Unit CF SF SF LF SF LF SF Cost $50. concrete crack sealing.00 $40. consultants. minor concrete patching and repair.4.00 $20. spot painting. an annual deck maintenance cost was utilized in place of the CSM. vegetation control. An annual cost of $0. patching.20 per square foot for Michigan bridges was obtained from a questionnaire sent to several DOT’s. Capital Scheduled Maintenance.00 $100.00 $45.6. joint repair/replacement.00 $35. Given that it is difficult to assume an average value for these tasks. includes superstructure washing. Table 6. flushing. and slope paving repair (MDOT 2007a).

A design life of 180 years was chosen for the UHPC structures because UHPC is expected to outperform traditional structural concrete by at least twice as much.5: Bridge Girder Maintenance Year 40 90 130 Control Maintenance Beam end rehab Beam replace Beam end rehab UHPC Year 90 Maintenance Beam end rehab 125 . Table 6. along with additional rehabilitation needs (MDOT 2007b). The Bridge Preservation Timeline was followed for all NSC components of the UHPC structures and a modified preservation timeline was used for the UHPC components. Michigan was one of twelve DOT’s that participated in the survey. at the end of which an evaluation of the bridge is made and replacement of the superstructure or bridge is decided. Tables 6. The Bridge Preservation Timeline obtained from MDOT suggests a 90 year design life for the control bridge.5 and 6. This assumption is based on results of mechanical and durability tests of the material performed herein. Additional rehabilitation needs included sub-structure repair of diagonal cracking and deteriorated concrete.Anido 1998). It was assumed that the deck and I-beams of the control bridge would need replacement at the end of the 90 year design life.6 compare the NSC and UHPC maintenance of the bridge girders and deck. and the price stated above reflects the individual response of MDOT.

Table 6. in comparison to a NSC deck. 6. It was assumed that little maintenance would be needed for a UHPC deck. 126 .1 details the method of reduction to NPV where r is the real discount rate and n is the difference in years from the present to future date. An analysis period of 180 years was chosen because it is the first incidence point of the control and UHPC structures.6: Bridge Deck Maintenance Control Deck Year Maintenance† 12 DP & JR 25 DP & JR 40 DO & JR 52 DP & JR 65 SO & JR 80 BIT Overlay 90 Deck Replace 102 DP & JR 115 DP & JR 130 DO & JR 142 DP & JR 155 SO & JR 170 BIT Overlay UHPC Deck Year Maintenance† 45 JR 90 JR 130 155 JR JR †DP = Deck Patch. JR = Joint Replace.4 Preliminary Life Cycle Costs The cost-benefit analysis for each bridge included the construction and maintenance costs over a 180 year period. DO = Deep Overlay. Maintenance costs and repair costs over the 180 year design life were reduced to a net present value (NPV) based on a seven percent discount rate and a three percent inflation rate. Equation 6. and the maintenance schedule was reduced to regular joint replacements. or the first point at which both structures are at similar maintenance needs. and SO = Shallow Overlay The UHPC structure in scenario two was also expected to not require annual deck maintenance.

000 $4. The changing cost for scenario two was found by taking the current unit cost of UHPC and assuming quantities for the bridge deck and beams.000 $71.000 50. if this cost were to drop.000 $69. given a change in the cost per yard of UHPC.000 $501. This is easily explained by the cost of the UHPC material.1 Table 6. The cost per cubic yard of the UHPC used in the experimental study for this research project was around 20 times more expensive than the cost per cubic yard of typical concrete ($2000/yd versus $100/yd). 2007 $ Activity Construction Maintenance Total (3% inflation/year & 7% discount rate) Control Bridge (NSC Girders and Deck) Scenario 1 (UHPC girders. based on their dimensions. Based on Figure 6-1.0% $497. along with 5% for additional charge that may be unforeseen.7 summarizes the construction and maintenance costs for the control and UHPC bridges for the 180 year common time increment. Figure 6-1 depicts the change in the cost of scenario two. which includes a total deck replacement.000 $516.000 $344. UHPC deck panels) $273. However. UHPC usage may become more advantageous in life cycle cost assessments. These unforeseen charges were not assumed to include plant modifications needed for UHPC or transportation costs for the bridge components.000 45. NSC Deck) Scenario 2 (UHPC Girders. When calculating the price for the beams a 10% design cost was added.NPV = (present $) / (1+r)n Equation 6. for it to be advantageous in 127 .7: Costs of Control and UHPC Bridges.6% Incremental increase (%) It is interesting to note that the control bridge still provides the lowest cost bridge over the 180 year period. Table 6.000 (baseline) $447.

scenario two the target cost would be about $1,750 per yard, or about a 12.5% decrease over current prices.

Target Price of Ductal

$350,000.00 $300,000.00 Scenario 2, 2007 $ $250,000.00 $200,000.00 $150,000.00 $100,000.00 $50,000.00 $0.00 $0 $1,000 $2,000 $3,000 Cost per yd, 2007 $
UHPC NSC

Figure 6-1: Target Cost of UHPC

6.5

Conclusion of the Preliminary Life Cycle Cost Analysis The preliminary life cycle cost analysis shows that the control bridge will cost $172,000

less than a UHPC bridge constructed of UHPC girders and a NSC deck, and $157,000 less than a UHPC bridge constructed of UHPC girders and UHPC deck panels, over a 180 year design life. However, the analysis performed only accounted for the construction and maintenance costs and does not incorporate additional user costs or benefits of either bridge. Also, the model bridge chosen incurred increased costs due to transportation of the bridge components. Furthermore, the UHPC bridge design and actual costs considered did not take full advantage of UHPC by using an optimized cross-section. After finding the target cost of UHPC it was apparent that if the unit cost of the UHPC material were to drop about 12.5%, life cycle costs may be less over the 180 year design life. This would entail that the transportation cost, cost of modifying the 128

current plant, and any additional charges would not raise the cost of UHPC above the stated target cost. It is acknowledged that the initial costs of a current NSC control bridge is much lower, but the reduced maintenance of the UHPC bridges, in particular for the bridge model using the UHPC deck panels, is significantly less over time. Additional costs and benefits to note relating to the user are the delays due to construction or maintenance, bridge safety, and the effects of deterioration on the performance of the bridge. The assumed maintenance of the UHPC bridges in this report can only be

considered a rough estimate; further observation of UHPC structures will be needed for a better assumption. Also, UHPC is a new material to the construction market, and little is known about the potential optimization of UHPC components and the future cost of the material after product familiarization has occurred. Component optimizations in this report are conservative estimates and may not truly reflect the capabilities of UHPC. As more UHPC structures are implemented, the limits of the material will be better understood and design of the components will improve. The future cost of the product is also hard to forecast, however, current prices may be higher due to research costs, low material production, and the cost of precast plants to modify current technologies to fit the needs of UHPC. 6.6 Future Work A thorough cost-benefit analysis is needed to gain more insight on the feasibility of UHPC bridges in Michigan. This analysis would include the user costs and benefits listed above, as well as other costs or benefits that may be pertinent, and a risk analysis to assess the sensitivity of the outcome to input variation. The risk analysis could be as simple as fluctuating the discount or inflation rates or as complicated as applying computer simulation of weighted 129

input variables. Monte Carlo Simulation is one method of computer simulation that could be utilized (Walls and Smith 1998). Eventually a detailed life cycle analysis is needed for UHPC structures. This analysis would include the environmental impact of the structure, costs and benefits of the structure, and sustainability of the materials used. As more UHPC structures are built, it is important that the proper data and observations are made throughout their lives, so that the life cycle analysis can be simplified.

130

Implementation and Future Work Recommendations are made for adapting current test procedures to properly evaluate UHPC properties.1. cylinders 7. This chapter summarizes information on modifications to the testing procedures used.1. 7. the spirit of ASTM and AASHTO standards were followed for testing UHPC but some modifications were necessary. cylinders were used instead of 4 x 8 in.000 psi. or 6 x 12 in.0 Recommendations. 3 x 6 in.1 Recommendations for UHPC Testing Procedures While ASTM and AASHTO Standards are accepted for normal strength concretes. 7. Implementation activities and suggestions for future work are also included in this chapter.2 Modulus of Elasticity and Poisson’s Ratio Base Procedure: ASTM C 469 using Compressometer-Extensometer Procedure Modifications for UHPC: • The load rate was increased from 35 psi per second to 150 psi per second. usually up to about 10.1 Compression Testing Base Procedure: ASTM C 39 Procedure Modifications for UHPC: • • The load rate was increased from 35 psi per second to 150 psi per second. As such. 131 .7. these procedures may not always be applicable for assessing UHPC performance.

5 times the first crack deflection so a greater number of toughness indices and residual strength factors could be calculated.4 Rapid Chloride Penetration Test Base procedure: ASTM C 1202 Procedure Modifications for UHPC: None Preparation Notes: • Specimens were cast in a 4 in. • The test was carried out to 20. 7.40 f `c from that test. per minute independent of curing regime tested. diameter x 3 in. Further research is needed to discover whether or not vacuum preparation actually saturates specimens. high cylinder and one day before treating were cut down from the top to 4 in.• Instead of breaking a companion cylinder and determining 0.003 in. diameter x 2 in.3 Flexural Strength Testing for First Cracking and Flexural Toughness Base Procedure: ASTM C 1018 Procedure Modifications for UHPC: • The midspan deflection rate was chosen as 0.1. The bottom surface of the cylinder should not be cut due to the fact that the finished surface of UHPC will normally be exposed. high a day before treating. 132 . 7. • Specimens may never be fully saturated during preparation.1. specimens were loaded to a predetermined load based on curing regime and age at time of testing.

• Visual inspection should be carefully performed to monitor for minute micro-cracking and erosion of the outer surface.1. 7.Testing Notes: • Finished bottom of specimens shall be exposed to the NaCl solution to simulate finished surface of UHPC. • Future freeze-thaw testing of UHPC should test the specimens for several months after testing is complete to monitor for changes in RDM.1. Testing Notes: • The optional length change test should not replace the RDM test due to the potential for autogenous healing of UHPC specimens while undergoing freeze-thaw cycling. but the significance must be checked.5 Freeze-Thaw Cyclic Testing Base procedure: ASTM C 666. • Length change tests should be performed vertically to allow for easier handling of the specimens while maintaining accuracy.6 Coefficient of Thermal Expansion Base procedure: AASHTO TP 66-00 Procedure Modifications for UHPC: 133 . 7. Both ends may be cut. Procedure B Procedure Modifications for UHPC: None Preparation Notes: • The time that UHPC specimens should be soaked in water prior to testing should be determined based on a balanced evaluation of the need to maintain curing regime integrity versus the need to have a saturated specimen for freeze-thaw cycling.

Design Recommendations for UHPC Ultra-High Performance Concrete (UHPC) is one of the latest advances in concrete technology and it addresses the shortcomings of many concretes today. As such. Both ends of the sample should be cut 1/2 in. but suggested for future testing) Specimens shall be cast in a 4 x 8 in. samples must be coated in an epoxy resin to allow for testing in the unsaturated state.2 Draft U.• Do not soak the UHPC samples in a lime bath prior to testing. Preparation Notes: • (Not performed during this research. Recommendations are based on testing conducted at Michigan Technological University and reported herein. Tests were conducted on 134 . Testing Notes: None 7.S.S. Samples must be completely coated in epoxy except for the points where the LVDT will be in contact with the sample and where the support frame buttons are in contact with the sample. innovative solutions can be created for long-term applications.000 psi (sometimes greater than 30. The purpose of these draft design recommendations is to begin the development of a design code for using the material in the U. to ensure a plane surface of contact for both the LVDT and the support frame buttons. Instead. This may hydrate additional cement particles and change the properties of the UHPC sample.000 psi) with exceptional durability performance. cylinder mold and then cut to 4 x 7 in. UHPC is typically defined as a concrete achieving high compressive strength in excess of 25. bridges.

Part I.1. only those sections with applicable results from the research study herein are drafted. market at the time of testing. (II) design and analysis of UHPC structures.50 in. The main benefits of thermal curing are: • • • Increased compressive and tensile strengths Improved durability Apparent reduced creep and shrinkage after curing 135 . While several sections of the recommendations are outlined below. the only UHPC currently available in the U. Recommended testing procedures for UHPC material and durability characterization were listed previously in Section 7. while other applications may not be able to allow for curing other than in ambient conditions. Part I – Material Characteristics  Characterization of material behavior is necessary to design systems using UHPC.1 – Mixing and Placing  The concrete mix is batched in a laboratory within manufacturer recommended procedures to attain design characteristics for the proposed application. length) was provided at a rate of 2% by volume.S. UHPC must be covered throughout the casting process and immediately after casting to avoid moisture evaporation that can lead to reduced hydration and excessive surface shrinkage.2 – Effects of Thermal Treatment  Some UHPC applications benefit from a thermal curing. Steel fiber reinforcement (8x10-3in. Other sections are noted as “future sections”. φ by 0. Trial batching at a plant should be conducted to ensure that all parties are familiar with the material prior to casting final elements. Part I.Ductal®. Typical recommendations would include three parts: (I) material characterization. and (III) durability.

Horizontal casting allows for parallel ends and eliminates the need for end grinding as neoprene pads and high-strength sulfer capping are not appropriate for UHPC’s high strengths. Test specimens are typically 3x6 in. a preliminary design compressive strength of 24 ksi at 28-days can be assumed.3.2. The ultimate compressive strength of ambient cured UHPC is time dependent and shall be measured for specific applications. Thermal treatment begins with a 6 hour ramp up period. The ultimate compressive strength of thermally treated UHPC can be expected to achieve 25-30 ksi independent of when thermal curing is applied.• Increased time to maturity so as to eliminate waiting for a 28-day or longer cure time Thermal treatment is defined as a 100% humidity steam treatment at 194°F for 48 hours.x – Compressive Strength Gain with Time (future section)  136 . cylinders cast vertically or horizontally. Part I. Part I. followed by a 6 hour ramp down period.x – UHPC Maturity (future section)  Comments: UHPC maturity to be defined by the set time. 48 hours at the specified humidity and temperature.3 – Compressive Strength  Compressive behavior is characterized most often by ultimate strength. If no other information exists at the design stage. Ambient air-curing is typically considered as 72°F at 30-50% humidity. such that thermal treatment is applied only after the set to avoid the potential of DEF (delayed ettringite formation). Part I.

5. a first crack flexural strength of 0. and 1. to date. Part I.4.4 – Tensile Strength  Tensile behavior has been characterized in two stages: an elastic stage and a postcracking stage.4.05 ksi for thermally cured specimens. tensile behavior of thin slabs may be different than beam elements. If no other information is known in the early design stages. Part I.5 – Modulus of Elasticity  Part I. design standards for mild steel reinforced concrete sections. The flexural tensile strength at first cracking can be compared to current U.4.x – Post­cracking Behavior (future section)  Part I.1 – Static Modulus of Elasticity  The following relationship can be used to estimate the modulus of elasticity for UHPC elements (AFGC 2002): 262.S. A direct tensile test is considered the most accurate measure of tensile strength.x – Splitting Tensile Strength (future section)    Part I.75 ksi can be assumed for ambient curing at 28-days.x – Direct Tensile Strength (future section)  Part I. 137 . Testing should be conducted to verify results for the intended curing method.Part I. The fiber strength and disbursement can strongly influence the tensile capacity of a UHPC member.S. Full size specimen testing may be needed for thin slab applications.4.000 (psi) where: f’c = compressive strength of UHPC (psi).x – Flexural Tensile Strength  The flexural tensile strength of UHPC at first cracking is dependent on the curing method applied. As additional test results become available. however no such test has proved successful for UHPC in the U.

9. Part I.7. an estimate of 7800 ksi after 14-days can be assumed. the modulus of elasticity for thermally cured fiber reinforced specimens can be estimated as 8150 ksi.8.2x10-6/°F can be assumed as the coefficient of thermal expansion for UHPC that has been thermally cured. For ambient cured specimens. Part I.7 – Thermal Characteristics  Part I.9.8 – Shrinkage Behavior (future section)  Part I.x – Early Age Behavior   Part I.21 within the elastic range.5.7x10-6/°F for 28-day ambient cured UHPC elements. Thermal Diffusivity.10 – Fatigue Strength (future section)  Part I.x – Thermal Conductivity.2 – Dynamic Modulus of Elasticity (future section)  Part I. Part I.1 – Coefficient of Thermal Expansion  A value of 8.If in the initial design stages and information is unknown.x – Post Thermal Treatment Behavior  Part I.8.6 – Poisson’s Ratio  Poisson’s ratio for UHPC can be generally assumed to be 0. This value should be reduced to 7. Specific Heat (future section)  Part I.9 –Creep Behavior (future section)  Part I.x – Early Age Behavior   Part I.x – Post Thermal Treatment Behavior  Part I.7.11 – Impact Strength (future section)  138 .

Applied loadings and induced stresses due to loadings shall be calculated in accordance with accepted methods and the current edition of the AASHTO Bridge Design Specifications. vibration Moment capacity and strain compatibility Shear capacity – one-way beam shear. and confinement steel for bursting zones Connection details Part III – Durability  Part III. 139 . punching shear.Part II – Structural Analysis and Design   Part II. capacities. and general detailing. Part II. cracking. spacing. interface (horizontal) shear Torsional capacity Fatigue resistance Buckling of slender members Anchorage – bond strength. The unit weight of UHPC shall be assumed to be 155 pcf.1 – Loads  Loading for UHPC elements does not vary from loadings applied to normal strength concrete members.1 Chloride Ion Ingress  UHPC has demonstrated superior resistance to chloride ion ingress (negligible penetrability) regardless of curing method. beveling Serviceability – deflections. The following outline is proposed and can be adjusted as research becomes available to support findings: • • • • • • • • • General detailing – cover.2 – Serviceability and Ultimate Limit States (future sections)  This section should include discussions for serviceability and ultimate limit states.

and coastal areas. End of DRAFT U.5 – Alkali­Silica Reactivity (future section)  Part III. 140 . As such.S. However. the direct implementation of UHPC into highway systems without a proper design code could result in an inefficiency of the material use.6 – Fire Resistance (future section)  Comment: No fire resistance testing has been performed on UHPC to date in the U. it is imperative to consider efficient designs through optimization of bridge girders sections and deck systems. especially those exposed to aggressive environments like those found in northern regions of the U. particularly in a decade in which engineers are well aware of the adverse effects that CO2 emissions (from cement production) can have on global climate change.Part III. Wasting material is not only expensive.S. it is irresponsible.S.3 – Carbonation (future section)  Comment: Carbonation has been found to be negligible (Japan 2006) Part III.01% length change) with no large cracking or spalling of the material regardless of curing method. Design Recommendations 7. making it a prime candidate for structural highway systems.2 – Freeze/Thaw Resistance  UHPC has demonstrated a high resistance to freeze-thaw damage (100+ durability factor and less than 0. because of the enhanced properties of UHPC (such as higher compressive strengths).4 – Chemical Attack (future section)  Part III. Part III. and advanced mechanical performance through increased compressive and flexural strengths.3 Implementation Activities UHPC has been shown to have extreme durability through high resistance to freeze-thaw cycling and chloride penetration.

The reality is that all prestressed elements would be creep loaded prior to production curing. Thermally curing the specimens while under load is an unanswered question which needs work. Poisson’s ratio. Also. and account for. Creep-testing research to date has been performed on small thermally cured specimens.UHPC appears to lock in some properties through the use of thermal treatment. 7.4 Suggested Future Work As a result of this work. UHPC exhibits increasing load carrying capacity beyond first-crack because of the fiber reinforcement. several items should be considered in future research. the practical limits of allowable cracking for use in the design of structural elements need to be determined. differences in first-crack flexural strength for specimens tested under several curing regimes need to be identified for proper implementation into design codes. Also. modulus of elasticity. Compressive strength. and if they do what is the limit of an Aircured specimen? The point at which a thermal cure does not have an impact needs to be located. In this research Air-cured specimens were tested at a maximum of 28 days. However. Furthermore. the increase in flexural capacity of UHPC as a function of crack growth/development. Suggestions for further study address many of these items. chloride penetration. Do the specimens continue to increase in strength or change modulus and approach an asymptotic ceiling. finding a relationship between the compressive stress and modulus of elasticity of nonthermally cured specimens is of interest. More research is needed to better understand. freeze-thaw resistance and the coefficient of thermal expansion exhibited no change in properties once thermally cured. It is of interest to see what happens over the period of months or years. several properties and their impact on design and performance are still not well understood. The 141 .

Performing an ESEM or petrographic analysis on freeze-thaw specimens post testing may shed more light on this interesting effect. Furthermore. then the element may be cured immediately or possibly stockpiled for some length of time before curing. There are two issues which need addressing.strands would be released and the element would have the compressive force applied. Supplemental work is necessary to develop a broad understanding of all types of concretes classified as ultra-high performance concretes. and other UHPC materials should be investigated to develop a comprehensive understanding of UHPC performance. The durability properties of UHPC may show great improvements and it is foreseeable that it could be used as a sacrificial or wearing course over normal strength concrete. and the second. Additional petrographic analysis on the continued hydration of Air-cured versus thermally treated UHPC specimens will also help in properly describing the self-healing phenomena of UHPC. 142 . UHPC’s resistance to freeze-thaw cycling in a saline environment should also be investigated. The first is the creep loading on a non-thermally cured specimen. what happens to the specimen when it is under load and exposed to the high temperatures of a thermal cure. The bond characteristics of these two surfaces would need to be investigated. The effects of self-healing in UHPC should be investigated further to determine whether long term benefits exist. A closer look at the dynamic response of UHPC undergoing freeze-thaw testing may provide information about the nature of the “skewed” effect on the frequency curves used to determine relative dynamic modulus. Currently only research on Ductal® has been performed in the United States. along with UHPC’s resistance to deterioration if cracked prior to freeze-thaw cycling.

Determining UHPC’s saturated CTE value. 143 . while UHPC shows promise as a material of choice for transportation infrastructure. Besides superplasticizer. the interaction between UHPC elements and NSC or HPC elements due to thermal gradients should be researched to provide practical recommendations when using UHPC with other concretes. Additionally. These studies can provide further flexibility when designing structures using UHPC. Because of the increased ductility afforded by UHPC. design codes may need to consider a crack-width based approach to design instead of the current stress-based limit states. and age effects are crucial to the implementation of UHPC as a viable structural material. A rigorous study of section optimization (such as stout double-tee shapes for girders or waffleslabs systems for slab bridges or deck panels) is warranted.S. Additional studies on the effects of specimen age on UHPC’s CTE value are valuable to prestressed concrete manufacturers to accurately estimating strand stress. thermal interaction with non-UHPC materials. such as water baths have also not been examined thoroughly in the U. The majority of the research to date has focused on the unsaturated CTE values of UHPC. Incorporating results from tests reported herein. Alternate curing regimes. design code development is an integral part of introducing any new material for application.Research into the coefficient of thermal expansion of saturated UHPC specimens should also be investigated. The effects on long term durability by including accelerators during the mixing process may be important for those looking for rapid strength gain and long term durability. no admixtures were used in this research. And lastly. will provide a comprehensive document for designing UHPC structures. in conjunction with other testing results by others as well as those suggested for further study.

(This page intentionally left blank) 144 .

Annual Book of ASTM Standards.M. Standard Test Method for Fundamental Transverse. E. Michigan. 551-552. Hartmann. ASTM C 617. J. American Concrete Institute-ACI (1997). Standard Practice for Capping Cylindrical Concrete Specimens. ACI Committee 318. (2006) ASTM C 39. Aitcin. ASTM C 666. ASTM C 192. Potter. Standard Practice for Use of Apparatus for the Determination of Length Change of Hardened Cement Paste.2R-92.. Steinberg. (1998) “High-Performance Concrete. Michigan. Graybeal. 2004. Farmington Hills.. ACI Committee 201. T. ASTM C 215. Ahlborn. Standard Method of Test for Coefficient of Thermal Expansion of Hydraulic Cement Concrete. Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens. and Torsional Frequencies of Concrete Specimens. New York.” Routledge. 2002. 145 . Farmington Hills. 549. Farmington Hills.ACI (2002).” Manual of Concrete Practice-Part I.References American Concrete Institute-ACI (2005). Orlando. Standard Test Method for Flexural Strength of Concrete (Using Simple Beam with Third-Point Loading).P. ASTM C 490. Longitudinal. ASTM C 469. 2003. “State-of-the Art Report on High-Strength Concrete. “Guide to Durable Concrete. (2003).L. 2002.” 3rd International Symposium on High Performance Concrete. and Concrete. Standard Test Method for Static Modulus of Elasticity and Poisson’s Ratio of Concrete in Compression. J. and Volygi.” ACI 201. Standard Test Method for Resistance of Concrete to Rapid Freezing and Thawing. 2002. American Society for Testing and Materials. Mortar. 1998. ACI Committee 363.. PCI. Michigan. American Concrete Institute ...A. pp. Florida. J. 2004. American Association of State Highway and Transportation Officials – AASHTO (2005) AASHTO Designation: TP 60-00. “Ultra-High Performance Concrete-Study Tour 2002. ASTM C 78. Building Code Requirements for Structural Concrete. Standard Practice for Making and Curing Concrete Test Specimens in the Laboratory. 2000. B. 2004. P.L.

(Withdrawn 2006)..P.J. Detroit. Banthia. L. 30(12). (2005). Dugat. pp. eds. (2000).. Prestressed Pedestrian Bridge-World’s First Reactive Powder Concrete Structure. and Pigeon. and Cheyrezy. Shah. and Aitcin. Behloul.. 60-71. (1997). C. September-October 1999. D. 1995. pp. Michigan. Banta.” Presentation at Bouygues Headquarters – Challenger. SETRA (Service d’études techniques des routes et autoroutes). 2001. O. (Bétons Fibrés à Ultra-Hautes Performacnes – Recommandations Provisoires). O. “Ductal Reference Projects. N. France. 41-75. 94.” ACI Materials Journal. Standard Test Method for Flexural Toughness and First-Crack Strength of Fiber-Reinforced Concrete (Using Beam With Third-Point Loading).E. “Precast... Dallaire.. (2002a). Vol. Lachemi.. M. P. AFGV (2002) Association Française de Génie Civil Interim Recommendations for Ultra High Performance Fibre-Reinforced Concretes.. No. Stevens. (2002b). Standard Practice for Use of Unbonded Caps in Determination of Compressive Strength of Hardened Concrete Cylinders. J.” Presentation at Bouygues Headquarters – Challenger. D. Gopalaratnam. “Characterization of the Granular Packing and Percolation Threshold of Reactive Powder Concrete. Morgan. Vernet. Johnston. M. France. July/August 1997. Behloul. 2000.” Testing of Fiber Reinforced Concrete. 286-290. M.” Cement and Concrete Research. 2005. 4. Chen. S.. “Mechanical Properties and Durability of Two Industrial Reactive Powder Concretes. Blais. ASTM C 1231. 5. ASTM C 1202.C..S. V. “Comparative Toughness Testing of Fiber Reinforced Concrete. M. Bonneau. “Horizontal Shear Transfer Between Ultra High Performance Concrete and Lightweight Concrete. No. American Concrete Institute.. 146 . St Quentin en Yveslines. “Seonyu Foot Bridge. and Aitcin. Moranville. 44. M.” PCI Journal. M.R. P. 1861-1867. Standard Test Method for Electrical Indication of Concrete’s Ability to Resist Chloride Ion Penetration. VA. St Quentin en Yveslines. Mindess. Bonneau. and Cheyrezy.. S.. and Tatnall P. T.” MS Thesis... ASTM C 1437. Blacksburg. France.D. 1997. E. pp. M. Standard Test Method for Flow of Hydraulic Cement Mortar. (1995). and Couture. pp. Virginia Polytechnic Institute and State University. M.. P.ASTM C 1018. Vol. C... (1999).

“Achieving the Promise of Ultra-High Performance Concrete. May 1996. 149-155. Ductal® Reference T 002. Collins. pp. G.R. Farmington Hills. C. “Mechanical Properties of Reactive Powder Concretes. Kassel University Press. (1998) “Reactive Powder Concrete (RPC) Application for Seismic Design. A. H.. (1996). and Dauriac. Ductal® Reference T 001. Lafarge North American. pp. Gatty. Roux. (1998).. Fehling. (2006) “Iowa Bridge gives Glimpse into the Future. A. “The destruction of concrete by frost. Germany. Clinard. C. Bunje. Bonnamy.” ACI Materials Journal. 41-46. J. P. L. 23(1).Cheyrezy. E. Federal Highway Administration (FHWA) (2006). “DASYLab 8. 26 Jan. 1-15. W. and VanDamme.” ACI Journal. “Prediction of the thermal coefficient of expansion of concrete. Vol. pp. 2007.. (1999).. Operating Procedure Cylinder & Prism Preparation. Dowd.A.. Dugat. C. N. Ductal® Reference T 009. Schreiber.” Ultra-high performance concrete (UHPC). Summer 2006. 448-454. pp. Emanuel. (1977). American Concrete Institute. Germany.” FOCUS. MI.” American Concrete Institute Convention.. Kassel. Operating Procedure Compressive Test. J. “Silica Fume Distribution and Reactivity in Reactive Powder Concretes.” Fly ash. 29. Feylessoufi.. Turner-Fairbank Highway Research Center Online. Thailand. SP 178. 29-41.” ver.04. Canada. (1944). 8. Edvardsen. Schmidt. silica fume. Ductal® Reference T 006.” Journal of the Institution of Civil Engineers.. W. “Water permeability and autogenous healing of cracks in concrete. Calgary. pp.0. No.. 147 . Inc. Behloul. Nov. J. 96-M56. Second International Symposium on Ultra High Performance Concrete. Ductal® Reference. pp. Operating Procedure Cylinder End Preparation. 233-240.. S. 188. DasyTech USA. W. and Hulsey. K.931-935. Richard. 1998. DASYLab (2004). (2008) “The “Gartnerplatzbrucke” Design of first hybrid UHPC-steel bridge across the river Fulda in Kassel. pp.. 581-588. October 29. Title No. Endicott...” Materials and Structures/Matériaux et Constuctions. M.00. Online. slag & natural pozzolans in concrete : proceedings : Sixth CANMET/ACI International Conference : Bangkok. Operating Procedure Flow Test. M. M.” Ascent. and Rernier.

Hartmann. June 2006. A. pp. (2005). January-February 2002..” PhD Dissertation.” 3rd International Symposium on High Performance Concrete. B. Harris.L. C. J. Virginia.. PCI. Federal Highway Administration. “Ultra-High Performance Concrete for Highway Bridges. FHWA. S. U. “Testing of Ultra-High-Performance Concrete Girders. D. (2007). Clearwater. B. Gilbertson. FL. Graybeal. “CO2 Emissions Profile of the U. Hartmann. Graybeal. Hanle. K. and Graybeal. and Sellevold. Jayaraman. Loukili. 37. E.” Journal of Materials in Civil Engineering. 148-149. Graybeal. J. Pijaudier-Cabot.” Virginia Transportation Research Council and Virginia Department of Transportation. and Roberts-Wollmann. Graybeal.” FHWA Structural Testing Laboratory.A.. Cement Industry. (2001). (2004). “Characterization of Punching Shear Capacity of Thin UHPC Plates.” Cement and Concrete Research. J.C. (2007).” Hartmann. G. (2004). B.” MS Thesis. Maryland. and Chanvillard. France.” 13th Annual Emission Inventory Conference. No.L. J.. Graybeal.” Rep. B.. Jacobsen. Graybeal. (2005).G. (2002). Fib Symposium 2004. “Nation’s First Ultra-High Performance Concrete Bridge Opens in Iowa. “Characterization of the Punching Shear Capacity of Thin Ultra-High Performance Concrete Slabs. Virginia. L. “U-HPC Testing: Shear. FHWA-HRT-06-103. 848-854. Orlando. 148 . V.. “Self healing of high strength concrete after deterioration by freeze/thaw. D.S. 47(2). “Wapello County UHPC Project. “Material property characterization of ultra-high performance concrete. J. Harris.” PCI Journal.. Charlottesville. G.Granger. University of Maryland. and Graybeal. Blacksburg. “Characterization of the Behavior of Ultra-High Performance Concrete. IDOT (2007). and Smith. (2006a). Washington. and Tanesi. 19(10). Avignon. Virginia Polytechnic Institute and State University. J.S.” AFGC: Concrete Structures: The Challenge of Creativity. “Durability of an Ultrahigh-Performance Concrete. Florida. College Park. pp. and Hartmann.” Research & Technology Transporter. EPA. D. (1996). (2004). 9 August. (2006b). and Perry.J. B. ASCE.A. B. (2003) “Strength and Durability of Ultra-High Performance Concrete. 55-62.” Cement and Concrete Research.. B. S.K. Email to C. “Experimental characterization of the self-healing of cracks in an ultra high performance cementitious material: Mechanical tests and acoustic emission analysis. 26(1). B. 519-527.K.

Loukili. “Influence of surface modified Carbon Nanotubes on Ultra-High Performance Concrete. Jungwirth.C. R. (1998). Michigan Technological University. Jamet. Chiu.. Bridge Using Ultra High Performance Concrete and PI Girders. N. P.-T. Lafarge North America (2006a). Gopalaratnam. Kollmorgen. T. Michigan.” <http://www. Lafarge North America (2006b).S.” Proceedings of the International Symposium on Ultra High Performance Concrete. Japan. Detroit.” Journal of ASTM International. “Structural Behavior of Tension Members in Ultra High Performance Concrete.” International Symposium on UHPC. Lopez-Anido. “The study of bond strength and bond durability of reactive powder concrete.” MS Thesis in Civil Engineering. eds.” Recent Advances in Concrete Technology : proceedings : Fourth CANMET/ACI/JCI International Conference : 136 Tokushima. V.. A. American Concrete Institute. SP-179. I. (2004).J. American Concrete Institute. “Phase transitions of adsorbates IV – Mechanism of frost action in hardened cement paste.A. and Wang.” Niwa. 149 . Banthia. “A Study on Delayed Deformations of an Ultra High Strength Cementitious Material.-C. A. Kassel. and Tatnall P. (2004). MI... Kassel. and Lamirault. J. Kowald. Gettu.pdf> (March 16.. 1995.. Keierleber. “The Life-Cycle Cost Evaluation Report of Fiber Reinforced Polymer (FRP) Composite Bridge Decks. August 2007.. “Recommendations for Design and Construction of Ultra High Strength Fiber Reinforced Concrete Structures (Draft). J. M. (2005).. Houghton. pp 195-202. Richard.” Proceedings of the 2007 Mid-Continent Transportation Research Symposium.” <http://www. Phares. 38-42... B. Germany. 2 June. pp. Bierwagen.-G. G. Kassel University Press.. V. “Wapello Bridge. Iowa. B. “Toughness of FiberReinforced High-Strength Concrete from Notched Beam Tests. (2007) “Design of Buchanan County. (1998).. 2008). R. Y. “Detroit-Columns. Stevens. Farmington Hills.JSCE (2006) Japan Society of Civil Engineers.imagineductal. “Impact of Age and Size on the Mechanical Behavior of and UltraHigh Performance Concrete.. and Couture. 23-39. and Aguado. F.. Litvan. D. J. 929-949. G.S. 2(7). ed. (2004).. (1995)..” University of Maine. D.” Journal of the American Ceramic Society. Iowa. (1972).G. Fanous. Gopalaratnam. Michigan.com/imagineductal/public/eng/pdf/Detroit-Columns. C. D. Germany.” Testing of Fiber Reinforced Concrete.pdf> (March 16. Ames.imagineductal.com/imagineductal/public/eng/pdf/WapelloBridge. Lee. pp. A. 2008). 55(1). and Muttoni.

Banthia. pp... SP-167. 15 July. Ohio LTAP Center (2007). Inc.F. Mindess. MDOT (2007b). “Preventative Maintenance/Repair Guidelines for Bridges and Culverts. Properties. O’Neil. A-94 Revised. American Concrete Institute. (1999) "Materials for Civil and Construction Engineers. and Zaniewski. CA.” Proceedings.M. Nemegeer. pp. Online.E. Office of Management and Budget (2007). D. Inc.. “Capital Scheduled Maintenance Bridge Manual.. Mitchell. pp.. Upper Saddle River. Donald Li (2008) “Influence of Curing Regimes on the Durability of an Ultra-High Performance Concrete. Michigan. Michigan Technological University. American Concrete Institute. Online. Email to Dr. “Michigan Engineers’ Resource Library. 249-261. Menlo Park. (1953).... MI. M.C. “Thermal expansion tests on aggregates. eds. (2006). (1995).J.C.E.Mamlouk. (1997).J.. Mehta.Strength Concrete : An International Perspective : papers presented at three halfday sessions at the ACI Convention in Montreal. “Ultra High Performance Concrete Highway Bridge. (2006). P. and Darwin. ASTM. “Measuring Toughness Characteristics of SFRC-A Critical Reviwe of ASTM C 1018.K. and concretes. OMB. P. V. Theresa Ahlborn.” Testing of Fiber Reinforced Concrete. “UHPC progress report . 2007. S. Michigan. D. Online. J. Misson. Gopalaratnam.” High. C. L. N. “Development of Reactive Powder Concrete (RPC) Products in the United States Construction Market.gov. J. B. 53. New Jersey. 150 ... 1995.F. 41-75. S. Inc.” Pearson Education. and Gililand.real discount rate and deck timeline. NY.. 26 Feb.. Detroit.” CTRE Online. and Tatnall. MDOT (2007a). Stevens.S.” The McGraw-Hill Companies.” Michigan. and Tatnall P. Moore.” MS Thesis in Civil Engineering. 23 July. and Monteiro." Addison Wesley Longman.gov.K. 963-977.J. “Concrete Microstructure. D. 15 July. and Bierwagen. Young.” Whitehouse. and Materials.” MERL (2006). P. Houghton.” ODOT On-line Bridge Maintenance Manual. “Concrete 2nd Edition. E. Farmington Hills. 25 April. “Circular No. D. Dauriac. Iowa State University. 31-192. (2003). Online. neat cement.” (Database of Michigan bid items that allows project managers and engineers make road and bridge project estimates). New York.

. American Concrete Institute. Perry. and Helmuth. V. (1953). Powers. 184-211. B. Richard. Perry. 2003. (1945). pp. Properties. Philadelphia. F.” Journal of the American Concrete Institute. 164-172. V. Perry. and Cavill. “A working hypothesis for further studies of frost resistance. (2003) “The World’s First Long-Span Roof Constructed in Ductal®. 151 . (1994). R.. STP 169C. PCA. Houghton.. 29. “Theory of volume changes in hardened portland cement pastes during freezing. and Future. “First Use of Ultra-High Performance Concrete for an Innovative Train Station Canopy. J.C. American Society for Testing and Materials. V. 25. and Pleau. and Cheyrezy. M. 507-518. “Reactive Powder Concretes with High Ductility and 200800 MPa Compressive Strength. Orlando. http://www. Orlando. London.” E & FN Spon. PA. Pigeon.pdf Viewed 04/17/2007. R.” Proceedings of the Highway Research Board. T. (1995). Michigan. (2004).” Nova Award. D. Peuse.” MS Thesis in Civil Engineering. and Zakariasen. 25(2). T. (2008) “Ultra-High Performance Concrete for Michigan Bridges – Material Behavior. “Ultra-High Performance Concrete: First Recommendations and Examples of Application. PCI.C. Richard.” AustRoads Conference.org/Nom2003/Nom22_03. Markets & Manufacturing. No. “Reactive Powder Concrete Bridges. (1996). 7 October 1995. pp. Pine Instrument Company (2006). Perth. (2006). pp. Florida. P. SP144. pp. and Cheyrezy.. P. PA. Michigan Technological University. pp. Vol. Rebentrost. Powers. M.Perenchio. 1501-1511. UK.” Significance of Tests and Properties of Concrete and Concrete-Making Materials. M.” Concrete Technology Today. Grove City. Farmington Hills.H.H.A. Materials. W.” 3rd International Symposium on High Performance Concrete. Australia.” Cement and Concrete Research.” 3rd International Symposium on High Performance Concrete.” Concrete Technology: Past. Erron J. D.. and Zakariasen. Florida. “Composition of reactive Powder Concretes. 16(4). Resplendino. PCI. 11-23. 245-272. “Corrosion of Reinforcing Steel. “AFCT1A / AFCT1C Coefficient of Thermal Expansion of Portland Cement Concrete Measurement System Operation Manual. Present. M. “Durability of Concrete in Cold Climates. (2003) “Overview of UHPC Technology. (1995). (2003).” Version 071206.H. MI. Publication Code Number 04169030-07.cif. and Petitjean. J.

D. Hooton. Ontario. Walker. Bristow.J. pp. (1999). R. 152 . 56-57.... 21(2).R. “Ultra-High-Performance Concrete: Research. <http://www. November.” Concrete Decor. Russell. Stanish. (2000). K. 2006." Concrete International. Department of Labor Online. 75-79. (2007). “Testing the Chloride Penetration Resistance of Concrete: A Literature Review.” FHWASA-98-079. S.G. 661-680.D.” Transportation Research Board Annual Meeting CD-ROM. 2008). University of Toronto. Bloem. “Life-Cycle Cost Analysis in Pavement Design. “Ductal Concrete by Lafarge. (2003).” Cement and Concrete Research. Sep. Sritharan. “Chloride Permeability and Impact Resistance of Polypropylene-Fiber-Reinforced Silica Fume Concrete. (2005). J. Orlando. (1998). U. Florida. Walls.” Concrete Precasting Plant and Technology. Teichmann. Kai.net/All_Access/703/CD703-Technology. ACI Special Publication. J.L.D. E.J. Fehling. and Bayasi... Schmidt.” FHWA Contract DTFH61-97-R-00022. and Perry. "ACI Defines High-Performance Concrete.S.” ACI Journal Proceedings.S. and Smith. Kingston: Reed.” Seventh International Symposium on the Utilization of HighStrength/High-Performance Concrete. D. D. (2003) “Characterizing an Ultra High Performance Material for Bridge Applications Under Extreme Loads. H. D. and Eggers.cfm> (March 16.” 3rd International Symposium on High Performance Concrete. Online. PCI. (2005).” 30 Nov. 28(7). B. 228. DC. and Mullen. pp 51-78 Searls. pp. Semioli. W. Torres. 2006 Toutanji. Toronto.A... H. M. Z. S. pp. B. Washington. “Ultra-High Performance Concrete: Perspective for the precast concrete industry. V. 24 July. “Effects of temperature changes on concrete as influenced by aggregates. (2001). W. Pavement Division: Washington.. 69(3). February.” Concrete International. and Thomas.. S. and Roland.concretedecor.. B. “The New Concrete Technology.. 2007. M. Schmidt. M. USA. (1952).H. Canada. 48(8). pp. T.. 961-968 U. 2(11). Department of Labor 2007. (1998). 16-29. M.RS Means (2005).. E.. 2004. FHWA.C. “Bureau of Labor Statistics. Development and Application in Europe. and Fehling. III. S. McNeil. “Capping Systems for High-Strength Concrete. Heavy Construction Cost Data 19th edition.

........ 3 Table A......8: RCPT specimen data sorted by curing regime................ 12 Figure A............ 8 Table A........................................................ 18 A-1 .... 16 Table A...........................................................5: Data for TT Flexural Specimens ..... 5 Table A.... 7 Table A.... Poisson’s Ratio....................3: Data for DTT and DTT Cylindrical Specimens....................................... 11 Table A................ 14 Figure A................................................................. Modulus of Elasticity.................. Coefficient of Thermal Expansion.......... and Flexural Tests ...........2: Data for TT Cylindrical Specimens .............. 13 Figure A............. 2 Rapid Chloride Penetration................11: Average length change of UHPC specimens undergoing freeze-thaw cycling .......................1: Data for Air-Cured Cylindrical Specimens ....9: Sample freeze-thaw cycle temperature – Position A17 ...................................APPENDIX A: EXPERIMENTAL TEST DATA Specimen naming scheme ................................................................................................................................................ 6 Table A.............................. 2 Compression Strength.................12: Michigan Tech CTE Summary Report on UHPC ................................ 9 Table A...........................10: Average mass change of UHPC freeze-thaw and side study specimens ... Freeze-Thaw Resistance.............................................................................7: RCPT specimen data sorted by batch ........................13: Mass change study on epoxy coated UHPC CTE specimens ........................................................................................... 15 Table A........................................................................ 10 Table A...6: Data for DTT Flexural Specimens .....4: Data for Air-Cured Flexural Specimens .........................................

x 11. This specimen would have been cast out of batch 3. and Flexural Tests The naming system followed the format of Batch number-Specimen geometry-Curing regime-Testing age-(optional character).25 in. diameter cylinders and B2 for 2 in. x 2 in. Poisson’s Ratio. air treated. beam.Specimen naming scheme This appendix includes data from all specimens tested herein. The curing regime followed the same designation as is laid out in Table 1. x 2 in. x 11.1 where A was Air-cured. and DDTT was double delayed thermal treatment.25 in. DTT was delayed thermal treatment.2 .3. An example of the specimen nomenclature would be B3-B2-A-28-B. was a 2 in. The testing age designation was the age in days at which the specimen was to be tested. The optional character was used only for the 2 in. beams because they were the only specimens in which more than one particular “geometry-curing regime-testing age” came from one batch. x 2 in. This specimen was cast from the second batch. In this case. Specimen geometry was designated by C3 for 3 in. x 2 in. tested at an age of 28 days and been the second of the 2 in. x 11. a A. Compression Strength.25 in. Another example would be B2-C3-TT-14. Modulus of Elasticity. beams. a letter A or B was included to facilitate record keeping. TT was thermal treatment. The batch number followed the form of B(number) where the number was 1 through 7 was used with 7 being the batch without fibers. A naming system was used to identify each specimen. beams made from the batch. x 11.25 in.

when a specimen was used for a side study test rather than actual testing. When referring to a specific specimen’s nomenclature. Additionally.3 in. cylinder.). RCP for the rapid chloride penetration test. Freeze-Thaw Resistance. x 6 in. cylinders thermally treated and being tested at 14 days came from batch 2. Rapid Chloride Penetration. A specimen nomenclature was used to identify batch number. FT for the freeze-thaw test. Supplementary CTE and RCPT specimens were also cast in batches M through S with excess material. the abbreviation “Air” is used. The test procedures were designated on each specimen as CTE for the coefficient of thermal expansion test. All specimens were marked with the following nomenclature after demolding: Batch number-Test procedure-(optional side study label)Curing regime-Testing age(optional letter within batch). x 6 in. the optional A or B tag was left off the specimen identifier. and C for the compression test. SS was added before the curing regime notation. and tested at 14 days. Because no other 3 in. and began again at M and ended at S (skipping the letter O label to avoid confusion) for batches comprising of primarily freeze-thaw test specimens. So a side study air cured specimen would have the label of SSA A-3 . when referring to the type of curing regime. test procedure. curing regime and testing age. the entire nomenclature will be used). Coefficient of Thermal Expansion A total of nine batches were cast for testing (with several more cast for specimen shakedown) and 79 samples were tested to observe the durability characteristics of UHPC (Error! Reference source not found. Batch numbers began at A and ended at D for the batches incorporating coefficient of thermal expansion and rapid chloride penetration test specimens. thermally treated. Curing regime designations followed the format of TT for thermal treatment curing and A for ambient air curing (Note: In this report.

Specimen Y marked N-FT-SSA-28 means that Specimen Y was cast from batch N for the purpose of freeze-thaw side study testing and was ambient cured for 28-days prior to testing. and S-C-TT-28C to distinguish between the individual specimens. S-C-TT-28B. For example. Batch S included three specimens designated as S-C-TT-28A. However. A. an additional letter (A – D) was added at the end of the specimen nomenclature. Similarly. when two or more specimens from a particular batch were cured the same. For example. Specimen X marked N-FT-A-28 would imply that Specimen X was cast from batch N for the purpose of freeze-thaw testing and was ambient air cured for 28-days prior to testing. – companion compression test cylinders). The final number in the nomenclature specifies the number of days after casting that a specimen was tested. In this study.g.4 . and tested under the same test (e. only companion compression test cylinders employed this additional nomenclature scheme.rather than A. for the same amount of time.

998 0.378 23.316 24.103 0.327 14.162 6859.109 0.092 Specimen B1-C3-A-3 B2-C3-A-3 B3-C3-A-3 B4-C3-A-3 B5-C3-A-3 B6-C3-A-3 B1-C3-A-7 B2-C3-A-7 B3-C3-A-7 B4-C3-A-7 B5-C3-A-7 B6-C3-A-7 B1-C3-A-14 B2-C3-A-14 B3-C3-A-14 B4-C3-A-14 B5-C3-A-14 B6-C3-A-14 B1-C3-A-28 B2-C3-A-28 B3-C3-A-28 B4-C3-A-28 B5-C3-A-28 B6-C3-A-28 B7-C3-A-28A B7-C3-A-28B B7-C3-A-28C B7-C3-A-28D B7-C3-A-28E B7-C3-A-28F A-5 .123 7795.332 0.2085 0.742 25.934 24.874 7537.223 0.2022 0.939 19.2095 0.073 7906.948 19.578 6912.635 7675.109 0.159 7899.160 23.195 20.2003 0.098 6954.149 0.542 24.327 0.299 7727.2065 0.126 0.716 24.2012 0.149 0.120 0.155 14.596 7993.1995 0.212 0.710 21.304 7666.141 19.972 21.602 7850.1900 0.427 7356.2062 0.1995 0.149 0.813 7540.2035 0.037 23.1952 0.2050 0.132 14.1989 0.172 0.354 25.2044 0.116 7736.809 7813.915 7777.049 22.2064 0.2109 0.1995 0.517 NR NR 0.080 0.1980 0.132 0.481 7376.074 0.109 0.Table A.149 0.656 NR NR 0.828 22.809 0.189 15.1: Data for Air-Cured Cylindrical Specimens Compressive Modulus of Poisson's Degrees out of Stress (ksi) Elasticity (ksi) Ratio Perpendicularity 13.652 7668.2114 0.1961 0.235 0.2031 0.606 23.585 24.667 6911.080 0.370 7701.1999 0.378 20.218 0.2020 0.747 7493.713 7688.501 25.399 7792.281 19.2048 0.418 14.2020 0.819 0.388 7905.2068 0.481 0.127 7756.160 0.511 7943.218 0.810 8056.796 23.115 0.727 21.2190 0.

046 31.241 8127.115 0.705 8020.650 8080.787 0.676 8231.944 33.2054 0.638 0.2029 0.884 30.2024 0.115 0.2060 0.2096 0.000 31.138 0.066 8127.1992 0.138 0.120 30.2124 0.575 8198.994 31.1968 0.2: Data for TT Cylindrical Specimens Compressive Modulus of Poisson's Degrees out of Stress (ksi) Elasticity (ksi) Ratio Perpendicularity 30.2063 0.119 8166.025 29.884 0.152 7879.014 30.2119 0.2047 0.986 8066.418 30.2025 0.2012 0.6 .2090 0.808 30.766 0.154 32.2064 0.097 0.1968 0.572 8151.503 8047.2063 0.310 0.338 31.464 8359.182 0.080 0.217 7858.143 0.172 0.005 0.015 7793.080 28.497 8348.577 8155.095 30.591 7948.081 28.2064 0.693 8124.074 0.180 8079.2050 0.2068 0.664 7844.063 8175.052 0.1991 0.120 0.153 29.456 30.092 30.2135 0.126 0.138 0.103 0.103 Specimen B1-C3-TT-7 B2-C3-TT-7 B3-C3-TT-7 B4-C3-TT-7 B5-C3-TT-7 B6-C3-TT-7 B1-C3-TT-14 B2-C3-TT-14 B3-C3-TT-14 B4-C3-TT-14 B5-C3-TT-14 B6-C3-TT-14 B1-C3-TT-28 B2-C3-TT-28 B3-C3-TT-28 B4-C3-TT-28 B5-C3-TT-28 B6-C3-TT-28 B7-C3-TT-28A B7-C3-TT-28B B7-C3-TT-28C B7-C3-TT-28D B7-C3-TT-28E A.126 30.2046 0.215 32.667 34.252 8048.115 0.845 30.Table A.103 0.390 7929.

103 30.989 8096.891 7807.726 29.120 0.424 29.115 0.2226 0.2063 0.955 0.572 8082.1980 0.282 30.086 0.124 8178.1987 0.Table A.2025 0.706 29.210 29.109 0.2065 0.2025 0.375 29.086 0.005 7786.340 32.166 28.2085 0.2005 0.1984 0.795 8180.298 0.431 8181.720 8144.2012 0.191 8129.630 0.278 30.097 0.2091 0.092 30.1977 0.166 30.097 29.378 30.098 30.098 7872.662 7544.103 0.761 24.2115 0.986 0.740 8128.1988 0.106 8001.063 7979.126 0.712 8098.420 30.154 29.644 0.2031 0.015 28.103 0.109 0.333 8026.2085 0.155 0.161 0.1994 0.097 28.2025 0.995 0.149 Specimen B1-C3-DTT-14 B2-C3-DTT-14 B3-C3-DTT-14 B4-C3-DTT-14 B5-C3-DTT-14 B6-C3-DTT-14 B1-C3-DTT-28 B2-C3-DTT-28 B3-C3-DTT-28 B4-C3-DTT-28 B5-C3-DTT-28 B6-C3-DTT-28 B7-DTT-28A B7-DTT-28B B7-DTT-28C B7-DTT-28D B7-DTT-28E B1-C3-DDTT-28 B2-C3-DDTT-28 B3-C3-DDTT-28 B4-C3-DDTT-28 B5-C3-DDTT-28 A-7 .086 0.926 8121.1995 0.2078 0.785 32.2233 0.080 0.310 8172.966 8164.057 0.869 8393.613 8239.3: Data for DTT and DDTT Cylindrical Specimens Compressive Modulus of Poisson's Degrees out of Stress (ksi) Elasticity (ksi) Ratio Perpendicularity 29.859 9003.276 0.

021 1.642 0.392 1.262 1.4 68.994 3.4 211 253 277 300 95.4: Data for Air-Cured Flexural Specimens Corrected First Crack Deflection at Equivalent Ultimate First Crack Strength First Crack Flexural Load Strength (ksi) (ksi) (in.484 1.757 0.308 4.601 1.00168 0.10 R10.2 78.4 16.482 4.7 40.716 0.5 18.2 77.322 4.00182 0.00146 0.778 0.1 222 285 323 333 109.747 0.00186 0.7 6.3 204 245 267 278 93.069 3.00176 0.00203 4.402 1.3 16.429 1.990 4.213 4.5 232 284 302 325 113.2 46.1 238 302 348 369 100.724 4.5 67.6 18.603 4.157 1.9 6.8 46.225 4.559 0.9 6.5 I20 44.00215 0.6 79.1 19.4 18.643 3.829 4.9 6.5 84.7 16.4 221 288 318 348 108.00185 0.1 6.127 3.445 4.3 41.2 I30 74.4 212 268 300 332 109.1 78.348 1.626 4.8 221 275 310 326 112.Table A.215 Specimen B1-B2-A-28A B1-B2-A-28B B2-B2-A-28A B2-B2-A-28B B3-B2-A-28A B3-B2-A-28B B4-B2-A-28A B4-B2-A-28B B5-B2-A-28A B5-B2-A-28B B6-B2-A-28A B6-B2-A-28B B7-B2-A-28A B7-B2-A-28B B7-B2-A-28C I5 6.291 1.00163 0.40 107.574 1.873 0.7 18.6 202 246 261 262 97.8 6.00203 0.829 3.365 1.878 4.3 17.7 198 247 275 290 Not Applicable Not Applicable Not Applicable A.8 41.4 44.7 I40 R5.8 76.8 7.772 0.391 1.566 0.9 46.20 R20.770 4.0 17.00168 0.2 225 288 327 347 121.6 I10 17.862 4.30 R30.7 45.9 6.5 76.308 4.3 6.862 4.864 1.00190 0.823 0.700 0.9 49.888 0.0 17.8 .3 67.575 4.) Strength (ksi) (kip) 0.637 5.6 6.741 0.772 0.336 1.2 70.00183 0.792 0.9 6.00198 0.2 46.4 42.9 214 273 317 334 114.00198 0.

140 2.6 40.230 4.539 5.00247 0.4 17.8 6.00230 0.4 15.3 91.9 42.783 2.5 16.743 1.0 6.943 1.051 1.) Strength (ksi) (kip) 1.3 98.124 5.873 1.078 1.5 7.622 5.878 1.180 2.997 4.739 1.1 R5.770 5.3 NR 88.967 1.8 16.4 95.889 1.989 1.986 1.00263 0.987 5.038 0.2 66.924 I5 6.1 59.6 42.8 16.00288 5.7 6.00229 0.026 2.649 4.965 0.00253 0.017 4.166 0.8 I40 92.40 188 233 258 277 194 236 259 259 199 242 269 NR 211 250 278 291 200 238 268 286 177 211 238 257 202 253 275 293 204 257 279 288 185 215 220 NR 184 227 246 262 178 210 234 248 194 229 250 272 Not Applicable Not Applicable Not Applicable A-9 .967 0.00240 0.1 42.4 70.7 15.5 6.9 NR 99.042 0.124 4.10 R10.5 59.30 R30.2 62.644 5.7 85.5 6.0 67.00203 0.2 38.Table A.209 1.00226 0.9 15.1 36.045 1.744 1.4 63.896 2.123 1.00258 0.00264 0.5 14.348 5.9 I20 39.7 6.444 4.764 5.101 1.047 1.1 6.00233 0.0 38.5: Data for TT Flexural Specimens Specimen B1-B2-TT-28A B1-B2-TT-28B B2-B2-TT-28A B2-B2-TT-28B B3-B2-TT-28A B3-B2-TT-28B B4-B2-TT-28A B4-B2-TT-28B B5-B2-TT-28A B5-B2-TT-28B B6-B2-TT-28A B6-B2-TT-28B B7-B2-TT-28A B7-B2-TT-28B B7-B2-TT-28C Corrected First Crack Deflection at Equivalent Ultimate First Crack Strength First Crack Flexural Load Strength (ksi) (ksi) (in.201 1.433 4.307 4.6 37.7 69.8 I30 65.2 6.0 16.00284 0.8 84.0 15.9 99.1 40.20 R20.367 5.960 4.4 67.978 6.134 1.9 91.0 6.6 70.873 4.00238 0.619 4.00248 0.197 5.6 6.4 35.2 I10 16.314 5.5 16.187 1.6 59.3 40.

365 4.322 2.00324 NR 5.961 1.2 39.3 51.534 4.284 1.3 14.315 2.255 2.077 4.094 2.7 6.609 4.067 4.1 6.925 5.0 6.6: Data for DTT Flexural Specimens Specimen B1-B2-DTT-28A B1-B2-DTT-28B B2-B2-DTT-28A B2-B2-DTT-28B B3-B2-DTT-28A B3-B2-DTT-28B B4-B2-DTT-28A B4-B2-DTT-28B B5-B2-DTT-28A B5-B2-DTT-28B B6-B2-DTT-28A B6-B2-DTT-28B B7-B2-DTT-28A B7-B2-DTT-28B B7-B2-DTT-28C Corrected First Crack Deflection at Equivalent Ultimate First Crack Strength First Crack Flexural Load Strength (ksi) (ksi) (in.7 NR NR 84.133 1.1 40.3 NR NR NR NR R5.40 192 231 NR NR 192 229 249 267 181 220 239 NR 175 205 217 NR 178 219 236 239 200 235 259 273 174 204 209 205 183 204 215 220 171 187 180 NR 175 199 212 NR 175 197 212 NR 183 213 231 NR Not Applicable Not Applicable Not Applicable A.20 R20.136 1.3 6.646 4.305 1.4 I10 16.184 1.773 5.141 2.832 4.2 16.9 60.7 34.197 2.185 4.250 1.4 34.732 4.6 6.10 R10.1 6.5 36.481 4.00249 0.5 35.2 35.059 2.7 79.00247 0.Table A.3 37.4 56.187 1.00277 0.642 NR 5.2 35.00276 0.331 0.7 66.00299 0.5 6.8 I30 NR 64.136 2.354 2.083 4.161 1.106 2.2 56.8 15.7 14.00256 0.7 15.013 0.00287 0.5 76.5 14.7 59.1 15.219 1.5 6.132 4.7 33.142 1.288 1.311 4.049 2.) Strength (ksi) (kip) 1.00314 0.8 14.227 5.30 R30.00267 0.2 57.7 93.0 37.4 55.475 5.405 5.1 6.015 I5 6.00271 0.293 2.827 1.9 55.00282 0.7 14.00295 0.768 4.043 1.10 .8 15.2 6.575 4.00285 0.2 6.6 I20 39.168 1.00267 0.733 2.1 16.0 61.9 I40 NR 90.

7: RCPT Specimen Data Sorted by Batch Charge Passed (coulombs) 12 93 16 9 78 13 57 11 73 19 10 Chloride Ion Penetrability Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Specimen ID 1A-RCP-TT-7 1A-RCP-A-28 1A-RCP-TT-28 1B-RCP-TT-7 1B-RCP-A-28 1B-RCP-TT-28 1C-RCP-A-28 1C-RCP-TT-28 D-RCP-A-28 D-RCP-TT-28 S-RCP-TT-7 Age 7 28 28 7 28 28 28 28 28 28 8 A-11 .Table A.

8: RCPT Specimen Data Sorted by Curing Regime Specimen age at time of testing (days) 7 7 8 Average St. COV (%) 28 28 28 28 Average St.12 . COV (%) Charge Passed (coulombs) 12 9 10 10 1. COV (%) 28 28 28 28 Average St.5 24 93 78 57 73 75 15 20 Chloride Ion Penetrability Negligible Negligible Negligible Specimen ID 1A-RCP-TT-7 1B-RCP-TT-7 S-RCP-TT-7 1A-RCP-TT-28 1B-RCP-TT-28 1C-RCP-TT-28 D-RCP-TT-28 Negligible Negligible Negligible Negligible 1A-RCP-A-28 1B-RCP-A-28 1C-RCP-A-28 D-RCP-TT-28 Negligible Negligible Negligible Negligible A.5 15 16 13 11 19 15 3. Dev.Table A. Dev. Dev.

9: Sample Freeze-Thaw Cycle Temperature – Position A17 A-13 .Figure A.

Figure A.14 .10: Average Mass Change of UHPC Freeze-Thaw and Side Study Specimens A.

Figure A.11: Average Length Change of UHPC Specimens Undergoing Freeze-Thaw Cycling A-15 .

1056 -0.8E-06 13.9E-06 13.5E-06 14.1137 -0.1093 CTE1 13.49 177.25E-06 7.1061 0.0E-06 14.7E-06 13.5E-06 13.5E-06 13.73E-06 7.0E-06 13.Table A.1072 0.1142 ΔLm2 0.6E-06 13.89E-06 7.53E-06 7.1E-06 13.9E-06 14.7E-06 13.1077 -0.44E-06 7.1061 0.1077 0.1040 -0.8E-06 14.1023 0.16 .6E-06 15.7E-06 14.25 177.4E-06 13.0E-06 13.3E-06 14.1023 -0.1067 -0.80 0.3E-06 CTEavg (oF) 7.0E-06 14.9E-06 14.07 177.80 ΔLm1 0.4E-06 13.31E-06 x x A.1083 -0.7E-06 13.6E-06 14.76 177.1018 0.1067 -0.6E-06 13.73E-06 7.12 177.3E-06 CTE2 13.59E-06 7.1088 0.44E-06 7.4E-06 13.1050 0.4E-06 CTEavg (oC) 13.1056 -0.8E-06 13.61E-06 7.1077 -0.09 177.7E-06 14.8E-06 13.7E-06 14.1066 0.20 177.8E-06 13.1040 -0.0E-06 7.8E-06 13.1066 0.73E-06 8.69E-06 7.0E-06 13.1050 -0.1050 0.73 177.8E-06 14.64 177.75 178.12: Michigan Tech CTE Summary Report on UHPC Date 5/22/2007 5/27/2007 6/3/2007 6/15/2007 6/16/2007 Core 1A-CTE-A-3 1A-CTE-A-7 1A-CTE-A-14 1A-CTE-A-28 1A-CTE-TT-28 Frame S/N 1135 1135 1135 1135 1135 Lo 175.1126 13.1023 0.05 176.0E-06 13.59E-06 7.6E-06 14.1067 0.70 177.99 177.9E-06 14.1E-06 13.1029 0.0E-06 13.1120 0.1050 -0.03 178.1040 0.1061 0.56E-06 7.53E-06 8.8E-06 15.5E-06 14.8E-06 13.82 176.4E-06 13.2E-06 13.1067 0.6E-06 15.7E-06 13.5E-06 13.97E-06 Not tested on correct day x 5/26/2007 5/30/2007 6/5/2007 6/19/2007 5/29/2007 6/1/2007 6/8/2007 6/22/2007 6/24/2007 6/6/2007 7/3/2007 7/23/2007 1B-CTE-A-3 1B-CTE-A-7 1B-CTE-A-14 1B-CTE-A-28 1C-CTE-A-3 1C-CTE-A-7 1C-CTE-A-14 1C-CTE-A-28 1C-CTE-TT-28 1D-CTE-A-7 M-CTE-A-7 M-CTE-TT-28 1135 1135 1135 1135 1135 1135 1135 1135 1135 1135 1135 1135 178.2E-06 13.79E-06 7.

1110 0.9E-06 CTE2 13.9E-06 14.19E-06 8.1110 -0.69E-06 8.1131 CTEavg (oC) 13.26E-06 Not tested on correct day Date 7/6/2007 7/26/2007 7/11/2007 7/14/2007 7/17/2007 Core N-CTE-A-7 N-CTE-TT-28 P-CTE-TT-7 R-CTE-TT-7 S-CTE-TT-7 Lo 176.1131 ΔLm2 0.36 177.7E-06 14.12 (continued): Michigan Tech CTE Summary Report on UHPC Frame S/N 1135 1135 1135 1135 1135 ΔLm1 -0.8E-06 14.1083 0.58 CTE1 13.1126 -0.8E-06 14.21E-06 8.1110 0.7E-06 14.9E-06 CTEavg (oF) 7.6E-06 14.09 178.14E-06 8.1072 -0.Table A.1115 -0.41 177.7E-06 14.8E-06 14.7E-06 14.9E-06 14.1137 0.85 177.9E-06 A-17 .7E-06 14.7E-06 14.

05 0.03 1C-CTE-A-14 14 0.03 Mean 0.00 0.03 N/A N/A 0.02 84.00 1A-CTE-A-7 7 0.02 COV (%) 66.18 .44 Thermally-treated UHPC Specimens Age of specimen at testing (days) 28 28 28 28 7 7 7 Mass Change (%) 0.05 1A-CTE-A-28 28 0.Table A.03 0.00 1B-CTE-A-28 28 0.00 1A-CTE-A-14 14 0.94 Specimen 1A-CTE-TT-28 1C-CTE-TT-28 M-CTE-TT-28 N-CTE-TT-28 P-CTE-TT-7 R-CTE-TT-7 S-CTE-TT-7 Mean Standard Dev. COV (%) A.03 D-CTE-A-7 7 0. 0.03 N-CTE-A-7 7 0.03 1C-CTE-A-28 28 0.13: Mass change study on epoxy coated UHPC CTE specimens Air-cured UHPC Specimens Age of specimen at Mass Change Specimen testing (days) (%) 1A-CTE-A-3 3 0.01 0.05 1B-CTE-A-14 14 0.03 1B-CTE-A-3 3 N/A 1B-CTE-A-7 7 0.03 M-CTE-A-7 7 0.03 Standard Dev.03 1C-CTE-A-3 3 N/A 1C-CTE-A-7 7 0.02 0.03 D-CTE-A-28 28 0.

APPENDIX B – CTE Test Procedure Modifications B-1 .

B-2 .

B-3 .

(This page intentionally left blank) B-4 .

Sign up to vote on this title
UsefulNot useful