Fresh Properties, Temperature Rise

and
Strength Development of High Strength Concrete
with
Binary and Ternary Blended Cements

Vassiem Sheikh
B.£ng (Hons.) MSc.

A Thesis submitted for the degree o f Doctor o f Philosophy

University o f London

D epartm ent o f Civil & Environmental Engineering

University College London

University o f London

2001
 ProQuest Number: U643746

All rights reserved

INFORMATION TO ALL USERS

The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.

uest.
ProQuest U643746

Published by ProQuest LLC(2016). Copyright of the Dissertation is held by the Author.

All rights reserved.

This work is protected against unauthorized copying under Title 17, United States Code.
Microform Edition © ProQuest LLC.

ProQuest LLC
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 48106-1346
 ABSTRACT

The use of high strength concrete in the construction industry has become more frequent
as both the knowledge of the behaviour of the material and the confidence in its
production have increased. An appropriate formulation of materials and mix proportions
can result in significantly enhanced performance such as high early strength, reduced heat
of hydration and increased durability. As a step towards obtaining optimum performance,
an investigation has been carried out on the fresh properties (workability), temperature
rise during hydration and strength development. This research was aimed at
understanding the role of supplementary cementing materials in binary (OPC+PFA,
GGBS, CSF) and ternary (OPC+ CSF/PFA, CSF/GGBS) combinations in these three
areas.

With respect to workability the use of binary mixes of PFA or CSF reduce the
superplasticiser dosage required to obtain a target slump, whereas GGBS increases it.
Optimum replacement levels of 10% CSF, 40%PFA+10%CSF and 60%GGBS+10% CSF
were found at a water/binder ratio of 0.26.

Binary mixes of 40% PFA or 60% GGBS reduce the peak semi-adiabatic temperature
rise compared to their equivalent OPC mix at 0.26 water/binder ratio. Ternary
combinations of 10% CSF with PFA or GGBS have shown significant reductions in peak
temperature rise compared to their equivalent binary mixes.

Measurement of the in-situ strength by temperature matched curing (TMC) has shown
higher early age strengths but lower long term strengths for both binary and ternary mixes
compared to cubes cured under standard conditions (20°C). Microstructural evaluation of
hardened cement paste indicates that these differences in strength are likely to be
associated with stresses generated at the paste/aggregate interface.

A novel non-destructive technique to assess the in-situ strength has shown good
correlation between conductivity and strength development of high strength concrete.
This thesis is dedicated in its entirety to my late parents

Ahmed Hassan Sheikh

&
Rashida Begum Sheikh
 ACKNOWLEDGEMENTS

I would like to thank my supervisor Dr. P.L.J Domone for his invaluable supervision,
advice and guidance throughout this research. The discussions we had, the remarks he
made and his patience made it possible to complete this work. I would also like to
express my thanks to all the technical staff at UCL for all their assistance during the
experimental work, in particular Owen Bourne, Les and Malcolm Saytch.

A special thanks to Rolf Laffan of Hydronix ltd, who made it possible for me to use the
dielectric strength sensor and include it in this research.

I would also like to express my gratitude to El-Mahadi Ahmed, for sharing his thoughts
and ideas on high strength concrete.

Finally, my thanks go to the Engineering and Physical Sciences Research Council for
funding the work presented in this thesis.

Ill
 CONTENTS

Abstract i
Dedication ii
Acknowledgements iii
Contents iv
List of Figures x
List of Tables xvii

Chapter 1 Introduction 1

1.1 Background 1
1.2 Definition of high strength concrete 2
1.3 Production of high strength concrete 3
1.4 Research programme and thesis structure 5

Chapter 2 Applications o f high strength concrete, advantages and

Disadvantages 6

2.1 Introduction 6
2.2 Structural Applications 6
2.2.1 High rise buildings 6
2.2.2 Bridges 9
2.2.3 Offshore platforms 11
2.2.4 Other areas of application 12
2.3 Disadvantages of High Strength Concrete 13
2.4 Conclusions 16

IV
CHAPTER 3 Literature review 17

3.1 Introduction - 17
3.2 Supplementary Cementing Materials 17
3.2.1 Pulverised Fuel Ash (PFA) 17
3.2.2 Ground granulated blast-furnace slag ( GGBS ) 19
3.2.3 Condensed Silica Fume ( CSF ) 19

3.3 Fresh properties of HSC 21

3.3.1 Definition of workability 21
3.3.2 Measurement of workability. 23
3.3.3 The effects of binary blends. 30
3.3.3.1 Pulverised Fuel Ash (PFA) 30
3.3.3.2 Granulated Glass Blastfumance Slag (GGBS) 36
3.3.3.3 Condensed Silica Fume (CSF) 39
3.3.4 The effects of ternary blends. 44
3.4 Temperature rise of high strength concrete 47
3.4.1 Heat evolution during hydration 47
3.4.2 Review of measuring methods. 48
3.4.3 Factors affecting the temperature rise of concrete. 50
3.4.4 The effect of binary blended cements. 55
3.4.4.1 Pulverised Fuel Ash (PFA) 55
3.4.4.2 GGBS 57
3.4.4.3 Condensed Silica Fume (CSF) 61
3.4.5 The effect of ternary blended cements. 63
3.5 Strength Development of High Strength Concrete 65
3.5.1 The effect of standard curing conditions on strength development 66
3.5.1.1 Binary blends 66
3.5.1.2 Ternary blends 68
3.5.2 Temperature development in high strength concrete. 71
3.5.3 The effect of in-situ temperatures on the strength development. 71
3.5.3.1 Temperature Matched Curing 71
 3.5.4 Proposed explanations for the strength development of temperature 87
matched cured concrete
3.6 Conclusions " 89

CHAPTER 4 Aims and Scope of investigation 90

4.1 Introduction. 90
4.2 Aims 91
4.2.1 Fresh properties of high strength concrete 91
4.2.2 Temperature Rise 92
4.2.3 Strength development 92
4.2.4 Dielectric properties 93
4.3 Scope of Investigation 93
4.3.1 Mix variables 93
4.3.2 Fresh Properties 94
4.3.3 Temperature Rise 95
4.3.4 Strength development 95
4.3.5 Dielectric Properties 95
4.3.6 Mix proportions used 97

CHAPTER 5 Materials and experimental procedures 98

5.1 Introduction 98
5.2 Materials 98
5.2.1 Ordinary Portland cement (OPC) 98
5.2.2 Supplementary cementing materials (SCMs) 98
5.2.3 Admixtures 100
5.2.4 Water 100
5.2.5 Fine and coarse aggregate 100
5.3 Mixing Procedures 102
5.3.1 Cement paste 102
5.3.2 Concrete 102

VI
5.4 Fresh Concrete Tests 103
5.4.1 Slump test 103
5.4.2 Two-point test ^ 103
5.5 Semi adiabatic temperature rise measurements 109
5.6 Tests on hardened concrete 109
5.6.1 Compressive strength measurements 109
5.6.2 Density measurements 110
5.7 Tests on cement paste 110
5.7.1 Porosity measurements 110
5.7.2 Compressive strength measurements 111

CHAPTER 6 Fresh properties o f high strength concrete 114

6.1 Introduction 114

6.2 Superplasticiser dosage required to produce a given slump 115
6.2.1 Binary blends - 0.26 w/b ratio 116
6.2.2 Ternary blends - 0.26 w/b ratio 117
6.2.3 Other water/binder ratios 119
6.3 Two point workability tests 122
6.3.1 Binary blends 122
6.3.2 Ternary blends 125
6.3.3 Other water/binder ratios 130
6.4 Comparison of slump and two-point test measurements 135
6.5 Conclusions 138

CHAPTER 7 Semi-adiabatic temperature rise of high strength

concrete 139

7.1 Introduction 139

7.2 Binary blends 140
7.3 Ternary blends 144
7.4 The effect of level of cement replacement on the peak temperature
at 0.26 water/binder ratio 146

Vll
7.5 Other water/binder ratios 146
7.6 Conclusions _ 152

CHAPTER 8 Strength development of high strength concrete 153

8.1 Introduction 153

8.2 Results and discussion 154
8.2.1 Binary blends -0.26w/b ratio 154
8.2.2 Temary blends - 0.26 w/b ratio 158
8.2.3 Other water/binder ratios 163
8.2.4 Summary of strength results 169
8.3 Practical Significance 169
8.4 Microstructural changes 175
8.4.1 Porosity of hardened cement paste 176
8.4.2 Calcium Hydroxide Content 177
8.4.3 Strength development of cement paste. 180
8.5 Conclusions 183

CHAPTER 9 Conclusions and recommendations for further work 184

9.1 Introduction 184

9.2 Conclusions 184
(i) Fresh properties of high strength concrete. 184
(ii) Estimated adiabatic temperature rise of high strength concrete. 185
(iii) Strength development of high strength concrete. 186
(iv) Micro-structural properties/characteristics. 187

9.3 Recommendations for further work. 188

References 190-208

VllI
Appendices
Appendix A Calibration of two-point test apparatus and workability results
Appendix B Temperature cycles and compressive strength results
Appendix C Dielectric Properties of High Strength Concrete

IX
 List of Figures

Chapter 3
Figure 3.1 Ritchie’s subdivisions 22
Figure 3.2: Two point test apparatus 26
Figure 3.3: BML Viscometer 29
Figure 3.4: (a) BT Rheometer (b) Testing container 29
Figure 3.5: Effect of partial cement replacement by PFA on the superplasticiser
dosage required for concrete mixes with a slump of 150 mm. 32
Figure 3.6: Relationship between yield (g) value and % replacement by PFA. 33
Figure 3.7: Relationship between plastic viscosity (h) and replacement by PFA. 33
Figure 3.8: The effect of % replacement by PFA on the yield (g) and plastic
viscosity (h) value for concrete mixes with water/binder of 0.26 and a slump of
approximately 150 mm. 35
Figure 3.9: The effect of GGBS on the superplasticiser dosage required for
150 mm constant slump. 37
Figure 3.10: The effect of % replacement by GGBS on the yield (g) and plastic
viscosity (h) value for concrete mixes with water/binder of 0.26 and a slump of
approximately 150 mm. 38
Figure 3.11: Variation of superplasticiser dosage at constant slump with various
levels of CSF replacement. 39
Figure 3.12: Relationship between water/binder ratio and superplasticiser
dosage for CSF mixes. 42
Figure 3.13: Effect of CSF on the rheological properties of fresh concrete 42
Figure 3.14: The effect of % replacement by CSF on the yield (g) and plastic
viscosity (h) value for concrete mixes with water/binder of 0.26 and a slump of
approximatelyl 50mm. 43
Figure 3.15: Relationship between water/binder ratio and superplasticiser
dosage for ternary GGBS/CSF mix. 45
Figure 3.16: Relationship between water/binder ratio and g and h values for
CSF and GGBS/CSF mixes. 46
Figure 3.17: Temperature rise in 1:2:4 concrete (w/c 0.6) cured adiabatically. 49
Figure 3.18: Temperature rise for mass concrete using different cements. 49
Figure 3.19: Relationship between heat evolution and time. " 53
Figure 3.20: The effect of casting temperature on the peak temperature rise of
OPC mixes at 0.26 w/b ratio. 54
Figure 3.21 : Temperature rise for PFA binary mixes. 56
Figure 3.22: The adiabatic temperature rise for PFA binary mixes. 56
Figure 3.23: Temperature rise for GGBS binary mixes. 58
Figure 3.24: The adiabatic temperature rise for GGBS binary mixes. - 58
Figure 3.25: Temperature rise of GGBS mixes under semi-adiabatic conditions. 60
Figure 3.26: BRE Cardington mass pour temperature readings. 60
Figure 3.27: The adiabatic temperature rise for GGBS binary and GGBS/CSF
ternary mixes. 62
Figure 3.28: The adiabatic temperature rise for a 10% CSF binary mix. 62
Figure 3.29: Adiabatic temperature rise for ternary mixes at 0.26 w/b ratio. 64
Figure 330: Strength development of standard cured CSF concretes with a
water/binder ratio of 0.26. 67
Figure 3.31: Compressive strength of CSF concretes with water/binder ratio
of 0.34. 67
Figure 3.32: Compressive strength of CSF concretes with water/binder ratio
of 0.28. 67
Figure 3.33: Strength development of standard cured PFA concretes with a
water/binder ratio of 0.26. 69
Figure 3.34: Strength development of standard cured GGBS concretes with a
water/binder ratio of 0.26. 69
Figure 3.35: Strength development of standard cured PFA/CSF concretes
with a water/binder ratio of 0.26. 70
Figure 3.36: Strength development of standard cured GGBS/CSF concretes
with a water/binder ratio of 0.26. 70
Figure 3.37: Temperature matched curing device. 73
Figure 3.38: Concrete wall results. 74
Figure 3.39 Temperature profiles. 77
Figure 3.40 Compressive strength results from Connell. 77

XI
Figure 3.41 Relationship between standard cured and TMC strengths for concretes
with 50% GGBS. 79
Figure 3.42 Relationship between standard cured and TMC strengths for concretes
with 30% PFA. 79
Figure 3.43: Compressive strength development of Grade 25 and Grade 40
concrete subjected to standard curing. 80
Figure 3.44: Compressive strength development of Grade 25 and Grade 40
concrete subj ected to temperature matched curing. 81
Figure 3.45 Comparison of strength results obtained by Bamforth and -
Wainwright. 83
Figure 3.46 Compressive strength results obtained by Price. 83
Figure 3.47: strength development of a 50% GGBS mix at 0.3 w/b ratio. 84
Figure 3.48: Strength development of high strength concrete mixes. 86

Chapter 5

Figure 5.1 : Two point test apparatus and Windograf recorder 105
Figure 5.2: (a) typical chart output from two-point test (b)idling test 106
Figure 5.3: Typical output from data analysis program. 108
Figure 5.4: Insulated plywood box. 112
Figure 5.5: Contest GDIOA compression testing machine. 113

Chapter 6

Figure 6.1: The effect of PFA,GGBS and CSF on the superplasticiser

dosage required to produce a 200 +/- 20 mm slump at 0.26w/b ratio. 117
Figure 6.2: The effect of partial cement replacement by 10% CSF + PFA
on the superplasticiser dosage required to produce a 200 +/- 20 mm slump. 118

XII
Figure 6.4: The effect of partial cement replacement by 10% CSF on the
superplasticiser dosage required to produce a 200 +/- 20 mm slump. 120
Figure 6.5: The effect of partial cement replacement by 40% PFA + 10% CSF "
on the superplasticiser dosage required to produce a 200 +/- 20 mm slump. 120
Figure 6.6: The effect of partial cement replacement by 60% GGBS + 10% CSF
on the superplasticiser dosage required to produce a 200 +/- 20 mm slump. 121
Figure 6.7: The effect of CSF on the yield (g) and plastic viscosity values
at 0.26 w/b ratio. 123
Figure 6.8: The effect of PFA,CSF on the yield (g) and plastic viscosity values .
at 0.26 w/b ratio. 124
Figure 6.9: The effect of GGBS on the yield (g) and plastic viscosity values
at 0.26 w/b ratio. 126
Figure 6.10: The effect of PFA(+10%CSF) on the yield (g) and plastic viscosity
values at 0.26 w/b ratio. 127
Figure 6.11: The effect of GGBS(+10%CSF) on the yield (g) and plastic viscosity
values at 0.26 w/b ratio. 129
Figure 6.12: The effect of 10%CSF on the yield (g) and plastic viscosity
values at 0.3-0.2 w/b ratios. 131
Figure 6.13: The effect of 40% PFA + 10%CSF on the yield (g) and
plastic viscosity values at 0.3-0.2 w/b ratios. 132
Figure 6.14: The effect of 60% GGBS + 10%CSF on the yield (g) and
plastic viscosity values at 0.3-0.2 w/b ratios. 134
Figure 6.15 : Relationship between yield value ( g ) and slump 136
Figure 6.16 : Relationship between plastic viscosity ( h ) and slump 136
Figure 6.17 : Relationship between yield value and plastic viscosity for both
NSC and HSC. 137

X lll
Chapter 7

Figure 7.1: Estimated adiabatic temperature rise of OPC and 10%CSFmixes

at 0.26 w/b ratio. 141
Figure 7.2: Estimated adiabatic temperature rise of PFA binary mixes
at 0.26 w/b ratio. 141
Figure 7.3: Estimated adiabatic temperature rise of GGBS binary mixes
at 0.26 w/b ratio. 143
Figure 7.4: Estimated adiabatic temperature rise of PFA ternary mixes
at 0.26 w/b ratio. 145
Figure 7.5: Estimated adiabatic temperature rise of GGBS ternary mixes
at 0.26 w/b ratio. 145
Figure 7.6: The effect of binary PFA and GGBS replacement levels
on the peak temperature at 0.26 w/b ratio. 147
Figure 7.7: The effect of tenary PFA and GGBS replacement levels
on the peak temperature at 0.26 w/b ratio. 147
Figure 7.8: Estimated adiabatic temperature rise of OPC and 10%CSF
mixes at 0.30 w/b ratio. 149
Figure 7.9: Estimated adiabatic temperature rise of 10% CSF and
PFA/CSF, GGBS/CSF ternary mixes at 0.30 w/b ratio. 149
Figure 7.10: Estimated adiabatic temperature rise of 10% CSFand
PFA/CSF, GGBS/CSF ternary mixes at 0.20 w/b ratio. 150
Figure 7.11: The effect of water/binder ratio on the peak temperature
rise. 150
Figure 7.12: The effect of binder replacement on the peak temperature
rise of binary mixes at 0.26 w/b ratio 151
Chapter 8
Figure 8.1: Strength development of OPC mix at 0.26 w/b ratio subjected to
standard and temperature matched curing. 155
Figure 8.2: Strength development of 10% CSF mix at 0.26 w/b ratio
subjected to standard and temperature matched curing. 155

XIV
Figure 8.3: Strength development of 20% PFA mix at 0.26 w/b ratio
subjected to standard and temperature matched curing. 157

Figure 8.4: Strength development of 40% PFA mix at 0.26 w/b ratio
subjected to standard and temperature matched curing. 157
Figure 8.5: Strength development of 30% GGBS mix at 0.26 w/b ratio
subjected to standard and temperature matched curing. 159
Figure 8.6: Strength development of 60% GGBS mix at 0.26 w/b ratio
subjected to standard and temperature matched curing. 159
Figure 8.7: Strength development of 40% PFA + 10% CSF mix
at 0.26 w/b ratio subjected to standard and temperature matched curing. 161
Figure 8.8: Strength development of 50% PFA + 10% CSF mix
at 0.26 w/b ratio subjected to standard and temperature matched curing. 161
Figure 8.9: Strength development of 60% GGBS + 10% CSF mix
at 0.26 w/b ratio subjected to standard and temperature matched curing. 162
Figure 8.10: Strength development of 70% GGBS + 10% CSF mix
at 0.26 w/b ratio subjected to standard and temperature matched curing. 162
Figure 8.11: Strength development of 80% GGBS + 10% CSF mix
at 0.26 w/b ratio subjected to standard and temperature matched curing. 164
Figure 8.12: Strength development of 10% CSF mix at 0.30 w/b ratio
subjected to standard and temperature matched curing. 164
Figure 8.13: Strength development of 40% PFA + 10% CSF mix
at 0.30 w/b ratio subjected to standard and temperature matched curing. 166
Figure 8.14: Strength development of 60% GGBS + 10% CSF mix
at 0.30 w/b ratio subjected to standard and temperature matched curing. 166
Figure 8.15: Strength development of 10% CSF mix at 0.20 w/b ratio
subjected to standard and temperature matched curing. 167
Figure 8.16: Strength development of 40% PFA + 10% CSF mix
at 0.20 w/b ratio subjected to standard and temperature matched curing. 168
Figure 8.17: Strength development of 60% GGBS + 10% CSF mix
at 0.20 w/b ratio subjected to standard and temperature matched curing. 168

XV
Figure 8.18 Summary of standard cured results in (a) binary, and
(b) ternary mixes (0.26 w/b ratio). 170
Figure 8.19 Summary of TMC cured results in (a) binary, and
(b) ternary mixes (0.26 w/b ratio) 171
Figure 8.20 : Comparison of cross-over points for binary mixes at 0.26 w/b ratio. 173
Figure 8.21 : Comparison of cross-over points for ternary mixes at 0.26 w/b ratio. 173
Figure 8.22 : Ultimate strength difference - binary mixes at 0.26 w/b ratio. 174
Figure 8.23 : Ultimate strength difference - ternary mixes at 0.26 w/b ratio. 174
Figure 8.24: Porosities of OPC and 10% CSF mixes at 0.26 w/b ratio. 176
Figure 8.25: Porosity of 40%PFA+10% CSF mix at 0.26 w/b ratio. 178
Figure 8.26: Porosity of 60%GGBS+10% CSF mix at 0.26 w/b ratio. 178
Figure 8.27: Calcium hydroxide contents of OPC and 10% CSF mixes
at 0.26 w/b ratio. 179
Figure 8.28: Calcium hydroxide contents of 40%PFA+10%CSF and
60%GGBS+ 10% CSF mixes at 0.26 w/b ratio. 179
Figure 8.29: Strength development of OPC paste specimen at 0.26 w/b ratio. 181
Figure 8.30: Strength development of CSF paste specimen at 0.26 w/b ratio. 181

XVI
 List of Tables
Chapter 2
Table2.1 Mix Proportions for River Plaza, Scotia Plaza and Pacific First Center. 8
Table 2.2: General Requirements for concrete on the Tsing Ma Bridge 10
Table 2.3: Summary of specification requirements for the Storeabaelt West Bridge 10

Chapter 3
Table 3.1 Chemical composition of typical Class F and Class C PFA 18
Table 3.2 Typical oxide compositions and physical properties of OPC, PFA,
GGBS and CSF. 20
Table 3.3: Rheological test results 34
Table 3.4: Mix proportions and workability results for the four mixes tested 34

Chapter 4
Table 4.1 : Summary of aims and scope of research. 96
Table 4.2: Basic mix proportions ( in kg/m^) used in the investigation. 97

Chapter 5
Table 5.1: Portland cement compositions. 99
Table 5.2: Compositions and physical properties of SCMs. 99
Table 5.3 : Properties of aggregates. 101

XVI1
C hapter 1_____________________________________________________________________________________________________________Introduction

CHAPTER 1

INTRODUCTION

1.1 Background

The acceptance of high strength concrete, and its use as a construction material is
growing. This has been due to an increased understanding of the requirements for
producing high strength concretes and to the substantial amount of research that has been
carried out on its engineering properties.

In the 1950s, concrete with a compressive strength of 35 N/mm^ was considered high
strength. In the 1960s, concretes with 40 and 50 N/mm^ compressive strengths were used
and by the early 1970s, 60N/mm^ concrete was used commercially In the late 1980s,
several countries were involved with major research programmes on high strength
concrete The increasing use of condensed silica fume (CSF) and superplasticisers
has today made it relatively easy to produce concrete with conpressive strengths of over
120 N/mm^

The advantages and uses of high strength concrete in civil engineering is documented in
several reports from the Japan Scandinavia Russia and the
In addition major publications documenting current research on the applications of high
strength concretes have been published in recent years

Whilst the state of knowledge on engineering properties of high strength concretes under
standard laboratory environments is increasing, data on in-situ performance is somewhat
limited. Regardless of the amount of laboratory data that has been obtained, in-situ
performance is ultimately the single most relevant criterion used to judge the
performance of high strength concrete.
Chapter 1_____________________________________________________________________________________________________________ Introduction

1.2 Definition Of High Strength Concrete

For the purposes of this study, the concrete being considered is that produced with readily
available materials and by conventional mixing, placing and curing techniques. Thus
more specialist products, such as reactive powder concrete, which can achieve ultra-high
strengths, are not considered.
Concrete is generally classified by strength grades, despite the fact that some properties,
such as durability, are not necessarily related to the compressive strength. The use of
superplasticisers in high strength concrete allows high fluidity to be achieved with lower
water/binder ratios and hence provides other improved characteristics, such as higher
elastic modulus, higher flexural strength, lower permeability, improved abrasion
resistance and enhanced durability. Hence the alternative term High Performance
Concrete is increasingly used in the concrete industry. However, this expression is not
universally accepted; for example, the name of ACI committee 363 still remains as High
Strength Concrete and not High Performance Concrete.

The FIP/CEB state-of-the-art report defines HSC as concretes having compressive

strengths between 60 to 130 N/mm^. The 130 N/mm^ ceiling was cited as the practical
upper limit for concretes produced with ordinary aggregates. Nagataki who defined
HSC as 50 to 120 N/mm^ shares a similar view, with the 120 N/mm^ ceiling considered
the practical upper limit. Saucier however, stated that the eventual ceiling on concrete
strength is virtually unlimited. He reported that very high compressive strengths could
only be achieved by changing production methods. In his report he stated that in the
United States concrete with strengths between 34.5 and 70 N/mm^ can be produced
readily by using conventional production techniques. He claimed that while it was
possible to produce concrete with a compressive strength of up to 103 N/mm^ by utilising
more expensive materials and improved production techniques, strengths greater than 103
N/mm^ would require ‘exotic’ materials and procedures. Today it is possible to produce
workable concretes with compressive strengths in the 100 to 150 N/mm^ range, by using
silica fume with readily available materials
Chapter 1_____________________________________________________________________________________________________________ Introduction

The definition of high strength concrete is however relative to current practice and
therefore varies with time and place; in the UK it is currently defined as concrete with
compressive strength in excess of 80 N/mm^ Further complications in defining high
strength concrete arise from the use of cylinder strengths in some cases and cube
strengths in others. Also whether specimens are tested at 28, 56, or 90 days, can make a
significant difference to the compressive strength.

1.3 Production of High Strength Concrete

The earliest attempts to produce high strength concrete were very different from the
techniques currently used. For example, in 1930 Yoshida^^^^ produced concrete with a
compressive strength of 104 N/mm^ by moulding a concrete with a water/binder ratio of
0.31 under a pressure of 1ON/mm^. Current methods of production are similar to those of
normal strength concrete, the main difference being in the constituent materials that need
to be of selected high quality. There is a gradual change in the mix proportions and the
resulting properties of concretes as strength increases. The large number of factors that
are involved in selecting appropriate materials for producing high strength concretes have
been summarised by various authors^^^’*^’*^\ These have generally been in relation to the
use of locally available materials that were unique to each case. Regardless of the
variations in materials, cost effective production of high strength concrete in construction
is achieved by selecting, controlling and combining cement, aggregates, supplementary
cementing materials, admixtures and water, using conventional mixing and placing
techniques.
Freedman stated that in order to achieve high strength concretes, it is vital for the
concrete producer to optimise the cement, aggregate quality, aggregate-paste interaction,
mixing and curing procedures. Mix proportions of high strength concrete have often been
selected empirically by following extensive laboratory testing such as the BRE and
ACI methods of proportioning normal concrete mixtures. Various researchers
( 17,23,24,25) produced guidelines on mix design.
High strength concretes are characterised by having a low water/cement or water/binder
ratio and carefully selected aggregates. In this thesis, the expression water/binder ratio is
used in the characterisation of the mixes. This is the ratio (by weight) of free water to
C hapter 1_____________________________________________________________________________________________________________ Introduction

total cementitious material i.e. Portland cement plus ggbs, pfa, and/or csf. The use of
crushed rocks such as granite and limestone with maximum size below 10mm in the U.K
and having good shape, surface texture and strength are normally preferred.
The fine aggregate should generally be free from clay or silt, and the use of relatively
coarse material (with a fineness modulus of around 3.0) is often recommended to reduce
stickiness of the fresh concrete.
The requirement of a low water/cement ratio has led to the use of high cement contents
(up to 550 kg/m^), but increasing the cement content will not in itself guarantee high
strength. Instead, it can lead to sticky/cohesive mixes^^^ leading to poor workability, and
many researchers^^^’^"*^ have criticised the slump test for assessing the workability of high
strength concrete and recommend the use of a two-point test which is able to measure the
two constants of yield stress and plastic viscosity of the Bingham model of behaviour.

The use of condensed silica fume (CSF) is used to improve the strength of the transition
zone between the aggregate surface and the hardened cement paste and its use is
considered essential in the production of high strength concrete.
Other supplementary cementing materials like pulverised fuel ash (PFA) and ground
granulated blast furnace slag (GGBS), are generally cheaper than ordinary Portland
cement (OPC) and in combination with possible improvements in workability, strength
development and moderation of the heat evolution of the cement-rich mixes tend to
encourage their use However, despite much recent research, their role in high strength
concrete is still relative unexplored especially when used in combination with CSF.
Recently metakaolin, which is manufactured by calcining kaolinite, has also gained
acceptance as a supplementary cementing material in high strength concrete^^^^l It is a
highly reactive and effective pozzolan that increases the early strength and also enhances
the durability of concrete^^^^lResearch into this material was considered beyond the
scope of this research.
Other methods used to produce high strength concrete have also been studied, such as
Fibre reinfbrcement^^^\ modification by polymers^^^\ compaction by pressure^^°\
autoclave curing^^^\ and mix proportioning using active or artificial aggregates^^^l
However, these are not widely used and are outside the scope of the present research.
C hapter 1_____________________________________________________________________________________________________________ Introduction

1.4 Research Programme and Thesis Structure

Many studies have been performed on the influence of a single supplementary cementing
material (binary blends), such as PFA, GGBS and particularly CSF on the properties of
high strength concrete. However, relatively few studies have dealt with the use of ternary
blends of supplementary cementing materials and their effects on the properties of high
strength concrete.

The research presented in this thesis aims to investigate the role of binary and ternary
cement blends on the properties of high strength concrete.

The main areas investigated are:

1. Assessing the fresh properties according to the Bingham model, with tests carried out
with TattersalFs two-point workability test.

2. Assessing the effectiveness of blended cements in reducing the semi-adiabatic

temperature rise of high strength concrete.

3. Measuring the effects of water/binder ratio and temperature matched curing on the in-
situ strength development.

4. Exploring a new technique based on measurement of the dielectric properties to assess

the in-situ strength.

In the thesis, chapter 2 illustrates some applications, advantages and disadvantages of

high strength concrete. The main literature review is in chapter 3, which then leads on to
chapter 4 where the aims and scope of the present research are given. Chapter 5 presents
the materials and methods used in the experimental programme. Chapters 6 to 8 present
and discuss the results in areas 1,2 and 3 listed above. The dielectric properties of high
strength concrete are discussed in Appendix C.
The thesis ends with conclusions and recommendations for further work.
Chapter 2 A pplications, A dvantages <fe D isadvan tages o f H S C

CHAPTER 2

APPLICATIONS OF HIGH STRENGTH CONCRETE,

ADVANTAGES AND DISADVANTAGES

2.1 Introduction

Some selected important applications of high strength concrete showing its technical and
economical advantages are presented in this chapter. Many of the applications presented
have concrete strengths less than 80 N/mm^; these are included to show how the strengths
have increased with time. Some of the disadvantages are discussed later in the chapter.

2.2 Structural Applications

The main structural uses of high strength concrete have been in:
• High rise buildings
• Bridges
• Offshore platforms
High strength concrete has also been used in other areas, including highway pavements,
piles and nuclear power stations.

2.2.1 High rise buildings

Concretes with progressively higher strengths, currently up to 130 N/mnf have been
used in high rise buildings, particularly in the USA. In 1972, the first 62 N/mm^ concrete
was used in the 50-storey Mid-Continental Plaza Building in Chicago Since then high
strength concrete has been used in many cities such as Houston, New York and Seattle
(38,1.40)

There are several advantages, the principle one being the economic benefits of its use in
primary structural members such as columns. Schmidt and Hoffmann were among the
first to publish data indicating that the most economical way to design a column was with
the highest available concrete strength and the least amount of reinforcing steel.
Chapter 2 A pplications, A dvantages & D isadvantages o f H S C

With an increase in concrete strength, the engineer can design a smaller member to carry
the same loads. This allows both the number of storeys to be increased and/or is
beneficial when there are architectural restrictions on column size.

Examples of use include:

• The Helmsley Palace Hotel which has a total of 53 floors. A strength increase
from 41.4 N/mm^ to 55.2 N/mm^ resulted in a 10% reduction of the reinforcing
steel in the columns of the lower five floors. These were designated for ballroom
and restaurant facilities and, therefore, required large column spacing. The
general limitations of 41.4N/mm^ concrete would have made the columns large
and uneconomical.
• The Richmond-Adelaide Centre in Canada in which the use of columns with
concrete design strengths of 60 N/mm^ enabled the architect to increase the usable
area of the underground parking garage by 30%.

Other advantages of using high strength concrete in high rise buildings include:
• Early stripping of formwork thus accelerating construction time. Construction of
the Texas Commerce Tower proceeded at the rate of two floors per week.
Concrete compressive strengths of 36 N/mm^ at three days allowed early stripping
of the formwork. During the construction of 311 South Wacker Drive high
strength concrete was used for the columns and the formwork was generally
stripped at an age of 16 to 18 hours after casting.

• A high modulus of elasticity hence reducing the amount of sway in the upper
storeys of a building, e.g. in the First Republic Bank Plaza high strength
concrete was specified to provide a high stiffness for the full height of the
structure, which had a height/width ratio of 7.24. The rigidity of steel was found
to be five to seven times that of concrete per unit volume, but the cost was 20 to
30 times the cost per unit volume of cast-in place reinforced concrete. Compared
to steel, the use of high strength concrete provided six times as much stiffness per
US dollar.
Chapter 2 A pplications, A dvan tages & D isadvan tages o f H SC

The use of superplasticisers in high strength concrete to obtain high fluidity has led to
innovative construction techniques. For example during the construction of the new
Shinagawa Prince Hotel high strength concrete was used inside steel tubular columns
to provide stiffness. The concrete contained 10% silica fume with a water/binder ratio of
0.28 and was filled in a single action to a height of 60 m by means of ‘bottom-up
concreting’.
The majority of the buildings made with high strength concrete have made use of the
economic and technical benefits of blended cements. Some examples are given in Table
2.1. In the River Plaza 10% of the cement was replaced by PFA t) obtain the desired
workability (100mm slump) and to achieve the design strength of 60 N/mm^ at 28 days.
Buildings such as Two Union Square and La Laurentienne Building have used up
to 10% cement replacement by CSF.
The Scotia Plaza concrete designers specified a ternary blended cement of, 65% OPC,
28% GGBS and 7% CSF. This project was one of the first to use GGBS in a high
performance concrete mixture. During the construction of the Pacific First Center
PFA was used in combination with OPC and CSF. Table 2.1 shows the details of these
high strength concrete mixes.

Table2.1 Mix Proportions for River Plaza, Scotia Plaza and Pacific First Centre

River Plaza^'*'*' Scotia Plaza^'**’^ Pacific First Centre

(Chicago ,USA) (Toronto, Canada) (Seattle, USA)
Cement 504 315 533
PFA 59 - 59
GGBS - 135 -
CSF - 36 40
Total binder (kg/m"’) 563 486 633
Coarse Aggregate 1026 kg/m^ 1130 kg/m^ 1068
Max. size 12.7 mm 10 mm 9.5 mm
Type Stone Limestone Gravel
Sand (kg/m’’) 616 745 622
W ater (k g W ) 195 145 130
Admixture 934 mL/m^ 6000 mL/m^ 4210 mL/m^
W ater/Binder ratio 0.35 038 0.22
Slump 115 mm 140 mm 250 mm
28 day strength 64.9 MPa 743 MPa 115 MPa
C hapter 2 A pplications, A dvantages & D isadvantages o f H S C

2.2.2 Bridges

The amount of published information on bridge structures utilising high strength concrete
is limited. According to Carpenter high strength concrete’s comparatively greater
compressive strength per unit weight and unit volume allows increases in span capability,
reduction of the girder depth, and lighter and more slender bridge piers. Several bridges
in North America, Europe and Japan have used high strength concrete. Nagataki
summarised the use of high strength concrete in Japanese prestressed bridges built
between 1968 and 1977; design strengths varied from 55 to 80 N/mm^.

The Tsing Ma Bridge has been recently completed between Kowloon and Lantau
Island, Hong Kong. It is a suspension bridge with slipformed concrete towers. High
strength concrete was specified to achieve a design life in excess of one hundred years;
the design specification highlighted two particular aspects of providing resistance to
chloride penetration and avoidance of thermal cracking. The general requirements for the
concrete are shown in Table 12 . During trials it was found that a water/binder ratio of
less than 0.4 was required to meet specifications, and the utilisation of a ternary blended
high strength concrete mix (OPC/GGBS/CSF) provided enhanced chloride resistance and
improved workability characteristics.

The Storebaelt West Bridge in Denmark is another example of using high strength
concrete to attain a design life in excess of one hundred years. The West Bridge is
approximately 6.6 km in length with 63 spans, the majority of which are 110m in length.
The main requirements for the concrete specification are shown in Table 23. A ternary
blended concrete mix (OPC/PFA/CSF) was specified to provide resistance to ASR and
chloride penetration.
It is interesting to note that in both of these examples a ternary blend was used to provide
high durability and not high strength.
C h apter 2 A pplications, A dvantages & D isadvantages o f H S C

(52)
lia b le 2.2: General Requirements for concrete on the Tsing Ma Bridge

REQUIREMENT LIMIT
C haracteristic 28 day cube
strength 50 Mpa minimum

Binder Types OPC/25-35% PFA

OPC/65-75% GGBS

C em ent C ontent Minimum - 350 kg/m^

Maximum- 550 kg/m^

W ater/Binder Ratio 0.40 Maximum

Maximum Alkali Content 3kg/m^ Na^C eg.

Maximum Chloride C ontent 0.06% of binder

(53)
T ab le 2 3 : Summary o f specification requirements for the Storeabaeit W est Bridge

CONCRETE TYPE
REQUIREMENT
1 II
C haracteristic 28-day cylinder strength 45 MPa

Total C em en t (C) Minimum 300 kg/m^

Microsilica (M) Min 4%, Max 8% of C
Fly Ash (PFA) Min 10% o fC
MS+PFA Max 25% of C
W ater (W) Max 140 kg/m^ Max 135 kg/m^
W/C Ratio Max 0.45 Max 0.35
Air co n ten t in p aste (<0.35mm) Min 8%, Max 20%
E ntrapped air in p aste Max 7%
Specified air surface Min 25 mm -1
Maximum Alkali content 3kg/m"^ Na20 eg.
Maximum Chloride content 0.1% OfC

10
Chapter 2 _________________________________________________________ Applications, Advantages & D isadvantages o f H S C

One of the most recent examples of the use of high strength concrete in bridge structures
is the Charenton Canal Bridge in Louisiana, USA. This bridge was completed late
1999, and marks the implementation of high performance concrete that -began with
research work in Louisiana in the early 1980’s. The bridge is a 111m long continuous
prestressed concrete structure. Each 22m span consists of five girders that are spaced at
3m centres and support a 203mm thick cast-in-place concrete deck. The specified
compressive strength of the girders and piles was 69 N/mm^ at 56 days. The use of high
strength concrete enabled the bridge to be designed with fewer girders than would be
required when using normal strength concrete. The additional strength in the precast piles
increased their resistance to compressive and tensile driving stresses and allowed casting
of longer lengths. A binary blended high strength concrete was used, with 30% PFA to
reduce the heat of hydration due to the high cement content of the mix. No further
concrete details were published. The internal concrete temperature of all precast members
was limited to 71° C and the use of temperature matched curing cylinder moulds was a
requirement for the precast fabricator.

2.2.3 Offshore platforms

High strength concrete has been utilised for offshore structures in the North Sea since
1973. The first concrete platform, Ekofisk l^^', was located in water 70 m deep and
contained 80,000 m^ of normal weight concrete with a specified 28-day compressive
strength of 45 N/mm^. Strengths between 45 and 50 N/mm^ were considered a near upper
limit at that time. Since then a total of 25 major structures containing over 2 million m^ of
concrete have been ordered for the North Sea oil fields and specified strengths have risen
to 80 N/mm^. The major advantage of using high strength concrete for offshore structures
is its potential for weight reduction.

One of the key features of concrete platforms is the high density of reinforcement (800-
1200 kg/m^ compared with 300-500 kg/m^ for normal strength concrete). Therefore the
placing of concrete in these congested sections requires extremely workable high strength
concretes.

11
C hapter 2 ________________________________________________________ Applications, A dvantages & D isadvantages o f H S C

During construction of the Gullfaks C oil platform the cement content was reduced to
400 kg/m^, the water content was reduced to 165 1/m^ and 6 litres of admixtures and 10kg
of silica fume were added to improve the pumpability and cohesiveness of the fresh
concrete. The average slump was 200mm, and the average compressive strength 79
N/mm^. The total volume of concrete necessary to build this platform was 224,000 m^,
with 70000 tonnes of reinforcing steel and 3500 tonnes of prestressing steel.

The Hibernia offshore platform^^^\ completed in Canada in 1996, is a gravity based

structure that has utilised 45000 tonnes of high strength concrete. The high strength
concrete had an average water/binder ratio of 0.30 for a design compressive strength of
69 N/mm^. As well as placing constraints the reinforcement density was 1000 kg/m^ to
satisfy design requirements of a density between 2200 and 2250 kg/m^ for buoyancy, and
have an elastic modulus greater than 32 GPa. To conply with these demands a binary
blended cement was used, containing 8.5% silica fume, with a naphthalene based
superplasticiser. In order to obtain a low unit mass, half of the coarse aggregate was
replaced by lightweight aggregate and some air was entrained. The tight quality control
procedures during construction led to an 86.7 N/mm^ average strength of the air-
entrained concrete.

2.2.4 Other areas of application

Zikeyev^^^ described how Russian precast pipe technology had been adapted to produce
high strength concrete tubular columns for industrialised buildings and for pipeline
supports. Centrifugal casting produced 80 N/mm^ concrete strengths, resulting in
substantial economies of reinforcing steel, cement and concrete to be made. Copen has
indicated that the use of 70 N/mm^ concrete in arch dams would result in greater
economy through reduced volume of concrete, and would tend to reduce deflections and
improve strength of construction joints.

12
C hapter 2_________________________________________________________ Applications, Advantages & D isadvantages o f H S C

The increased abrasion resistance of high strength concrete has been exploited in the
surfacing of the stilling basin of Kinzua Dam. Recent inspection of the 96 N/mm^
concrete overlay suggests that it will have a much longer service life than normal strength
concrete. Gjorv et al state that high strength concrete also has good abrasion resistance
in highway applications.
High strength concrete piles have been used for the Selmar-Sande building in Oslo
Around 250 piles, 275mm square, were driven in clay down to the rock at 30m depths. A
characteristic strength of 75 N/mm^ was used which increased the pile capacity, leading
to a reduction in the required number of piles.

Various other applications have been described by the CEB/FIP^^^^ report on the
applications of high strength concrete, including grandstand roofs, marine foundations,
prestressed concrete poles and avalanche shelters.

2.3 Disadvantages of High Strength Concrete

Most of the limitations concerning high strength concrete result from a lack of research
and available information on its behaviour under actual field conditions. Some of these
are summarised below.
One of the major drawbacks of using low water/binder ratios is that the workability is
adversely affected. Even with substantial doses of superplasticiser the low water/binder
ratio mixes tend to be very cohesive and sometimes quite viscous, leading to the
progressive collapse of fresh concrete during the slump test. Due to this effect, many
researchers have questioned the suitability of the slump test for high strength
concrete. The workability also often declines rapidly with time after mixing making it
difficult to handle and place. In general, there are no set standards on the minimum
acceptable slump for site placing. It has been suggested by Schmidt that a minimum
slunp of 65 mm is required for site placing, whereas the ACI Committee report 363^^^
states that slumps less than 75 mm require special consolidation equipment and
procedures. It advocated that a slump higher than 100 mm generally provide the required
workability; however, details of forms and reinforcement bar spacing need to be
considered before designing the mix.

13
Chapter 2 _________________________________________________________A pplications, A dvantages & D isadvantages o f H S C

Slumps in excess of 200mm are commonly required, especially in areas of congested

reinforcement To achieve this, high superplasticiser dosages are used and strict quality
control procedures need to be adopted, leading to extra cost.

The use of high cement contents (ranging from 500 to 600kg/m^) can lead to high heat of
hydration temperature rises in large pours, which can lead to thermal cracking due to the
induced tensile or flexural stresses when the concrete element is restrained It has
been suggested that for normal strength concrete a differential temperature below 20°C is
required to avoid cracking This has been questioned by Browne who criticises the
temperature differentials as guidelines, and stated that thermal cracking is a stress
concentration problem and temperature gradients are more important. Temperature
gradients generate differential strains which can result in failure if they exceed the tensile
strain capacity of the concrete. Aitcin^^'^ states that while the internal temperature of a
concrete element can rapidly reach 65°C, if the thermal gradient does not exceed 20°C/m,
then network cracking due to unequal cooling will not occur.

Many conqjuter applications based on the finite-difference model are used when
designing high strength concrete elements to predict the temperature gradients. These
enable designers to specify safe limits for thermal gradients, taking into account the
optimum mix proportions together with thermal and mechanical coefficients.

Another consequence of using high cement contents is that the in-situ strength
development may be different to the laboratory specimens cured at a constant
temperature. Subjecting concrete to a temperature cycle during early ages similar to that
which would occur at the centre of a large pour can provide an estimate of the strength of
the actual concrete in the structural element

14
C hapter 2 A pplications, A dvantages & D isadvan tages o f H S C

Other disadvantages of the hardened properties of high strength concrete include:

• Increasing brittleness.
Shah compared the brittleness of normal and high strength concretes and showed that
the difference increases with strength. The brittle behaviour in high strength concrete
columns is characterised by spalling of the concrete cover. In order to enhance the
ductility of high strength concrete the amount of confinement reinforcement is generally
increased.

• Autogeneous shrinkage.
The combination of low water contents and condensed silica fume leads to self-
dessication and hence shrinkage. In one study it was found that the autogenous
shrinkage of a 100 N/mm^ concrete was 110 microstrain compared with only 40
microstrain for a 40 N/mm^ concrete.

• Fire resistance
High strength concrete at high temperatures is more vulnerable compared to normal
strength concrete. Diederichs et al have shown that for high strength concrete at
150 °C there is a distinct loss of strength (30%), while normal strength concrete retained
its strength up to a temperature of 250° C. The moisture content and the permeability of
the concrete govern the extent of spalling. At high temperatures, the moisture vaporises
causing pore pressures near the surface, consequently forcing the free water into the
concrete. This flow of water is prevented in high strength concrete due to its low
permeability; hence the pore pressures exceed the tensile capacity of the concrete which
induces spalling. Introducing polypropylene fibres into the concrete can increase the
resistance to fire.

15
Chapter 2 ________________________________________________________ Applications, A dvantages & D isadvantages o f H S C

2.4 Conclusions

High strength concrete has successfully been used in high rise buildings, bridges and
offshore structures. Its use as a construction material has come about from its technical
and economical advantages.
• High rise buildings - With an increase in concrete strength, the engineer can design
a smaller member to carry the same loads. This allows both the number o f storeys to
be increased and/or is beneficial when there are architectural restrictions on column
size. High strength concrete is often economically viable in buildings greater than
150m in height.
• Bridges - High strength concrete’s comparatively greater compressive strength per
unit weight and unit volume allows increases in span capability, reduction o f the
girder depth, and lighter and more slender bridge piers.
• Offshore structures - The major advantage o f using high strength concrete fo r off
shore structures is its potential for weight reduction.

Many proponents of high strength concrete also indicate that it has certain disadvantages,
which in some cases can negate its advantages.

• One o f the major drawbacks o f using low water/binder ratios is that the
workability o f high strength concrete is adversely affected, even with the use o f
superplasticisers. In particular the workability can decline rapidly with time after
mixing. This makes it difficult to handle and place.
• The use o f high cement contents (ranging from 500 to 600kg/m^) can lead to high
heat o f hydration temperature rises in large sections, which can lead to thermal
cracking.
• High strength concrete also exhibits increased brittleness and autogeneous
shrinkage and has poor fire resistance compared to normal strength concrete.

Some of these disadvantages were investigated in this research programme (as outlined in
section 1.4).

16
C hapter 3 L iteratu re Review

CHAPTER 3

LITERATURE REVIEW

3.1 Introduction

As highlighted in the previous chapters, high strength concrete has many advantages and
it is necessary to acquire the maximum performance out of all its constituent materials.
The majority of applications cited in the literature showed the use of at least one
supplementary cementing material, CSF, and in some cases the use of PFA or GGBS in
conjunction with CSF to improve the properties.

The chapter begins with a brief description of each supplementary cementii^ material
and then reviews the effects of these on the fresh properties, temperature rise and strength
development of high strength concrete.

3.2 Supplementary Cementing Materials

3.2.1 Pulverised Fuel Ash (PFA)

PFA is a by-product of the combustion of pulverised coal in thermal power plants. It is

removed by the dust collection system as a fine particulate residue from the combustion
gases before they are discharged into the atmosphere. The particles are spherical, ranging
in diameter from less than 1 to more than 150 pm, the majority being less than 45 pm.

PFA can have different chemical compositions which are related to the type and amount
of impurities contained in the coal. More than 85% of most PFAs comprise compounds
and glasses formed from the elements silicon, aluminum, iron, calcium and magnesium.
Generally, coal from the same source will produce the same type of PFA.

17
C hapter 3 Literature Review

The different PFAs that are available can be classified into broad categories.
recognizes two general classes of PFA. Class C is produced by burning lignite or sub-
bituminous coal and Class F is usually produced in power plants burning anthracite or
bituminous coals. The most notable difference between the two classes is that Class C
PFAs contain high levels of calcium. Table 3.1 shows typical oxide compositions of the
two types of PFA. In the UK, bituminous coals are used, the PFA is classified according
to BS 3892:Part 1^^^^^ or EN 450 The latter permits a greater range of fineness (up to
40% residue on a 45 |im sieve). Table 3.2 shows a typical composition of PFA found in
the UK.

Table 3.1 Chemical composition of typical Class F and Class C PFA^^^^

Oxide Class F Class C

SiOî 47 55

AI2O3 23 23

Fe203 21 3.5

CaO 1.8 12.5

MgO 1.1 1.2

Na20 1.5 1.6

K2O 3.15 0.7

PFAs exhibit pozzolanic activity. A pozzolan is defined as a siliceous or siliceous and

aluminious material which in itself possesses little or no cementitious value but which
will, in a finely divided form and in the presence of moisture, chemically react with
calcium hydroxide at ordinary temperature to form compounds possessing cementitious
properties.

The action of a pozzolan in concrete may be simplified as follows:

Portland cement + Pozzolan + Water —> C-S-H.

18
Chapter 3 L iterature R eview

3.2.2 Ground granulated blast furnace slag (GGBS)

GGBS is the by-product of the manufacture of iron in a blast furnace. It is formed as a

liquid at 1350- 1550 °C and consists primarily of silica and alumina (from the iron ores),
combined with the calcium and magnesium oxides (from the fluxing stone), and with
some impurities from the coke charged into the blast furnace. If allowed to cool slowly, it
crystallizes to give a material having no cementing properties. If cooled sufficiently
rapidly to below 800° C, it forms a glass which is a latent hydraulic cement. In practice,
the material is cooled by spraying droplets of the molten slag with high-pressure jets of
water. This gives a wet, sandy material which when dried and ground is called GGBS.
The major oxides of silica, alumina, lime and magnesia constitute 95 % or more of the
total oxides. Table 3.2 shows the typical properties of GGBS.

In an alternative treatment, called pelletization, the molten slag is partially cooled with
water in a rotating drum. The resulting pellets vary in size from a few mm to around 15
mm and have a high glass content, and can be used as a cementitious material (after
grinding) or as lightweight aggregate.

The reaction mechanism of GGBS with Portland cement largely depends on the amount
of calcium hydroxide present and is activated by the strong alkaline conditions. The
material contains sufficient lime and silica to form calcium silicate hydrates however,
lime is also available from the hydrating Portland cement and some will normally be
incorporated in the products of hydration of the slag. At high GGBS replacements
(>50%), there is more silica and less lime than Portland cement alone, and hydration of
the blended cement produces more C-S-H and less lime than Portland cement alone. This
results in a dense microstructure and hence high levels of GGBS replacements are often
specified for improving the durability of concrete.

3.23 Condensed Silica Fume (CSF)

CSF is a by-product of the production of silicon or silicon alloys by reducing quartz in an
electric furnace. It consists of fine spherical particles of glass with diameters ranging

19
C hapter 3 Literature Review

from less than 0.1 |im to about 1 |im and is collected from the gases escaping from the
furnaces. CSF is mostly composed of silica, the SiO] content varies, depending on the
type of alloy produced. CSF produced during the manufacture of silicon metal generally
contains more than 90% SiO]. CSF produced during the manufacture of 75% Fe-Si alloy
has a SiOi content greater than 85%.

The specific gravity of CSF is about 2.2, compared to 3.1 for normal Portland cement.
The specific surface area of CSF cannot be measured in the same way as for Portland
cement owing to its extreme fineness, and is determined by nitrogen adsoiption. Typical
values range from 15,000 to 25,000 m^/kg, compared to 1500 m^/kg for Portland cement.

Compared with the other supplementary cementing materials, the characteristics that
make CSF a very reactive pozzolanic material are its high SiO: content, its amorphous
state and its extreme fineness, as illustrated in Table 3.2.

Table 3.2 Typical oxide compositions and physical properties of OPC, PFA, GGBS and CSF.
Oxide PFA GGBS CSF (Silicon Metal) OPC

SiOz 48 36 97 20

AI2O3 27 9 2 5
FezOa 9 1 0.1 4
MgO 2 11 0.1 1

CaO 3 40 - 64
NazO 1 - - 0 .2

KzO 4 - - 0.5
Typical oxide compositions (% by weight)
PFA GGBS CSF OPC
Specific gravity 2.1 2.9 2.2 3.15

Particle size 10-150 3-100 0.01-0.5 0.5-100

range (microns)
Specific surface 350 400 20,000 350

area (mVkg)

20
C hapter 3 L iterature Review

The rapid reactivity of CSF enables it to be used as a replacement for a small proportion
of Portland cement, generally up to 10%. The effects of CSF on the properties of concrete
are due not only to a rapid pozzolanic reaction, but also to the physical effect of the fine
particles, which is known as the ‘filler effect’. The CSF particles fill the voids between
the larger cement particles during mixing, resulting in a dense solid matrix. They also act
as nucléation sites for the hydrates.

3 3 FRESH PROPERTIES OF HIGH STRENGTH CONCRETE

33.1 Definition of workability

There is no single precise definition of workability in the literature; most definitions are
qualitative in nature and based on the personal viewpoint of individuals rather than on
scientific precision.

Newman^^^^^ for example has proposed that workability should be defined by at least three
separate properties:
i. Compactibility, or the ease with which the concrete can be compacted and the air
voids removed.
ii. Mobility, or the ease with which concrete can flow into moulds around steel and
be remoulded.
iii. Stability, or the ability of concrete to remain a stable coherent homogeneous
mass during handling and vibration without the constituents segregating.

Glanville et aF^^ who exhaustively examined the field of compaction and workability,
defined workability as the amount of useful internal work necessary to produce full
compaction. The ASTM C 125-93 definition is: “ that property determining the effort
required to manipulate a freshly mixed quantity of concrete with minimum loss of
homogeneity”. The ACI definition of workability, given in ACI 116R-9(P^^^ is: “that
property of freshly mixed concrete or mortar which determines the ease and
homogeneity with which it can be mixed, placed, consolidated and finished”.

21
 jsQ

THE RHEOLOGY OF FRESH CONCRETE.

STABILITY. COMPACTIBILITY. MOBILITY.

Stability is defined as the flow of fresh Compactibility m easures the ea se with Mobility of fresh concrete can be
concrete without applied forces and is which fresh concrete is compacted. The described in terms of its viscosity,
measured by bleeding and segregation process of consolidation occurs in two cohesion, and internal resistance to
characteristics. stages; the first comprises the major shear.
subsidence or slumping of the concrete;
the second involves de-aeration (removal
of entrapped air).

BLEEDING. SEGREGATION.
Bleeding occurs when the mortar is Segregation is defined as a mixture's
unstable and releases free water. instability, caused by a weak matrix that
cannot retain individual aggregate
particles in a homogeneous dispersion.

VISCOSITY COHESION ANGLE OF INTERNAL FRICTION

As conceived by Ritchie, the viscosity of Internal friction occurs w hen a mixture is
Cohesion is defined by Ritchie as the
the matrix contributes to the ease with displaced and the aggregate particles
force of adhesion between the matrix and
which the aggregate particles can move translate and rotate. Tfie resistance to
the aggregate partides. It provides the
and rearrange themselves within a deformation depends on the s h a p e and
tensile strength of fresh concrete that
m ix tu re. To a c h ie v e a b e tte r texture of the aggregate, the richness of
resists segregation and is m easured by a
understanding of a mixture’s flow tfie mixture, the w ater-oem ent ratio, and
direct tension test, which w as first used
characteristics, it is important to be able the type of cem ent used. Thus, th e angle
Hallstrom.
to m easure tfie viscosity of the cement of internal friction plays an im portant part
paste fraction of the mixture and to study in the mobility of a concrete mixture.
to its stiffening with time.
to

Figure 3.1: Ritchie’s subdivisions of Rheology m)

Chapter 3 Literature R eview

Ritchie^^^^ considers these definitions too restrictive, and relates the workability of fresh
concrete to its rheological properties. He has further subdivided the rheology of fresh
concrete into stability (bleeding and segregation) compactability and mobility, (viscosity,
cohesion and angle of internal friction) as shown in figure 3.1. Ritchie, however, did not
point out whether any relationships exist between these subdivisions.

33.2 Measurement of workability

There is no test that provides a complete description of the workability of fresh concrete,
but the importance of workability is reflected in the number of tests that have been
proposed for its measurement. The most common standard tests for the assessment of
workability are described in British Standard 1881 Tarts 102 to 1 0 5 these are
the slump test, compacting factor test, Vebe test and the flow test.

• Standard Tests

The slump test measures by how much a concrete slumps after it has been placed in a
standard maimer in a standard mould and the mould has been removed. It is used
extensively on site all over the world, but results from the test have been seen to vary for
the same concrete. Glanville et aF^\ found that, for one particular mix, the addition of a
small amount of water would result in a collapse or shear slumps, while only a little less
water would result in having uniform slumps less than 25 mm.

The compacting factor test was developed at the Road Research Laboratory^'^ and is
used to assess the ‘compactability’ of the concrete by measuring the compaction
produced by imparting a given amount of energy by dropping the concrete through a
standard height. The test suffers from the disadvantage that cohesive concrete tends to
stick in the hoppers of the apparatus and mixes with low workabilities produce wide
variations in results.

23
Chapter 3 Literature R eview

Cusens^"^'^ criticised the test by showing that the energy applied in the test is much less
than that used in compacting a concrete by vibration and the density ratio achieved was
smaller than that produced by the mild vibration conditions.

The Vebe consistometer test was developed in Sweden. In this test a cone of concrete is
first allowed to slump normally and then vibrated to collapse to a given end point, the
time for which is taken as an inverse measure of workability. The main criticism is that
both end-points are badly defined. The start of the test is vague because it is related to the
beginning of a vibration process that takes time to build up, and the finish is difficult to
assess because it is approached at a decreasing rate. In order to overcome the end-point
difficulties, showed that the use of settlement-time recorders may be helpful, this view
was contradicted by Hughes and Bahramian^^^^ who found that the resulting curve does
not facilitate more accurate assessment of the V-B time. They do, however, suggest that
the area under the curve can be used to give an indication of the cohesiveness of the
concrete.

The Flow table test was developed in Germany in 1933. It is mainly used for flowing
concretes which exhibit collapse slumps. The test involves measuring the average
maximum spread of the fresh concrete after the slump mould is removed and the table
has been dropped. The test is appropriate for mixes having a flow of 500 to 650 mm. If
the concrete at this stage does not appear uniform and cohesive, this is an indication of a
lack of cohesiveness of the mix.

Of these standard tests, the slump and the flow test are the only tests that have gained
general acceptance on site. The compacting factor test is very occasionally used for
control of a practical job but the Vebe test is well renowned as a laboratory based test but
is very rarely used nowadays.

One of the major criticisms regarding these tests, as suggested by Tattersalf^"^ is that they
each give only one measurement. The practical outcome of this deficiency is that a given
test may classify as identical two concretes that are subsequently found to behave quite

24
C hapter 3 L iterature R eview

differently on a construction site. His argument is that the flow of fresh concrete can be
described by the Bingham model, using the equation

X= To +
This equation lepresents-the relationship between the applied shear stress (x) and the
resulting rate of strain (y), where xq and p are the Bingham constants, called yield stress
and plastic viscosity respectively. These both need to be measured to describe the
rheology fully.
There are several techniques available to measure these rheological properties, these are
now discussed.

• Rheological Tests

The advent of superplasticisers and the wider use of high strength concretes has led to an
increasing interest in rigorous rheological testing. Such mixes behave very differently
rheologically to conventional mixes, one consequence of which has been that the
inadequacies of single point testing (standard tests) have been even more apparent.

The development of Tattersalls two point test apparatus, shown in figure 3.2, has been
well documented The concrete under test is contained within a cylindrical bowl
and is sheared by an impeller which is driven by an electric motor through an infinitely
variable hydraulic transmission and a reduction gear. There are two impellers that can be
used, a helical impeller for concrete with medium to high workability (MH system); and
a H-impeller (offset to give a planetary motion) for concrete with low to medium
workability {LMsystem).

The torque is measured indirectly via the pressure developed in the oil in the hydraulic
transmission, which is measured by a Budenburg pressure gauge. The pressure produced
by concrete shearing is obtained from the total pressure by subtracting the pressure
produced whilst the machine is idling. This net value is then converted into impeller

25
Chapter 3 Literature Revieyv

torque by the relationship obtained from prior calibration. The impeller torque T and
speed N are found to be related by the linear equation:

T = g + hN

This equation is similar to that of the Bingham model, where g and h are proportional to
To and p. respectively and thus provide a measure of the two fundamental quantities
characteristic of a material that conforms with the Bingham model.

By simply measuring the torque produced on an impeller rotating in fresh concrete at

various speed settings, the values of g and h can be determined. Both g and h are
dependent on the measuring system and concrete properties.

ICX'frOCOOtfcrf ___ ^

Figure 3.2: Two point test apparatus (75)

26
C hapter 3 L iterature Review

Various researchers have carried out several modifications to this test apparatus. Cabrera
and Hopkins^™^ measured the torque directly by adding a transducer to the impeller drive
shaft, with the output signal being taken off by slip rings. Wallevik and Gjorv^'^^^ added
an electronic tachometer to record both the impeller speed and oil pressure for analysis.
Similarly Wimpenny and Ellis^’^®^added a pressure transducer to enable the oil pressure to
be recorded.

More recently, Domone et took into account the various modifications outlined
above and developed a new compact version of Tattersalls two point test apparatus. The
main feature of this apparatus is that a single gearbox for both the helical and H impeller
systems is used.

The BML viscometer has been developed in Norway^®”’®’’; this is a coaxial cylinder
system. Figure 3 3 shows a cross section through the cylinders. The inner cylinder is
deeply ribbed to prevent slippage and has a radius of 100mm and the outer cylinder has a
radius of 200mm (the sample mix container), and hence there is a 50mm annular gap.
The two part inner cylinder, with the bottom part fixed, eliminates the effect of three
dimensional shearing that normally occurs in concentric cylinder viscometers at the
bottom of the cylinder.

Analysis of the behaviour of the concrete with and without plug flow in the annulus has
resulted in equations from which yield stress and plastic viscosity can be obtained from
the torque/speed flow curve^^^\ Hue et al^^^^ have criticised the apparatus by showing that
it is not very suitable for concrete with a slump below 100mm, however, Mork®'^ claims
that reliable results can be obtained with slumps down to 50mm. Recent modifications of
the test apparatus has made it possible to assess the rheology of mortars and cement paste
successfully^^^\

27
Chapter 3 ______________________________________________________________________________________________________ L iteratu re R eview

The BT Rheom, which has been developed in France^^*’^’^ is essentially a parallel plate
rheometer. A specimen in the shape of a hollow cylinder is sheared between its (fixed)
base and its top section (rotating about a vertical axis), as shown in figure 3.4. The shear
may be applied at various speeds of rotation. The blades within the apparatus ensure an
efficient shearing action with no slippage. The test is controlled by a microcomputer,
which delivers the measured quantities directly to the user.

Concretes with slumps in excess of 100mm can be successfully tested. The testing speed
is such that the flow is laminar, and hence the values of yield stress and plastic viscosity
can be obtained from the torque/speed relationship. Generally, there is a good correlation
between the yield stress and plastic viscosity values and g and h values measured in the
two-point workability apparatus with a helical impeller^^*\ BanfilF'^^ has expressed some
reservations about the choice of test geometry, and TattersalF^ has criticized some of the
claims made for the apparatus.

The tests reported in this thesis have been carried out using Tattersalls two point
workability apparatus. This test was the only one available at the start of the test
programme, and has a good and proven track record for both normal^^^'^^^ and high
strength concrete^^^ and the results obtained have been compared with previous data
obtained by other researchers.
It was beyond the scope of this work to carry out any instrument development, or any
comparative tests with the other instruments such as the BML viscometer and BT
Rheometer, both of which became commercially available during the course of the
research.

28
C hapter 3 Literature Review

L oad c«l

(80)
Figure 3 3 : BML Viscometer

1 0
rt V : //.'■ -.'./////.■ /•> \ y , w '.y . V /.-y.-yy |{

i5 \ \ b : . d t c [I A iO i

( i ) 3-D o f Û : ipp i.-arjs (b ) 2 D dlaçrj.'n o f ll'.c cor.

d m : r : s l o n i ; H 30 y ^ 30 cm
CG-.Li;.-.:r d im c r.:;o ,-j ; H 2-1 x <(, 27 c m )

Figure 3.4: (a) BT Rheometer (b) Testing container

29
C hapter 3 L iterature R eview

333 The effects of binary blends

The lower chemical activity of supplementary cementing materials, particularly PFA and
GGBS, means that a partial replacement of cement is beneficial in controlling the fresh
properties of high strength concrete^^"^ This section reviews the general effects of
supplementary cementing materials on the fresh properties as measured by the slump and
two point tests.

3 3 3 .1 Pulverised Fuel Ash (PFA)

The small size and spherical form of low calcium PFA particles cause a reduction in the
amount of water required for a given degree of workability from that required for an
equivalent paste without PFA^®®\ This view is also shared by NagatakP’^who states the
same principle applies to concrete. Hansen^’’^ states that PFA acts as a water-reducing
agent, much like an air- entraining agent or plasticizer.

Coleman et ah*®^ reported that data available on the effect of different PFA replacement
levels are somewhat conflicting. They found that small percentages of PFA were not
beneficial in improving the fluidity of concrete mixes, but large replacement levels of the
order of 35% were moderately helpful.

Owens^*^^ reported that the major factor influencing the effects of PFA on the workability
of concrete is its proportion of coarse material ( > 45 pm ). By replacing 50 % by mass of
the cement by fine particulate PFA reduced the water demand by 25 % and a similar
substitution using PFA with 50 % of the material greater than 45 pm there was no effect
on water demand.

Besari et af^^^ however found that considerable improvements in slump were made using
PFA regardless of the amount of coarse material. Instead they found that the replacement
level has more effect. The slump of a concrete mix containing 25% PFA was 124 mm

30
Chapter 3 L iterature R eview

compared to 52 mm of that with 15 % PFA. They stated that this improvement in

workability is due to a combined effect of an increased paste volume, the smaller particle
size of PFA compared to Portland cement, and the lubricating ‘ball bearing’ effect due to
the spherical shape of the PFA particles.

Soutsos’s^^) work on the effects of PFA on the workability of high strength concrete,
found that partial replacement by PFA causes a reduction in the amount of
superplasticiser dosage required for a given degree of workability from that required for
an equivalent OPC concrete, as shown in figure 3.5.

Tattersalls two point workability test has been used by various researchers to investigate
the rheology of PFA blended concretes. Ellis^^^^ reported that by increasing the PFA
replacement level both the yield value (g) and plastic viscosity (h) are reduced, (figures
3.6-3.7). Banfilh^'*^ however showed that with an increasing level of cement replacement
by PFA, the yield value of cement pastes was reduced while the plastic viscosity was
unchanged (Table 33). He found a similar trend on the rheological properties of concrete
(Table 3.4). However, mixB showed a reduction of both g and h values. No explanation
was given for this.

Soutsos^"^"^ found that cement replacement by PFA at levels of 20,30 and 40 % reduced
the yield value (g) at a 0.26 w/b ratio, as shown in figure 3.8. However, the plastic
viscosity was higher than that of OPC concrete, especially at higher water/binder ratios.
Gjorv^^^) arrived at a similar conclusion, a mix containing 20 % PFA showed reductions in
the yield value and increases in plastic viscosity compared to the OPC concrete.

31
C hapter 3 Literature Review

0.4

0.38 -

0.36 -

? 0.32 - OPC

10 % PFA
0.28 -
20 % PFA
5 0.26

0.24 -
3 0 % PFA
0.22
*0 % PFA
0.2 -

0 0.5 1 1.5 2 2.5

Superpiastlclser d o tag e (% by weight of binder).

Figure 3.5: Effect of partial cement replacement by PFA on the superplasticiser dosage

required for concrete mixes with a slump of 150 mm.

32
Chapter 3 L iterature Review

8 .0
I 90 Limi

4 .0 .

0 2 0 4 0 a 0 8 0 100
% OPC R « p la t* m * n f wilh pia

Figure 3.6: Relationship between yield (g) value and % replacement by

3. 0

90 Confident# Limitt

h •y

0 20 4 0 80 8 0 100
O PC R*pl a<«m«nl wuh pta

Figure 3.7: Relationship between plastic viscosity (h) and replacement by

33
C hapter 3 Literature Review

Cement pfa Y ield PI a a t l c

% Value V l a c o a 1ty
N/m- Na / m -

100 0 71.9 0 .29

90 10 67.9 0 .30
80 20 61.0 0 .31
60 40 31.5 0 .34
40 60 15.7 0 .30

Table 3..3 R h e o l o g i c a l test resu lts

Cement Water Aggregate Sand pfa Sl u mp & b

Mi x
kg/m3 kg/m^ kg/m4 % •4 ■ua Ne Noia

A 250 165 2025 35 0 22 10 . 4 5. 3

10 33 7. 4 6. 7
20 50 4 .5 6 .3
40 • 2. 9 7.0
60 - 1 .7 5. 3

B 250 180 1985 45 0 28 6 .7 7 .2

10 25 7. 7 6. 4
20 90 2. 8 3. 0
40 180 1 .5 3. 2
60 180 1.6 2. 3

C 450 165 1825 35 0 27 7. 3 4 .7

10 30 4 .4 5. 8
20 46 6 .7 4.1
40 55 5, 5 4.9
60 95 4 ,. 3 3. 8

D 450 180 1785 45 0 24 5 .8 2. 3

10 45 7 ,. 2 2 . ,1
20 72 3 .4 2 .. 8
40 90 3 .3 2. , 3
60 120 2.4 2 ,. 7
* Oeoocas collapse y l u m p , mix a l s o segregated

Table 3.4 Mix proportions and workability results for the four mixes tested

34
C h apter 3 L iteratu re R eview

4.5

I
UJ
3
%
S
o
d
>-
3.5 -

0 10 20 30 40
PER CENT CEMENT REPLACEMENT BY PFA.

(a) Yield (g) value.

-O'4.5 -

LU

Ü 3.5

2.5
0 10 20 30 40
PER CENT CEMENT REPLACEMENT BY PFA.

(b) Plastic viscosity (h) value.

Figure 3.8: The effect of % replacement by PFA on the yield (g) and plastic viscosity (h)
value for concrete mixes with water/binder of 0.26 and a slump of approximately 150
(90)
mm

35
Chapter 3 Literature R eview

3.3.3.2 Granulated Glass Blastfurnance Slag (GGBS)

When part of the cement is replaced by GGBS, the characteristics of the concrete are
affected by its fineness and replacement ratio. Nagatakf®’^ found that the unit water
content necessary to obtain a constant slump decreases with an increase in GGBS content
and fineness. Meusel and Rose^’^^ reported a similar finding, that in non-superplasticised
concrete when the GGBS content is increased there is an increase in slump.

Tachibana et aP’^ states that GGBS can improve the properties of high strength concrete
including its workability. Read et aF*^ showed that at a 35% GGBS replacement level a
0.28 water binder ratio high strength mix gave adequate workability as a result of adding
the GGBS. In contradiction to these findings Taylor^’’^ stated that for GGBS concretes the
workability is similar to that obtained with Portland cements. The reasons for these
differences are not clear, but maybe associated with variations in morphology and/or
composition of GGBS.

There is limited information available in the literature on the rheological properties of

GGBS concretes. As part of a study on the fresh properties of high strength concrete,
Soutsos^^) found that for GGBS replacement levels of 10, 20, 30,45 and 60 % a reduction
in the amount of superplasticiser dosage is required for a given slump compared to an
equivalent OPC concrete (figure. 3.9). However, inconclusive results were obtained as to
whether GGBS increases or reduces the yield (g) and plastic viscosity (h) of concretes
with equal slumps, as shown in figure. 3.10. From these observations he concluded that
GGBS has only a small effect on the rheological properties of concrete.

36
Chapter 3 L iteratu re Review

0.4

0.38
OPC
0.36

0.34 10 % GGBFS

d 0.32
20 X GGBFS

0.3

0.28
30 X GGBFS
5 0.26
45 X GGBFS
0.24

0.22
60 X GGBFS
0.2

0.18
0 0.5 1.5 2 2.5
Superpfasdcisef d o sa g e (% by weight of binder).

F ig u re 3.9; The effect of GGBS on the superplasticiser dosage required for 150 mm
constant slump

37
C h apter 3 L itera tu re R eview

4.5

UJ

3.5

15 30 45 60
PER CENT CEMENT REPLACEMENT BY GGBFS

(a) Yield (g) value.

4.5

UJ

CO

3.5

CL

2.5
0 15 30 45 60
PER CENT CEMENT REPLACEMENT BY GGBFS.

(b) Plastic viscosity (h) value.

Figure 3.10: The effect o f % replacement by GGBS on the yield (g) and plastic viscosity
(h) value for concrete mixes with water/binder of 0.26 and a slump of approximately 150
(90)
mm

38
Chapter 3 Literature Review

33.3.3 Condensed Silica Fume (CSF)

It has been well documented that the use of CSF in high strength concrete reduces the
workability as a result of an increased water demand and requires a consequent addition
of superplasticiseF^°°’'°‘-‘°^\ Yogendran^'°^^ reported different superplasticiser dosages at
constant slump for concretes with water/binder ratios of 0.34 and 0.28. Figure 3.11
shows that, at the 0.34 water/binder ratio, the amount of superplasticiser required to
maintain a constant slump increases linearly from 10 to 30 % replacement, while at 5 %
replacement no superplasticiser is required. However, at 0.28 water/binder ratio the
amount of superplasticiser required at 5 and 10 % dosage was nearly the same as the OPC
concrete. A similar conclusion was reached by Sellevold and Radjy*°^\ Hjorth^*®'*^ and
Larrard et af'^\

- w o l c r / c e me n l i l i o u s r o l i o = . 2 8

It
/ w o l e f / c e me n l i l i o u s fotio * 3 4

5 10 15 20 25
Pe r c e nt oge of C e m e n t R e pl oc e d by Silica Fume

Figure 3.11: Variation of superplasticiser dosage at constant slump with various levels of
CSF replacement

39
C hapter 3 L iteratu re R eview

Soutsos^^) investigated the effect of CSF at replacement levels of 5, 10 and 15%, on the
relation between water/binder ratio and superplasticiser dosage required for a 150 mm
constant slump, as shown in figure. 3.12. He found that at low water/binder ratios below
0.28 the CSF enhances the slump and lower dosages of superplasticiser are required. He
attributed this behaviour to the ultrafine spherical particles of CSF being sufficiently
dispersed to act as a workability aid. Soutsos also stated that the concrete, even though of
high slump, became increasingly cohesive with a reduction in water/binder ratio and
increase in CSF replacement level.

This finding is in agreement with the results from two research bodies in France^^°^’^°®\
They claim that the use of ultrafine particles, i.e. grain size smaller than of cement,
facilitates the production of low water/binder ratio concrete by the filler effect: the grains
fill the voids between those of cement thus reducing the water requirement.

Yu rugi et al^^°^^ have reported that satisfactory properties for handling and placing could
not be obtained at water/binder ratios below 0.29 without using CSF. Tachibana et al^^^^
found that with a water/binder ratio of 0.25, concrete without the use of CSF required
more mixing time than concrete containing CSF. Furthermore, they showed that the
electric energy consumed during mixing decreased by incorporating CSF in high strength
concrete mixes. This effect was attributed to the better dispersion of cement grains by the
use of CSF. In addition, they observed that the superplasticiser dosage was less when
using powdered CSF than using densified CSF. This observation was also reported by
Yonezawa^^^°\ who states that undensified CSF gives greater increase in fluidity and
strength of concrete than densified CSF.

Djellouli et ah'") found that at a water/binder ratio of 0.27, compared to a reference

Portland cement mix, the superplasticiser dosage required to produce a 200 mm slump
was reduced by 18 % with a 10 % CSF replacement. A similar conclusion was reached
by Baalbaki et ah"^'.

40
Chapter 3 L iteratu re Review

More recently, Duval and Kadari^’®^^ have shown that at a 10% CSF replacement level
there is a continual reduction in the required superplasticiser dosage at water/binder ratios
of 0.45-0.25.

Another effect of CSF on fresh concrete is an increase in internal cohesion. Generally, it

is possible to design CSF concretes which exhibit no bleeding and have no segregation
tendencies. Gjorv^^^^ states that due to this increased cohesiveness of CSF mixes. The
slump test does not predict the flow or response to compaction and therefore, an
increased slump for CSF concrete is normally recommended compared to that of normal
strength concrete. Also when comparing workabilities using the slump test, equal slumps
do not necessarily indicate equal workabilities. This emphasises the need for rigorous
rheological testing, i.e. two-point testing, for CSF concretes.

Gjorv^^^) reported that if Portland cement is replaced by an increasing amount of CSF the
plastic viscosity (h) is strongly reduced down to a certain threshold level, while the yield
stress (g) is almost unaffected, as shown in figure 3.13. These threshold values are
approximately 2, 4 and 6 % for a cement content 200, 300 and 400 kg/m^ respectively.
For higher contents of CSF, above the threshold levels, the yield stress is substantially
increased and the plastic viscosity also starts to increase.

In contradiction to these findings Soutsos^^^ investigated the effects of CSF replacement

on the g and h values of high strength concrete mixes, as shown in figure 3.14.
He concluded that both g and h values were lower than the control OPC concrete with a
510 kg/m^ cement content.

Yonezwa^'^°^ investigated the effect of undensified CSF on the flow rate of high strength
concrete at water/binder ratios of 0.23, 0.28 and 0.33 by using a L-shaped flow meter. He
found that the addition of CSF leads to a lower plastic viscosity, since a higher flow rate
means a lower viscosity.

41
Chapter 3 Literature Review

0.4

0.38

0.36 - 15% CSF

10% CSF
3 0.32 -

0.3
OPC
0.28

5 0.26 -
5% CSF
ul \v
0.24

0.22 -

0.2 -

0.18
0 0.5 1 1.5 2 2.5
Superpiastidsef dosage (% by weight of binder).

Figure 3.12: Relationship between water/binder ratio and superplasticiser dosage for
CSF mixes

400

200
jk O cm cnt

siucA .

Figure 3.13: Effect of CSF on the rheological properties of fresh concrete (95)

42
C h apter 3 L itera tu re Revieyv

4.5

UJ
4

o>
3.5
UJ

2.5
0 5 10 15
PER CENT CEMENT REPLACEMENT BY CSF.

(a) Yield (g) value.

4.5

3.5
o
>■
3

2.5
0 5 10 15
PER CENT CEMENT REPLACEMENT BY CSF.

(b) Plastic viscosity (h) value.

Figure 3.14: The effect of % replacement by CSF on the yield (g) and plastic viscosity
(h) value for concrete mixes with water/binder of 0.26 and a slump of approximately 150
(9 0 )
mm

43
C hapter 3 L iteratu re R eview

33.4 The effects of ternary blends

In ternary blends incorporating CSF and GGBS. Djellouli et aF"^ investigated a

combination of 10% CSF and 36% GGBS at a water binder ratio of 0.27. Compared to a
reference OPC concrete, the amount of superplasticiser required to produce a 200 mm
slump was reduced by 36% by using a ternary blend. They state that the need for less
superplasticiser dosage is due to the substitution of cement by a less reactive cementitious
material that results in a decrease of the quantity of C3A, thereby reducing the formation
of ettringite.

Read et af*^ states that the use of GGBS in combination with CSF contributes to
increased workability at very low water/binder ratios. This phenomenon can be used to
advantage in reducing the total water content, thus further decreasing the water/binder
ratio. Similarly, Baalbaki et found, using a reground low C3 A content cement, that
at a water/binder ratio of 0.25 significantly less superplasticiser was needed to achieve a
170 mm slump with a mix containing 10% CSF + 60% GGBS.

Soutsos^^'^ found that a ternary blend with 50 % GGBS + 10 %CSF used at low
water/binder ratios of 0.2 to 0.38 reduced the amount of superplasticiser required to
produce a 150 mm slump. For example, a GGBS/CSF ternary blended concrete at
water/binder ratio of 0.2 requires only 1.3 % superplasticiser dosage as compared to 2.3
% for a 10% CSF binary blend, figure 3.15.

Soutsos^^^^ also investigated the effects of this ternary blend on the rheological properties
(g and h values) as measured with Tattersalls two point workability test, figure 3.16.
From the results he concluded that g and h values increased as the water/binder ratio
decreases, but were in each case lower than those of the control OPC mix.

44
Chapter 3 Literature Review

In contrast, very limited data exists on temar>' mixes of PFA/CSF. One example is
Bayasi^"^\ who investigated by means o f the slump test the efficiency of ternary blends of
PFA + CSF + OPC, and found that with a 10% CSF + 20% PFA mix the slump increased

compared to the control concrete.

Combinations of OPC with PFA and GGBS have also been used in self-compacting

concrete. Miura et al^'"^ reported that, a ternary combination of

OPC+30%PFA+50%GGBS gave the highest flowability (or slump) at 0.36 w/b ratio

compared with a binary mix with PFA or GGBS.

0.4

0.38 -

0.36
10 % C S F
0.34

50 % GGBFS i 10 % CSF

0.3
OPC
? 0.28 -

5 0.26 -
60 % GGBFS
0.24

0.22 -

0 0.5 1 1.5 2 2.5

Superplasticiser dosage (% by weight of binder).

Figure 3.15: Relationship between water/binder ratio and superplasticiser dosage for

ternary GGBS/CSF mix

45
Chapter 3 Literature Review

0.4

0.30

0.36
O P C MIXES

y-
Π0.32

m
S 0.28 -

5 0.26
C S F & G G B F S /C S F
0.24
MIXES

0.22

0.2
1 1.5 2 2.5 3 3.5 4 4.5 5
YIELD (g) VALUE (Nm).

(a) Yield (g) value.

0.4

0.30

0.36

s 0 .34
O P C MIXES
cr 0.32
IXI
0.3

28

26

0.24 CSF 4
G G B F S / C S F MIXES
0.22

0.2
1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
VISCOSITY (h) VALUE (N ms ).

(b) Plastic viscosity (h) value.

F igure 3.16: Relationship between water/binder ratio and g and h values for CSF and

GGBS/CSF mixes

46
C hapter 3 L iteratu re R eview

3.4 TEMPERATURE RISE OF HIGH STRENGTH CONCRETE

Portland cements are combinations of complex chemical compounds and their hydration
reactions have been the subject of considerable research. A simplistic description is that
Portland cement is essentially a mixture of four principal components, tricalcium silicate
(C 3 S), dicalcium silicate (C 2 S), tri calcium aluminate (C 3 A) and tetracalcium
aluminoferrite (C 4 AF) and small amounts of minor components such as alkali, sulphate
and magnesium oxide. These four main constituents are in their impure form in Portland
cement and are normally referred to as alite (C 3 S), belite (C 2 S), aluminate (C 3 A ) and
ferrite (C 4 A F). The relative proportions of these compounds in cement determine the
characteristics it will possess; these characteristics include the rate of gain of strength,
resistance to chemical attack and the rate and total quantity of heat evolved during the
exothermic hydration reaction.

3.4.1 Heat evolution during hydration

The rate at which heat is generated during hydration depends on the composition,
temperature, fineness, water/cement ratio and age of the cement and is of considerable
importance when concrete is poured in large volumes. Temperatures in the interior of
such large pours up to 60°C above ambient temperature have been recordecf"‘*\ A general
rule of thumb used by the construction industry is to expect a 12.5° C temperature rise for
every 100 kg/m^ of Portland cement used. In the case of high strength concrete where
cement contents in the region of 500 kg/m^ are used, this temperature can be a major
concern. Firstly, this will affect the ultimate properties of the hardened concrete, and
secondly the strain gradients, resulting from the thermal gradients within the structural
element when the concrete cools down to ambient temperature, may induce cracking.

47
C hapter 3 Literature R eview

3.4.2 Review of measuring methods.

There are three main established methods for measuring the heat output of cement during
hydration. The first method is based on the application of Hess’s law of constant heat
summation to the difference between the heats of solution of unhydrated cement and
cement hydrated for 7 and 28 days. This method is described in BS 4550:1978 and
ASTM C186-78 which are similar except for minor details. The test is relatively
inexpensive, but suffers from several chemistry related complications^"^’”^\ Its main
advantage is its applicability for long term measurements.

The second method is the adiabatic calorimeter method. This was developed to simulate,
in samples of concrete, the temperature rise in the centre of large masses of the same
concrete, where conditions approximate to adiabatic. Many researchers have judged this
to be a major advantage of this method, since it enhances the possibility of direct
application of the results for practical purp o ses^'D avey and Fox^^^’^ described their
calorimeter in which the temperature of the environment was controlled to follow closely
that of the sample it contained. Curves showing the temperature rise under adiabatic
conditions up to 3 days are shown in figure 3.17. The method can be used satisfactorily
up to 7 or 28 days, though the effect of slight cumulative errors becomes apparent.

Meissner^'^*^ has commented on the high accuracy that can be reached with adiabatic
calorimetry, especially in the early stages of the hydration process up to 7-days. Beyond
this time the accuracy decreases, but for the majority of practical purposes this does not
constitute a serious problem. Some typical values for mass concretes with different types
of Portland cement up to 28 days are shown in figure 3.18. This method is not applicable
to small masses and Forrester^'referred to its limited application for reaction rate
studies due to the multiple unknown activation energies. Bamforth^'^°^ used the same
principles as the adiabatic calorimeter and suggested a similar approach to measuring the
adiabatic heat generation of concrete.

48
 C hapter 3 L ii e r a i u r e K e v ie w

SOC
aisIG i^
amCtnfnt
7SC
Ais./G
nA MC
eM tH
T.
40 1^.

S4Ca
l 3.I6kamCtnenr.
47C
A Ls/6ifAMCtmenT.
O
SCaL
sJGA
k MC
lMtH
T.

X 40 51
H n ,,o t A rrco P lA C IN G .

Figure 3.17; Temperature rise in 1:2:4 concrete (w/c 0 .6 ) cured adiabatically

90
Type of P o r t la n o Ce m e n t
80 f-lARDEoaSi^-------

70
Normal

5 0

30'

20

0 3 7 74 21
Age (Days)

Figure 3.18: Temperature rise for mass concrete using different cements

49
Chapter 3 L iteratu re R eview

The third method is designed so that the heat evolved from the hydrating sample of
cement paste is rapidly conducted to a constant temperature sink, measurement being
made of the rate of heat output at the constant temperature. A conduction calorimeter,
which was originally developed by Tian in 1923, is used. In this, the temperature gradient
along the heat paths and their conductance may be used to determine the rate of heat
evolution in the specimen. Integration with respect to time gives the total quantity of heat
evolved at a given time. Forrester^”’^ outlined the development of a conduction
calorimeter based on that of Tians. Danielsorf'^^^ also described the principle and
operation of similar calorimeters.

3.43 Factors affecting the temperature rise of concrete

With any Portland cement, the amount of heat evolved is directly related to the amounts
of the clinker phases that have reacted, and thus depends on such factors as the cement
composition, particle size distribution, water/cement ratio and temperature.

• Composition
Attempts have been made by several researchers to derive factors representing the
contribution of each of the four main cement compounds to the heat evolved during
hydration^"^’"^’"'^^ Particular attention was given to the heat of hydration of C3 A, which
reacts with gypsum with the formation of ettringite and the liberation of about 624 J/g
This amount of heat should be added to the 8 6 6 J/g liberated during the
conversion of ettringite into stable monosulphate, totaling 1,490 J/g. The fact that not all
of the C3 A will react with gypsum an apparent heat of hydration of 1,350 J/g is generally
adopted. This value is in close agreement with the results of an analysis carried out by
Verbeck et al^'^^l

50
C hapter 3 L iteratu re R eview

Lerch and Bogue^'^*^ examined the heat evolution characteristics of hydrating Portland
cements using the method of heat of solution on unhydrated and completely hydrated
samples of the pure compounds; the values obtained were:

C3 A 865 J/g
C3 S 502 J/g
QAF 418 J/g
CzS 260 J/g

Woods et al performed a similar analysis on 13 hydrating Portland cements at ages of

3,7,28,90 and 180 days. They suggested that the most probable contributions of each
component to the total quantity of heat evolved at 180 days was:

C3 A 913 J/g
C3 S 505 J/g
C 4 AF 306 J/g
C2 S 220 J/g

These values indicate the relative contributions of the various compounds and may be
used to estimate the total heat evolution from cement. Similar results have been obtained
by other researchers^’^ but the agreement between the relative values for the
four compounds is far from close. The differences are particularly large in the case of the
alumina and iron compounds.

• Fineness
The reaction between cement and water takes place only at the surface of the solid
particles, and increasing the fineness of cement therefore increases its reactivity.
Keienburg^’^’’ states that the addition of gypsum during cement manufacture increases the
specific surface by 5 to 6 m^/kg per percent of added gypsum. Forrester^"’^ compared the
rate of heat evolution from a Portland cement ground to four grades of fineness.

51
Chapter 3______________________________________________________________________________________________________ L iterature Review

The specific surfaces of the four samples were 250, 350, 450 and 550 m^/kg. As expected
the maximum heat evolution from each sample were 209, 292, 376 and 502 J/g
respectively.

• Water/cement ratio.
Verbeck and Foster^*^^^ and Danielsson^*^^^ showed relationships between heat of hydration
of pastes containing Portland cement and the combined free water content. They both
established that a minimum amount of water is necessary, about 25%, for complete
hydration and heat evolution.

Danielsson^^^^) also demonstrated by the use of conduction calorimetry that at 20 °C there

was little variation in the total quantity of heat evolved up to the age of 2 days for pastes
of water/cement ratio between 0.25 and 0.5, but at the age of 30 days the total quantity of
heat evolved from the paste with water/cement ratio of 0.5 was approximately 50 %
greater than that with water/cement ratio of 0.25.

Parrotf'^^) conducted tests on various high strength concrete mixes to investigate the role
of water/cement ratio and its influence on heat of hydration. His conclusion was in
agreement with Danielsson, that the rate of heat evolved is reduced at low water/cement
ratios (figure 3.19). Aarsleff et suggested that the low heat evolved is due to the
greater amount of unhydrated cement compounds at water/cement ratios below 0.35.

• Temperature
Verbeck and FosteF’^^^ conducted tests on a selection of 27 Portland cements to
investigate the effect of initial temperatures on the resulting heat of hydration. They
found that an increase in temperature results in an acceleration of the early hydration rate
but decreases more rapidly after the maximum rate is achieved. Davey and Fox^^^’^ found
a similar relationship for their adiabatically cured specimens.

52
Chapter 3 Literature Review

Bamforth’s^'-‘” work on normal strength concrete investigated initial mix temperatures

between 5 and 35 °C, and S o u t s o s w o r k on high strength concretes also confirmed that

the maximum temperature rise reduces with increasing initial temperatures, as shown in
figure 3.20. Bamforth attributed this effect to the higher initial rate of hydration at higher
mixing temperatures which may have significantly modified the hydration process.

50

40

% 30
2
0>
01
ra
OJ
X 20

20 50 100

Age ( h o u rs I

Figure 3.19: Relationship between heat evolution and time (1:8)

53
Chapter 3 Literature Review

W ATER-ΠM ENT RATIO = 0 J 6

CEMENT CONTENT = 510 kg/m-*

50 (...) INDICATES CASTING TEMPERATURE

Ü
1 ^ 0

UJ
CE
Ü
cr
3(-
(30*Q
LU
Q.
2
LU
t-

1 2 5 10 20 50 100 200
TIME (hours).

Figure 3.20: The effect of casting temperature on the peak temperature rise of OPC
mixes at 0.26 w/b ratio

54
Chapter 3______________________________________________________________________________________________________ Literature R eview

3.4.4 The effect of binary blended cements

The rate of temperature rise in concrete can be substantially reduced by partially

replacing Portland cement with PFA or GGBS. This is most significant within the first
few days after casting, where they can reduce the risk of thermal cracking in Targe pour’
type construction. However, this is not always the case when using CSF.

3.4.4.1 Pulverised Fuel Ash (PFA)

Various reports^^^'^^*^'"^''^^'^^^) recommend the use of PFA for reducing the early age
temperature rise of normal strength concrete. Bamforths^’^®^ extensive study on the effect
of using PFA at levels up to 50% replacement, for concretes with a total cement content
of 400 kg/m^, showed that at a 15% replacement the final temperature of the PFA
concrete was marginally greater than the OPC concrete, but, high replacement levels
reduced the adiabatic temperature rise by up to 30% (figure 3.21).

Soutsos(^) reports a similar trend in high strength concrete where PFA at low replacement
levels increased the rate of temperature rise (figure 3.22), and at a 40% replacement level
a reduction of 18% was observed. This increase in temperature rise at low replacement
levels is normally attributed to the contribution of the pozzolanic reaction in the presence
of excess lime from Portland cement hydration. The increase in temperature provides
sufficient activation energy for the pozzolanic reaction to begin at early ages.
Temperature reductions at higher replacement levels result from the lower reactive
amounts of lime produced.

The nature of the PFA (Class C or F) also has an affect. Crow and Dunstan^'^"*^ found that
the adiabatic temperature rise for a 25% low calcium PFA (Class F) showed a reduced
temperature rise whilst concrete with 25% of high calcium PFA (Class C) produced as
much heat (at similar rate) as an OPC control.

55
 C h apter 3 Literature Review

TEMPERATURE (‘CJ
90

80

70

60

50

40

30

20
Cement content = 400 kg/m’.
Casting temperature = 15°C.
10

T:,V.E(t*1 ) KCUR

1 2 3 5 10 30 50 100 2 0 0 30C 5 0 0
( 120)
F igure 3 2 1 : Temperature rise for PFA binary m ixes

60
WATER-CEMENT r a t io = 0.26
CENtENT CONTENT - 510 k&An’

50
OPC

T 30
40% PFA
20% PFA
20

10

1 2 5 10 20 50 100 200
T IM E ( h o u n ) .

F igure 3.22: The aciiabatic temperature rise for PFA binary mixes

56
C hapter 3 L iterature Review

3AA.2 GGBS

In contrast to PFA, there is a substantial amount of literature on the effect of GGBS on

the temperature rise of concrete and GGBS has been successfully used to reduce the
temperature rise in large pours for many years^‘^^’^^®\

Bamforth’s study also included the effect of GGBS at replacement levels up to 75%
for normal strength concretes with a total cement content of 400 kg/m^ (figure 3.23). At
an initial mix temperature of 15°C the effect of a 25% replacement level was to
marginally reduce the early rate of hydration, but to increase the temperature rise after 8

days, indicating that the GGBS has a potentially higher level of hydration than Portland
cement. At higher levels of replacement the temperature was reduced both in the short
and long term indicating a reduction at 8 -days of up to 40% for the 75% replacement mix
compared with the control concrete. This suggests that the ratio of OPC to GGBS is
critical in determining the heat generating characteristics of the GGBS itself; the higher
the ratio, the greater the temperature rise attributable to the GGBS in the long term.
Similarly, Wainwright and Tolloczko’s^^^^ data indicate net temperature reductions of
approximately 10 °C could be obtained with a 70% GGBS mix for Grade 25 and Grade
45 concretes.

Coole^'^^) investigated the heat release characteristics of various OPC/GGBS blends,

measured under isothermal test conditions, by conduction calorimetry. He expresses the
importance of proper selection of GGBS due to different reactivities. He states that peak
temperatures higher than that reached by a comparable Portland cement can be expected
in pours with a minimum dimension of 3 m, unless the GGBS replacement is above 75%.
The higher level of replacement for GGBS is due to its own hydraulicity, which also
results in a significant increase in temperature rise at low replacement levels.

57
Chapter 3 Literature Review

TEMPERATURE (*CJ
90

80

70

60
75
50

40

30

2 0

HOURS

2 3 5 10 20 30 50 100 2 00 300 500

( 120)
F igure 3.23: Temperature rise for GGBS binary m ixes

60
WATER-BINDER RATIO = 0.26
BINDER CONTENT - 510 k&/m'
50

OPC

10% GGBFS

30
LU

30% GGBFS
60% GGBFS

LU

10

1 2 5 10 20 50 100 200
T IM E ( h o u r s ) .
(90)
Figure 3.24: The aidiabatic temperature rise for GGBS binary m ixes

58
C hapter 3 L iteratu re R eview

Soutsos^^"’ examined the effect of GGBS replacement in reducing the adiabatic

temperature rise (figure 3.24), 60% replacement reduced the temperature rise by up to
14%. He suggests this should be the minimum limit for reducing the adiabatic
temperature rise in high strength concrete.

Yurugi et ah'"^^ published data for tests carried out on mortars at water/binder ratios of
0.25; these show significant temperature reductions when the GGBS replacement level
exceeded 40%, which is consistent with Bamforths data for lower binder contents. This
substantiates Coole’s^*^’^ suggestion that significant variations in performance can be
obtained with GGBS from different sources. Tachibana et aP^ reported that with a 55%
GGBS replacement level, the adiabatic temperature rise was 50,4 °C, which was 24%
less than that for an equivalent OPC mix with a cement content of 540 kg/m^.

As part of a study on temperature effects of GGBS in high strength concrete, Mak and
investigated the semi-adiabatic temperature rise at varying GGBS contents of 0, 50
and 70% at a 0.3 water/binder ratio. They show that the temperature development is
progressively reduced with increasing GGBS content. As well as the reduction in
temperature rise, the dormant period also increased significantly with increasing GGBS
content (figure 3.25). These results are in agreement to those reported by Soutsos^^°\

Similarly, Sioulas and Sanjayarf'”^’ investigated GGBS contents of 30,50 and 70% at a
0.3 water/binder ratio. Their results show a progressive reduction in the peak temperature
with increasing replacement levels. In contrast to these findings, Kokubu et aP®®^ found
that the adiabatic temperature rise increased at GGBS replacement levels of 35% and
55% while a significant reduction in temperature was found at a 70% replacement level.

The Appleby Group^'-’*^^ have carried out a vast amount of work on the benefits of using
GGBS. They state that when casting small to medium sections up to 50% GGBS should
be used in the mix, with large sections a minimum of 70 % GGBS replacement is
recommended.

59
C hapter 3 Literature Review

50

uj 40

h p L oo/ .
3 30

HPL50/.

2 20
Ui H PL 70/.
H-
z 10

0 1 2 3 4 S 6 7
TIME AFTER CASTING (days)

Figure 3 2 5 : Temperature rise of GGBS mixes under semi-adiabatic conditions^“°®^

60

50
(1) C en tre
40
(2) S urface
30 (3) D ifferential

20

10

6 12 18 24 36 62 91 134 226 253 277

Hours from T h erm o co u p le cover

C a r d i n g t o n m a s s p ou r te m p e r a tu r e r e a d i n g s - T h e r m o c o u p l e 2 ( 5 m fro m e d g e )

60
(1) C entre
50 (2) S urface
(3) Differential
40

30
(1)
20
(2 )
10
(3)
0

_ 6. 12 18 24 . _36 ..6 2 9L 134 226 2.53 277

Figure 3 2 6 : BRE Cardington mass pour temperature readings

60
Chapter 3 Literature R eview

A recent study on temperature effects in a GGBS concrete was conducted at BRE

Cardington^‘^’\ A massive concrete slab floor measuring 6 8 m x 44 m x 1.2 m was cast,
using a 75% GGBS replacement level at 0.5 water/binder ratio. A maximum temperature
differential between the surface and the centre of the slab was specified as 20° C. These
results highlight the benefits of using GGBS to reduce the risk of early age thermal
cracking (figure 3.26).

3.4.43 Condensed Silica Fume (CSF)

The effects of CSF on the temperature rise of concrete are conflicting. Tank^’'^®^reported
greater heat of hydration with increasing CSF contents in pastes; he attributed the higher
heat evolution rate to the enhancement and acceleration of the C3 S hydration by the CSF.

Smeplass and Maage^'"'^ investigated the heat of hydration for a number of high strength
concrete mixes by means of the adiabatic calorimeter test. They found that at
water/binder ratios less than 0.4 the CSF decreases the peak adiabatic temperature. The
reduction in temperature with an increase in CSF content from 5 to 10% at a water/binder
ratio of 0.27 was 3.1°C. A similar conclusion was reached by Hellan(T^^\ in which he
states that at low water/binder ratios CSF gives very little contribution to the heat of
hydration. In agreement, Tachibana et ab^’^found that with a 0.25 water/binder ratio, the
adiabatic temperature rise for a 10% CSF replacement level was 60.2° C compared to
66.4° C for an OPC mix as shown in figure 337.

Soutsos^^^ however contradicts these findings and reported that at 10% CSF replacement
high strength mixes had higher peak temperatures than the control OPC mix (figure
3.28). His view is shared by de Larrard'"^^ who found that higher peak temperatures could
be attained provided that the CSF dosage is less than 10%. However, no explanations
were given for these observations.

61
Chapter 3 Literature Review

SF/C B B FS/C B Empirical e q u a t i o n

(%) (.%) (R egression)
ki (5 0 0 Q(t) = 6 5 . 4 ( 1 - e - ' ’“ ‘)

(U (D 10 0 Q ( t ) - 6 0 . 2 ( 1 - e - ’ ‘” ‘)
in @ 10 35 Q(t) = 5 5 . 9 ( 1 - e - ’ ” ")
0 10 55 Q(t) = 5 0 . 4 ( 1 - e - ' ” ^')
CJ 10 0 ( t ) = 3 8 . 5 ( 1 - e - ‘ “ “')
0 70
W /C B - 2 5 î f e .C B = 5 4 0 k g / m ’
<D
Q.
E
CD

(D
r:
(Z
<

Age (d)

F igure 3 2 7 The adiabatic temperature rise for GGBS binary and GGBS/CSF ternary
:

mixes

60
WATER-BINOER RATIO - 0.26
BINDER CONTENT - 510 kQ/rrf
50

10% CSF

LU
30

LU 20 OPC

2 5 10 20 50 100 200
TI M E ( h o u r s ) .

Figure 3 2 8 : The adiabatic temperature rise for a 10% CSF binary mix

62
C hapter 3 L iteratu re R eview

3.4.5 The effect of ternary blended cements

There is limited information available on the effect of ternary blends on the adiabatic
temperature rise of high strength concrete.

Tachibana et ah^’^investigated the adiabatic temperature rise of high strength concrete

made using ternary blends of GGBS and CSF (figure 3.27). They found that at 55% and
70% GGBS replacement levels in combination with 10% CSF, the adiabatic temperature
rise, as compared to CSF concrete, is reduced by 9.8P C and 14° C respectively. Yurugi et
ah'"^) found similar but more substantial effects; their tests on concrete with 70% GGBS
in combination with 10% CSF at a water binder ratio of 0.26 exhibited a temperature rise
around 30° C lower than that for an equivalent OPC mix.

Soutsos^^^^ investigated the role of two ternary blends, 54% GGBS + 10% CSF and 36%
PFA + 10% CSF at 0.26 water/binder ratios (figure 3.29). He showed that in both mixes
there was a significant reduction in temperature rise compared to the equivalent binary
mixes of GGBS and PFA. Similar reductions in temperature have also been reported by
Alshamsh’®’\ who tested a 50%GGBS+10%CSF cement paste at 0.35 water/binder ratio.
It is possible that the lower temperature rise is due to insufficient lime produced at very
early ages for the rapid reactivity of CSF to be effective.

More recently, work reported by Sioulas and Sanjayan^^®^^ contradict these findings.
Their results show that for a 45%GGBS+10%CSF ternary blend at 0.3 water/binder ratio
a peak temperature rise of 34° C is attained compared to 30° C for a binary 50% GGBS
mix.

The information in the literature on the temperature rise of high strength concrete on
ternary combinations is limited and clearly shows scope for further work in this area.

63
Chapter 3 Literature Review

60
WATER BINDER RATIO - 0.26
BINDER CONTENT - 510 K^/m*

50

OPC

54% GGBFS - 10% CSF

UJ 20 50% GGBFS

10

2 5 10 20 50 100 200
TIME (hours).

60
WATER-BINDER RATIO - 0.26
BINDER CONTENT - 510 kp/m'

50

Ü
UJ OPC
V) 40
iS
tr
o 40% PFA
Z
UJ 30
cc
ZD 36% PFA - 10% CSF

20

10

5 10 20 50 100 20C
nUC /Sniircl

Figure 3.29; Adiabatic temperature rise for ternary m ixes at 0.26 w/b ratio.
(a) 54% G G B S + 10% CSF

(b) 36% P F A +10% CSF

64
Chapter 3 L iteratu re R eview

3 5 STRENGTH DEVELOPMENT OF HIGH STRENGTH CONCRETE

One of the main characteristics of high strength concrete compared to that of normal
strength concrete is the more uniform and homogeneous microstructure. The high
water/binder ratio of normal concretes results in a porous microstructure, specifically
around the aggregate particles. The transition zone around the aggregate which is 20 to
10 0 |xm wide has a very different microstructure compared to that of the bulk matrix^*®®^
It has a porosity greater than that of the bulk paste, and thus leads to a poorer bond
between the aggregate and the cement paste. When CSF is introduced, and in particular in
high strength concrete, considerable changes in the microstructure of the transition zone
take place. Regourd et ah'®^^ and Aitcin et ah‘^^ observed that high strength concrete with
CSF was not as crystalised and porous as normal strength concrete, and all of the
hardened cement paste in the vicinity of the aggregate was occupied with amorphous and
dense calcium silicate hydrates. Also, direct contact was formed between the aggregate
and the calcium silicate hydrates rather than with calcium hydroxide as in normal
strength concrete. Scrivener et af”'^ quantified the interfacial microstructure and showed
that in high strength concrete with CSF, the porosity of the transition zone was practically
eliminated.

There are two other factors relating to compressive strength development that are
particularly important:

• The influence of the maximum temperature reached during hydration on the

strength
• The strength of in-situ concrete compared to standard laboratory cast specimens.

This section of the review focuses on the effects of standard curing (20°C) and in-situ
temperature rise on the compressive strength development, and highlights the methods
used for assessing the in-situ strength of high strength concrete.

65
C hapter 3 L iterature R eview

3.5.1 The effect of standard curing conditions on the strength development

The compressive strength development in concrete depends on the materials used, mix
proportions, efficiency with which the mix is compacted into a mould, as well as the
temperature, humidity and time of curing. Curing is generally carried out at 20° C, and
the strengths obtained are referred to as the standard strengths and BS 12 classifies
cements on the basis of the level attained. Aitcin and Riad^^^^^ observed that the 28 day
strength of specimens cured under standard conditions gave a fair representation of the
actual strength when the water/binder ratio was below 0.30. In contrast, Carrasquillo and
Carrasquillo reported that the strength under field conditions was lower than that of
the standard cured samples. The use of CSF, PFA and GGBS also significantly affect the
strength development under standard curing conditions.

3.5.1.1 Binary blends

• CSF
It is generally accepted that CSF improves the compressive strength of high strength
concrete by enhancing the transition zone. Soutsos’s standard cured strength results
indicate that at a water/binder ratio of 0.26, an approximate strength ceiling is observed at
about 10% CSF replacement level, as shown in figure 330. The results also show that
the 28 day strength improvement obtained at 10% CSF replacement is only 3.2%, while a
significant increase was obtained at 56 days. Sarkar and Aitcin reported a similar
observation and this was attributed to a delay in the dissolution of CSF at low
water/binder ratio. In contrast to these results. Read et al^^^^ have reported that at CSF
replacements of 8 and 1 2 % high early age strengths were achieved (up to 28 days), at
0.27 water/binder ratio. They state that this benefit is lost at ages beyond 28 days, a
similar observation was also reported by Djellouli et al

Yogendran et al investigated the effect of different CSF replacement levels at 0.34

and 0.28 water/binder ratios. For concretes at a water/binder ratio of 0.34, the strength
increased up to 15% replacement and subsequently decreased, (figure 331). However,
for CSF concretes at 0.28 water/binder ratio, the strengths at 28, 56 and 91 days
decreased compared to the control concretes (figure 332).

66
 Chapter 3 Literature Review

15 0
15% CSF
140 WATER-BINDER RATIO - 0.26
10% CSF
BINDER CONTENT - 5 1 0 kg/m
13 0

120 ■ 5% C S F ■

110 OPC
^ 100

I 90
Z

b 70
I 60

M 50
40

30

20

7 28 56 91 18 0
AGE (days).

Figure 330: Strength development of standard cured CSF concretes with a water/binder
ratio of 0.26^^^

70

10500

a
2

9000 - I
I

C I 9000
50
I
50

7 dor»

40 '6000 7500
0 5 10 20 25 20
PefCtnKJQe o( Ct*n*nl flfo k jce o by S * c o Fvm «
PfTceoiog» ol Cfmrnl RfoWcrO by Sil<0 Fxttt

F i g u r e 3 J I: C o m p r e s s i v e strength o f Figure 3 3 2 : C o m p ress iv e strength o f C S F

C S F c o n cret es with w a t e r /b in d e r ratio c o n c r e t e s w i t h w a t e r / b i n d e r ratio
of 0 3 4 '" ':' o f 0.28^ '°-’

67
Chapter 3 L iteratu re Review

• PFA
In general the main contribution of PFA to the strength of concrete, under standard curing
conditions, occurs between 28 and 90 days As part of Soutsos’s^^®^ investigation into
the effect of 10,20 and 40% PFA replacements on strength, he found that at a
water/binder of 0.26, even at low replacement levels the PFA concretes exhibited
significantly lower rates of strength gain at early ages, but considerable strength increases
beyond 28 days (figure 333). However, these mixes do not equal the strength of OPC
concrete even after 180 days.

• GGBS
The effect of GGBS on the strength development under standard conditions gives a
reduction in concrete strength at one day, but an increased gain from about seven days
onwards, depending on slag reactivity and content This effect has been confirmed by
Soutsos^^^^ at 10, 30 and 60% GGBS replacements. The strengths reduced with increasing
GGBS content, but show an accelerated strength g^in after 7 days. The data indicates that
in the 10%GGBS mix there is a reduction in strength between 28 and 91 days (figure
334). No explanation was given for this. Parrott observed reduced 28 day strengths
with GGBS concretes, but at 90 days all mixes achieved strengths of about 100 N/mm^ at
water/binder ratios of 0.26 to 0.2. However, Mphonde^^^^^achieved similar strengths at 28
days with a superplasticised 50% GGBS concrete at 0.27 water/binder ratio.

3.5.1.2 Ternary blends

Djellouli et al^^^^^ investigated the strength development of a 30%GGBS+10%CSF
ternary blend at 0.27 water/binder ratio. They reported low early age strengths (up to 5
days) compared to a 10% binary CSF mix. However, between 28 and 105 days they
found the strength gain was greater than the control OPC mix. Similarly, Soutsos^^^^
investigated the strength development of two ternary mixes of, 36%PFA+10%CSF
(figure 335) and 54%GGBS+10%CSF (figure 336) with water/binder ratio of 0.26. The
results show an improvement in strength compared to the equivalent binary PFA and
GGBS mixes but comparable strengths to OPC concrete at 56 days.

68
 Chapter 3 Literature Review

150

140

130
20% PFA
120

110 OPC

80
10% PFA

40% PFA

40

30
WATER-BINDER RATIO » 0.26

BINDER CONTENT - 5 1 0 kg/m

28 56 91 180
AGE (days).

Figure 3 3 3 : Strength develooment of standard cured PFA concretes with a water/binder

ratio of 0.26

150

140 W ATER-BINDER RATIO - 0.26

BINDER CONTENT - 510 kg/m
130

120 OPC
10 30% G G BFS
00

E 90
10% G G BFS 60% G G BFS
80

70
g
cc 60

50

40

30

20
10

0
7 28 56 91 180
AGE (days).

Figure 3 3 4 : Strength development of standard cured GGBS concretes with a

water/binder ratio of 0.26

69
Chapter 3 Literature Review

150

140

130
10% C S F
120 OPC
o-
110

1z 90
36% PFA & 10% C SF
•40% PFA
b 70

i 60
50

40

30

W ATER-BINDER RATIO - 0 .2 6
BINDER CO NTENT - 5 1 0 kg/rrr*

28 56 91 180
AGE (days).

Figure 3 3 5 : Strength development of standard cured PFA/CSF concretes with a

water/binder ratio of 0.26^^^

150

140 W A T E R - B I N D E R R A T IO - 0 . 2 6
B I N D E R C O N T E N T - 5 1 0 k g /m
130 10% C S F
120
OPC
10

00
"E 90 60% G G B F S

80
54 % G G B F S & 10% C SF
70

60
Π58% G G B F S & 5% C S F
50

40

30

20
10

0
7 28 56 91 180
AGE (days).

Figure 3 3 6 : Strength development of standard cured GGBS/CSF concretes with a

water/binder ratio of 0.26^^^

70
Chapter 3 L iterature Review

3.5.2 Temperature development in high strength concrete.

In the overall context of concrete performance the early age properties are extremely
important, one reason being because the long term performance may be governed by the
early age in-situ temperature development. These temperatures are governed by many
factors, in addition to those arising from the mix constituents, such as the size of a
structural element, the thermal conductivity of the concrete, the insulation properties of
formwork, the casting temperature and the prevailing ambient temperature^’'^’^''^'

One of the earliest measurements of in-situ temperature development of high strength

concrete was reported by Aitcin et in 1984. They found that the maximum
temperature within a Im square column was 6 8 °C. The mix had a total binder content of
530 kg/m^, SOkgW of silica fume and a water/binder ratio of 0.25. Christie et
reported temperature measurements in columns 800 x 800 mm cross section containing
silica fume with a water binder ratio of 033. The maximum temperature at the centre was
approximately 65° C. Yuan et ab''*®^ reported maximum temperatures in excess of 80° C in
1800 X 1800 mm blocks cast containing PFA with a water binder ratio of 0.28.

It is amply evident that temperature rises above 6 CP C are common in high strength
concretes under either adiabatic or semi adiabatic conditions. The development of high
in-situ temperatures result in a complex combination of effects that have many
implications for the development of in-situ strength and other engineering properties.

3.53 The effect of in-situ temperatures on the strength development.

3.53.1 Temperature Matched Curing

Investigations on the in-situ strength of normal strength concrete date back to the early
1900s, and much of the previous research undertaken over a period of 70 years has been
periodically reviewed by various authors^"'^"'^'^^\ The work primarily focused on
differences between in-situ strengths and standard cube strengths based on sampling and
testing procedures and temperature development. The development of high in-situ

71
Chapter 3 L iteratu re R eview

temperatures means that both the rate and extent of cement hydration of the in-situ
concrete would differ significantly from those of laboratory specimens cured at standard
temperatures.

In general, the greater in-situ temperatures tend to accelerate early age hydration leading
to enhanced early in-situ strength but may be detrimental to long term strength. The
effects are likely to be greater and potentially more critical in high strength concrete. One
of the ways that this can be assessed is by using temperature matched curing (TMC). A
set of reference cubes in a curing tank follow the same temperature as the in-situ
concrete, with thermocouples placed in the structure controlling the temperature of the
curing water, thus allowing a more realistic estimation of the strength of the actual
structure. This has obvious advantages for determining formwork removal times and
verifying suspect cube or cylinder strengths.

Wainwright and Tolloczo^’^°’ proposed a TMC system, which was too large and complex
to use on a construction site. Harrison^'^’^also proposed a similar system. It was not until
Blakey^'^^) who introduced a portable system for use on site that TMC gained serious
practical recognition. In the UK, interest in TMC led to a publication of a British
Standard (DD92: 1984^^^) which details the specific requirements of a temperature
matching system. The portable unit shown in figure 337 was designed to meet the
requirements of DD92 for use on site. Basically, cubes are made from samples of the
concrete placed in the structural element and a glass or steel plate is placed on the
exposed cube surface. The covered cube moulds are then submerged in the curing bath.
The water in the bath is matched in temperature to a pre-selected location in the in-situ
element by means of a heating element situated in the curing bath. Temperatures are
measured by platinum resistance thermocouples housed in stainless steel probes, one
positioned in the bath and the other in an insertion probe in the in-situ concrete.

72
t^Hapcer J L ite r a tu r e R e v ie w

(3) Pr i nc i pl e o l the

&

SI

lb) Con (roller/Recorder

Ic) W n ie r B . ii h

Figure 3 3 7 : Temperature matched curing device

73
Chapter 3 Literature Review

25
bvIQW # u r f # c« o f # 3 0 0 Id*
I nmu I # t e d « t * « l f o r w n ^ o r k

C fn trt of cub# cured a 1o n g #!d

20 ■ • — ■ AmbI#nI

>5

10

o
0 10 20 30 60 70

Time a f t e r c* * 11 n g - hour*

Aver*o* * tl-*ngth of J cube* - K / mr a *

T l x ! a f te r c o atin g
Curing avthod
19 h 26 h 67 h

Cube* cured alongside 1. 6 " 3.9 13.3

Tempera tu re-eiatch cd 3.6 7.1 16.0

cube* to a poin t 50 «**
below surface of the
wall

ThI• do* * n o t ** t I * f y * 2 N/»t*^ s tr ik in g c r i te r io n for v e r t i c a l form *.

Figure 338: Concrete wall results

74
C hapter 3 L iterature R eview

In normal strength concrete, Harrison^^^^’ carried out tests on a 300mm wide concrete wall
to investigate the effect of in-situ temperature on the strength development of normal
strength concrete. The wall was cast in insulated steel formwork and showed an increase
in temperature of around 17 °C some 23 hours after casting. Results showed that TMC
cubes were approximately twice as strong as the standard cured cubes at 19 hours wifti
the difference subsequently reducing (figure 338).

Connelh'^^) also reported results from a wall pour on the M63 motorway in Stockport.
Investigations were carried out to determine the temperature cycle and early age strength
development of the concrete. The wall pour dimensions were 73m x 4m x 600mm. A
C40 grade of concrete was used with an OPC content of 228 kg/m^, GGBS content of
152kg/m^ and 20mm crushed limestone. The concrete had a water/cement ratio of 0.43.

Tests were carried out on cubes by different laboratories in order to obtain a comparison
of the results. Twelve 100mm cubes were cast by Frodingham Cement and were
immediately covered by a top plate and immersed in the TMC bath. A further twelve
150mm cubes were cast by Sandbergs staff, six of which were cured on site and the
remainder cured in the laboratory at 20 °C. Twelve 150mm cubes were also cast by
Tarmac Topmix staff and cured in their laboratory; for which no temperature record is
available. The temperature profiles are shown in figure 3 3 9 and the compressive
strength results are plotted on figure 3.40.

As can be seen, the concrete structure reached a peak temperature of 343 °C at around 36
hours after casting. The strength data can be used to demonstrate the practical
significance of TMC by using CIRIA report No.l36^^^^\ This states that formwork
striking times are based on two strength levels, 2 and 5 N/mm^. The 5 N/mm^ level is
required to resist frost and mechanical damage. From the strength development curve for
the TMC cubes and using the values in CIRIA 136 recommendations, it can be seen that
the side forms could have been released after 16 hours for the 5 N/mm^ requirement and
less than 12 hours for the 2 N/mm^ requirement. The compressive strengths of the site
cured cubes and Sandberg laboratory cured are 10-25% and 50-60% respectively of the

75
Chapter 3 L iteratu re R eview

TMC cubes. It is worth noting that the site cured cubes were air cured and followed
almost exactly the same ambient temperature cycle, resulting in significantly lower
strengths.

As reviewed earlier, the use of supplementary cementing materials changes the rate of
temperature rise and the peak temperature attained during hydration, and therefore the
strength of the concrete, Bamforth’s^*^^ extensive study on the effects of cement
replacement materials, such as PFA and GGBS in reducing the temperature rise of
normal strength concrete also highlighted the effect of the temperature cycle on the
strength of concrete.

76
Chapter 3 L iterature Review

40

35

30
U
25
2
3
20
0)
CL
15
d)
I-

10
Centre
5
TMC tank
0
0 10 20 30 40 50 60
Time (hours)

( 156)
Figure 3 3 9 Temperature profiles

25

E TMC
E Tarmac
z
Sanbergs
Site
{/)
d>)
(/)
(/)
2

E
o
O

0 10 20 30 40 50
Age (hours)

Figure 3.40 Compressive strength results from Connelh'^^'

77
C hapter 3 L iteratu re R eview

The strength development of both standard cured and heat cycled concrete was measured
using 100mm cubes. Standard curing was in accordance to BS 1881; the temperature
cycled cubes were sealed in plastic bags and stored in a temperature controlled cabinet.
The imposed temperature regime was calculated from the measured adiabatic temperature
rise for each individual concrete to match the temperature cycle occurring at mid-height
in a 2.5m deep pour.

For Portland cement concretes the effect of the temperature cycle increased early age
strength by up to 50% compared to standard cured cubes but, at 28 days by up to 20%.
When GGBS was used to replace 50% of the Portland cement, the heat cycled strength
was marginally higher, of the order of 5%, than the standard cured strength at 28 days
(mix 2). The effect of the temperature cycle also reduced the strengths at ages beyond 56
days, as shown in figure 3.41.

The use of PFA to replace 30% of the Portland cement accelerated the early strength
development of the heat cycled strengths at 28 days which were 2 0 % higher than the
standard cured value, but again resulted in lower strengths beyond 56 days (figure 3.42).
The results show in general that the effect of subjecting concrete to a heat cycle
accelerates the early strength gain but impairs the long term strength development.
Similar effects have been reported by Kleiger^'^^^ for curing at constant temperatures.

In a similar study on normal strength concrete, Wainwright^’^®^used GGBS replacement

levels of 50 and 70 %. Two basic mixtures were used with nominal cement contents of
300 and 450 kg/m^. These mixtures were of grade 25 and grade 45 as specified in CP
110. The results showed that with standard curing the grade 45 concrete at 7 days the 50
and 70% GGBS concretes reached 61% and 41% of the strength of the OPC concrete
respectively, and at 28 days the values were 97% and 70%. A similar trend was found for
grade 25 concrete (figure 3.43). The temperature matched grade 45, 50% GGBS concrete
cubes were found to exceed the strength of the OPC concrete by 9% after 2 days and after
7 days both the concretes containing GGBS exceed the strength of the OPC concrete
(figure 3.44).

78
C h a p te r 3 L ite ra tu re R e view

MIX CASTIN G
N o. TEM PER A TU RE
Mu 3

300
Mu 1 400
500
Mu 2 400
1.6

Mû 4
H
X
Cu
O
0
1

0.6
3 7 28 56 180 365
AGE (days).

Figure 3.41 Relationship between standard cured and TMC strengths for concretes with

50% GGBS

MIX B INDER C A ST IN G
N o. C O N TEN T TEM PER A TU RE
kg/m’
3 00
400
5 00
g 400
s

Mu 2

Mu 4

0.6
3 7 28 56 180 365

Figure 3.42 Relationship between standard cured and TMC strengths for concretes with

30% PFA

79
C h a n te r 3 L ite r a tu r e R e v ie w

N OMMi k L C U R I N O
ORAOC 4 5 C O N C R E T E

«1 lO V C I H t * V I •*

L O G TI ME S I N C E C A S T I N G ( D A T S )

norm al CU R IN G
GRADE 2 5 CONCRETE

(b)

IO C TI ME S I N C E C A S T I N G ( O A T S )

COMPRESSIVE STRENGTH vs TIME FOR STANDARD 20*C CURING

Figure 3.43: Compressive strength development of Grade 25 and Grade 40 concrete

subjected to standard curing

80
C h a p te r 3 Literature Review

A O U I A T I C CU niM O
OAADC 4 * C O M C M C n

• • ••«CtM tATI

(a)

LO G TIME t I M C E C A S T I N G ( O A T S )

A D IA B A T IC CU BIN G
OAAO C 2 ft C O N C B C T C
■ 10 0% C I N » « * f ■■ »
■ii«e% C(M iA V t —- e —

(b)

L O G T I ME S I N C E C A S T I N G ( D A T S )

CO M PR ESSIV E STRENGTH v s TIME FOR ADIABATIC CURING

Figure 3.44: Compressive strength development of Grade 25 and Grade 40 concrete

subjected to temperature matched curing

81
C hapter 3 L iteratu re R eview

It can be seen from this investigation that the concretes containing GGBS are far more
sensitive to increases in temperature than equivalent OPC concretes. The rate of gain of
strength at early ages for GGBS concrete increases far more with an increase in curing
temperature than an equivalent OPC concrete and therefore TMC is even more important
for a realistic assessment of in-situ strength than for plain OPC mixes.

In a study conducted by the Appleby Group^*^^^ during the construction of two large
unreinforced foundations to support the Hulme Arch Bridge in Manchester, a 70% GGBS
replacement level and a cement content of 365 k g W at w/c 0.5 was used. Table 3.6
shows the comparison of TMC and standard cured results from a trial block. The TMC
strengths were marginally higher at 28 days than the standard cured cubes, but are lower
at 56 days. These results follow the same trend as the results obtained by BamfortH*^^ in
his investigation of a 75% GGBS mix, as illustrated in figure 3.45.

As part of a study to assess the in-situ properties of high strength concrete, Pricé^^^
conducted large scale structural testing of a 15m x 15m slab with a depth of 250mm. A
Clio grade high strength CSF concrete was used with a free water/binder ratio of 0.24.
The maximum temperature rise recorded in the slab was 31 °C. The high strength
concrete was characterised by an extremely rapid development of in-situ strength, with
temperature matched cubes giving values of 100 MPa in less than 24 hours. This rapid
strength development led to early removal of formwork and falsework. Standard cured
cubes gained strength at a slower rate than the TMC cubes but achieved a mean 28 day
strength of 119.5 Mpa (figure 3.46). No long term data was reported in this case.

Mak and conducted tests on five high strength concretes to establish the
relationship between TMC and standard cured conditions. All mixes had nominal binder
contents of 500 kg/m^ and water/binder ratio of 0.3. Levels of cement replacement up to
70% GGBS were used. The experimental procedure adopted was very similar to the one
used by Bamforth^'^"\ They found that when compared to standard temperature
conditions, high hydration temperature generally increased strengths at early ages.

82
Chapter 3 L ite r a tu r e R e view

70
60
A p p leb y
50 ■Standard
cn
c
V 40 •TMC
s •Standard
0)
> 30
(/) •TMC
in
(U 20 Bam forth
a.
E
o 10
Ü
0
0 Age (days)
60

Figure 3.45 Comparison of strength results obtained by Bamforth ^*-°’and

Wainwright

120

nj
CL
z 100
I

b 80
z
LU
cc
t—
in
LU
> tem perature - m atched
in
in 40 s t a n d a r d cu red
LU
cn
CL
S
o
o

0 5 10 15 20 25 30
AGE - days

Figure 3.46 Compressive strength results obtained by Price (i«))

83
Chapter 3 Literature Review

100

S. HPLOO-S23 -

80
X HPL50-TMC
J—
CD
z
LU
CC HPL50-S23
60
k
UJ
>
Ui
(A
X 40
LU
o_
5
O
Ü
20
0 20 40 60 80 100
AGE (days)

Figure 3.47: strength development of a 50% GGBS mix at 0.3 w/b ratio

Figure 3.47 shows the strength development of a 50% GGBS mix (HPL 50). The 7 day
TMC strengths were within 5% of their 28 day strengths for mixes with 50 and 70%

GGBS. In contrast, at standard temperature (20°C), the 7 day strengths of the two
standard cured cylinders were approximately 30% lower than their 28 day strengths for
the two GGBS mixes. More importantly the authors discovered that the 7 day TMC
strengths for the two GGBS mixes actually exceeded the 28 day strength of cylinders
cured at standard temperatures. The impact of high hydration temperature on compressive
strength reduced markedly with age. At 28 days the increase in strength of TMC

cylinders over standard cylinders for all the mixes did not exceed 10%. It is interesting to
note that the results obtained in this investigation follow a similar trend to those reported

by Wainwright^'^*^’ and Bamforth^'-"' for GGBS normal strength concrete mixes.

In a study to investigate the effect of CSF on the in-situ strength development, Mak and

Torri^'^-’ found that high in-situ temperatures of up to 70 °C significantly increased the 7

day strength of a 10%CSF high strength concrete, compared to standard and sealed cured
specimens. An increase in compressive strength of the TMC specimens between 28 days

84
C hapter 3 L iteratu re R eview

and 1 year was only 5 MPa, compared to 10 MPa for the standard and sealed cured
specimens.

Similarly, Price and Hynes^'®’^ investigated the in-situ strength of a CSF high strength
concrete. In order to assess the development of in-situ strength, two trial columns were
constructed in 18mm plywood formwork with internal dimensionsof Im x 0.5m x 0.5m.
Two concrete mixes were used, one with a granite aggregate and the other with a gravel
aggregate. Temperature matched cured cubes were cast and placed in an oven whose
temperature was controlled by a temperature probe in the column. Cores were also taken
from the columns for strength assessment. The results from this investigation are shown
in figure 3.48. It is interesting to note that although the early age strength development is
high there is little strength gain after 28 days, but between 28 and 90 days the strength
gain of the standard and sealed cured cubes is greater than the TMC specimens. It can
also been seen that the ultimate strengths reached by the granite aggregate mix exceeded
that of the gravel mix.

It is important to note the limitations of the information found in the literature on

temperature matched curing. The majority of the published work focuses on binary mixes
of CSF and GGBS, but no work was found for PFA binary mixes or ternary mixes of
GGBS/CSF and PFA/CSF. Further work is therefore needed to ascertain the TMC effects
of these blends in high strength concrete.

85
Chapter 3 Literature Review

120

100 -

? Tomp.Mnl.
w 40 Coroi€CS)
Standard cure
O 20 Seat cure

0 14 28 42 56 70 98
(a) Age: days

120

100

S 80

o) 60
Terrp f/ d
w 40 Core{ECS)
Standard cure
Ü 20 - j Seal cure

0 14 23 42 56 70 84 98
(b) Age: days

Figure 3.48: Strength development of high strength concrete mixes

(a) Granite aggregate
(b) Siliceous gravel agrregate

86
C hapter 3 Literature R eview

3.5.4 Proposed explanations for the strength development of temperature matched

cured concrete

It is evident from the work presented that despite the early acceleration in strength
development, curing at high temperatures and TMC generally leads to lower long term
strengths. The reasons for this are however not very clear, and possible causes that have
been put forward by various researchers are conflicting.
Bamforth^*^°^ suggested that this effect may be due to a fundamental change in the
hydration products formed.

Verbeck and Helmuth^'^^^ postulated that the strength reduction could be caused by the
formation of a very dense and impermeable shell around the hydrating cement grains
resulting from the accelerated hydration of high temperature curing. This reduces the
extent of hydration and restricts the diffusion of hydration products, resulting in a non-
uniform distribution of hydration products in the interstitial spaces between the cement
grains.

Based on microstructural evidence, Kjellsen et ab^“ ^ suggested that the strength loss of
OPC mortars under high temperature curing (50°C) is consistent with the formation of a
coarse pore structure compared to mortars cured at standard 20° C. A similar view was
expressed by Ikabata^’^’^who showed that when cured at 65° C, the hydration products in
a blended cement paste are fundamentally different to those of paste cured at 20° C (in
terms of porosity and pore size distribution).

In contradiction to the above findings Laplante and Aitcin^'^^ state that the loss of
strength in high strength concrete is not governed by the physical and chemical changes
in the hydrated cement paste but is more a problem involving paste/aggregate interaction.

87
C h apter 3 L iteratu re R eview

Mak and have explained their results of lower ultimate TMC strengths at 1 year
as attributable to self-dessication effects. However, Price and Hynes^^^’^ state that on the
account of the low water/binder ratio high strength concrete is relatively impermeable to
water. They stress that temperature matched cubes are sealed against moisture loss or
gain. The cubes experience a similar moisture condition as well as temperature history to
that of the in-situ concrete.

These explanations are somewhat contradictory, and show scope for further investigation.

88
C hapter 3 L iteratu re R eview

3.6 Conclusions

This literature review has focussed on key properties of high strength concrete, namely
the fresh properties, the temperature rise and the corresponding effects of temperature on
the strength development characteristics.

The review has highlighted that despite the amount of research conducted in these areas,
some uncertainties and conflicting arguments still exist. Most of the studies have been on
mixes with binary blends of OPC with PFA, GGBS or CSF, but the use of ternary
blended cements (OPC+CSF+PFA or OPC+CSF+GGBS) have been shown to have
beneficial effects on the rheological properties, temperature rise and compressive strength
development of high strength concrete. However the amount of information available on
such ternary cement blends and their effects on the properties of high strength concrete is
limited.

The questions which have been be addressed in this research programme on high strength
concrete are:

• What effects do binary and ternary blended cements have on the fresh properties?
• How effective are binary and ternary blended cements in reducing the semi-
adiabatic temperature rise?
• How do binary and ternary cements influence the in-situ strength development?
• Why do temperature matched cured cube generally give lower long term strengths
when compared to standard cured cubes?

89
Chapter 4 A im s and scope o f in vestigation

CHAPTER 4

AIMS AND SCOPE OF INVESTIGATION

4.1 Introduction

The previous two chapters have shown that there has been much research and
development of high strength concrete. However, there are still many areas where
research is needed to ensure economical and proper application of the material.

The literature review has shown that the use of CSF is considered essential for producing
high strength concrete, but also the use of PFA and GGBS can enhance the fresh
properties and reduce the temperature rise during hydration. The majority of applications
have been in using these materials as a binary blend with Portland cement, although some
limited use has been made of ternary blends. However, the effect of these blended
cements on the workability, temperature rise and strength development has not, however,
been thoroughly investigated.

The research presented in this thesis is a continuation and extension of previous research
at UCL(^\

The overall aim was to investigate the role of blended cements in high strength concrete
so that their use can be extended beyond current applications.

In this chapter the main points from the literature review are summarised, followed by the
specific aims and scope of the research.

90
Chapter 4 A im s and scope o f in vestigation

4.2 Aims

4.2.1 Fresh properties of high strength concrete

As mentioned earlier, high strength concrete is characterised by low water/binder ratios

and high cement contents, and the use of superplasticisers is therefore essential to provide
adequate workability. The use of PFA and GGBS is reported in general to cause a
reduction in the amount of water required for a given degree of workability

It has been well documented that the use of CSF in high strength concrete reduces the
workability as a result of an increased water demand and requires a consequent addition
of superplasticiser^"^'"'''"^). However it has been shown that above a certain
superplasticiser dosage CSF has a water reducing effect This has been attributed
to the ultra fine spherical particles of CSF being sufficiently dispersed to act as a
workability aid. Furthermore, Gjorv has stated that with increasing amount of CSF the
plastic viscosity is reduced, while the yield stress is almost unaffected.

Very limited amount of information is available on the workability of mixes with ternary
blends; some tests have been performed by Soutsos^^\

Many researchers have stated that the slump test is inadequate for measuring the
workability of high strength concrete, and it should be expressed according to classical
rheological models such as those of Bingham or Newton. The general Tattersall’s two
point test apparatus has been well documented (^5 .76.90.134)^ but, the results from tests carried
out by different researchers on high strength concrete with this are somewhat limited, and
in some cases conflicting.

91
C hapter 4 A im s an d scope o f investigation

Therefore the present work aims to investigate the effect of binary and ternary blended
cements on:
( 1 ) the superplasticiser dosage requirements at a constant slump of 200 ± 20 mm,

(2) the fresh properties of high strength concrete by using Tattersails two-point
workability test, which measures the Bingham, constants of yield stress and plastic
viscosity.

4.2.2 Temperature Rise

The literature review showed that the high cement contents in high strength concrete can
lead to high temperature rises during hydration. This can lead to thermal cracking and
also has implications for the hardened properties of the concrete. As mentioned in section
3.4.4, the use of supplementary cementing materials can significantly reduce the rate of
temperature rise, but there is limited information on the effect of ternary blended cements
on the temperature rise of high strength concrete.

Therefore, the aim was:

To measure the semi-adiabatic temperature rise of high strength concrete mixes during
the first few days after mixing.

4.23 Strength development

The development of high in-situ temperatures in high strength concrete means that both
the rate and extent of cement hydration of the in-situ concrete would differ significantly
from those in specimens cured at standard temperatures. This can lead to enhanced early
in-situ strength development, but either retarded or regressed long term compressive
strengths

92
Chapter 4___________________________________________________________________________________ A im s an d scope o f investigation

Suggested possible causes for this behaviour have been put forward by various
researchers, which are in some cases conflicting. Also, the data available are limited and
in particular no data are available on ternary blended cements.

The aims, therefore, were:

1. To measure the effects o f typical early age in-situ temperature cycles on the short and
long term strength development of high strength concrete, and to compare this with
the strength development o f standard, 20 ° C, cured specimens.

2. To provide a possible explanation of the results obtained in (1) by examining the

microstructure o f the cement paste fraction of the concrete, specifically the porosity
and determining the calcium hydroxide content.

4.2.4 Dielectric Properties

During the latter part of the research programme, a novel technique to determine the in-
situ strength development became available and was used in the investigation for
evaluation purposes in high strength concrete. This technique, developed at the Technical
University of Delft in the Netherlands, is based on measuring the dielectric constants of
the concrete, i.e. the permittivity and conductivity and relating them to strength.

4.3 Scope of Investigation

4.3.1 Mix Variables

The main variables investigated in this testing programme are water/binder ratio and
levels of cement replacement by PFA, GGBS or CSF in both binary and ternary
combinations. Three water/binder ratios i.e. 0.30, 0.26 and 0.20, were investigated. These
cover the range of values for most high strength concretes, and enabled direct comparison
of the results from the previous programme at

93
Chapter 4 ___________________________________________________________________________________ A im s an d scope o f investigation

Table 4.1 summarises the aims and scope of the investigation as outlined in sections
4.32 to 4.3.5.

4.3.2 Fresh Properties

The fresh properties of high strength concrete were measured by using Tattersall’s two
point test, which gave the Bingham constants of yield stress and plastic viscosity. Since
the slump test is the most widely used field and laboratory test for assessing the fresh
properties of concrete, it was therefore used to enable comparisons with the two point test
measurements and with results reported by other researchers. According to BS 1881 Part
102(176) ii^g slump test is invalid for collapsed slump mixes i.e for slumps in excess of
175mm, and the flow table test should then be used. In contrast, ASTM C 143-98^^^^^ has
no upper limit on slump and does not specify use of the flow table test. Consequently
many researchers worldwide, report slumps well in excess of 2 0 0 mm and sometimes
over 250mm and there is widespread data in the literature on high (>200mm) slump
mixes.
In the present study it was difficult to carry out both the slump and flow table tests
alongside the two point tests due to the lack of technician support available and slump
was considered suitable as the sole single-point test.
It is also worth noting that despite the increasing acceptance of the flow table test on
construction sites utilising high strength concrete, Tattersall^^"^^ has widely criticised this
test as well as the other single point tests as providing an inadequate measure of
workability

The variables in the concrete compositions investigated are shown in Table 4.1.

94
Chapter 4 ___________________________________________________________________________________ A im s a n d scope o f investigation

4.3.3 Temperature Rise

Semi-adiabatic temperature rises were measured for all mixes at 0.30, 0.26 and 0.20
water/binder ratios. The levels of cement replacement investigated are shown in Table
4.1.

4.3.4 Strength development

Three water/binder ratios of 0.20, 0.26 and 0.30, were used to determine the contribution
of supplementary cementing materials to the strength of concrete and also whether their
contribution is affected by the water/binder ratio.

The concrete cubes were subjected to a temperature cycle to simulate in-situ conditions
and duphcate specimens were cured under standard conditions according to BS 1881.
The temperature cycles were determined from the semi-adiabatic temperature rise data.
In order to assess the porosity and determine the calcium hydroxide content the cement
paste fraction of the concrete was used and a series of mixes were investigated at a
water/binder ratio of 0.26. The porosity was measured by using the alcohol resaturation
technique and the calcium hydroxide content was determined by Thermogravimetric
analysis (TGA).

4.3.5 Dielectric Properties

In order to determine the relationship between dielectric properties and strength, a

dielectric measurement sensor was used. Normal strength concrete mixes, similar to those
used by researchers at Delft, were initially investigated to gain confidence in the use of
the apparatus. A series of mixes at 0.30 and 0.26 water/binder ratios were then
investigated.
The theory, practical significance and discussion of experimental results are provided in
Appendix C.

95
 T able 4.1 : Summary of aims and scope of research
ÿ ^ T e s tg fe r J fe i Experimental details
series- ^ ^ ^ ^ e 'r jîè n t réplaoe'nfî|nt
B inary b le n d s:
Exam ine the fresh p r o p e r tie s with 0. 10% C S F Slump
1 binary and ternary blended cem en ts 0, 20. 40% PFA and
at w/b ratios of 0 .3 0 -0 .2 0 . 0, 30. 60% G G BS Tattersall's two-point test
0.26 Ternary b le n d s:

10 % C S F w ith 4 0 . 50% PFA

To a s s e s s the se m l-a d la b a tic 10% CSF with 60. 70, 80% G G BS T herm ocouple,
2 tem p er a tu re rise with binary and ternary 1 m^ Ply-w ood
blended cem en ts at w/b ratios of 0.3 0 -0 .2 0 insulated box
0.3 0, 10% C S F binary mix
10% CSF with 40% PFA. 60% G G BS ternary m ixes
To m easu re the co m p ressiv e C u b es ca st
str e n g h d e v e lo p m e n t according to
3 with binary and ternary 0.2 10% C S F binary mix B S 1881
blended cem en ts at w/b ratios of 0 .30-0.20 10% CSF with 40% PFA. 60% G G BS tenary m ixes

M icrostructural p r o p e r tie s of Alcohol re-saturation

I
cem en t p a ste s involving : m ethod. TGA. A
4 m easu rem en ts of p o r o sity an d CaOH; c o n t e n ts 0.26 0, 10% C S F binary mix and 5 0 mm c u b e s

and
10% CSF with 40% PFA, 60% G G B S ternary m ixes I%
s tr e n g th d e v e lo p m e n t 0, 10% C S F binaary mix S'

\o
o\
5
A s se s sm e n t of d ie le c tr ic p r o p e r tie s In
NSC,
SCO
and
0.5 an d 0.60
0.30
100% O PC
100% OPC
Dielectric strength
se n s o r
I
a

HSC 0.26 0, 10% C SF, 10% CSF with 40% PFA, 60% G GBS
Chapter 4 A im s and scope o f investigation

43.6 Mix proportions used

In this programme, high strength concrete mixes that were designed by Soutsos*'^®^ were
used. These were then modified to accommodate different levels of cement replacement
by supplementary cementing materials. Soutsos adopted a method method for optimising
the mix proportions based on the ‘Maximum Density Theory’. This requires the
aggregate to occupy as large a relative volume as possible. The relative proportions of the
aggregates are chosen to produce the minimum void content, and produce a minimum
required volume of cement paste.
This method does not consider the effect that the aggregate surface area has on the
requirement of excess paste for lubrication; therefore it was modified by Soutsos to
account for this.
The mix proportions used to assess the variables are summarised in Table 4.2.

Note: Two additional binary mixes of CSF (5 and 15%) and four ternary blends
(20%PFA+10%CSF, 36%PFA+10%CSF) and (30%GGBS+10%CSF, 54%GGBS+10%
CSF) were tested at 0.26 w/b ratio to determine optimum levels of workability. These
proportions were obtained from Soutsos’s work on binary and ternary blends.

Table 43: Basic mix Proportions ( in kg/m^ ) used in the investigation

# w /b Binder: Granite JSand^ ..W a te r.; ■ . . - s p . r

ratio .a":''" D o sa g es

0.30 459 1115 670 142 Variable

0.26 510 1115 670 133 Variable

0.20 590 1115 670 118 Variable

* Dependent on Type and content of CRM used

97
Chapter 5_________________________________________________________________________M aterials an d experim ental procedu res

CHAPTERS

MATERIALS AND EXPERIMENTAL PROCEDURES

5.1 Introduction

This chapter presents the details of the materials and experimental procedures used in the
research. The materials used are reported in section 5.2, the mixing procedures are
presented in section 5.3 and the main test procedures are described in section 5.4. The
dielectric test procedure is given in Appendix C.

5.2 Materials

5.2.1 Ordinary Portland cement (OPC)

In total eight batches of typical Portland cement, obtained from Rugby Cement pic, were
used during the testing programme. A large number of batches had to be used because
other students/researchers working on concrete related projects were also using the same
cement. The cement used was in compliance with British Standards, BS 12: 1989
Table 5.1 shows the chemical composition of the cements.

5.2.2 Supplementary cementing materials (SCMs)

The following types of SCMs were used:

• Condensed silica fume (CSF), obtained as a 50/50 slurry in water, supplied by Elkem
Chemicals.
• Pulverised fuel ash (PFA), complying with BS 3892: Part 1:1992 supplied by Ash
Resources Ltd.
• Ground granulated blast furnace slag (GGBS), complying with BS 6699:1992
supplied by Civil and Marine Ltd.
The compositions and physical properties of the three types of SCMs used are shown in
Table 5.2.

98
C h a p te r 5 M a te ria ls a n d e x p e r im e n ta l p r o c e d u r e s

Table 5.1 : Portland cement compositions

Sample C3S . - .'C4AF ;

No. W : (% )<
PC -1 59 12 9.6 9.0 2.57 0.61 376 62.5

PC-2 55 14 10.6 9.1 2.95 0.63 355 59.4

PC-3 50 18 11.0 9.0 2.75 0.63 380 55.0

PC-4 57 13 9.5 9.2 2.69 0.62 395 60.5

PC-5 59 10 9.6 9.2 2.50 0.62 370 59.4

PC-6 53 15 9.8 9.9 2.75 0.69 380 53.2

PC-7 54 15 9.0 9.3 2.73 0.71 345 55.1

PC-8 57 14 10.1 8.6 2.51 0.62 330 60.0

* 28 day mortar compressive strength ( by suppliers

Table 52 : Compositions and physical properties of SCMs

'^CaO - .^-.AUOStS
rtype%:
CSF 0.3 92 1.0 1.0 0.6 - 15000 - 2.2
20000
PFA 1.4 51.4 25.0 9.4 1.4 4.80 87.50% 2.40
< 45 um
GGBS 41.3 33.7 11.5 1.8 9.0 - 400 - 440 2.90

99
C hapter 5_________________________________________________________________________ M aterials an d experim ental procedu res

5.2.3 Admixtures

Only one type of commercially available superplasticiser was used (Conplast SP 435).
This was supplied by Fosroc Expandite Ltd, and complied with BS 5075^'^^^ and ASTM
This superplasticiser is essentially a sulphonated naphthalene formaldehyde
and was supplied as a liquid and had a solid content of 40% by weight and a specific
gravity of 1.19. It has been reported that naphthalene based superplasticisers provide
better performance in terms of workability compared to a melamine based product^^l
In this study, the superplasticiser dosage is expressed as a percentage of solids by weight
of binder. The manufacturers recommended dosages are 0.7 - 2.0 litres/100 kg binder for
high strength/ high workability concrete, which is approx 0.3 - 1.0 % solids by wt of
binder. The superplasticiser dosages used in this study to obtain a target slump of 200±
2 0 mm were generally higher than this.

5.2.4 W ater

Ordinary tap water was used throughout for mixing. Its temperature varied between
17 °C in winter to 22° C in the summer months.

5.2.5 Fine and coarse aggregate

Throughout the testing programme Thames Valley sand was used with a fineness
modulus (FM) ranging from 2.20 to 2.60. The coarse aggregate was mainly 10mm
Granite obtained from Tarmac Roads tone, Middx. In total five batches were used during
the testing period, these had aggregate crushing values (ACV) of 11.3 and 13.1.
For producing conventional normal strength concrete, for the dielectric tests, 20 and 10
mm Thames Valley gravel was used.

The coarse and fine aggregate gradings together with the physical properties in
accordance with BS 812^^®®^ are shown in Table 5.3.

100
T able 5 3 : Properties of aggregates
# # # Percentagéivyeight retained^
f f Siève . r /M W S an d s • ' #KLGraviel Granite \ r
Batch 1' {Batch 3 <Batchm 2Q10 fnnr .10-5 mm 3 "iÇ ?
: 3"- ^
25.4 mm 100 100 100 100 100.00 100.00 100.0 100.00 100 100.0

19.1 mm 100 100 100 100 84.16 100.00 100.0 100.00 100 100.0

10.0 mm 100 100 100 100 8.80 93.84 86.0 91.00 90 91.0

5.0 mm 99.1 99 98.0 97.73 2.00 16.48 2.7 3.80 3.9 4.1

2.36 mm 88.03 87 87.0 86.97 0.00 0.00 0.0 0.10 0.0 0.0

1.18 mm 78.5 77 75.0 77.07 0.00 0.0 0.0 0.00 0.0 0.0

600 um 69.9 69 55.0 63.36 0.00 0.0 0.0 0.00 0.0 0.0

300 um 28.71 41 19.0 25.74 0.00 0.0 0.00 0.0 0.0 0.00

150 um 4.91 8 4.0 2.53 0.00 0.0 0.00 0.0 0.0 0.00

P.M. 2.31 2.2 2.6 - 2 .5 - * -

fG
aA.
I
Other properties
S.G. 2.64 2.64 2.64 2.64 2.60 2.60 2.70 2.70 2.70 2.70 I
Absorption 1.5% 1.2% 1.2% 1.2% 1.00 1.10 0.7% 0.7% 0.7% 0.7%
I
at SSD
Bulk density
( kg/m3 )
- - 1756.0 1760.0 1590.00 1560.0 1595.0 1578.00 1584.0 1592.0 !
ACV 11.31 11.72 12.85 11.95
^ 0
C hapter 5 M aterials an d experim ental procedu res

5.3 Mixing Procedures

The object of mixing is to coat the surface of all aggregate particles with cement paste,
and to blend all the ingredients of concrete into a uniform mass; this uniformity must
furthermore not be disturbed by the process of discharging from the mixer.

At the start of the testing programme, oven dry aggregates were used in the laboratory
which meant that the mixing water included the absorption water of the aggregates. From
the latter part of 1996, wet aggregates were used, which required the moisture content of
the aggregates to be measured and adjustments made to the final mixing water. This
decision was made based on other ongoing research work at UCL, which preferred the
use of wet aggregates. Preliminary concrete mixes at 0.26 w/b ratio were prepared using
dry aggregates. However all the mixes reported in this thesis (0.3,0.26 and 0.2 w/b ratio)
were batched using wet aggregates.

5.3.1 Cement paste

The cement paste used for the porosity tests (test series 4) was mixed using a Silverson
high shear mixer, initially at the maximum speed for two minutes after which a palette
knife was used to clean off the agglomerated paste which had adhered to the mixer and
mixing bowl. The Silverson was then restarted and mixing continued for a further three
minutes.
The mixing water was premixed with the superplasticiser and added at the start of
mixing.

5.3.2 Concrete
Concrete was mixed using two types of mixers. Conventional mixes were mixed using a
Soroto L-33 mixer, which has a capacity of 30 litres. High strength mixes were produced
using a pan mixer, which has a capacity of 110 litres. In order to minimise wastage of the
materials, batch weights of 30 or 40 litres were made using the pan mixer.

102
Chapter 5 M aterials an d experim ental procedu res

As mentioned earlier, the moisture content of the aggregates was measured before
mixing, using a microwave oven. The 20 mm aggregate, if used, was placed in the mixer
followed by the 10mm aggregate, the sand, the cement, together with any PFA or GGBS.
When CSF was used it was added 30 seconds after the start of initial mixing.
The superplasticiser was also added, after 30 seconds, premixed with the mixing water
and mixing continued for up to 5 minutes.

5.4 Fresh Concrete Tests

5.4.1 Slump test

The slump test was performed on all of the mixes in accordance with BS 1881: Part 102
The test was carried out immediately after mixing when using wet aggregates and
after 5 minutes for the dry aggregates. After the slump test, the concrete was remixed for
1 minute, and two-point testing commenced.

5.4.2 Two-point test

As previously mentioned in section 3.3.2, to assess the fresh properties of high strength
concrete, single point tests such as the slump test are thought to be inapplicable.
Tattersall’s two-point test to assess the workability of concrete has been well documented
(75,76.77) is thought to be more suitable when assessing the workability of high strength
mixes, as reviewed in section 3.3.2. In the present study, an earlier version of the two-
point test was used as the new version was developed during my research.

The concrete under test is contained within a cylindrical bowl and sheared by an impeller.
In this programme, only the MH system was used; and a Windograf digital recorder was
used to record the two-point test data in the form of flywheel voltage and pressure
transducer voltage. The test data was recorded on floppy disc and chart-outputs from the

103
Chapter 5 M aterials an d experim ental procedu res

Windograf. The two point test apparatus (MH) and the Windograf digital recorder are
shown in figure 5.1.

• Test procedure

1) The two- point test wass initially warmed up for one hour at speed setting 4.
2) The Windograf digital recorder was connected to the apparatus, and set to glitch
mode with a chart speed of 2.5 mm/s.
3) Voltage traces from the flywheel and pressure transducer were zeroed.
4) The test bucket was then raised in position, the impeller speed is reduced to speed
setting of Vz, the concrete is then poured in the bucket to about 75mm from thetop.
5) The impeller speed was then increased to speed setting 4, and recording of data
proceeds.
6 ) A continuous sweep method of recording was adopted through the speed settings of
4, 3.5, 3.0, 2.5, 2.0, 1.5, 1.0 and 0.5, for 5 seconds at each stage. The total recording
time was around 35 seconds.
7) The test bucket was emptied, the impeller speed is increased to 4, and the recording
of the idling pressure is started, steps (5-6) are repeated.

Figure 5.2 shows typical chart outputs for a two-point test and idling data.

104
C h a p te r 5 M a te ria ls a n d e x p e r im e n ta l p r o c e d u r e s

Figure 5.1: Two point test apparatus and Windograf recorder

105
Chapter 5 Materials and experim ental procedures

tiû
G

0
çu
1
B
o

g.
G
O
c(Q
-G
O
13
o
'Oh

ui
ît
3D.
O
IZ

106
C hapter 5 M aterials an d experim en tal pro ced u res

• Data analysis

The software supplied with the Windograf digital recorder allows the conversion of the
data recorded on floppy disk to ASCII code using Dasa Utilities. These ASCII files are
then analysed by means of a computer program. A FORTRAN 77 program was used
to process the data, based on the following algorithm.

The impeller speed (rps) and the torque can be determined from the fly wheel and
pressure voltages respectively, by applying the following equations :

Helical impeller speed (rps) = fly wheel voltage(volts) / 0.5 / 4.6815 (gear ratio) (5.1)

Idling pressure (psi) = 176.44 x idling pressure voltage - 155.5 (5.2)

Total pressure (psi) = 176.44 x pressure voltage - 155.5 (53)

Nett pressure (psi) = Total pressure - Idling pressure (5.4)

Torque, T (Nm) = 0.0215 x nett pressure (psi) (53)

These equations were determined using the relevant calibration relationships given in
Appendix A l.
The two Bingham constants g and h from the best fit flow curve and the correlation
coefficient are then found by linear regression. It is important to realise that different
two-point test apparatus may have different coefficients.

Torque (Nm) = g + h N (rps) (5.6)

Typical outputs from the data analysis program are shown in figure 5 3 .

107
 C hapter 5 M aterials an d experim en tal procedu res

Run me .exe under DOS s y s c e m (<0.15,kl = 1 0 0 ,<0 .40 out T = 0 .0215P, u s e i

Sept/9 Ô)
Test : C200597 C2 T r i g g e r Date 20/05/97
Test ; 1200597 Tri gger Date: 20/05/97
I dling p res s u r e n and cor r e lat ion coef ficient r==> 74 0. 9930
I dli n g p res s u r e ai bi==> 1.55284 .09261
n flywheel speed voltage p r e ssure voltaae idling p r e s s u r e
1200 3.250 2 .731'' 1.854
800 2 .911 2.646 1.822
600 2.607 2.548 1.794
700 2.299 2 .451 1.766
600 1.964 2.326 1.735
600 1.654 2.220 1.706
400 1.269 2.055 1 .670
600 .892 1.918 1.635
300 .618 1.794 1.610

T'orque (Nm) Shaft speed(rps)

3.329 1.388
3.124 1.244
2 .860 1 .114
2.598 .932
2.243 .839
1.949 .707
1.459 .542
1.073 .381
.696 .264

Two p o i n c test constant g(Nm)= .18

Two p o i n t test constant h(N m s) = 2.3 7
L i n e a r regression correlation coefficient r= .9959

Figure 53: Typical output from data analysis program.

108
C hapter 5 _________________________________________________________________________ M aterials an d experim ental procedures

5.5 Semi adiabatic temperature rise measurements

The temperature rise of the concrete was measured under semi adiabatic conditions . The
method adopted was similar to the approach used by other researchers Fresh
concrete was placed and compacted in a plywood mould insulated with polyreuthane and
heavy duty polythene to prevent moisture loss. The internal dimensions of the mould
were 210x210x210 mm. A thermocouple was inserted into the concrete through a hole in
the top of the mould, linked to a Chessell chart recorder. The mould was then placed in
the centre of a Im square heavily insulated plywood box , to achieve near adiabatic
conditions, as shown in figure 5.4.

The rate of temperature drop, occuring after the maximum temperature rise had been
reached, was found to be 0.2 °C per hour and was used to correct the semi adiabatic
temperature measurements into estimated adiabatic curves. This was achieved by adding
0.2° C to all of the semi-adiabatic temperature readings from the start of the test.

5.6 Tests on hardened concrete

5.6.1 Compressive strength measurements

All compresive strength measurements were obtained on 100mm cubes. These were cast
in steel moulds and were filled in three layers on a vibrating table. After casting, the
cubes were covered with a polythene sheet and demoulded after 24 hours. These cubes
were cured under standard conditions in water at 20 °C until the test date.

The temperature matched cured cubes were placed in a Fisons climatic cabinet, a
predetermined temperature cycle was calculated (as described in Chapter 8 ) and fed into
the cabinet by means of scribing out the profile on a set of cams that controlled the
temperature. After the cycle was completed the cubes were demoulded and sealed in
polythene and stored in a 20° C constant room.

109
Chapter 5 _________________________________________________________________________ M aterials an d experim ental procedures

In total fifteen cubes were usually cast under each curing condition to monitor the
strength at 1, 7,28,90 and 180 days.
Sets of three cubes were tested at each test age, according to BS 1881 :Part 106 using a
Contest GDIOA compression machine, shown in figure 5.5 which complied to BS
1881:Part 115'^'"'.

5.6.2 Density measurements

The densities of the concrete cubes were measured by weighing them in air and then in
water. The density (p) in kgW is given by:

p = (Wa)/(Wa-W^)xlOOO

where : W a is the mass in air (kg),

W w is the mass in water (kg).

5.7 Tests on cement paste

5.7.1 Porosity measurements

Paste specimens of 125 x 50 x 26 mm were cast, for determining the porosity. These were
demoulded after 24 hours and cured under standard and temperature matched curing
conditions same as the concrete.
At test dates of 1, 7, 28 and 90 days the specimens were weighed in the SSD (saturated
surface dry) condition and then oven dried at 105 °C until constant weight was achieved.
The specimens were then subsequently immersed in Propan2ol until constant weight was
achieved. The increase in weight from the oven-dried condition was used to determine
the Propan2ol resaturation porosity (expressed as a percentage of bulk volume).

110
C hapter 5 M aterials and experim ental proced u res

5.7.2 Compressive strength measurements

The compressive strength of cement paste was measured on 50 mm cubes. These were
cured under standard and temperature matched curing conditions similar to the concrete.
In total nine cubes were usually cast under each curing condition to monitor the strength
at 1,7 and 28 days.

I ll
Chapter 5 Materials and experimental procedures

Figure 5.4: Insulated plywood box.

112
Chapter 5 M aterials and experimental procedures

Figure 5.5: Contest GDIOA compression testing machine.

113
C h apter 6 F resh p ro p erties o f H S C

CHAPTER 6

FRESH PROPERTIES OF HIGH STRENGTH CONCRETE

6.1 Introduction

The literature review has highlighted the major criticisms of the standard tests used to
measure the workability of high strength concrete. Although often considered
adequate to assess the workability of normal strength concrete, these tests are limited
in their application because

a) a single point test cannot be applied effectively over the full workability range of
concretes encountered in practice and
b) they cannot measure the fundamental rheological properties of concrete.

There is general agreement that fresh concrete conforms reasonably well to the
Bingham model. The use of Tatersall’s two point workability test is well documented,
and was used in addition to the slump test to assess the workability.

This chapter presents the results of the tests on the fresh properties of binary and
ternary blended high strength concretes. The main areas are;

• The effect of superplasticiser dosages to produce a given slump.

• The relationship between the Bingham parameters, and the level of cement
replacement in binary (PFA, GGBS & CSF) and ternary (PFA/CSF, GGBS/CSF)
blends.

• To establish the optimum replacement levels for each combination.

• The effect of PFA/CSF and GGBS/CSF mixes at different water/binder ratios.

• To compare the slump and two point test measurements for assessing the
workability.

114
C h apter 6 F resh p ro p e rtie s o f HSC

6.2 Superplasticiser dosage required to produce a given slump.

Despite the number of drawbacks associated with the slump test, it is still the most
popular standard test used on site, throughout the world. This test was therefore used
to investigate the relationship between the type of supplementary cementing material
and the superplasticiser dosage required to produce a 2 0 0 +/- 2 0 mm slump at various
replacement levels. This range of slump values was chosen as it is representative of
typical high strength concretes used in the field.

The superplasticiser dosage was based on adjusting dosages reported by Soutsos^^)

who used similar materials in his work. However, in his work the mixing procedure
that he adopted was based on a direct addition of superplasticiser. Many researchers
have shown that the most efficient performance of a superplasticiser is obtained with
delayed addition. For example, Pentalla^^^^^ reported that in order to obtain the
optimum workability, the superplasticiser should be added 1/2 - 3 minutes after the
water and cement are first brought into contact. In the present study a 30 seconds
delayed addition rate is used. The superplasticiser used is a calcium based sulphonated
naphthalene which is different to that previously used by Soutsos^^^ (sodium based).

The mixing and experimental details were described in Chapter 5.

115
Chapter 6______________________________________________________________________________ Fresh properties o f H SC

6.2.1 Binary blends - 0.26 w/b ratio

The relation between the superplasticiser dosage for a 200 ± 20 mm slump and the
level of PFA, GGBS and CSF at 0.26 w/b ratio is shown in figure 6.1. Partial cement
replacement by PFA causes a reduction in the amount of superplasticiser for a given
slump compared to the control OPC mix. These results are in agreement to those
reported by Soutsos^^^ who found a similar trend using 10, 20, 30 and 40% PFA
replacement levels (figure 3.5). Besari et aF^^ have attributed these improvements in
workability to the combined effect of an increased paste volume due to the lower
specific gravity compared to OPC and the lubricating effect of the spherical shape of
the PFA particles.

The effect of GGBS causes an increase in the required superplasticiser dosage with
increasing replacement level. These results contradict the findings reported by
Soutsos^^), who reports the opposite, the higher the replacement level (up to 60%), the
greater the reduction in the superplasticiser dosage. A possible reason for the
differences with the present results may be due to the different reactivities and
physical characteristics of GGBS which came from different sources. The GGBS used
by Soutsos^^®^ was supplied by Frondingham Cement, the GGBS used in this study was
supplied by Civil & Marine Ltd.

The results for CSF indicate that lower dosages of superplasticiser are required
compared to the OPC control mix to produce the required workability. In the literature
review, the majority of the work cited on CSF highlights that its use in high strength
concrete reduces the workability and requires a consequent addition of
superplasticiser. Domone and Soutsos^'®^^ have found a similar trend and reported that
at water/binder ratios below 0.28 the ultrafine spherical particles of CSF are
significantly dispersed and act as a workability aid.

116
Chapter 6 Fresh properties o f H SC
2.6

2 .4

2.2

2
0
U))
TO
U) 8

■oO
1.6
C/D

1 .4 PFA :
GGBS I
1.2

0 10 20 30 40 50 60 70
% replacement
Figure 6.1: T h e e f fe c t o f P F A ,G G B S and C S F o n th e s u p e r p la s tic is e r
d o s a g e req u ired to p r o d u c e a 2 0 0 + /- 2 0 m m s lu m p at .
0 .2 6 w /b ratio.
The results are also in agreement with the results of two research teams in
France<'°* '“ ’, who claim that the use of ultra fine particles facilitates the production of
low water binder ratio concrete by the filler effect, i.e. the grains fill the voids
between those of cement reducing the water requirement.

6 2 2 Ternary blends - 026 w/b ratio

The effect of PFA/CSF ternary blends on the relation between superplasticiser dosage
and replacement level at 0.26 water/binder ratio is shown in figure 62. The results
show a continual decrease in the required superplasticiser dosage with increasing PFA
replacement with no further reduction above 40% replacement. In the literature, there
has been limited work performed in this area. In one study, Bayash"^^ found that with
a 20% PFA + 10% CSF at a w/b ratio of 0.41, the slump increased compared to the
CSF control mix.

As can be seen the superplasticiser dosage requirements for the ternary blends are
significantly lower than the control 10% CSF as well as the equivalent binary PFA
mixes. However, the proportional change is similar to the PFA binary mixes, (figure
6 .1 ).

117
 Chapter 6 Fresh properties o f HSC

G G BS/CSF

0
O)) PFA/CSF
[E
o
■D
CL
C/5

0 .6-
0.5 -
0 10 20 30 40 50 60 70 80 90
% replacem ent by PFA, GGBS (+10% CSF)

Figure 6.2: The effect of partial cement replacement by 10% CSF +

PFA, GGBS on the superplasticiser dosage required to produce a 200
+/- 2 0 mm slump.

18
Chapter 6______________________________________________________________________________ Fresh properties o f HSC

At the 50% PFA+10% CSF replacement level a trial superplasticiser dosage of 0.9%
did not give a slump in the specified target range, but gave a value of 165 mm. This
mix was discarded and a dosage of 1.0% was used, same as the 40% PFA+10% CSF
blend, giving a slump of 200 mm. This implies that a plateau is reached with respect to
PFA replacement.

The effect of GGBS/CSF ternary blends on the relation between superplasticiser

dosage and replacement level at 0.26w/b ratio is shown in figure 6.2. The results show
a reduction in the superplasticiser dosage with increasing GGBS content up to 60%
replacement. At the 70% and 80% GGBS replacement levels, the superplasticiser
dosage had to be increased in order to obtain the required slumps, the mixes were also
found to be very sticky. However, these mixes require a lower superplasticiser dosage
compared to the equivalent GGBS binary mixes.
There is limited data in the literature. Soutsos^^^ found with a 50%GGBS+10%CSF
mix that the superplasticiser dosage is reduced compared to a 10% CSF mix (see
figures. 15). This may be attributed to the small spherical size of the CSF particles
which may act as a lubricant when dispersed by the superplasticiser.

6.2.3 Other water/binder ratios

It was found that it was not possible to produce the 100% OPC (control) mix at 0.20
w/b ratio, even with excessive doses of superplasticiser.
The effect of 10% CSF replacement on the relationship between water-binder ratio
and superplasticiser dosage for a 200 ± 20 mm slump is shown in figure 6.4. As the
water/binder ratio is reduced there is a significant increase in the superplasticiser
dosage required to obtain the target slumps. As before, the supeiplasticiser dosage is
reduced for a 10% CSF mix at 0.3 w/b ratio compared to the control OPC mix. This is
in disagreement with Soutsos’s^^^ findings, who found that at water-binder ratios
greater than 0.28 there is an increase in the required superplasticiser dosage for a 10 %
CSF mix.
More recently, Duval and Kadiri^’^^^ have however found that at a 10% CSF
replacement level there is a continual reduction in the superplasticiser dosage at water-
binder ratios of 0.45-0.25 when compared to a control OPC mix.

119
 Chapter 6 Fresh properties o f H SC

0.34

0.32 -

0.30 -
100% OPC
2 0.28 -

.Ç 0.26 -

I 0.24 -
10%CSF
0.22 -

0.20 -

0.18
0.5 1.0 1.5 2 .0 2.5 3.0
SP d osage (%

Figure 6.4: The effect of partial cement replacement by 10% CSF on the
superplasticiser dosage required to produce a 2 0 0 +/- 2 0 mm slung).

0.34

0.32 -

0.30 -
o
2 0.28 -
0
Ç 0.26 -
€
0.24 - 10% CSF
1 0.22 -
40%PFA+10%CSF
0.20 -

0.18
0.5 1 .0 1.5 2 .0 2.5 3.0
SP D osages (%

Figure 6.5: The effect of partial cement replacement by 40% PFA + 10% CSF
on the superplasticiser dosage required to produce a 2 0 0 +/- 2 0 mm slump.

120
 Chapter 6 Fresh properties o f H SC

This effect may be attributable to the small particle size and spherical shape of the CSF
particles that leads to an enhanced filling of void space between the larger cement
grains, which would otherwise trap free water and reduce workability.

The effect of the 40%PFA/10%CSF ternary replacement level on the relationship

between water-binder ratio and superplasticiser dosage is shown in figure 6.5. At all
water/binder ratios the required superplasticiser dosage was reduced compared to the
control 10% CSF mix, the greatest reduction being at the 0.20 w/b ratio. An identical
relationship was found for the optimum GGBS/CSF ternary mix (figure 6 .6 ). The
same superplasticiser dosages were used as for the PFA/CSF mixes to obtain slumps in
the specified range. This improvement in workability could be attributed to the low
C]A content of the binder and to a reduction in the amount of ettringite formed
together with the filler effect.

These results are in agreement with observations by several w o r k e r s ^ " a n d clearly

show one of the advantages of using such blends in high strength concrete.

0.34
0.32 -
0.30 -
o
2 0.28 -
0
1 0.26 -
fB 0.24 - 10% CSF
i
0.22 -
^ 60%GGBS+10%CSF
0.20 -

0.18
0.5 1 .0 1.5 2 .0 2.5 3.0
SP dosage (%)

Figure 6 .6 : The effect of partial cement replacement by 60%

GGBS + 10% CSF on the superplasticiser dosage required to produce a
2 0 0 +/- 2 0 mm slump.

121
Chapter 6______________________________________________________________________________ Fresh properties o f HSC

6.3 Two point workability tests

The development and theory behind Tattersalls two point workability test apparatus
have been discussed in section 3.3.2. The mixing procedures and experimental method
has been given in Chapter 5. The details of mixes investigated are given in Table 4.1.
The superplasticiser dosages of these mixes are the same as those reported in section
. .
6 2

6.3.1 Binary blends

The two point test results for the CSF binary mixes at 0.26 w/b ratio are shown in
figure 6.7. The results show that both yield (g) and plastic viscosity (h) values for CSF
replacement levels up to 15 % are lower than those of the control OPC mix. This is in
general agreement with Soutsos’s^^^ work. However, in his work the results did not
show a systematic trend in (h) plastic viscosity versus replacement level. The results in
figure 6.7, also indicate that at the 15% CSF level there are no substantial
improvements in the workability compared to the 10% CSF mix. From this and the
superplasticiser dosage requirements discussed in section 6 .2 . 1 , it can therefore be
concluded that the optimum replacement level by CSF is 10%.
The present results are in contradiction with those reported by Gjorv^^^\ who showed
that increasing the CSF replacement level the plastic viscosity (h) is greatly reduced,
while the yield stress (g) is almost unaffected (figure 3.13). The reasons for these
differences are not clear but may be associated with variations in mixing procedures
and/or the superplasticiser dosage levels. Details of these were not reported by Gjorv.

Partial cement replacement by PFA at 0.26 w/b ratio has been found to reduce the
yield value (g) of the fresh concrete, as shown in figure 6 .8 . There are significant
reductions in the g value up to 20% PFA replacement, with little change occuring up
to the 40% replacement level. The plastic viscosity (h) was found to be higher than
that of the control OPC concrete, up to the 40% replacement level. These tendencies
are in agreement with Soutsos’s results (see figure 3.8). Although it can be agued that
the significant reduction in g value may outweigh the small increases in plastic
viscosity, strictly speaking the improvements in workability as reported in the
literature^^^’^^^ would be expected to reduce both of the bingham parameters.

122
 C h apter 6 Fresh p ro p erties o f H S C

10
9
8
E 7
Z
6
3) 0

I 5
4
3
2
1
0
0 2 4 6 8 10 12 14
% Replacement by CSF
(a) Yield (g) value

12

10

(0
o
uw 4
■>
o
w
ro 2

0 -4
0 5 10 15
% Replacement by CSF

(b) Plastic viscosity (h) value

Figure 6.7: The effect of CSF on the yield (g) and plastic viscosity values
at 0.26 w/b ratio.

123
 Chapter 6 Fresh properties o f HSC

10

(D 6
3 PFA
5
I
3 4

;
T3
3

2
C SF

0
0 5 10 15 20 25 30 35 40 45

% Replacement by PFA, CSF

(a) Yield (g) value

13 -
12 -

I
Z
PFA

I
>
C SF
Ü
%
CD
CL

0 5 10 15 20 25 30 35 40 45
% Replacement by PFA,CSF
(b) Plastic viscosity (h) value
Figure 6 .8 : The effect of PFA, CSF on the yield (g) and plastic viscosity
values at 0.26 w/b ratio.

124
C h apter 6 F resh p ro p erties o f H S C

Gjorv^^^) also found that PFA increases the plastic viscosity, but did not provide an
explanation for this. Tattersall & BanfilF^^ attribute this effect to the changes within
the paste.

The results for the GGBS mixes at 0.26 w/b ratio are shown in figure 6.9. This shows
an increase in both the yield (g) and plastic viscosity (h) values, with increasing
replacement level. However, Soutsos^^^ stated that his results with GGBS mixes were
inconclusive but nonetheless concluded that GGBS has only a small effect on the
rheological properties of concrete. Furthermore, other researchers^®’’®^ have reported
that the use of GGBS can lead to improvements in workability.

Observations during mixing revealed that by increasing the amount of GGBS

replacement, the mix becomes less workable and difficult to handle 10-15 seconds
after mixing. The poor workability properties of the GGBS mixes were noticed whilst
filling the two point test bucket; this explains the erratic points on the graph. The rate
of stiffening was very severe at the 60% replacement level compared to that at 30%.
This led to the use of lower shear rates during testing, in order to prevent damage
occurring to the two-point test apparatus i.e. speed settings 2 - 0 rather than 4 - 0 .

The lower workability of these mixes, despite their constant slump is consistent with
the higher superplasticiser demand found in these mixes, in the previous section.

63.2 Ternary blends

The two point test results for the PFA/CSF ternary blends at 0.26 w/b ratio are shown
in figure 6.10. There is a noticeable reduction in both the yield (g) and plastic
viscosity (h) values compared to the control CSF binary mix. The reductions are
however not as significant as those for binary CSF mixes (figure 6.7). For example,
with the 40% PFA + 10% CSF mix the yield (g) value is reduced from 4.4 to 2 5 Nm
(43%) and the corresponding plastic viscosity (h) is reduced from 6.0 to 4.5 Nms
(25%).

1 25
 C h apter 6 F resh p ro p e rtie s o f H S C

18

16

14

12

10

O) 8
■D
)
0
>-

10 20 30 40 50 60 70

% Replacement by GGBS

(a) Yield (g) value

V)
E
z

V)
8(0
>
o
w
m
Q.

0 10 20 30 40 50 60 70

% Replacement by GGBS

(b) Plastic viscosity (h) value

Figure 6.9: The effect of GGBS on the yield (g) and plastic viscosity
values at 0.26 w/b ratio.

126
C h a p ter 6 Fresh p ro p erties o f H S C

4.5

E
z
(U
D
5
O)

0 10 20 30 40 50 60

% R eplacem ent by PFA (+10% CSF)

(a) Yield (g) value

<n
1

10 20 30 40 50 60
% R eplacem ent by PFA (+10% CSF)

(b) Plastic viscosity (h) value

Figure 6.10: The effect of PFA(+10%CSF) on the yield (g) and

plastic viscosity values at 0.26 w/b ratio.

127
C h apter 6 F resh p ro p e rtie s o f H S C

In the case of the 10% CSF and the OPC mix, the g value was reduced from 9.5 to
4.37 Nm (54%) whilst the h value is reduced from 10 to 6.1Nms (39%). The results in
figure 6 .1 0 also show that there are no further improvements in workability beyond
the 40% PFA+10%CSF level. It is also interesting to note that the plastic viscosity
with the 40% PFA+10% CSF ternary mix is almost 10 Nms lower compared to that
with the 40% PFA binary mix and there was no signs of segregation occurring. This
therefore clearly highlights the excellent workability characteristics of ternary mixes.

The two point test results for the GGBS/CSF ternary blends at 0.26w/b ratio are
shown in figure 6.11. The results show continuous reductions in both the yield values
and plastic viscosities up to the 60% GGBS+10% CSF level, followed by slight
increases at higher replacement levels. Soutsos’s^^^ limited results with a single 50%
GGBS+10% CSF mix similarly showed that the both the Bingham parameters are
lower than the control 10% CSF mix. According to Djellouli et aF"'\ such workability
improvements are due to a reduction in Q A content and the small amount of ettringite
formed as a consequence of the less reactive binder.

The increase in both g and h values beyond the 60%GGBS+10%CSF level may on the
other hand be related to the GGBS/CSF replacement ratio. That is, the effect of the
10% CSF on the workability of ternary blends with high GGBS replacement levels is
perhaps not sufficient. Further work would therefore be required to explore the effect
of higher CSF replacement with high GGBS replacement levels.

128
C h a p ter 6 Fresh p ro p erties o f H S C

3
05

0 20 40 60 80 100

% R eplacem ent by GGBS (+10% CSF)

(a) Y ie ld (g) v alu e

U) 7
E
z
6
i
1 5
!c
4
V)
8 3
g
2
I
iS 1
Ql

20 40 60 80 100

% R eplacem ent by GGBS (+10% CSF)

(b) Plastic viscosity (h) value

Figure 6,11: The effect of GGBS(+10%CSF) on the yield (g) and

plastic viscosity values at 0.26 w/b ratio.

129
C h apter 6 F resh p ro p erties o f H S C

633 O ther w ater/binder ratios

• Binary blends

The rheological properties were further investigated at different water/binder ratios of

0.20 and 0.30.
The two point test results for the 10% CSF mixes are shown in figure 6.12. These
show that both yield (g) and plastic viscosity (h) values are significantly lower than
those with neat OPC. At the 0.20 w/b ratio a comparison cannot be made with the
control OPC mix since this could not be produced. These results are in agreement
with those reported by Soutsos^’"^ The nature of the CSF mixes during mixing and
handling were noticeably better than the neat OPC mixes, for example the amount of
effort required for the tamping rod to penetrate the CSF concrete during the slump test
was lower than OPC concrete although similar slumps were obtained. This partly
explains the lower g and h values found in this investigation. This is also in agreement
with Tachibana et al’s^’’^ work, which showed that the electric energy consumed
during mixing decreased by using CSF in high strength concrete mixes.

The distinct rheological benefits of using CSF in high strength concrete mixes as
demonstrated in this section, can to some extent eliminate the uncertainties that
currently exist in the literature.

• T ernary blends

The two point test results for the 40%PFA+10%CSF mix at different water binder
ratios is shown in figure 6.13. At all water binder ratios the yield (g) value is lower
than the 10% CSF mixes, however, the reduction in g is greater at 0.26 and 0.20 w/b
ratios. An interesting feature of the plastic viscosity (h) is that it is very similar to the
10% CSF control mixes at 0.30 and 0.20 w/b ratios, but there is a noticeable reduction
with the ternary blend at 0.26 w/b ratio.

130
C h apter 6 F resh p ro p e rtie s o f H S C

0.32 -

0.3 -
100%
0.28 J

0.26 4

0.24 -

0.22 -
10%CSF
0.2 -

0.18 4- 1 — :— i— T— I— I— I— I— 1— p - i — I— I— :— | -

0 4 5 6 7 8 9 10
g value (NM)
(a) Yield (g) value

0.32

0.3 100% OPC

0.28
o
0.26 ^
2

I 0.24 -

0.22 - 10%CSF

0.2

0 1 2 3 4 5 6 7 8 9 10 11 12

h value (Nms)

(b) Plastic viscosity (h) value

Figure 6.12: The effect of 10%CSF on the yield (g) and plastic
viscosity values at 0.3-0.2 w/b ratios.

131
C h a p ter 6 F resh p ro p erties o f H S C

0.32

0.3

0.28

0.26
10% CSF
0.24
40%PFA+10%CSF
0.22

0.2

0.18
0 1 2 3 4 5 6 7 8

g value (NM)

(a) Yield (g) value

0.32

0.28 -
o
0.26 i
2
10% CSF
I 0.24 -
40%PFA+10%CSF
0.22

0.2

0 1 2 3 4 5 6 7 8 9
h value (Nms)

(b) Plastic viscosity (h) value

Figure 6,13: The effect of 40% PFA + 10%CSF on the yield (g) and
plastic viscosity values at 03-0.2 w/b ratios.

132
C h apter 6 F resh pro p erties o f H S C

The two point test results for the 60%GGBS+10%CSF mix at different w/b ratios is
shown in figure 6.14. The yield (g) values are again lower than the 10% CSF binary
mix. At 0.30 w/b ratio the g values are similar for both binary and ternary mixes, but
the difference again increases as the water binder ratio is reduced. In comparison with
the PFA/CSF ternary blended mixes, the plastic viscosity (h) values are found to be
higher at 0.3 and 0.2 w/b ratios and lower at the 0.26 w/b ratio compared to the 10%
CSF mix. Similarly, Soutsos reported that for a 50%GGBS+10%CSF mix, both g
and h increased as the water/binder ratio was reduced compared to the OPC control
mix.

Observations during mixing revealed the improved workability of ternary blends,

these were noticed whilst carrying out the slump test and filling the two point test
bucket. In comparison with the binary mixes with GGBS, ternary blends eliminated
the difficulties experienced with rapid reductions in workability after mixing and
during testing.

In general, the use of ternary combinations of PFA/CSF and GGBS/CSF can

considerably reduce the problems associated with the workabilty of high strength
concrete mixes and further provide economical advantages during production.

133
C h apter 6 F resh p ro p e rtie s o f H S C

0.32

0.3

0.28
o
0.26
2
n
3 0.24 10% CSF
0.22
60%GGBS+10%CSF-
0.2

0.18
0 1 2 3 4 5 6 7 8

g value (Nm)
(a) Yield (g) value

0.32

0.3 -

0.28 -
0 0.26 J
2
10% CSF
1 0.24

0.22 4 60%GGBS+10%CSF

0.2

0.18
0 1 2 3 4 5 6 7 8
h value (Nms)

(b) Plastic viscosity (h) value

Figure 6.14: The effect of 60% GGBS + 10%CSF on the yield (g)
and plastic viscosity values at 03-0.2 w/b ratios.

134
Chapter 6_______________________________________________________________________________ Fresh properties o f HSC

6.3.4 Comparison of slump and two-point test measurements

The relation between the slump and the g and h values are shown in figures 6.15 and
6.16 respectively. The figures include some typical results for normal strength
concrete from Tattersall and BanfilF^l

As can be seen there is a reasonable correlation between slump and the yield (g) value
in normal strength concrete. Although some of the results in high strength concrete fall
in the normal strength concrete relationship, particularly ternary blends, most of the
data does not show a correlation with slump. The wide scatter of the results
particularly with the low workability GGBS mixes highlights the inadequacy of the
slump test and the importance of using the two-point test for assessing the fi'csh
properties at low water/binder ratios. The reasons for the scatter in the results are not
clear, many researchers have however commented that the rheological properties of
high strength concrete are distinctly different from those of normal strength
concretes^*Clearly this is an area where further research is required.

From the relation between slump and the plastic viscosity (h) value, it can be seen that
there is no obvious correlation. The fact that the h value varies over a wide range, in
high strength concrete, again, highlights the importance of the two point test. It is also
worth noting that there is a lack of information in the literature on correlations
between the Bingham constants of high strength concrete with single point tests, such
as the flow table test.
Figure 6.17 shows the relation between the two point test measurements (g and h
values). It is interesting to note that there is no overall correlation between the g and h
values, which indicates the importance of determining both constants.
In normal strength concrete, the h values vary over a much narrow range, 1-3 Nms,
and the g values can range between 2-8 Nm. This demonstrates that the workability of
normal strength concrete is mainly governed by the changes in the yield value, hence
the success of the slump test for assessing workability. In contrast to this, in high
strength concrete g and h values can vary in opposite directions, as demonstrated by
the PFA binary mixes at 0.26 w/b ratio. This indicates the inadequacy of the slump test
for assessing workability changes.

135
 Chapter 6 Fresh properties o f H SC

♦ HSC (my m ixes)

16 -
□ NSC ( By Tattersall & Banfill (75) )
♦
♦
14 - ♦ ♦

12 -

o) 10

I s In N SC : r = 0 .9 3 5
•o
0)
6 -

4 - ♦♦

2 -

50 100 150 200 250

Slump (mm)

Figure 6.15 : Relationship between yield value ( g ) and slump

18 -
♦ HSC
16 -
□ NSC ( By Tattersall & Banfill (75))
14 -

12 - ♦ ♦
♦

10 . ♦

8 - ♦
♦ ♦
♦
♦ ♦
6 - ♦
♦ ♦ ♦
♦
♦ ♦
4 ♦

□ ♦
□ □ °
o □ □
□□

50 100 150 200 250

Slump (mm)

Figure 6.16 ; Relationship between plastic viscosity ( h ) and slump

136
 C hapter 6 Fresh properties o f H SC

18 -

I ♦ HSC ( my m ixes ) I
16 -
: i
I □ NSC ( By Tattersall & Banfill (75) ) i

14
1
12
? 1
z 1
O) 10 J
Qj"
3
m
> 8 -
T3
0)
♦
>
6 -
♦
□
4 J
i
2 -! §

0 -

0 2 4 6 8 10 12 14 16 18

Plastic viscosity, h (Nms)

Figure 6.17 : Realationship between yield value and plastic viscosity

for both NSC and HSC

137
Chapter 6_______________________________________________________________________________ Fresh properties o f HSC

6.4 Conclusions

• The use of binary blended cements incorporating CSF and PFA result in the
improvement of the workability, by allowing a reduction in the amount of
superplasticiser (by up to 2 0 %) required to maintain a similar slump to the
OPC concrete. The use of GGBS, however increases the amount of
superplasticiser required (by up to 30%) to produce a similar slump to the OPC
concrete.

• The use of ternary blends (PFA/CSF and GGBS/CSF) enhances the

workability of high strength concrete and enables a significant reduction in the
superplasticiser dosage required (by up to 30%) to produce a given slump,

• Binary blends of CSF increase the workability by reducing both the g and h
values (by 58 and 38 % respectively). In contrast the GGBS binary blends
reduce the workability, giving up to 40% increases in yield value and 30%
higher plastic viscosities

• The ternary blended cements (60%GGBS+10%CSF) and (40%PFA+10%CSF)

also show considerable improvements in workability at constant slump and
0,3-0,2 w/b ratios. In the PFA ternary blend these improvements are
represented by reductions in the g and h values up to 35 and 20% respectively.
The corresponding workability improvements in the GGBS ternary blend are
reflected by reductions in g and h values of up to 25 and 13% respectively

• Optimum replacement levels of 60%GGBS + 10% CSF and 40% PFA + 10%
CSF at 0,26 w/b ratio were found with respect to workability.

Having determined the effect of binary and ternary blends on the fresh properties of
high strength concrete, their effectiveness in reducing the semi-adiabatic temperature
rise is reported in the next chapter.

138
Chapter 7 Semi-adiabatic temperature rise o f H SC

CHAPTER 7

SEM I-ADIABATIC TEMPERATURE RISE OF HIGH

STRENGTHCONCRETE

7.1 Introduction

The use of high cement contents in the production of high strength concrete can lead
to high heat of hydration temperature rises which, as discussed in section 3.4, can lead
to three problems:

• The development of large thermal gradients within the structural element can
generate tensile stresses that can result in cracks.
• The temperature increase is not uniform throughout the whole mass of the
concrete, leading to different setting and strength gain rates within the concrete.
• The long term strength may be lower than that of standard cured specimens, and
therefore lower than anticipated.

Work reported in this chapter investigates the semi-adiabatic temperature rise of

binary and ternary blended cements. The temperature rise data obtained was then used
to determine the typical in-situ temperature cycles by using a fmite-difference heat
flow program. This was then used to determine the relation between temperature
cycled compressive strengths with those cured under standard temperature conditions.
Calculation of the temperature cycles is given in Appendix B and the compressive
strength development of the mixes tested are presented in the next chapter.

The main areas of investigation are:

• The effect of the level of cement replacement in binary (PFA, GGBS & CSF) and
ternary (PFA/CSF, GGBS/CSF) blends.
• The optimum replacement levels for each combination.
• The effect of PFA/CSF and GGBS/CSF ternary mixes at different water/binder
ratios.

139
C h apter 7 S em i-adiabatic tem perature rise o f H S C

1 2 Binary blends - 026 w ater/binder ratio

The estimated adiabatic temperature rise of an OPC mix and a 10% CSF replacement
mix is shown in figure 7.1. The rise is small during the so-called dormant period of
the first few hours after mixing. Domone and Soutsos^^’^ in their investigation
quantitatively defined this period as the time taken for the adiabatic temperature to
increase by 10°C. The 10% CSF mix has a shorter dormant period compared to the
neat OPC mix, but there is no significant difference in the peak temperatures. The
shorter dormant period may be due to the very high specific surface area of the CSF
particles causing a very rapid pozzolanic reaction. Soutsos attributes this effect to
the CSF particles acting as nucléation sites for the cement hydration products to form.

These results are in agreement to those reported by Soutsos and contradict those
reported by Tachibana et al and Helland who concluded that at low water/binder
ratios CSF gives very little contribution to the heat of hydration (see Section 3.4.4.3.)

The temperature rise for the control OPC mix is equivalent to 9.6° C per 100 kg/m^ of
cement. This is lower than the general rule of thumb used by the construction
industry, 12 °C per 100 k g W of Portland cement. Bamforth^^^®^ reported a value of 14
°C for mixes with cement contents ranging from 150 kg/m^ at a water/binder ratio of
1.27 to 500 k g W at a water/binder ratio of 0.39. This reduction could be due to an
increasing amount of unhydrated cement compounds, due to insufficient water
available for the continuing hydration. A similar reduction with reducing water/binder
ratios has also been reported by Smeplass and Maage^^'*^\

140
Chapter 7 Sem i-adiabatic tem perature rise o f H SC

60

50 - OPC ;
107oCSF !
U
Q
w)
§ 30 -
5q5
a
E 20 -
0)
I-

1 10 100

Time (hours)

Figure 7.1: Estimated adiabatic temperature rise of OPC and

10%CSFmixes at 0.26 w/b ratio.

60
OPC
50 - - 20% PFA
- " 40% PFA

30

20

10

1 10 1 00

Tim e (hours)

Figure 72: Estimated adiabatic temperature rise of PFA binary

mixes at 0.26 w/b ratio.

141
C h a p ter 7 S em i-adiabatic tem perature rise o f H S C

The temperature rise of binary PFA mixes is shown in figure 1 2. Partial replacement
with 20 and 40% PFA has increased the dormant period. The peak temperature at
20% replacement is comparable to the neat OPC mix whilst at 40% the temperature
rise is reduced by 20%. Soutsos reports a similar finding, he observed that at 20%
PFA replacement the peak temperature was comparable to the OPC control and a 40%
PFA replacement the adiabatic temperature is reduced by 18.4%.

The high peak temperatures obtained at low PFA replacement levels can be attributed
to the contribution of the pozzolanic reaction in the presence of excess lime from the
Portland cement hydration. The increase in temperature provides sufficient activation
energy for the pozzolanic reaction to begin at an early age.

Bamforth has also reported a similar value for normal strength concretes with a
total cementitious content of 400kg/m^ and a 30% PFA replacement level. The
reduction in adiabatic temperature rise was 10-15%, increasing to 30% with 50% PFA
(see fig. 3.21). However, Crow and Dunstan found that the adiabatic temperature
rise for 25% low calcium PFA (Class F) showed a reduced temperature rise compared
to an OPC control.

The effect of partial replacement by GGBS at levels up to 60% is shown in figure 7 3 .

A similar trend to the 20% PFA mix is seen at a GGBS replacement level of 30%,
where the peak temperature is comparable to the OPC mix. The dormant period is
significantly longer than the OPC control, increasing from 4 hours to 10 hours. This is
in agreement with Mak and Lu’s^’®’^findings and can be attributed to the nature of the
reaction of GGBS with water. Taylor states that the initial reaction of GGBS
particles with water forms a layer that inhibits any further reaction. However, the
hydration of cement releases activators (mainly calcium hydroxide), which trigger the
subsequent pozzolanic reaction. Based on this explanation, the dormant period in the
GGBS mixes depends on the physical and chemical composition of the GGBS. This
view is shared with Coole who attributes the effect to different reactivities of
GGBS.

142
C h a p te r 7 S e m i-a d ia b a tic te m p e r a tu r e r is e o f H S C

60
OPC
50 30%GGBS
60%GGBS

aa
E
t-
m

1 10 100

Time (hours)

Figure 73: Estimated adiabatic temperature rise of GGBS binary

mixes at 0.26 w/b ratio.

It is clear that in order to achieve an appreciable reduction in the temperature rise a

high replacement level of GGBS is necessary, i.e. 60% GGBS replacement reduces
the peak temperature by 25% compared to the OPC mix (figure 7.1). This is in
agreement with Soutsos’s results in high strength concrete and with Bamforth^^^\
who has also recommended levels higher than 60% to achieve any significant
reductions in the adiabatic temperature rises of normal strength concretes.

143
C h a p ter 7 S em i-adiabatic tem perature rise o f H S C

1 3 T ernary blends - 026w/b ratio

The results of the PFA/CSF mixes are shown in figure 7.4. It can be seen that in both
the ternary blends the peak temperature is, however, significantly reduced, compared
to the binary 40% PFA mix. The greatest reduction in the peak temperature was found
in the 50% PFA+10% CSF mix, which is explained by the low cement content of the
mix. The dormant periods of the ternary blends are shorter than the binary PFA mix,
similar to the 10% CSF mix. This again could be due to the high reactivity rate of the
CSF.

In the literature, there has been limited work reported in this area of ternary blends.
Soutsos found a similar trend using a 36%PFA + 10%CSF mix at 0.26 w/b ratio, he
attributed the effect of shorter dormant periods to the CSF particles acting as
nucléation sites for the cement hydration products to form.

Partial replacement by 10% CSF in combination with different replacement levels

with GGBS is shown in figure 13. A similar but more exaggerated trend to that of
PFA/CSF mixes is found. It can be seen that despite the very low cement content of
these mixes, the dormant periods are much shorter compared to the 60% GGBS
binary mix, the peak temperatures are lower as the level of GGBS replacement
increases.

Soutsos investigated one ternary mixture of 54%GGBS+10%CSF, he also found

that the peak temperature was less than the 60% GGBS binary equivalent (figure
3.29).

144
Chapter 7 Sem i-adiabatic tem perature rise o f H SC

50 ---10%CSF !
— — 40% PFA '
Ü
40% PFA +10% C SF i
40 40% PFA
(/))
0 50% PFA +10% C SF I

2
3
Q)
Q.
E
0
h-

1 10 1 0 0

T im e (h o u rs)
Figure 7.4: Estimated adiabatic temperature rise of PFA ternary
mixes at 0.26 w/b ratio.

• - ■ 10%CSF
60%GGBS
60%GGBS+10%CSF
u • * — 70%GGBS+10%CSF
(<U
/) 80%GGBS+10%CSF

I 30 -
2
&
o 20 -

60%GGBS

1 10 1 0 0

Tim e (hours)

Figure 7.5: Estimated adiabatic temperature rise of GGBS

ternary mixes at 0.26 w/b ratio.

145
Chapter 7 Semi-adiabatic temperature rise o f HSC

The incorporation of CSF in the ternary blends seems to act as an advantage in the
sense that it enables low peak temperatures to be obtained, but accelerates the rate of
reactivity in the early ages (short dormant periods) and has lower subsequent
reactivity compared to the control binary mixes.
This could be beneficial during early age strength development, where traditionally
GGBS mixes are known to contribute to strength at later ages. The compressive
strength development of these mixes is described in the next chapter.

7.4 The effect of level of cement replacement on the peak temperature at 0.26
water binder ratio

Figures 7.6 and 7.7 show the effects of binary and ternary blended cements on the
peak temperature of high strength concrete respectively. The binary PFA and GGBS
mixes appear to reduce the peak temperature with increasing cement replacement. A
similar trend is obtained with ternary blends these mixes appear to be heavily
dependent on cement content. By comparing a 40%PFA and 60%GGBS binary blend
with their ternary equivalents, the ternary blends show a lower peak temperature.
Further work is needed to determine the exact mechanisms responsible for the
temperature rise of ternary blended cements.

7.5 Other water/binder ratios

The semi-adiabatic temperature rise at water/binder ratios of 0.3 and 0.2 was also
investigated. The mixes that were chosen were those that were regarded as optimum
mixes as found in the previous fresh properties chapter ie.10% CSF,
40%PFA+10%CSF and 60%GGBS+10%CSF.

The semi-adiabatic temperature rise of the OPC and a 10% CSF replacement mix at
0.3 w/b ratio is shown in figure 7.8. A similar trend to that found when using 0.26
water/binder ratio is apparent. However, the CSF mix has a shorter dormant period
and a lower peak temperature compared to the neat OPC mix.

146
C h a p te r 7 S em i-adiabatic tem perature rise o f H S C

50

49

47
PFA
GGBS
ffl 46
45

44

43

42

0 10 20 30 40 50 60 70
% replacement

Figure 7.6: The effect of binary PFA and GGBS replacement

levels on the peak temperature at 0.26 w/b ratio

38

36

34

3 32

30

28

26 PFA/CSF
GGBS/CSF
24
40/10 50/10 60/10 70/10 80/10
% Replacement level
Figure 7.7: The effect of tenary PFA and GGBS replacement
levels on the peak temperature at 0.26 w/b ratio

147
C h apter 7 S em i-adiabatic tem perature rise o f H S C

The fact that lower peak temperatures are obtained compared to that at 0.26
water/binder ratio could be attributed to the lower cement contents of these mixes.

The temperature rise for the control OPC mix is equivalent to 9.6 °C per 100 kg/m^ of
cement. As mentioned earlier, this is lower than the values reported by Bamforth
in normal strength concrete, again suggesting that at low water binder ratios there is
insufficient water present for all of the cement to hydrate.

Partial replacement by 10% CSF in combination with 60%GGBS and 40%PFA at 0.3
water/binder ratio is shown in figure 7.9. From these curves, it can be seen that there
is a similar trend to that obtained at 0.26 water/binder ratio (figs.7.4 and 7.5),
however, the peak temperatures at 0.3w/b ratio are 4 °C (60%GGBS+10%CSF) and
2 °C (40%PFA+10%CSF) lower than at 0.26 w/b ratio. In contrast, recent work
published by Sioulas and Sanjayan^^°^^ shows an increase in the peak temperature for a
45%GGBS+10%CSF mix compared to the 45%GGBS binary mix at a water/binder
ratio of 0.30.

The semi-adiabatic temperature rise data for a neat OPC mix at 0.20 water/binder
ratio could not be obtained, since the mix could not be produced. The results of the
10% CSF and the two ternary mixes are shown in figure 7.10. These curves show a
significantly high peak temperature compared to the 0.3 and 0.26 water/binder ratios.
This is probably due to the fact that these mixes had higher cement contents.

Figure 7.11 shows the effect of the different water/binder ratios used in this
investigation on the peak temperature. The total cementitious contents were 590
kg/m^, 510kg/m^ and 459 kg/m^ at 0.20, 0.26 and 0.30 water/binder ratios. Unlike the
ternary blends, the peak temperature of the CSF binary mix at all water/binder ratios
is very similar and does not appear to be influenced by the cement content. However,
the ternary blends are heavily dependant on the cement content, the peak temperature
is higher at the low water/binder ratio.

148
 Chapter 7 Sem i-adiabatic tem perature rise o f H SC

60
OPC

50 -
10%CSF

u 40

1 10 1 0 0

Tim e (hours)

Figure 7.8: Estimated adiabatic temperature rise of OPC and

10%CSF mixes at 0.30 w/b ratio.

60

10% CSF
50 -
40% PFA +10% CSF
60% G GBS+10% CSF

3
2
0)
a
20

1 10 1 0 0

Time (hours)
Figure 7.9: Estimated adiabatic temperature rise of 10% CSF and
PFA/CSF, GGBS/CSF ternary mixes at 0.30 w/b ratio.

149
 C h apter 7 S em i-adiabatic tem perature rise o f H S C

- 10%CSF
50 --40% PFA+10% CSF
60%GGBS+10%CSF
40

30

20

10

1 10 1 0 0

Time (hours)
Figure 7.10: Estimated adiabatic temperature rise of 10% CSF
and PFA/CSF, GGBS/CSF ternary mixes at 0.20 w/b
ratio.

54
52
50

3 44
2 10%CSF I
40%PFA+10%CSF I
60%GGBS+10%CSF I
OPC !

30
0.2 0.22 0.24 0.26 0.28 0.3 0.32

water/binder ratio

Figure 7.11: The effect of water/binder ratio on the peak temperature

rise.

150
 Chapter 7 Semi-adiabatic temperature rise o f H SC

Figure 7.12 shows the effect of binder replacement on the peak temperature of binary
mixes at 0.26 w/b ratio. As can be seen that the peak temperature increases with
increasing CSF content (0-51 kg/m\ corresponding to 0 and 10% cement
replacement). In contrast the PFA and GGBS binders indicate systematic reductions
in peak temperature with increasing cement replacement. It is interesting to note the
presence of 306kg/m^ of GGBS and 204 kg/m^ of PFA have identical peak
temperatures of 42®C. By adding 51 kg/m^ of CSF to these mixes (ternary blends)
lowers the peak temperature to 36°C (figure 7.11). This maybe attributed to the rapid
consumption of calcium hydroxide by the CSF, implying that the PFA and GGBS
ternary components are essentially inactive during the early stages of hydration. This
is further investigated in the next chapter.

10%CSF
50 -
opd 30%GGBS

20%PFA
D 45

40%PFA 60%GGBS

0 50 100 150 200 250 300 350

(510 kg/m3)
Binder replacement (kg/m )

Figure 7.12: The effect of binder replacement on the peak

ten^erature rise of binary mixes at 0.26 w/b
ratio (with a total binder content of 510 kg/m^ )

151
Chapter 7 Semi-adiabatic temperature rise o f HSC

7.6 Conclusions

• Binary blended cement high strength concrete mixes incorporating high levels
of cement replacement by PFA and GGBS, i.e. 40 and 60% respectively,
reduce the semi-adiabatic temperature rise by up to 15%. In contrast the
incorporation of 10%CSF in the binder increases the peak temperature by 4%
at 0.26 w/b ratio.

• The use of ternary blended mixes in high strength concrete have been shown
to reduce the adiabatic temperature rise at all water/binder ratios compared to
the binary blends. The temperature reductions in the 40%PFA+10%CSF and
60%GGBS+10%CSF ternary blends are up to 14%.

• The peak temperature rise of binary and ternary blended high strength
concrete is more dependant on the cement content than the amount of water in
these mixes.

The main purpose of these tests was to provide data for the calculation of temperature
cycles. The compressive strength development of these mixes is described in the next
chapter.

152
Chapter 8 _______________________________________________________________________________________ Strength development o f HSC

CHAPTER 8

STRENGTH DEVELOPMENT OF HIGH STRENGTH CONCRETE

8.1 Introduction

This chapter presents the results of the tests on the con^ressive strength development of
high strength concretes subjected to early age in-situ temperature cycles and standard
curing conditions (20°C). It is divided into three parts:

1. The effect of the level of cement replacement in binary (PFA, GGBS & CSF) and
ternary (PFA/CSF, GGBS/CSF) blends on the compressive strength development
under standard (20°C) and temperature matched curing (TMC) conditions at 0.26 w/b
ratio.

2. The effect of different water/binder ratios on the strength development.

3. An investigation into the microstructure of the cement paste, specifically the porosity,
to provide an explanation of the strength results obtained.

153
Chapter 8 ________________________________________________________________________________________Strength development o f HSC

8.2 Results and discussion

The mixes tested and experimental details were given in chapters 4 and 5 respectively.
The temperature cycles obtained and the method used for their prediction, for each mix,
are given in Appendix B

8.2.1 Binary blends -0.26w/b ratio

The compressive strength development for the control OPC mix cured under standard and
temperature matched curing is shown in figure 8.1. Only the average strengths (of set of
3 cubes) are shown for clarity, and this has subsequently been adopted for all of the
figures in this chapter. The individual cube strengths are included with all the detailed
results in Appendix B.

It can be seen from figure 8.1, that the temperature-matched cubes show the highest early
strength gain. However, by an age of 7 days, the standard cured cubes have attained the
same strength as the temperature-matched cubes, and between 28 and 180 days the
strength gain of the standard cured cubes is greater than the temperature-matched
specimens. This trend has been observed in many studies on normal strength concrete.
For example, Bamforth^*^®^ found that for an OPC mix undergoing temperature-matched
curing, the early age strength was 50% higher compared to standard cured cubes (section
3.5.3).

Mak and Torri^‘“ ^ found a similar relationship in high strength concrete, showing that for
an OPC mix, the temperature matched cured cylinders exhibited high early strengths
when compared to those cured at standard temperature. The temperature matched cured
cylinders continued to develop strength up to 1 year but at a slower rate than standard
water cured cylinders.

154
 Chapter 8 Strength developm ent o f HSC

130
120 -

100 -

D)

80 -

Q.
SC
TMC
50 -

1 10 100 1000
Age log(days)

Figure 8.1: Strength development of OPC mix at 0.26 w/b ratio

subjected to standard and temperature matched curing.

140
130 -
I 120-
z 110 -
% 100 -

90 -

60 - SC
TMC
50 -

1 10 100 1000
Age log(Days)

Figure 8.2: Strength development of 10% CSF mix at 0.26 w/b

ratio subjected to standard and temperature matched
curing.

155
Chapter 8_______________________________________________________________________________________ Strength development o f HSC

Figure 8.2 shows a similar trend for the 10%CSF mix. The temperature-matched cubes
show the highest strength gain up to 28-days, at which the standard cured cubes reach a
similar strength. The long-term strength development up to 180 days is greater in the
standard cured specimens. The early age strengths during the first 28 days are
significantly higher compared to the OPC mix. For example, the temperature-matched
cube strengths at 1,7 and 28 days are 8 6 , 104 and 106 N/mm^ respectively. This trend is
not reflected in the standard cured cube results, which show near identical values between
the OPC and 10%CSF mixes.
These results suggest that the filler effect and the pozzolanic reaction of CSF is enhanced
by increasing the curing temperature.

Mak and Torri^^^^^ also showed that the temperature matched cured strength development
of a binary CSF high strength concrete was appreciably enhanced, up to 28 days, when
compared to concrete cured at standard temperature (page 84). In agreement, Price^'^^
found a similar trend in high strength concrete containing CSF, where similar results
were obtained between the two curing conditions at 28-days (page 83).

The compressive strength development of the 20%PFA and 40%PFA binary mixes is
shown in figures 8 3 and 8.4 respectively. From these results it can be seen that, again,
the temperature-matched cube strengths exceed those of the standard cured cubes at early
ages. However, between 28 days and 180 days the standard cured cubes show the greatest
strength gain at both replacement levels. This is in agreement with Domone and
Soutsos^^®^ and Berry and Malhotra^^^ who found that for high strength concrete, PFA
acts by providing increased strength at late ages (56 to 91 days).

Bamforth^'^”^ found that with a 30%PFA normal strength concrete mix, there was an
accelerated early age strength gain of temperature-matched cubes, at 2 8 -days these were
20% higher than the standard cured value. The long-term strength development was
impaired in the temperature-matched cubes. A similar result is also reported by Coole and
Harrison^'^^) for PFA concrete.

156
 C hapter 8 Strength development o f H SC

120

110

E 100

D)

Q. SC
TMC

40
1 10 100 1000
Age log(Days)
Figure 8.3: Strength development o f 20% PFA mix at 0.26
w/b ratio subjected to standard and tenperature
matched curing.

1 2 0

"“e 1 0 0
E
z
£O)
c
CO

SC
£ TMC
o
O

1 10 100 1000
Age log(Days)

Figure 8.4: Strength development o f 40% PFA mix at 0.26

w/b ratio subjected to standard and temperature
matched curing.

157
Chapter 8_______________________________________________________________________________________ Strength developm ent o f HSC

The compressive strength development for the 30%GGBS and 60%GGBS binary mixes
are shown in figures 8.5 and 8 .6 . The trends in each case are very similar to each other.
It is interesting to note that although these trends are comparable to those seen in
previous mixes, the temperature-matched cube strengths are only higher up to 7 days.
The standard cured cubes subsequently show the highest strength gain up to 180 days.

In the previous chapter, it was found that a binary mix of 60%GGBS gives a longer
dormant period during the eary stages of hydration, implying a slow reaction rate.
Conversely, the strength development in the GGBS mixes is poor compared to the OPC
control mix under standard curing conditions.

Wainwright^^^^^ carried out a similar investigation with GGBS replacements of 50 and

70% in normal strength concretes. He concluded that concretes containing GGBS were
far more sensitive to increases in curing temperature than OPC mixes. In contradiction to
these findings, Mak and investigated high strength concrete mixes containing
GGBS replacements up to 70%. They found that the 7-day temperature matched cylinder
strengths exceeded the 28-day strength of cylinders cured at standard temperatures. These
differences are not clear, but may be attributed to the different water/binder ratios and
types of GGBS used in these two investigations.

8.2.2 Ternary blends - 0.26 w/b ratio

It has been shown in the previous chapters that high levels of cement replacement by
GGBS or PFA in combination with CSF are beneficial in:
1. Increasing the workability of fresh concrete without an excessive dosage of
superplasticiser, and
2. Reducing the estimated adiabatic temperature rise of concrete.

This section reports the compressive strength development of mixes with CSF containing
high levels of PFA (40 and 50%) and GGBS (60, 70 and 80%).

158
Chapter 8 Strength developm ent o f HSC

140

100

g 80

2 40
SC
TMC

1 1000
A gelog(D ays)

Figure 8.5: Strength development of 30% GGBS mix at 0.26

w/b ratio subjected to standard and temperature
matched curing.

120

O)

Q.

1 10 100 1000
Age log(days)

Figure 8.6: Strength development of 60% GGBS mix at 0.26

w/b ratio subjected to standard and temperature matched
curing.

159
C h a p te r 8 S tre n g th d e v e lo p m e n t o f H S C

Figures 8.7 and 8.8 show the compressive strength development under the two curing
conditions for the 40%PFA+10%CSF and 50%PFA+10%CSF mixes respectively. Again
the early age temperature matched cube strengths are significantly greater (between 2 0

and 40 N/mm^) than the standard cured strengths. Furthermore, the temperature-matched
cube strengths continue to show a rapid strength gain until 56 days. Between 90 and 180
days the standard cube strengths only marginally exceed the temperature-matched
strengths.

These trends are different to those observed in the PFA binary mixes and may be
attributable to the rapid reactivity rate of CSF under the influence of high temperature in
addition to the pozzolanic reaction of PFA resulting from the use of these binders. The
influence of these effects is explored in the next section. There is very little data in the
literature concerning the influence of in-situ temperatures (TMC) on the strength
development characteristics of ternary blends, implying limited scope for direct
comparisons to be made. However, figure 8.7 also shows the standard cured cube results
of a 36%PFA +10%CSF mix at 0.26 w/b ratio tested by Soutsos^’°^. It can be seen that
there is close agreement between the two standard cured results.

The strength development of the GGBS ternary mixes are shown in figures 8.9- 8.11. In
the case of the 60%GGBS+10%CSF mix, a benefit of using temperature matched curing
is seen at 1-day. There is a significant increase in the 1-day strength (29 N/mm^) which
creates a definite advantage for early formwork removal. The standard cured cube
strengths achieve a similar strength to the temperature-matched cubes beyond 7 days. In
comparison to the 60%GGBS binary mix, the temperature-matched cube strengths are 30
N/mm^ higher at 28 days. This again suggests that the CSF particles react very rapidly,
contributing to the early age strength development up to 28 days, as seen in the 10%CSF
mix. Figure 8.9 also includes standard cured strength data for a 54%GGBS+10%CSF mix
at 0.26 w/b ratio tested by Soutsos^^l Despite lower 7 day strength, as shown in Soutsos’
results there is a general trend between the two plots.

160
C h a p te r 8 Strength d e v e lo p m e n t o f H S C

140

120 -

Z 100 -
cn

Data from Soutsos for

36% PFA+ 10%CSFmix.

CL
40 — #— SC
— ■— TMC
■- TÉT- - Soutsos (90)

1 10 100 1000
Age log (days)

Figure 8.7: Strength development of 40% PFA + 10% CSF

mix at 0.26 w/b ratio subjected to standard and
temperature matched curing.

140

120

I 100
■£
O) on

SC
TMC

1 10 100 1000
Age log(Days)

Figure 8.8: Strength development o f 50% PFA + 10% CSF mix

at 0.26 w/b ratio subjected to standard and temperature
matched curing.

161
 C h a p te r 8 Stren gth d e v e lo p m e n t o f H S C

140

E 120-

Z 1 00 -

O)
Data by S ou tsos for
54%GGBS + 10%CSFmix.

— ■— TMC
- - -A - • S ou tsos(90)

1 10 100 1000
Age log(days)

Figure 8.9: Strength development of 60% GGBS + 10%

CSF mix at 0.26 w/b ratio subjected to standard and
temperature matched curing.

120

E 100 -

80 -
O)
60 -

40 -

SC
TMC

1 10 100 1000
A ge log(days)

Figure 8.10: Strength development o f 70% GGBS + 10%

CSF mix at 0.26 w/b ratio subjected to standard and
temperature matched curing.

162
Chapter 8_______________________________________________________________________________________ Strength development o f HSC

A more distinct trend is observed in the 70 and 80%+10%CSF mixes, as shown in

figures 8.10 and 8.11 respectively. The temperature-matched cube strengths are up to 15
N/mm^ higher at 1-day for both mixes, this suggests that the CSF component of these
mixes significantly accelerates the early age strength gain due to the rapid rate of the
pozzolanic reaction occurring in the temperature-matched cubes. However, between 7
and 180 days the results show large differences in strength development between the two
curing conditions compared to the previous mixes. This may be attributable to the
consumption of large quantities of calcium hydroxide at early ages due to the higher
curing temperature under temperature matched curing.

These observations clearly show a distinct advantage for ternary mixes, where
traditionally these mixes are known to contribute to strength at later ages to be used in
the practical application of high strength concretes. The early strength development opens
up the possibility of significant reductions in the time before formwork and falsework can
be removed. This in tum can result in considerable increases in speed of construction.

8.2.3 O ther water/binder ratios

The strength development characteristics of the mixes tested at 0.30 w/b ratio are shown
in figures 8.12 to 8.14. Conparison between the two curing conditions for the binary
10%CSF mix (figure 8.12), indicates that the temperature matched cube strengths are
higher than the standard cured cubes up to 28 days. Between 28 and 180 days the strength
gain of standard cured cubes is greater. Figure 8.12 also shows data from Price and
Hynes(^^^\ Their data, determined for a 9%CSF mix at 0.33 w/b ratio, also shows that
temperature matched cubes exhibit higher strengths up to 28 days. These trends are also
similar to the results found at 0.26 w/b ratio.

163
 C h a p te r 8 S tren gth d e v e lo p m e n t o f H S C

1 0 0

90
80
70
60
I« 50
>
CO
40
ÏQ. 30
E
o 2 0
SC
o
1 0 TMC
0

1 10 100 1000
Age log(days)

Figure 8.11: Strength development of 80% GGBS + 10% CSF mix

at 0.26 w/b ratio subjected to standard and temperature matched
curing.

140

120

100

O) 80
c *- X
0)

60 sc
« TMC
g
Q. 40
E — -A- - SC
o (Priœ+Hynes(169))
O 20 -TMC
(Price+Hynes(169))
0

1 10 100 1000
Age log(Days)

Figure 8.12: Strength development o f 10% CSF mix at 0.30 w/b

ratio subjected to standard and temperature matched curing.

164
Chapter 8 Strength development o f HSC

Figure 8.13 shows the strength development of a ternary 40%PFA+10%CSF mix,

subjected to the different curing regimes. The temperature matched cube strengths are
higher up to 56 days; thereafter the standard cured cubes are only marginally greater.
This behaviour is very similar to that found at 0.26w/b ratio and suggests that the general
mechanisms of hydration are the same.

The strength development of a 60%GGBS+10%CSF ternary mix, subjected to the

different curing regimes, is shown in figure 8.14. This shows a similar trend to that found
at 0.26 w/b ratio. The standard cured cube strengths achieve a similar strength to the
temperature-matched cubes at 7 days, but the temperature-matched cubes maintain their
higher strength gain up to 28 days.

The strength development characteristics of the mixes tested at 0.20 water/binder ratio
are shown in figures 8.15 to 8.17. An interesting feature in the strength development of
the 10%CSF mix is seen in figure 8.15. The difference in the 1-day temperature matched
cube and standard cube strengths is as much as 40 N/mm^. This sharp increase in strength
suggests that the CSF reacts rapidly under temperature matched curing and may also be
dependent on the cement content of the mix. It is worth noting that the binder content of
this mix was 590 kg/m^ which gave a peak semi-adiabatic temperature rise of 51° C.
It can be seen that beyond 7 days there is hardly any strength gain in the temperature
matched cubes, whilst the standard cured cubes continue to gain strength up to 180 days.

Figure 8.16 shows the strength development of the ternary 40%PFA+10%CSF mix. A
similar trend to that obtained at 0.3 and 0.26 w/b ratios is observed. The main difference
is that the overall strengths under both curing conditions are higher at 0.2 w/b ratio
compared to the other water binder ratios tested.

165
C h a p te r 8 Strength development o f HSC

140

120 -

100 -

? 80 -

SC
20 - TMC

1 10 100 1000
Age log(days)
Figure 8.13: Strength development of 40% PFA + 10% CSF mix at
0.30 w/b ratio subjected to standard and temperature matched
curing.

140

120 1

z 100 -

^ 80 -

Ë 40 -
SC
2 0 -
TMC

1 1 0 1 0 0 1000
Age log(days)

Figure 8.14: Strength development o f 60% GGBS + 10% CSF mix at

0.30 w/b ratio subjected to standard and temperature matched curing.

166
C h a p ter 8 Strength development o f HSC

160

140
E
E
I 120
O)
1 0 0

>/)
<
Q.
I
o sc
TMC
40
1 1 0 1 0 0 1000
Age log(days)

Figure 8.15: Strength development of 10% CSF mix at 0.20 w/b

ratio subjected to standard and temperature matched curing.

167
 Chapter 8 Strength development o f HSC

140

1 2 0 -

I 100 -

I 80 -

SC
2 0
TMC

1 1 0 1 0 0 1000
Age log(days)

Figure 8.16: Strength development o f 40% PFA + 10% CSF mix at

0 . 2 0 w/b ratio subjected to standard and temperature matched

curing.

140

"Ie 1 2 0 -

e 100 -
05

SC
TMC

1 1 0 1 0 0 1000
A ge log(days)

Figure 8.17: Strength development o f 60% GGBS + 10% CSF mix

at 0.20 w/b ratio subjected to standard and temperature matched
curing.

168
Chapter 8_______________________________________________________________________________________ Strength developm ent o f HSC

The strength development of 60%GGBS+10%CSF ternary mix is shown in figure 8.20.

Although a similar trend is found as in the 0.26 and 0.3 w/b ratios, the overall cube
strengths are higher, up to 180 days, compared to that at 0.3 and 0.26 w/b ratios(figures
8.8 and 8.14).
It is useful to point out that, the PFA/CSF ternary blend mix exhibited lower strengths
than the GGBS/CSF at 0.2 w/b ratio, a feature that was also reported by Soutsos^^®^ for
similar ternary mixes cured under standard conditions at 0.2 water/binder ratio. From the
binder compositions given in Tables 5.1 and 5.2, it can be concluded that there are larger
amounts of lime available in the 60%GGBS+10CSF mix compared to the
40%PFA+10%CSF mix that enable the cementitious reaction to continue at later ages.

8.2.4 Summary of strength results

The results presented so far have shown higher early age but lower long-term strengths in
the TMC specimens compared to the standard cured specimens. The influence of
supplementary cementing materials on the strength development of binary and ternary
blends at 0.26 w/b ratio are summarised in figures 8.18 (a-b) and 8.19 (a-b) The TMC
cube results generally show higher early age strength differences between the control
mixes (OPC and 10%CSF) and their binary and ternary counterparts.

The next section discusses the practical significance of the compressive strength results
obtained.

8.3 Practical Significance

The results obtained from this investigation have significance for both early age and long
term behaviour. The use of temperature-matched curing to determine the in-situ strength
is beneficial, for example in early formwork stripping. Price and Hynes^^^’^ showed that
the relationship between in-situ core strength with cubes cured using different techniques,
indicates that the standard cured specimens overestimate the in-situ compressive strength
and that temperature-matched cubes compare well with cores at 7 and 28 days, and are
closest to the core strengths at 90 days.

169
 C h a p te r 8 S trength d e v e lo p m e n t o f H SC

140

120

E 100

O)

10% C SF

20% PFA

40% PFA

30% G G B S

60% GGBS

1 10 100 1000
(a ) Age log(days)

1 40

120

E
E 100
z

O
c
)
2
CO
-# - 10% C SF

40% +10% C SF
I
Q . 50% +10% C SF
E
<3 60% G G B S+10% C SF

70% G G B S+10% C SF

80% G G B S+10% C SF

1 10 100 1000
Age, log(days)

Figure 8.18 Summary o f standard cured results in (a) binary, and (b)
ternary mixes (0.26 w/b ratio)

170
 C h a p ter 8 S tren gth d e ve lo p m e n t o f H S C

140

120 -
E
I 100 -

p5 80 -
I -# - OPC
I 10%CSF
CO
20%PFA
£
Q.
40%PFA
I 30%GGBS

60%GGBS

1 10 100 1000
(a Age, log(days)

140

120 -
E
^ 100 - e-
z
80 -

CO 10%CSF
^ 60 -
CO 40%+10%CSF
2 50%+10%CSF
g- 40 ,
60%GGBS+10%CSF
Ô ^
20 - 70%GGBS+10%CSF

80%GGBS+10%CSF

1 10 100 1000
^ Age, log(days)

Figure 8.19 Summary of TMC cured results in (a) binary, and (b)
ternary mixes (0.26 w/b ratio)

171
Chapter 8 Strength development o f HSC

The results obtained in this investigation show that the age at which temperature matched
cube strengths are equal to the standard cured cube strengths, ie. ‘cross-over’ point (on
the graph), varies and is largely dependant on both the binder content and type of binder
used (Figures 8.20 and 8.21). In the binary mixes at 0.26 w/b ratio the results show that
the cross-over point shifts from 7 days in the OPC mix to 14 days (20%PFA mix), 10
days (30%GGBS mix) and 28 days (CSF mix).

In ternary blends of PFA, the cross-over point occurs at 72 days. However, in the
60%GGBS+10%CSF ternary blend, this cross-over point occurs at 56 days. This is
significantly reduced to about 7 days for the 70%+I0%CSF and 80%GGBS+10%CSF
ternary blends, which is similar to the OPC mix.

In practice most design codes specify 28 days compressive strength as guidance for
structural design. The current observations suggest standard cured specimens of high
strength concretes exhibiting cross-over points less than 28 days may grossly over
estimate the in-situ compressive strength, and in all mixes the long term strength will be
overestimated. For example, in the 70%GGBS+10%CSF and 80%GGBS+10%CSF
ternary mixes, which have a cross-over at 7 days, their standard cured cube strengths are
around 15 N/mm^ higher than the temperature matched cured cube strengths at 28 days.
This may raise a safety issue in structural design, as this value is higher than the in-situ
strength. These differences should be taken into consideration when specifying strengths
as they may reduce the assumed partial safety factors.

The ultimate strength difference between standard and temperature matched cured
conditions at 180 days also varies with binder composition. Figures 8.22 and 8.23 show
this difference (SC-TMC) in the binary and ternary mixes respectively. Both binary
mixes show a reduction. In ternary PFA blended concretes, the difference is significantly
reduced, and may be due to the filler effect of CSF promoting the strength gain in the
temperature-matched cubes.

172
 Chapter 8 Strength development o f HSC

20 -

1? 15 -

O)

PFA
GGBS

0 10 20 30 40 50 60 70
% Replacem ent

Figure 8.20: Comparison of cross-over points for bunary mixes at

0.26 w/b ratio.

70 -

60 -
U 50 -

S . 40 -

< 30 i PFA/CSF
GGBS/CSF
20 -
10 -

0 10 20 30 40 50 60 70 80 90
% R e p la c em e n t

Figure 8.21 : Comparison of cross-over points for ternary mixes at

0.26 w/b ratio

173
C h a p te r 8 S tren gth d e v e lo p m e n t o f H S C

16

14

12

10

Û
CO 6
D
4
PFA
2 GGBS

0
0 10 20 30 40 50 60 70
% Replacement

Figure 8.22: Ultimate strength difference - Binary mixes at

0.26 w/b ratio

30

25

20

15
O
CO
D 10

PFA/CSF
5
GGBS/CSF

0 20 40 60 80 100

% Replacement (+10%CSF)

Figure 8.23: Ultimate strength difference - Ternary mixes at

0.26 w/b ratio

174
Chapter 8 _______________________________________________________________ ________________________ Strength development o f HSC

The GGBS binary and ternary relationships show a distinct feature. The ultimate strength
difference is the same in both 60%GGBS binary and 60%GGBS+10%CSF ternary mix.
However, at very high levels of replacement (>60%GGBS) in the ternary blends, the
difference in ultimate strengths significantly increases. It is possible that the lower
cement content in these mixes leads to insufficient amounts of calcium hydroxide at later
ages in the temperature-matched cubes. This maybe consumed by the rapid pozzolanic
reaction at early ages.

8.4 Microstructural changes

Most researchers who have investigated the effects of temperature on normal strength
concrete have explained the loss of long term strength, for concrete subject to heat curing,
as the result of physical and chemical changes in the hydrated cement paste, (c.f. Review
section 3.5.4).
To assist in the understanding of the experimental results, a series of tests were
performed on cement paste to assess their physical and chemical properties, namely the:

• Porosity
• Calcium Hydroxide content
• Strength development

The main objective of these tests was to obtain additional useful information on the
mechanisms of hydration. These tests were considered simple and capable of producing
useful information. Other tests e.g. microscopy. X-ray diffraction etc. could not be
carried out within the timescales and are beyond the scope of this research.

175
 C h a p ter 8 Strength development o f HSC

8.4.1 Porosity of hardened cement paste

It is well known that the strength of concrete is intimately associated with the porosity
and pore size distribution of the hardened cement^^^^’^^^l
The porosity experimental details, based on alcohol re-saturation, were given in Chapter
5 and the mixes tested in Chapter 4.

Figure 8.24 shows the porosity of the two control mixes cured under standard and
temperature matched curing conditions up to 90 days at 0.26 w/b ratio. It is evident that
the temperature matched cured cement paste specimens exhibit lower porosities. This is
in agreement with results reported by other workers, who found that high curing
temperatures reduce the porosity of cement paste^^^^’^^^’’^^\ Of particular significance is
the tendency of the 10%CSF paste to exhibit no change in porosity after 7-days under the
two curing conditions.
This suggests that the rapid pozzolanic reaction of CSF blocks pores, or encapsulates
areas of porosity restricting any further alcohol intrusion. A similar observation has also
been reported by Day and Marsh^^'^\

OPC mix

d: 10 10% CSF mix

sc
TMC
SC
TMC

0 20 40 60 80 100
Age (days)

Figure 8.24: Porosities of OPC and 10% CSF mixes at 0.26 w/b ratio.

176
Chapter 8 ________________________________________________________________________________________Strength developm ent o f HSC

Figures 8.25 and 8.26 show the porosity of 40%PFA+10%CSF and

60%GGBS+10%CSF ternary mixes at 0.26 w/b ratio respectively. It can be seen from
these plots that there is a general reduction in porosity with time and both mixes have
lower porosities than the control 10%CSF mix. The values obtained for the PFA/CSF
ternary mix beyond 28 days are lower compared to the GGBS/CSF mix. This change in
pore structure between the two mixes may be attributed to the combined pozzolanic
reaction of PFA and CSF with the calcium hydroxide. The consumption of the calcium
hydroxide by this pozzolanic reaction may possibly be responsible for the reduction in
porosity. This is further investigated in the next section.

These results can to some extent explain the strength development characteristics found
in the CSF binary and ternary concrete mixes. The enhanced early age temperature
matched cured strengths of these mixes may be directly related to the rapid pozzolanic
reaction and the filler effect of CSF.

8.4.2 Calcium Hydroxide Content

In order to further substantiate the effect of CSF on the strength development

characteristics of high strength concrete, the calcium hydroxide Ca(0 H) 2 contents were
determined for the OPC and 10%CSF mixes by using Thermogravimetric Analysis
(TGA). Unfortunately, due to experimental problems, only the results from standard
cured paste specimens are reliable. Figure 8.27 shows the Ca(0 H) 2 contents up to 28
days. The OPC sample shows an increase in the Ca(0 H) 2 content up to 28 days. In
contrast the Ca(0 H) 2 content in the binary IO%CSF mix is lower compared to the OPC
mix, and remains near constant up to 28 days.

177
 Chapter 8 Strength development o f HSC

20 -
OPC mix

15 -
SC
(A TMC
2
SC
£
TMC

0 10 20 30 40 50 60 70 80 90 100
Age (days)

Figure 8.25: Porosity of 40%PFA+10% CSF mix at 0.26

w/b ratio.

25

20 O P C mix

15

SC

10 TMC
SC
TMC

0
0 10 20 30 40 50 60 70 80 90 100

Figure 8.26: Porosity of 60%GGBS+10% CSF mix at 0.26

w/b ratio.

178
 C h a p te r 8 S tren gth d e v e lo p m e n t o f H S C

16
14

c 12
Sc
8 10
0
•O
X 8
2
6
4
OPC
2 CSF
0
0 5 10 15 20 25 30
Age(days)

Figure 8.27: Calcium hydroxide contents of OPC and 10% CSF mixes at
0.26 w/b ratio.

2 .7

^ 2.6

2.3

2.2
GGBS/CSF
PFA/CSF

0 5 10 15 20 25 30

Age (days)

Figure 8.28: Calcium hydroxide contents of 40%PFA+10%CSF and

60%GGBS+ 10% CSF mixes at 0.26 w/b ratio.

179
Chapter 8________________________________________________________________ _______________________ Strength development o f HSC

This suggests that the rate of consumption of Ca(0 H) 2 by the pozzolanic reaction of CSF
maybe equivalent to the rate of its formation by the hydration reaction. Sun and Young^^"^
have shown by using the NMR technique that there is very rapid consumption of CSF
over the first 28 days and particularly within the first 7 days. This further substantiates
the porosity results and explains the rapid strength development characteristics (up to 28
days) of the CSF binary high strength concretes. Traetteburg^^'^) determined the Ca(0 H) 2
content for a 10%CSF paste at a water/binder ratio of 0.5. He found that the Ca(0 H) 2
reached a maximum at 6 days and then declined from 7 to 15 days and was interpreted as
being due to the rapid reaction of CSF with lime to give additional C-S-H gel. Similar
results were also obtained by Chattel]i et aP ’^\

Figure 8.28 shows the Ca(0 H) 2 contents up to 28 days for the 40%PFA+10%CSF and
60%GGBS+10%CSF ternary blends. It can be seen that there is a substantial decrease in
the overall calcium hydroxide contents compared to the 10%CSF binary mix (fig.8.27)
The Ca(0 H) 2 content in the PFA/CSF mix is seen to gradually reduce with time
compared to the GGBS/CSF mix which levels off. This is in agreement with predicted
variations in the Ca(0 H) 2 content with hydration time for binary PFA cement paste made
by Parrot^ from his mathematical models and those calculated by Dalziel and
Gutteridge^^^^^Halse^^^'^^ has also observed a substantial drop in Ca(0 H) 2 for binary PFA
cement paste at and beyond 28 days. This further substantiates the porosity results for
these mixes and indicates that the consumption of Ca(0 H) 2 by the pozzolanic reaction is
more pronounced in ternary mixes.

8.4.3 Strength development of cement paste.

Two control paste mixes, OPC and 10%CSF, at 0.26w/b ratio were subjected to standard
and temperature matched curing. The temperature cycle used for each paste mix was
identical to that used in the respective concretes. Figures 8.29 and 8.30 show the
strength development for the OPC and 10%CSF paste mixes. The results show that the
temperature matched cured paste specimens, for both mixes, are higher compared to the
standard cured specimens at all ages.

180
 Chapter 8 Strength development o f HSC

130
120
£ 110
5 100

CL

SC
TMC

0 5 10 15 20 25 30
Age (days)
Figure 8.29: Strength development of OPC paste specimen at 0.26 w/b ratio.

140
_
N 130
I 120
& 110
f 100
CD
% 90
I 80
CD

g 70
E
o
60 SC
^ 50 TMC
40
0 5 10 15 20 25 30
Age (days)

Figure 8.30: Strength development of CSF paste specimen at 0.26

w/b ratio.

181
Chapter 8 ________________________________________________________________________________________Strength development o f HSC

These observations suggest that high strength paste specimens perform better than high
strength concrete when cured under temperature-matched conditions (figures 8.4 and
8.5). This is in contradiction to the behaviour found in heat cured normal strength
concretes, where most researchers who have explained the loss of strength as a result of
physical and chemical changes in the hydrated cement paste^^“ '^^’*“\

The current results therefore suggest that changes in the hydrated cement paste are not the
sole reason for the reduced strength obtained in the temperature matched cured
specimens. A possible explanation may lie in the nature of the transition zone between
the hydrated cement paste and aggregate particles. It is well documented that CSF
improves the strength of concrete by enhancing aggregate-matrix bond (i89.190.214.215)
However, when concrete is subjected to a temperature cycle, the heating and cooling
parts of the cycle may cause internal stresses between the paste/aggregate interface, due
to the differing coefficients of thermal expansion. This can lead to the development of
micro-cracks that may effect the long term strength development. Laplante and AitcW**)
have also arrived at a similar explanation.

182
Chapter 8_______________________________________________________________________________________ Strength development o f HSC

8.5 Conclusions

Standard curing at 20° C underestimates the early age in-situ strength development of
high strength concrete; but overestimates the long term strength.

Binary blended cements vyith high levels of cement replacement by PFA and
GGBS, i.e. 40% and 60% respectively reduce the strength of standard cured high
strength concretes by up to 15%.

The strength development for the 10% CSF TMC cubes is higher than the
equivalent OPC concrete by as much as 20%. Binary mixes of PFA and GGBS
exhibit lower early-long term TMC strengths compared to the equivalent OPC
concrete.

The use of ternary blends significantly enhance the 1-day strength of high strength
concrete when subjected to temperature matched curing. The strength
enhancement in the 40%PFA+10%CSF and 60%GGBS+10%CSF blends is
equivalent to 75 and 80% respectively.

• The paste porosities of the binary 10%CSF mix, the ternary 40%PFA+10%CSF
and 60%GGBS+10%CSF mixes at 0.26 w/b ratio are up to 50, 80, and 75% lower
than the equivalent OPC mix respectively.

The calcium hydroxide contents of the binary 10%CSF mix and the ternary
40%PFA+10%CSF and 60%GGBS+10%CSF mixes at 0.26 w/b ratio are up to
55, 85, and 84% lower than the equivalent OPC mix respectively.

A possible explanation for the difference between long term standard and
temperature matched cured strengths lies in the stresses generated in the
paste/aggregate interface as determined by comparison of paste and concrete
strengths.

183
Chapter 9 _____________________________________________________ Conclusions an d recom m endations f o r fu rth e r work

CHAPTER 9

CONCLUSIONS AND RECOMMENDATIONS

FOR FURTHER WORK

9.1 Introduction.

This chapter presents the conclusions from the experimental work carried out in Chapters
6 to 8 , followed by recommendations for further work.

9.2 Conclusions.

(i) Fresh properties of high strength concrete.

1. Partial cement replacement by CSF and PFA reduces the quantity of

superplasticiser required to obtain a 2 0 0 ± 2 0 mm slump compared to the
equivalent OPC concrete. The 15% CSF binary mix gives no further
reductions in superplasticiser dosage compared to the 10% CSF mix,
(Figure 6.1). In contrast, the use of GGBS increases the superplasticiser
dosage required to obtain an equivalent slump.

2. The combination of 10% CSF with 40% PFA or 60% GGBS, at 0.26 w/b
ratio, reduces the superplasticiser dosage compared to their equivalent
binary mixes (Figure 6.2).

3. As the water/binder ratio decreases from 0.30-0.20, the superplasticiser

dosage required to obtain the target slump increases (Figures 6.4-6.6 ).

4. Partial cement replacement by CSF (5-15%) at 0.26 w/b ratio, reduces

both the yield value (g) and plastic viscosity (h) compared to the OPC
concrete, (Figure 6.7). At 15% CSF the-e are no substantial improvements
in workability compared to 10% CSF.
184
 Concl usions and recom m endations f o r fu rth e r work

Partial cement replacement by PFA (20,40%) reduces the yield value

(g), but increases the plastic viscosity (h) compared to the OPC
concrete, (Figure 6 .8 ).
Partial replacement by GGBS (30, 60%) increases both the yield value
(g) and plastic viscosity (h) compared to the OPC concrete (Figure 6.9).

5. Ternary blends of 10% CSF with PFA or GGBS increase the workability
by reducing both the yield value (g) and plastic viscosity (h) at 0.26 w/b
ratio. The optimum replacement levels are 40% PFA + 10% CSF and 60%
GGBS + 10% CSF, (Figures 6.10 and 6.11).

6 . Both the yield value (g) and plastic viscosity (h) increase as the w/b ratio
decreases from 0.30 to 0.20 (Figure 6.12).

(ii) Estimated-adiabatic temperature rise of high strength concrete.

1. Partial cement replacement by 10% CSF reduces the dormant period, but
does not significantly affect the peak temperature (Figure 7.1).
PFA and GGBS, at low replacement levels of 20 and 30% respectively,
give comparable peak temperatures to the OPC mix (Figures 7.2 and 7.3).
At higher replacement levels of 40% PFA and 60% GGBS the peak
temperatures are significantly reduced .

2. Ternary blends of 10% CSF with PFA (40,50%) or GGBS (60,70,80%)

reduce the peak temperature compared to the binary control mix. The
highest temperature reductions are obtained with the 50% PFA + 10%
CSF and 80% GGBS + 10% CSF mixes (Figures 7.4 and 7.5).

3. As the w/b ratio decreases from 0.30 to 0.20, the peak temperature for
binary and ternary blended mixes increases (figure 7.11).

185
Chapter 9______________________________________________________ Conclusions a n d recom m endations f o r fu rth e r work

(iii) Strength development of high strength concrete.

1. The strength development of temperature matched cured (TMC) cubes are

typically higher than the equivalent standard cured (SC) cubes at early
ages, but lower at later ages.

2. The cross-over point (where the TMC and SC strengths are equal) in
binary mixes of CSF, PFA or GGBS occurs at 28, 14 and 10 days
respectively. In ternary blends of PFA with 10% CSF, the cross-over point
is at 72 days, and is independent of the PFA replacement level. In ternary
blends of 10% CSF with GGBS (60, 70 or 80%) the cross-over is at 56, 7,
and 7 days respectively (Figures 8.20 and 8.21).

3. The strength development for the 10% CSF TMC cubes is higher than the
equivalent OPC concrete. Binary mixes of PFA and GGBS exhibit lower
early-long term TMC strengths compared to the equivalent OPC concrete
(figures 8 .2 -8 .6 ).

4. Ternary blends of 10% CSF with PFA or GGBS have lower TMC
strengths compared to the control 10% CSF concrete (figures 8.7-8.11).

5. The early age TMC strengths increase with decreasing water/binder ratio.
The ultimate TMC strengths at 0.30, 0.26 and 0.22 w/b ratios in a 10%
CSF mix are however similar.

186
 Conclusions an d recom m endations f o r fu rth e r work

(iv) Micro-structural properties/characteristics.

1. The porosity of TMC and SC cement paste specimens decreases with

time. The TMC specimens exhibit lower porosities than the SC
specimens.
The porosity of a 10% CSF mix, decreases during the first 7 days and is
lower than the OPC mix (Figure 8.24).
The ternary blends of 10% CSF with 40% PFA or 60% GGBS have
comparable porosities to the 10% CSF mix during the first 7 days, but
exhibit continued porosity reductions with time (Figures 8.25 and 8.26).

2. The calcium hydroxide content for an OPC mix at 0.26 w/b ratio increases
with time. In contrast, the 10% CF mix shows lower and near constant
calcium hydroxide contents up to 28 days (Figure 8.27).
Ternary blends of 10% CSF with 40% PFA or 60% GGBS exhibit lower
calcium hydroxide contents than the 10% CSF control mix (figure 8.28).

3. The strength development of standard and temperature matched cured

cement paste specimens is higher than the equivalent concrete at 0.26 w/b
ratio. This shows that the paste/aggregate interface plays a contributing
role in reducing the strength of temperature matched high strength
concrete.
The temperature matched cured cement paste strengths are higher at all
ages (Figures 8.29 and 8.30).

187
Chapter 9_____________________________________________________ Conclusions an d recom m endations f o r fu rth er w ork

9.3 Recommendations for further work.

1. The contribution of superplasticiser dosage to the workability of high strength

concrete has been investigated by using one type of superplasticiser. It would be
beneficial to assess the effects of different superplasticisers in reducing the yield
value (g) and plastic viscosity (h) of fresh concrete.
The use of ternary blends in increasing the workability of fresh concrete has been
demonstrated. It would be of interest to compare the effects of varying CSF
contents.
Further work is also required to evaluate the effects of different two-point test
devices (c.f. section 3.2.1) on the measured workability properties.

2. The action of a superplasticiser disperses the cement particles and exposes a

greater surface area to hydration. It is not clear whether this has any intrinsic
effect on the temperature rise of concrete. This and the effect of using different
cement replacement materials (Rice husk, calcined paper sludge, metakaolin etc.)
on reducing the temperature rise should be explored.
Alternative test methods for assessing the adiabatic temperature rise could also be
investigated.

3. The mixes at 0.3 w/b ratio have shown marginally lower strengths, reduced peak
temperatures but much better workability characteristics than at 0.26 and 0 . 2 0 w/b
ratios. Further is therefore needed to determine these properties at higher
water/binder ratios (0.3 to 0.4).

188
Chapter 9______________________________________________________ Conclusions an d recom m endations f o r fu rth e r work

4. The role of supplementary cementing materials in reducing the porosity of cement

pastes has been investigated using only one technique (alcohol re-saturation). A
more reliable and accurate measurement of porosity can be obtained by using MIP
(mercury intrusion porosimetry) and should be used in any further work. However
the results should be interpreted with caution as it has been shown that the
pressures required can damage the structure of the hydrates.

5. Other techniques, such as SEM and NMR, reduce the effect of carbonation and
should be used to determine the calcium hydroxide contents.
The influence of the transition zone in reducing the compressive strength in TMC
specimens needs to be validated by micro-structural evidence using SEM to
measure the thickness of this zone.

189
 ____________________________________ References

REFERENCES

1. ACI Committee 363, State-of-the-art Report on High Strength Concrete, ACI Journal,
Proceedings V.81,No.4, 1984, pp.364-411.

2. Burnett, I. High Strength Concrete in Melbourne, Australia, Concrete International,

V .ll,N o.4,1989,pp.l7-25.

3. Aitcin, P.C. and Baalbaki, M. Canadian Experience in Producing and Testing HPC,
ACI SP -I59,1996, pp. 295-308.

4. Chem, J.C., Hwang, C.L. and Tsai, T.H. Research and Development of HPC in
Taiwan. Concrete International, V.17, No. 10,1995, pp.71-77.

5. Nagataki, S., The use o f Superplasticisers, Seminar on Special Concretes, 8 ^ FIP

Congress, Wexham Springs, 1978, p. 15.

6 . Murata, J., Kawai, T., and Kokubu, K., Studies on the utilization of water-reduced
high strength concrete for monorail piers. Development in the use of
superplasticisers, ACI SP-6 8 , 1981, pp. 41-60.

7. Gjorv, O.E., High Strength Concrete, Advances in Concrete Technology, CANMET,

1994, pp. 19-82.

8 . Bernhardt, C., and Fynboe, C., High Strength Concrete Beams, Nordic Concrete
Research, 1986, No.5, pp. 19-26.

9. Zikeyev, L.N., and Petshold, T.M., Strength and Crack Resistance o f Centrifuged
Columns o f 1 Storey Industrial Buildings, FIP/CPCI Symposium, V.3, 1984, pp. 83-
89.

10. Clarke, J.L., and Pomeroy, C.D., Concrete Opportunities for the Structural Engineer,
Structural Engineer, V.63A, No.2, 1985, pp.45-53.

11. Ryan, W.G., and Potter, R.J., Application o f High performance Concrete in Australia,
High Performance Concrete Proceedings, ACI Conference, Supplementary papers,SP-
159, 1994,pp.l-15.

12. Lew, H.S., and Frohnsdorff, G.J., U.S Programme on High Performance Concrete,
High Performance Concrete Proceedings, ACI Conference, Supplementary papers,
SP-159,1994,pp.67-74.

13. Yoshida, T., Making Concrete with the highest strength. Journal of the Japan Society
of Civil Engineers, V.26, N o .ll, 1930, pp. 997-1016.

190
________________________________________________________________________ References

14. FIP/CEB High Strength Concrete, State-of-the-art Report, CEB Bulletin d’

information, No. 197, 1990.

15. Nagataki, S., High Strength Concrete - History and Progress, Technical Report No.
40, Dep. of Civil Engineering, Tokyo Institute of Technology, 1989.

16. Saucier, K., High Strength Concrete, past, present andfuture. Concrete International,
V.2, No.6 , 1980, pp. 46-50.

17. Domone, P.L.J, and Soutsos, M.N., An Approach to the Proportioning of High
Strength Concrete mixes. Concrete International, October 1994, pp. 26-31.

18. Carrasquillo, R.L., Production o f High Strength Pastes, Mortars and Concretes, Very
High Strength Cement-Based Materials, Symposium Proceedings, V.42, 1984,
pp.145-156.

19. Mehta, P.K., and Aitcin, P.C., Microstructural basis for Selection of Materials and
Mix Proportions for High strength Concrete, High Strength Concrete, ACI SP-121,
1990, pp. 265-286.

20. Freedman,S., High Strength Concrete, Modem Concrete, V.34, No.6,1970, pp.29-36,
No.7, 1970, pp. 28-32; No.8 , 1970, pp. 21-24; No. 10, 1971, pp. 16-23.

21. Teychenne, D.C., Franklin, R.E., and Emtroy, H.C., Design of Normal Concrete
Mixes, Building Research Establishment Report, 1988.

22. ACI Committee 211, Standard Practice for Selecting Proportions for Normal,
Heavyweight and Mass Concrete, ACI Manual of Concrete Practice, Part 1, 1991.

23. De Larrard, F., Hu, C. and Sedran, T. Best packing and specified rheology: two key
concepts in high performance concrete mixture propotioning, Adam Neville
Symposium on Concrete Technology, CANMET, 1995, pp. 109-127.

24. De Larrad, F and Sedran, T., Optimization of Ultra High Performance Concrete By
the Use o f the Packing Model, Cement and Concrete Research, V.24 No.6,1994, pp.
997-1009.

25. Gopalan, M.K., Mix Design and Microstructure Effects On High Strength Concrete,
Utilization Of High Strength Concrete, Conference Proceedings, Lillehammer,
Norway, V.2, 1993, pp. 722-743.

26. Detwiler, G., High Strength Silica Fume Concrete-Chicago Style, Concrete
International, V. 14 No. 10, 1991, pp.32-36.

191
________________________________________________________________________ References

27. Parrott, L.J., High Strength Concrete Properties: A Bibliography, Cement and
Concrete Association Services, 1988, pp.46.

28. Naaman, A.E., High Strength Fibre Reinforced Cement Composites, Very High
Strength Cement-Based Materials, Symposium Proceedings, V.42, 1984, pp.96-105.

29. Eden, N.B., and Bailey, J.E., The Flexural Strength and Fracture Toughness of a
Normal and a High Strength Polymer Modified Portland Cement, Very High Strength
Cement-Based Materials, Symposium Proceedings, V.42, 1984, pp.79-88.

30. Roy, D.M., Gouda, G.R., and Bobrowsky, A,. Very high Strength Cement Pastes
prepared by Hot p ressing andother High Pressure Techniques, Cement and Concrete
Research, V.2, No.3, 1972, pp. 349-366.

31. Roy, D.M., Nakagawa, Z.E., Scheetz, B.E., and White, E.L., Optimized High Strength
Mortars: Effects of Chemistry, Particle Packing and Interface Bonding, Very High
Strength Cement-Based Materials, Symposium Proceedings, V.42, 1984, pp. 179-188.

32. FIP Commission, Tentative Interim Report On High Strength Concrete, ACI Journal,
Proceedings, V.64, No.9, 1967, pp.556-557.

33. Tattersall, G.H. Relationship between the British Standard Tests for Workability and
the Two-Point Test, Magazine of Concrete Research, V.28, No.96,1976, pp. 143-147.

34. Tattersall, G.H.and Banfill, P.F.G., Rheology o f Fresh Concrete, Pitman Publishing,
London, 1983.

35. Wainwright, P.J. and Tolloczko, J.A. Early and later age properties of blast furnace
slag cements, mortars and concrete, ACI Journal Vol. 79, Nov-Dee, 1982, pp. 444-
457.

36. High Strength Concrete in Chicago High Rise Buildings, Task Force Report No.5,
Chicago Committee on High Rise Buildings, Feb. 1977,p.63.

37. Diederichs, U., Jumppanen, U.M., Penttala, V., Material properties of high strength
concrete at elevated temperatures, LABSE, 13* Congress, Helsinki, 1988.

38. Colaco, J.P., 75-Storey Texas Commerce Plaza, Houston: The Use of High Strength
Concrete, High strength Concrete, Special Publication SP-87, ACI, pp. 1-8.

39. de Larrard F., and le Roy, R. The influence of mix composition on mechanical
properties o f high performance silica fume concrete. Fly ash, silica fume and natural
pozzolans in concrete, ACI SP-132, pp.965-986, 1993.

192
________________________________________________________________________ References

40. Randall, V., and Foot, K., High Strength Concrete for Pacific First Centre, Concrete
International, V.IO, No.4, 1989, pp. 14-16.

41. Schmidt, W., and Hoffman, E.S., 9000-psi concrete- why?, why not?. Civil
Enginnering, ASCE, V. 45, No. 5, 1975, pp.52-55.

42. Cook, J.E,. 10,000psi Concrete, Concrete International, Oct. 1989, pp. 67-75.

43. Concrete Giant Rises on Chicago Skyline, Concrete Construction, Jan. 1990, pp.5-10.

44. Godfrey, K., Concrete Strength Record Jumps 36%, Civil Engineering, Oct. 1987,
pp. 84-86.

45. Aitcin, P.C., Laplante, P., and Bedard, C., Development and Experimental Use o f 90
MPa Field Concrete, High Strength Concrete, SP-87, ACI, Detroit, 1986, pp. 14-17.

46. Ryell, J., and Bickley, J.A., Scotia Plaza: High Strength Concrete for Tall Buildings,
Utilization of High Strength Concrete, Proc. Symposium in Stavanger, Norway, June,
1987, pp.641-654.

47. Carpenter, J.E., Applications of High strength Concrete for Highway Bridges, Public
Roads, V.44, No.2, 1980, pp.76-83.

48. Burgess, A.J., Ryell, J., and Buting, J., High Strength Concrete for the Willows
Bridge, ACI Journal, V.67, No.8 , 1970, pp. 611-619.

49. Malier, Y., Brazillier, D., and Rio, S., The Bridge of Joigny, Concrete International,
V.13,No.5, 1991,pp.40-42.

50. Yonezawa, T., The Contribution o f Fluidity Improving Technology to the Widespread
use o f High Strength Concrete, Radical Concrete Technology, Proceedings of
International Conference, Dundee, 1996, pp.525-542.

51. Aitcin, P.C., High Performance Concrete, E&FN SPON, 1998.

52. Price, W.F., Two Examples of High Performance Concrete in Practice, Part 2,
Quality Concrete, May, 1995, pp. 127-130.

53. Price, W.F., Two Examples of High Performance Concrete in Practice, Part 1,
Quality Concrete, April, 1995, pp.99-102.

54. Federal Highway Administration., Bridge Views, Issue No.8 , March 2000,USA.

193
________________________________________________________________________ References

55. Jakobsen, B., Eidenes, A., and Olsen, T.O., Recent Developments and Potentials for
High Strength Offshore Concrete Platforms, Utilization of High Strength Concrete,
Proc. Symposium., Norway, June, 1987, pp.585-596.

56. Ronnenberg, H. and Sandvik, M., High Strength Concrete for North Sea Platforms,
Concrete International, V.12,No.l, 1990, pp.29-34.

57. Hoff, G.C., and Elimov, R., Concrete Production for the Hibernia Platform,
Supplementary Papers, 2“^ CANMET/ACI International Symposium on Advances in
Concrete Technology, Nevada, June 1995, pp.717-739.

58. Thomas, M., Mukheijee, P.K, Sato, J. and Everitt, M. Effect o f fly ash composition on
thermal cracking in concrete. Fly Ash, Silica Fume, Slag and Natural Pozzolanas in
Concrete. 5^ CANMET/ACI Conference, 1995, pp.81-98

59. Copen, M.D., Problems Attending Use of Higher Strength Concrete in Thin Arch
Dams, ACI Journal, Proc., V.72, No.4, 1975, pp. 138-140.

60. Gjorv, O.E., Baerland, T., and Ronning, H.R., High Strength Concrete for Highway
Pavements and Bridge Decks, Utilization of High Strength Concrete, Proc. Symp.,
Norway, June, 1987,pp.l 11-122.

61. Lea., The chemistry of cement and concrete, 3^^ edition, Edward Arnold Publishers,
London, 1988

62. CEB Bulletin d ’information No.222., Application of High Performance Concrete,

Report of the CEB/FIP Working Group on High Strength/High Performance Concrete,
6 6 pp, 1994.

63. Yonezawa, T., Okuno, T., Mitsui, K., Numakura, N., Ohura, T., and Sato, M.,
Bottom-Up Concreting into Steel Tube Column Filled With Ultra High Strength
Concrete Using Silica Fume, Concrete Journal, Japan Concrete Institute, V.31,No. 12,
1993, pp.22-33.

64. Fitzgibbon, M.E., Large Pours 2, Heat Generation and Control, Current Practice
Sheet No.35, Concrete Magazine, Dec. 1976, pp.33-35.

65. General Discussion, Proceedings of the Symposium on Large Pours for Reinforced
Concrete Structures, Concrete Society, 1973, Paper 2, p.87.

6 6 . Bamforth, P.B., In-Situ Measurement of the Effect of Partial Portland Cement

Replacement Using Either PFA or GGBS on the Performance of Mass Concrete,
Proceedings of the Institution of Civil Engineers, Part 2, 1980, pp.777-800.

194
________________________________________________________________________ References

67. Domone, P.L.J., and Soutsos, M.N. Development of high strength concrete mixtures
with low heat generation. Concrete for Environment Enhancement and Protection.
E&FN Spon, London, 1996.

6 8 . Shah, S.P., Fracture Toughness for High Strength Concrete, ACI Materials Journal,
V.87, No.3, 1990, pp.260-265.

69. Ritchie, A.G.B. The Rheology o f Fresh Concrete. Journal of the Constmction
Division, Proceedings of the American Society of Civil Engineers, January 1968, pp.
55-74.

70. Cabrera J.G., and Hopkins C.J., A modification of the Tattersalls two-point test
apparatus for measuring concrete workability. Magazine of Concrete Research,
Vol.36 No. 129, pp.237-240, 1984.

71. Glanville,W.H, Collins ,A.R, and Matthews, D.D. The grading of aggregates and the
workability o f concrete. 2°^ edition. Road Research Technical Paper no. 5, 1947.

72. Cusens, A.R. The measurement o f the workability o f concrete mixes. Magazine of
Concrete Research, Volume 8 No. 22, 1956, pp.22-30

73. Hughes, B.P and Bahramian, B. Workability of Concrete: a comparison of existing

tests. Journal of Materials, Volume 2 No. 3, 1967, pp.519-536.

74. Tattersall,G.H. Relationship between the British Standard Tests for Workability and
the Two-Point Test. Magazine of Concrete Research, Volume 28 No. 96, 1976, pp.
143-147.

75. Tattersall, G.H and Banfill, P.F.G. Rheology of Fresh Concrete, Pitman Publishing,
London, 1983.

76. Tattersall, G.H. Workability and Quality Control o f Concrete. E&FN Spon
Publishing, London, 1991.

77. Tattersall, G.H and Bloomer, S.J. Further development of the two point test for
workability and extension o f its range. Magazine of Concrete Research, Volume 31
No. 109, 1979,pp.202-210.

78. Hu, C, de Larrard, F, Sedran, T, Boulay, C, Bose, F and Defiorenne, F. Validation of

BTRHEOM, the new rheometerfor soft-to-fluid concrete. Materials and Structures,
Volume 29, 1996, pp.620-631.

79. Larrard, F, Hu, C, Sedran, T, Szitkar, J C, Joly, M, Claux, F and Derkx, F. A new
rheometer for soft-to-fluidfresh concrete. ACI Materials Journal, May-June 1997,
pp.234-243.

195
 References

80. Wallevik, O.H and Gjorv, O.E. Development of a coaxial cylinder viscometer for
fresh concrete. Properties of Fresh Concrete, Chapman & Hall, London, 1990, pp-
213-224.

81. Mork, J.H. A presentation o f the BML - viscometer. Production methods and
workability o f concrete, ed. P.J Bartos, D.L Mams and D.J Clelland, E&FN Spon,
London, 1996, pp.369-376.

82. Hu, C, Larrard, F, and Gjorv, O.E. Rheological testing and modelling of fresh high
performance concrete. Materials and Stmctures, Volume 28, 1995, pp. 1-7.

83. Wallevik, O.H and Jankovic, D. Rheological measurements on fresh mortar with a
new measuring system on the BML viscometer. Icelandic Building Research Institute,
Iceland, 1996.

84. Uchikawa, H, Uchida, S and Okamura,T. The influence of blending components on

the hydration of cement minerals and cement. Review of the 41®* general meeting. The
cement association of Japan, pp.36-39.

85. ASTM C595-80 Standard specification for blended hydraulic cements. Annual book
of ASTM standards. Part 13, page 353.

8 6 . Berry, E.E and Malhotra, V.M. Fly Ash in Concrete, Supplementary Cementing
Materials for Concrete, Canadian Government Publishing Centre, Ottawa, Canada,
1987,pp.35-163.

87. Nagataki O. Mineral admixtures in concrete: state of the art trends, American
Concrete Institute, SP 144-22 pp.447-482

8 8 . Coleman E and Diamond, S. Studies of low porosity concretes designed for bridge
deck applications: Introductory remarks, mix designs, strength development and
rheological properties. Cement and concrete Research, Vol. 14, 1984, pp.670-678.

89. Owens, P.L. Fly Ash and its usage in concrete. Journal of the Concrete Society,
Volume 13 No.7, 1979, pp.21-26.

90. Soutsos, M.N. Mix Design, Workability, Adiabatic Temperature and Strength
development o f High Strength Concrete. PhD thesis. University of London, 1992.

91. Hansen ,M. Modified DOE mix design method for high volume fly ash concretes and
controlled low strengths. Magazine of Concrete Research, Vol.44 No. 158, 1992,
pp.39-45.

196
________________________________________________________________________ References

92. Besari, M.S, Munaf, D.R and Iqbal, M.M Stability of mechanical properties and
interface density of high performance fly ash concrete. Radical Concrete
Technology, Proceedings of the International Conference, Dundee, 1996, pp. 47-56.

93. Ellis, C. The application of the two-point workability test and British Standard tests
to OPC/PFA concretes. Proceedings of International Symposium on the use of PFA in
concrete, 1982, pp. 121-132.

94. Banfill, P.F.G. An experimental study o f the effect of PFA on the rheology offresh
concrete and cement paste. Proceedings of the International Symposium on the use of
PFA in concrete. 1982, pp. 160-171.

95. Gjorv, O.E. Concrete workability: A more basic approach needed. Selected research
studies from Scandanavia - Report TVBM 3078, Lund Institute of Technology, Lund,
1997, pp.45-56.

96. Meusel, J.W, and Rose, J.H. Production of blast furnace slag at Sparrows Point, and
the workability and strength potential o f concrete incorporating the slag. Special
Publication SP-79 American Concrete Institute, Detroit, 1979, pp.867-890.

97. Tachibana, D, Imai, M, Yamazaki, N, Kawai, T and Inada, Y. High strength concrete
incorporating several admixtures. Special Publication SP-121, American Concrete
Institute, Detroit, 1990, pp.309-330.
'V
98. Read, P, Carrette, G.G, and Malhotra, V.M. Strength development characteristics of
High strength concrete incorporating supplementary cementing materials. SP-121,
1990, pp. 527-547.

99. Taylor, H.F.W, Cement Chemistry, 2^^ Edition, Thomas Telford Publishing, London,
1997.

100. ACI Committee 226, Silica Fume in Concrete, ACI Materials Journal, March-April,
1987,p p .158-166

101. Malhotra, V.M.and Carrette, G.G. Silica Fume Concrete —properties, applications
and limitations. Concrete International, May 1983, pp. 40-46

102. Yogendran, V, Langan, B.W, Haque, M.N and Ward, M.A, Silica Fume in High
Strength Concrete. ACI Materials Journal, March-April 1987, pp. 124-129.

103. Sioulas, B and Sanjayan, J.G. Hydration temperatures in large high strength
concrete columns incorporating slag. Cement &Concrete Research, Vol.30 No.l 1,
2000, pp. 1791-1799.

197
 References

104. Hjorth, L. Microsilica in Concrete. Nordic Concree Research 1982, Publication 1,

Nordic Concrete Federation, Oslo, 1983, Paper No. 9, p. 18.

105. Sellevoids, E.J, and Radjy, F. Condensed Silica Fume in Concrete : Water demand
and strength development. Fly Ash, Silica Fume, Slag and other Mineral by­
products in Concrete. Special Publication SP-79, American Concrete Institue,
Detroit, pp.677-694, 1983.

106. De Larrard, F., Moreau, A., Buil, M.,and Paillere, A. Improvement of Mortars and
concretes really attributable to condensed silica fume. Supplementary Paper- 2°^
International Conference on Fly Ash, Silica Fume, Slag and Natural Pozzalanas in
Concrete, Madrid, 1986, pp. 1-13.

107. Domone, P.L.J and Soutsos, M.N. Properties of High Strength Concrete containing
PFA and GGBS. Concrete International, 1995.pp.30.

108. Ollivier, J.P and Carles-Gibergues. Activitépouzzalanique et action de

remplissage d ’unefumee de silice dans les matrices de béton de haute resistance.
Cement & Concrete Research Vol 18 No.3, 1988, pp.438-448.

109. Yurugi M, Mizobughi, T and Terauchi, T. Utilization of GGBS and silica fume for
controlling temperature rise in high strength concrete. Fly Ash, Silica Fume, Slag
and other Mineral By Products in Concrete. ACI Special Publication S P -132,
Detroit, 1993, pp. 1438-1450.

110. Yonezawa, T. The contribution of fluidity improving technology to the widespread

use of high strength concrete. Radical Concrete Technology, Published by E&FN
Spon, 1996, pp.526-542

111. Djellouli, J. Aitcin, P.C, and Challal. The use ofggbs in high performance concrete.
American Concrete Institute, Special Publication S P -121, Detroit, 1990, pp. 351-
358.

112. Baalbaki M, Sarkar S.L, Aitcin, P.C and Isabelle. Properties and microstructure of
high performance concretes containing silica fume, slag andfly ash. Fly Ash, Silica
Fume, Slag and other Mineral By- Products in Concrete, ACI, Special Pulication
S P -132, Detroit, 1993, pp. 921-942.

113. Bayasi, Z. Effects offly ash on the properties of silica fume concrete. Concrete
International, April 1992, pp.52-54.

198
________________________________________________________________________ References

114. ACI Committee Report 273 . Mass concrete for dams and other massive structures
Proceedings 67, 1970.

115. Killoh, D C. Advances in Cement Research, Volume 1, N o.3,1988, pp. 180-186.

116. Woods, H. The effect o f composition of Portland cement on heat evolved during
hardening. Engineering Chemistry, Vol. 24, No. 11, 1932, pp. 1207-1214.

117. Davey, N and Fox, E. Temperature rise in hydrating concrete. Building Research
paper. No. 15, HMS0.1933.

118. Meissner, H.S. ACI Journal, sep-oct 1933, pp. 21-26.

119. Forrester, J. Conduction calorimeter for the study of cement hydration. Cement
Technology, Vol. 1, No. 3, 1970, pp. 95-99.

120. Bamforth, P. An investigation into the influence of partial Portland cement

replacement using either pfa of ggbs on the early and long term behaviour of
concrete. Taylor Woodrow, Research Report, Oct. 1978.

121. Danielsson, U. Conduction calorimeter studies o f the heat of hydration

of Portland cement. Swedish C&C Research Institute, Stockholm, 1966, pp. 121.

122. Lerch, W. and Bogue, R. Heat o f hydration o f Portland cement pastes. Bureau of
Standards Journal of Research. Vol. 12. 1934, pp. 645-664.

123. Verbeck, G. and Foster, C. Long-time study o f cement performance in concrete.

ASTM Proceedings, Vol. 50, 1950, pp. 1235-1262.

124. Woods, H. The effect of composition o f Portland cement on heat evolved during
hardening. Engineering Chemistry, Vol. 24, No. 11, 1932, pp. 1207-1214.

125. Bogue, R.H. Calculation o f composition in Portland cement. Industrial Engineering

Chemical Annual, Edition 1, 1929, pp. 192-197.

126. Lerch, W. Portland Cement Association, Bulletin no. 12, 1946.

127. Keienburg, R. Particle size distribution and normal strength of Portland cement.
PhD Thesis, Germany, 1976.

128. Parrott, L. The production and Properties of high-strength concrete. Concrete,

November 1969, pp. 443-448.

199
 References

129. Aarsleff, L., Paulsen, E. and Jorgensen, J. The properties of ultra high strengh
concrete particularly heat o f hydration. Materials Research Society symposia
proceedings. Vol. 42. 1984, pp. 201-210.

130. FIP/CEB High Strength Concrete, State-of-the-art report. CEB Bulletin,

information. No. 197 August, 1990.

131. Parrott, L.J. A Review o f High Strength Concrete Properties. C&CA services,
Jan.1988.

132. Wallevik, O.H and Gjorv, O.E. Modification of the two-point workability
apparatus. Magazinre of Concrete Research, Vol.42 No. 152, pp. 135-142, 1990

133. Coole, M.J. The effect o f simulated large pour conditions on the temperature rise
and strength growth o f pfa concrete. Proceedings of 8 ^ International Conference on
Chemistry of Cement, Rio, Volume 6 , 1986, pp. 351-355.

134. Domone, P.L.J, Xu,Y, and Banfill, P.F.G. Developments of the two-point
workability test for high performance concrete. Magazinre of Concrete Research,
Vol.51 No.3, pp.171-179, 1999.

135. Newman, K. Properties of concrete. Structural Concrete. Reinforced Concrete

Association, V ol.2,N o.ll, 1965.

136. Wimpenny, D.E and Ellis, C. Oil pressure measurement in the two-point workability
apparatus. Magazine of Concrete Research, Vol.39 No.40, pp. 169-174, 1987.

137. Coole, M.J. Heat release characteristics o f concrete containing GGBS in simulated
large pours. Magazine of Concrete Research, Volume 40, No. 144, 1988.

138. Appleby Group Ltd. GGBS Data Sheets, series 2.1-2.4, Scunthorpe, North
Lincolnshire, June 1997.

139. Building Research Establishment, BRE Cardington - a study on temperature

monitoring. Data sheet No. 2.4.3, July 1997.

140. Tank, S.B. The use o f condensed silica fume in Portland cement grouts. PhD Thesis,
University of London, 1987, pp. 377.

141. Smeplass, S. and Maage, M. Heat o f hydration of high strength concretes.

Proceedings of LABSE colloquium on High Strength Concrete, Berkeley,^1990,
pp. 433-456.

20 0
 References

142. Helland, S. Temperature and Strength Development in Concrete with W/C Less
Than 0.40. Utilization of High Strength Concrete, Proceedings of the Symposium in
Stavanger, Norway, June 1987, pp. 473-486.

143. de Larrard 4* International Symposium on Utilization of High-strength/high

performance concrete, Paris, 1996, pp.415-421.

144. Lachemi, M., Lessard, M. and Aitcin, P.C. Early-Age Temperature Developments in
a High Performance Concrete Viaduct, ACI Special Publication SP-167, 1996,
pp. 149-174.

145. Schaller, I., de Larrard, P., Sudrat, J.P., Acker, P. and LeRoy, R. Experimental
monitoring of the Joigny Bridge. High Performance Concrete: From Material to
Structure, E&FN Spon, London, 1992, pp. 432-457.

146. Aitcin, P.C., Bedard, C., Plumat, M. and Haddad, G. Very high strength cementfor
very high strength concrete. Proceedings of Synq)Osium of the Materials Research
Society, Boston, USA, November, 1984, pp.210.

147. Christie, E.A., Harrison, R.S. and Ho, D.W.S. Hydration temperatues of high
strength concrete columns at Melbourne Central. Report to the Cement and
Concrete Association of Australia, highett, Victoria, 1991.

148. Yuan, R.L., Ragab, M., Hill, R.E. and Cook, J.E. Evaluation of core strength in high
strength concrete. Concrete International, Volume 13 No. 5, May 1991, pp.30-34.

149. Davey, N. Influence of temperature upon the strength development of concrete.

Building Research technical paper No. 14, HMSO,1933.

150. Wainwright, P. and Tolloczo, J. A Temperature Matched Curing system controlled

by micro computer. Magazine for Concrete Research, Volume.35 No. 124. 1983,
pp. 164-169.

151. Harrison, T. Temperature Matched Curing of concrete. Civil Engineering

magazine, August 1983, pp. 26-27.

152. Blakey, H. Temperature matching curing bath - an aid to earlierformwork striking.

Concrete, May 1976, pp.25-26.

153. British Standards Institution. Methodfor temperature matched curing of concrete

specimens. BSI, DD 92: 1984.

154. Dean, M. Temperature matched curing trial, Full Sutton Prison, North Yorkshire.
Appleby Group, Internal Report - TSS FIG 50,1985.

201
 References

155. British Standards Institution. Methods of making and curing test specimens.
BS 1881: part 3: 1970.

156. Connell, M. Report on in-situ temperature measurement and temperature matched

strength development of a wall pour, M63 Stockport. Frodingham Cement,Report
TSS 53, 1987.

157. Kleiger, P. Effect o f Mixing and Curing Temperature on Concrete Strength.

Proceedings of ACI Journal, Volume 29, No. 12, 1958.

158. Wainwright, P. Early and later age properties of temperature cycled slag-OPC
concretes. ACI SP- 114, 1986, pp.1293-1322.

159. Appleby Group Ltd. Hulme Arch, Manchester - A study on temperature monitoring
and temperature matched curing (TMC). Data Sheet, 2.4.0, October 1997.

160. Price, W. Stronger, bigger, better. Concrete, Magazine of the Concrete Society
Jan. 1996, pp. 28-29.

161. Mak, S. and Lu, A. Engineering properties of high performance concretes

containing GGBS under in-situ temperature conditions. ACI, Special Publication
SP- 149, Singapore, 1994.

162. Mak, S and Torii, K. Strength development of high strength concretes with and
without silica fume under the influence of high hydration temperatures. Cement and
Concrete Research, Volume 25, No. 8 , pp.1791-1802,1995.

163. Price, W.F. High Strength Concrete. Current practice Sheet No. 118, Concrete
magazine, June 1999

164. Wenzel, T.H., Browne, T.M., Scherzer, K.H. and Hassaballah, A. The effects of
temperature on the performance o f high strength concrete with fly ash. Radical
ConcreteTechnology, E&FN Spon, 1996, pp.401-411.

165. Verbeck, G.J. and Helmuth, R.H. Structures and physical properties o f cement
paste Proceedings, Chemistry of Cement, 5*^ International Symposium, Tokyo,
pp. 1-32 1968.

166. Kjellsen, K.O., Detwiler, R.J and Gjorv, O.E. Development of microstructures in
plain cement pastes hydrated at different temperatures. Cement and Concrete
Research, Volume. 21 No. 1, 1991, pp. 179-189.

202
 References

167. Ikabata, T., Yoshida, Matsuoka, Y, and Uchida, K. Effect of curing temperature on
the microstructure o f cement pastes. Proceedings of Cement and Concrete, Concrete
Association, Japan, No. 44, 1990, pp. 46-51.

168. Laplante, P. and Aitcin, P.C. Compressive strength of concrete: from cylinders to
structures. 4^ International Symposium on Utilization of High-strength/high
performance concrete, Paris, 1996, pp.645-653.

169. Price, W.F and Hynes, J.P. In-situ strength testing of high strength concrete.
Magazine of Concrete Research, Volume 48, No. 176, 1996, pp. 189-197.

170. Thuraintanam, H. Heat flow analysis. Internal Report, Department of Civil

Engineering, University College London, 1987

171. Coole, M.J and Harrison, A.M. The effect of simulated large pour curing conditions
on the temperature rise and strength growth of PFA containing concrete. Materials
research Society Synposia Proceedings, Vol.8 6 , 1987. pp.277-289.

172. Parrott, L.J. Modelling the development o f microstructure, research on the

manufacture and uses o f cement. Proc. Eng. Foundation Conference, Henniker, 1985

173. Dalziel, J.A and Gutteridge, W.A. The influence o f PFA upon the hydration
characteristics and certain physical properties of a Portland cement paste. Technical
Report No.560, C&CA, 1986.

174. Halse, Y.H. The effect o f the addition offly ash on the hydration of cement, PhD
thesis. Imperial College, London, 1985.

175. Breugel K.van. Simulation o f hydration andformation of structure in

hardening cement-based materials, PhD thesis. University of Delft, Delft, 1991.

176. British Standards Institution BS 1881: Part 102. Testing concrete: Methodfor
determination o f slump. British Standards Institution, 1983 London.

177. British Standards Institution BS 1881: Part 103. Testing concrete: Methodfor
determination of compactingfactor. British Standards Institution, 1993 London.

178. British Standards Institution BS 1881: Part 104. Testing concrete: Methodfor
determination ofVebe time. British Standards Institution, 1983 London.

179. British Standards Institution BS 1881: Part 105. Testing concrete: Methodfor
determination of flow. British Standards Institution, 1984 London.

203
 References

180. Bye, G.C. Portland Cement, Composition, Production and Properties. 2“^ edition,
Thomas Telford Publishers, 1999.

181. De Almeida, I.R. Non destructive testing o f high strength concretes: Rebound
(Schmidt Hammer) and ultrasonic pulse velocity. Proceedings of 2°^ International
Symposium on Quality control of concrete structures, Ghent, June 1991, pp.387-397.

182. Duval, R and Kadari, E.H. Influence o f silica fume on the workability and
compressive strength of HPC. Cement and Concrete Research, Vol28, No.4, pp.
533-547, 1998.

183. Miura, N, Takeda, N., Chikamatsu, R., and Sogo, S. Application of super workable
concrete to reinforced concrete structures with difficult construction conditions.
ACI SP-140-8,pp. 163-186, 1984.

184. Crow, R.D and Dunstan, E.R. Properties o f fly ash concrete. Proceedings
symposium on fly ash incorporation in hydrated cement systems. Edited by
S.Diamond, Materials Research Society, 1981, pp.214-225.

185. CIRIA Report 136. Formwork striking times, criteria, prediction and methods of
assessment. CIRIA, London, 1995.

186. Kokubu, S., Takahashi, H and Anzai, H. Effect o f curing temperature on the
hydration and adiabatic temperature characteristics of Portland cement-blast
furnace slag concrete. American Concrete Institute SP-114, Detroit, 1989,
pp.l361- 1375

187. Alshamsi, A.M. Microsilica and GGBS effects on hydration temperature. Cement
and Concrete Research, Vol.27 No. 12, pp. 1851-1859, 1997.

188. Diamond, S. The microstructure of cement paste in concrete. Proceedings of 8 ^

International Congress on Chemistry of Cement, Vol. 1, pp. 122-147, Rio de
Janeiro, 1986.

189. Regourd, M., Mortureaux, B., Aitcin, P.C. and Pinsonneault, P. Microstructure of
field concretes containing silica fume. Proceedings of 4^ International Symposium
on Cement Microscopy, Nevada, USA, pp.249-260, 1983.

190. Aitcin, P.C, and Sarkar, S.L. Comparative study o f the microstructure of very high
strength concrete. Cement, Concrete and Aggregates, Vol. 9 No.2, pp. 57-64,1987.

191. Scrivener, K.L., Bentur, A and Pratt, P.L. Quantatative characterization of the
transition zone in high strength concretes. Advances in Cement Research, Vol.l,
No.2, pp.230-237, 1988.

204
 References

192. Mak, S L and Lu, A. Temperature effects on the early age properties of high
strength binder systems containing blast furnace slag. 2nd International Symposium
on Blended Cements, pp. 71-76, Malaysia, 1994.

193. British Standards Institution BS 12. Testing concrete: Specification for Portland
cement. British Standards Institution, 1996, London.

194. British Standards Institution BS 1881: Part 106. Methodfor determination of

compressive strength of concrete cubes. British Standards Institution, 1983, London.

195. British Standards Institution BS 3892: Part 1. Specification of PFA for use with
Portland cement. British Standards Institution, 1997, London.

196. British Standards Institution BS 6699. Specification for GGBSfor use with
Portland cement. British Standards Institution, 1992, London.

197. British Standards Institution BS 5075: part 3. Specification for superplasticising

admixtures. British Standards Institution, 1985, London.

198. American Society for Testing Materials. Standard specifications for chemical
admixtures for concrete, ASTM book of standards, 1999.

199. Taylor, H.F.W. Private Communication.

200. British Standards Institution BS 812. Testing Aggregates. British Standards

Institution, 1990, London.

201. Carlson, R.E. A simple method o f computation of temperatures in concrete

structures. ACI Journal, Vol.34, pp. 89-90, 1937.

202. Schmidt, E. Foppls Festchrift, pp. 179-180, Springer, Berlin, 1924.

203. Ross, A.D and Bray, J.W. The prediction of temperatures in mass concrete by
numerical computation. Magazine of Concrete Research, pp. 9-20, January, 1949.

204. Saucier, K., Tynes, W. and Smith, E. High compressive strength concrete- report 3,
summary report. US Army Engineer Waterways Experiment Station, 1965.

205. Zoldners, N. Thermal properties of concrete under sustained elevated temperatures.

ACI- Special Publication 25, Temperature and Concrete, 1971.

206. Campbell- Allen, D. The thermal conductivity of concrete. Magazine of Concrete

Research, Vol. 15, No.43, 1963.

2 05
 References

207. Browne, R. Properties of concrete in reactor vessels. Institution of Civil Engineers

Conference on P.C.P.V, paper 13, 1962.

208. Carmen, A. The thermal conductivity and diffusivity of concrete. University of

Illinois, Bulletin No. 122, 1981.

209. Birch, F. The thermal conductivity and diffusivity of rocks. Handbook of physical
constants, London, 1949.

210. Weaver, J. Temperature development in hydrating concrete. PhD thesis. University

of London, 1972.

211. Sun, G. and Young, J.P. Cement and Concrete Research, Vol 23, pp.480, 1993.

212. Traetteberg, A. Cemento, Vol.75. pp.369, 1978.

213. Chattel]i. S., Collepardi, M and Moriconi, 0 . Proceedings of International

Conference, Use of fly ash, silica fume, slag and other by-products, SP-79, pp.221,
1983.

214. Bentaur,A, and Goldman, A. Curing effects, strength and physical properties o f high
strength silica fume concrete. ASCE Journal of Materials in Civil Engineering,
Vol.l.No.46, 1989.

215. Xie, J., Elwi, and McGregor, J.G. Mechanical properties o f three high strength
concretes containing silica fume. ACI Journal of Materials, Vol. 92, pp. 135, 1995.

216. Laplante, P. and Aitcin, P.C. Compressive strength of concrete: from cylinders to
structures. 4^ International Symposium on Utilization of High strength/high
performace concrete, pp.645-653, 1996.

217. Day, R. and Marsh, B.K. Measurement o f porosity in blended cement paste. Cement
and Concrete Research, Vol. 18, pp.63-73, 1988.

218. Malhotra, V.M. In situ/non~destructive testing of concrete- A global review. In

situ/non-destructive testing of concrete, ACl-SP-82, Detroit, USA, pp. 1-16, 1984.

219. Hilhorst, M.A. Dielectric characterisation o f soil. DLO- Institute of Agricultural

and Environmental Engineering, thesis publication, Netherlands, 1998.

220. Hasted, J.B. Aqueous dielectrics. Published by Chapman and Hall, London, 1973.

206
 References

221. Grant, E.H. Dielectric behaviour o f biological molecules in solution. Published by

Oxford University Press, London, 1978.

222. Van Beek.A, Dielectric properties of young concrete. PhD thesis

223. Rhim, C., Buyukozturk, O. Electromagnetic properties of concrete at microwave

frequency range. ACI materials Journal, Vol.9, No.3, pp.262-271, 1998.

224. Haddad, R.H, Al-Qadi, I.E. Characterization of Portland cement concrete using
electromagnetic waves over the microwave frequencies. Cement and Concrete
Research, Vol. 28,No.lO. pp. 1379-1391, 1998.

225. Al-Quadi I.L, Hazim, 0.A, Su, W, Riad S.M. Dielectric properties of Portland
cement concrete at low frequencies. Journal of materials in civil engineering, Vol.7
No. 3, pp. 192-198, 1995.

226. K. van Breugel, and Hilhorst, M.A. In-situ measurement o f the dielectric properties
of hardening concrete as a basis o f strength development. Non destructive
Evaluation of Civil Structures and Materials, Colorado, USA, pp. 7-21, 1986.

227. Beek, A. van, Breugel K van, Hilhorst, M.A. Monitoring system for hardening
concrete based on dielectric properties. Utilizing ready-mixed concrete and mortar.
Creating with concrete, Dundee, pp.303-312, 1999.

228. Chai, H.W. Design and Testing of Self Compacting Concrete. PhD thesis, UCL,
University of London, March 1998.

229. Penttalla, V.E. Effects of delayed dosage of superplasticiser on high performance

concrete. High Strength Concrete Proceedings, Lillehammer, Norway, June 1993,
pp.874-881.

230. Odler, I. And Rossler, M. Investigations on the relationship between porosity,

structure and strength of hydrated Portland cement pastes. Cement and Concrete
Research Vol 15 No.2, 1985, pp.320-325.

231. Mindess, S. Relation between the compressive strength and porosity o f autoclaved
calcium silicate hydrates. Journal of American Ceramics Society, Vol.53 No.l 1,
1970, pp.621.

232. Aitcin P C. and Riad N, Curing temperature and very high strength concrete.
Concrete International, Vol. 10, No. 10, 1988, pp.69-72.

207
________________________________________________________________________ References

233.Carrasquillo, P.M and Carrasquillo, R.L. Evaluation of the use of current concrete
practice in the production of high strength concrete. ACI Materials Journal, Vol.85,
No.l, 1988, pp.49-54.

234. Sarkar, S.L. and Aitcin, P.C. Dissolution rate of silica fume in very high strength
concrete. Cement and Concrete Research, Vol.l7, 1987, pp.591-601.

235. Parrott, L.J. Use of milled slag in high strength concrete. Unpublished data. C&CA
services, 1970, p.3.

236. Mphonde, A.G and Frantz, G.C. Shear tests of high and low strength concrete
beams without stirrups. ACI Journal, July-August 1984, pp.350-357.

237. Kostuch J.A., Walters G.V. and Jones, T.R. High performance concrete
incorporating metakaolin- a review. Concrete 2000, 1993, Vol2. pp,1799-1811.

238. Wild S, Khatib ,J and Jones, A. Relative strength, pozzolanic activity and cement
hydration in superplasticised metakaolin concrete. Cement and Concrete
Research1996, Vol.26, No. 10 pp. 1537-1544

239. ASTM C 143-98 Standard test methodfor slump ofhydraulic-cement concrete.

Annual book of ASTM standards, 2000.

240. European Committee for Standardisation, Specification for Fly Ash, ENV 450.

241. British Standards Institution BS 1881: Part 115. Specification for compressive
testing machines for concrete. British Standards Institution, 1986, London.

208
 APPENDIX A

1. Two point test calibration

2. Workability results
Table A l: Torque/Pressure calibration data

Idling .Total ^ Spring lever toKfüe

.Pressure Balance. :
im (W (Nm)
0.5 130 136 6 0.55 0.2 0.11
216 86 9.35 0.2 1.87
329 199 20.75 0.2 4.15
360 230 27.80 0.2 5.56
375 245 26.34 0.2 5.27
431 301 32.04 0.2 6.41
459 329 36.01 0.2 7.20
437 307 33.43 0.2 6.69
137.5 155 17 1.60 0.2 0.32
165 27 2.26 0.2 0.45
233 95 10.00 0.2 2.00
245 107 11.72 0.2 2.34
319 181 18.71 0.2 3.74
338 200 18.71 0.2 3.74
345 207 21.61 0.2 4.32
495 357 38.38 0.2 7.68
507 369 39.67 0.2 7.93
143 154 11 1.18 0.2 0.24
152 9 0.97 0.2 0.19
198 55 8.06 0.2 1.61
192 49 5.38 0.2 1.08
284 141 16.23 0.2 3.25
353 210 22.58 0.2 4.52
470 327 35.15 0.2 7.03
150 155 5 0.54 0.2 0.11
227 77 9.03 0.2 1.81
333 183 19.89 0.2 3.98
307 157 16.56 0.2 3.31
520 370 39.78 0.2 7.96
453 303 33.33 0.2 6.67
162 285 123 13.44 0.2 2.69
279 117 10.86 0.2 2.17
432 270 27.50 0.2 5.50
473 311 34.72 0.2 6.94
487 325 34.94 0.2 6.99
517 355 38.16 0.2 7.63
493 331 37.09 0.2 7.42
175 177 2 0.30 0.2 0.06
182 7 0.75 0.2 0.15
216 41 3.98 0.2 0.80
452 277 29.35 0.2 5.87
527 352 37.63 0.2 7.53
542 367 38.81 0.2 7.76
 T = 0.0215 P

r = 0.9975

0 4
►
E
£ 3
......... ^
r

100 200 300 400

Net pressure, P ( psi )

Figure A1 : Torque-Pressure calibration of two-point test apparatus

 Table A2: Pressure/Voltage data (obtained from idling and mortar tests)

II1ÎÜ B
200
w 2.05 388 3.11
192 1.99 370 3.02
185 1.96 358 2.94
178 1.92 345 2.86
172 1.88 330 2.78
165 1.84 315 2.71
158 1.80 298 2.59
150 1.75 275 2.50
143 1.71 262 2.40
138 1.68 245 2.30
130 1.64 225 2.20
123 1.60 - -

450

400 - P = 176.6 V - 1 5 5 . 5
= 0 .9 9 9 6

350 -

300 -

CL 2 5 0 -

200 -

CL 150 -

100 -

50 -

0 0 .5 1 1.5 2 2 .5 3 3 .5

Voltage, V (volts)

Figure A2: Pressure/Voltage Calibration of two-point test

apparatus
Table A3: Shaft speed / speed setting data for determination of gear ratio.
Speed Shaft speed Flywheel Gear
Setting Tachom eter /60 Speed ratio
rps (rps)
0.5 8 0.13 0.32 0.64 4.80
1.0 18 0.30 0.74 1.49 4.95
1.5 28 0.47 1.17 2.34 5.01
2.0 39 0.65 1.60 3.20 4.92
2.5 52 0.87 2.03 4.07 4.69
3.0 64 1.07 2.49 4.98 4.67
3.5 78 1.30 2.92 5.84 4.49
4.0 89 1.48 3.35 6.71 4.52
4.5 103 1.72 3.79 7.58 4.41
4.0 91 1.52 3.38 6.75 4.45
4.5 103 1.72 3.81 7.62 4.44
4.0 87 1.45 3.38 6.75 4.66
3.5 80 1.33 2.94 5.88 4.41
3.0 64 1.07 2.51 5.01 4.70
2.5 52 0.87 2.04 4.07 4.70
2.0 40 0.67 1.61 3.22 4.83
1.5 29 0.48 1.17 2.34 4.85
1.0 19 0.32 0.75 1.49 4.71
0.5 8 0.13 0.31 0.63 4.70
Average 4.6815
* Hand-held tacheometer used to measure shaft speed
+ Digital volmeter ( DVM )
1 volt of the DVM is equivalent to 2 rps of the flywheel

2.0

9-

! CO

%
.c
CO

0.0
0.0 1.0 2.0 3.0 4.0 5.0
Speed setting
Figure A3: Calibration of shaft speed / speed setting
 Binary Mixes at 0.26 w /b
Table 1 : Slump and two-point test results for binary m ixes of CSF

m m ■ s s r ,
100% OPC* 200 9.5 10 0.9714
[1.8%]

5% CSF 180 6.3 7.91 0.9822

[1.6%]

10% CSF 180 4.37 6.11 0.9901

[1.5%]

15% CSF 195 3.95 6.21 0.9893

[1.5%]

Table 2 : Slump and two-point test results for binary mixes of PFA

^lüm
^ ^pr S
Vî^ld value
( mmf ' a sT c
100% OPC* 200 9.5 10 0.9714
[1.8%]

20% PFA 180 5.17 12.1 0.9792

[1.5%]

40% PFA 190 4.91 13.21 0.9823

[1.4%]

T able 3 : Slump and two-point test results for binary mixes of GGBS
" TW<>rpointtôetJeeults
YieWvalue Correlation
m coefRaentV
100% OPC* 200 9.5 10 0.9714
[1.8%]
30% GGBS 200 12.5 12.22 0.9417
[2.0%] 185* 10.17 13.79 0.9389
180* 14.23 11.4 0.9288
60% GGBS 190 15.1 13.77 0.9377
[2.5%] 205* 14.2 15.84 0.9503
180* 15.81 12.9 0.9401
All values in brackets [ ] respresent superplasticizer dosages used, * repesent repeat mixes
 Ternary Mixes at 0.26 w/b ratio
Table 4 : Slump and two-point test results for Ternary m ixes of PFA + CSF

; : Two-pointiest results v r^ V

10% CSF 180 4.37 6.11 0.9901

[1.5%]

20% PFA + 10%CSF 180 3.5 5.5 0.9923

[1.3%]

36%PFA + 10%CSF 205 3.1 5.19 0.9941

[1.1%]

40% PFA + 10%CSF 210 2.82 4.83 0.9987

[1.0%]

50%PFA + 10%CSF 220 3.04 4.71 0.9975

[1.0%]

Table 5 : Slump and two-point test results for Ternary mixes of GGBS + CSF

i l ü a , . f a r s s r ,
10% CSF 180 4.37 6.11 0.9901
[1.5%]

30%GGBS + 10%CSF 200 3.98 5.73 0.9927

[1.4%]

54%GGBS + 10%CSF 215 3.42 5.41 0.9953

[1.1%]

60%GGBS + 10%CSF 220 3.21 5.3 0.9911

[1.0%]

70%GGBS + 10%CSF 220 4.67 6.88 0.9844

[1.1%]

80%GGBS + 10%CSF 210 5.22 7.51 0.9907

[1.3%]
Binary and Ternary Mixes at 0.30 and 0.20 w /b ratio
Table 6 : Slump and two-point te st results for m ixes at 0.30 w/b ratio

Slump H wihttestm sult s4 ^

^Correlation ;
coefficient,<r

100% OPC 195 6.04 7.1 0.9879

[1.4%]

10% CSF 190 2.51 3.17 0.9945

[1.1%]

40% P F A + 10%CSF 220 1.72 3.11 0.9976

[0.8%]

60%GGBS + 10%CSF 210 2.18 4.04 0.9972

[0.8%]

Table 7 : Slump and two-point test results for mixes at 0.20 w/b ratio

m '^ m

10% CSF 180 7.22 7.85 0.9802

[2.7%]

40% PFA + 10%CSF 220 4.25 7.93 0.9877

[1.5%]

60%GGBS + 10%CSF 200 5.69 8.27 0.9831

[1.5%]

All values in brackets [ ] respresent superplasticizer dosages used, * repesent repeat mixes
 APPENDIX B

1. Prediction of Temperature cycles

2. Compressive strength results

B.l Prediction of temperature cycles by numerical computation

B 1.1 Description of the one-dimensional rectangular model

Step by step mathematical models for predicting temperature distributions have been
devised by Carlson^^®'^ and Schmidf^°^\ and first applied to concrete by Ross and
Bray(203)^
Schmidt’s method, which is based on the calculations of temperature differences, was
devised for solving problems where there is heat flow in only one direction and forms
the basis for the method used in this work.
The analysis involves considering two stages.

1) Cooling only

2) Cooling with heat generation

Knowledge of the heat of hydration characteristics, thermal properties and the

boundary conditions of the concrete is required.

1) Cooling only

Consider a prism of material having a uniform cross sectional area. A, divided up into
a number of equal elements of length, x, with its axis parallel to the direction of flow.
Figure B.l shows three adjacent elements A, B and C, with temperatures at their
centres of Ta, Tb and Tc respectively.

Figure B.l: One dimensional heat flow model

The heat flow from A to B in time At, is

Hi =Jç^ ( T a -T b ) At (1)
AX

The heat flow from B to C is

H2 = kA_(TB-Tc) At (2)
AX

Therefore the heat gained by element B is :

AH = H i-H 2 = ±A (Ta+Tc - 2 Tb) A t (3)

AX

where k = Thermal conductivity

A = Cross sectional area
T = Temperatures at the centre of each element
At = Interval of time during which the heat flow is assumed
constant

The temperature rise resulting from the heat gain, AT, is given by

AH= p V cA T (4)

Where p = density of the material

c = specific heat
V = Volume of element
Substituting for AH, from equation 3.

AT= k A At (T a + Tc -2T b ) (5)

p c V Ax

This can be simplified to:

AT = D At (Ta + Tc-2Tb) (6 )
Ax'

where D = Thermal diffusivity = k

pc

2) Cooling with heat generation

With simultaneous heat generation, it is necessary to know the adiabatic temperature

rise that would be produced in the concrete by the heat generated during each time
interval. This is then simply added to the final temperature of each individual element
at the end of that time period. For example, if the adiabatic temperature rise in At is
Tad and the temperature at the start of the time increment At is Tb, then the
temperature at the end of the increment, Tb ’ is given by :

T ’b = Tb + DA 1 (Ta + T c -2 T b ) + Tad (7)

A x:
B.1.2 Boundary condition analysis for heat loss through form work and the
surronding air

At the boundary (formwork/air) the heat flow may be due to conduction or

convection. This will depend on what conditions are experienced at the boundary and
the parameter that is used to represent the heat flow at the boundary is surface
conductance, h.
If we assume that in the previous example (figure B.l) element C is in contact with
the surface element, S, which in turn is in contact with asurrounding airelement, F. A
necessary assumption is that the surface element is sosmall thatwe can equate heat
coming in and flowing out of it.

Heat transfer from B to C = kA ( Tb - Tc ) At (8 )

AX

Heat transfer from C to S = kA f Tr - Tq 1 At (9)

A X/2
Therefore, heat gained by element C is :

= kA (Tb+ 2Ts - 3 Tc) a t (10)

Ax
And as before,
T ’c = Tc + D A l (TB + 2 T s-3T c) + Tad (11)
Ax'

So if we consider the following:

Heat transfer from C to S = kA ( Tc - Tg ) At (12)

A X/2

Heat transfer from S to F = h A (Tg - Tp) At (13)

where h = surface conductance.
Due to the small thickness of S, we can equate heat flowing into and out of S :
 2kA_(Tc-Ts) = h A (Ts-Tf) (14)
AX

Therefore
T s = (B T c + If ) (15)
1+B

where B = 2k
hAx

Therefore after time increment At,

T ’s . ( B T ’r + T f ) (16)
1+B

B.1.3 Radial heat flow model

The one-dimensional rectangular heat flow model can be modified to a two-
dimensional radial heat flow model, which is more appropriate for the current study.
The same principles apply, the only difference being, the quotient A/V is not constant
throughout. Instead it is a function of the radius from the centre to the element in
question (figure B.2) .In addition the two areas across which heat flows in and out
respectively are functions of radii.

Figure B.2: Radial heat flow model

Heat flow from 0 to 1 = Jç_. V2 Ax .0 (To-Ti) A t (17)
Ax

Heat flow from 1 to 2 = ^ . 1.5 Ax .0 (T1-T2 ) A t (18)

Ax
Therefore AH = 1 . . ‘A Ax .0 . { To + 3Tz - 4Ti } . At (19)
Ax
But,
AH = p. V. 0 . AT
= p.c. A T .{% (1 .5 A x )^0 - ^2 ( 1/2 Ax 0}
= p . c . AT . Ax^. 0

AT = k. At . To + 3 T2 - 4Ti } (20)
p.c. Ax^

This expression varies with the position of the element.

B.1.4 Selection of the required therm al properties of the concrete

The analysis is clearly dependent on the choice of appropriate thermal properties.

These are more complex than for most other materials, because not only is concrete a
composite material the components of which have different thermal properties, but its
properties also depend on moisture content and porosity. Data on thermal properties
of high strength concrete are limited but it has been shown that these are
approximately within the same range as those of normal strength concrete^^^^’^®'*\

The properties required are: thermal conductivity, thermal diffusivity and specific
heat.
• Thermal conductivity (k)

The thermal conductivity of concrete depends on:

(i) the thermal conductivity of the aggregate and paste phase.
(ii) the moisture content.

Hardened saturated cement paste made with OPC has a k value at 20° C between 1.1
and 1 .6 W/m°C.(^°^)
The thermal conductivity for aggregate varies widely, between less than 1.0 W/m °C
to greater than 4.5 W/m°C'^^®^^ However it is largely dependent on rock type; in my
research granite is used, which has a k value between 2.2 and 2.6 W/m°C .
An average value of 2.4 W/m° C was used in the analysis.

• Specific heat ( c )

Specific heat represents the heat capacity of the concrete and is governed mainly by
its moisture content A typical value for saturated concrete at 20 °C is 1.0 kJ/kg °C,
however this value can vary between 0.7 kJ/kg °C and 1.5 kJ/kg °C It has been
shown that changes in aggregate content or type have a negligible effect on the
specific heat of concrete The typical value of 1.0 kJ/kg °C was used in the
analysis.

• Thermal Diffusivity (D)

The thermal diffusivity, which gives a measure of the rate of heat flow under transient
conditions, can be measured directly or calculated from other properties of the
concrete using the expression :

D = k / cp
 D = Diffusivity
k = thermal conductivity
p = density
c = specific heat

The thermal diffusivity of concrete will clearly depend on the same factors that
influence the other properties. Typical values of D for concrete range from 0.7 to
1.89x10"

The calculated value of 1.0 xlO'^mVs was used for the analysis.

B.1.5 Selection of the required constants for the surface characteristics

When forced convection of air takes place over a plane surface the surface
conductance h can be obtained from:
h = 5.68 (1+0.730) ) W W °C (17)
Where x>is the wind velocity in m/s.
For a wind speed of 1.1m/s the surface conductance is therefore 10 W/m^ °C.

Davey and Fox^^'^ measured the surface conductance for various concrete-faces,
obtaining values ranging from 8 to 10 W/m^ °C . Generally, at the formwork/concrete
boundary the surface conductance is less than the value obtained for a concrete
surface, since the formwork acts as an insulating medium. Weaver^^'®^ estimated the
surface conductance of a 19mm plywood formwork, to be 5.5 W/m^ °C . However, in
practice the effective thickness of plywood formwork can be assumed to be greater
than 19mm due to the use of plywood stiffeners. This generates a lower surface
conductance value of 3.5 W/m^ °C.
CIRIA report 136^'®^^ states surface conductance values for various concrete/formwork
boundaries. For a 37 mm thick plywood sheet under normal and severe exposure
conditions equivalent to wind speeds of 3 and 9 m/s, the surface conductance values
were 3.1 and 3.6 W/m^ °C respectively.
For the analysis, a value of 3.5 W/m^ °C was used for the formwork in place and a
value of 10 W/m^ °C when the formwork was stripped at 24 hours.

B.1.6 Description of the spreadsheet

Based on Ross and Brays^^®^^ method of numerical computation, Thurairatnam^^^®^

developed a spreadheet program, at UCL, as a tool to predict the temperature
distribution in grouts. In this study, the adiabatic temperature rise, thermal properties
for concrete and the surface characteristics for formwork were entered into the
program.

The spreadsheet calculates the temperature at various distances from the surface of a
typical Im diameter column to the centre (elements) due to the adiabatic temperature
rise and cooling when the formwork is removed. Each column in the spreadsheet
represents an element. The temperature distribution is symmetrical about the centre
line (0.5m diameter) and hence it is only necessary to calculate temperatures for one
half of the column.
The temperature at the centre of a Im diameter column was taken as the temperature
cycle to be used.

Tables B.l and B.2 show typical outputs from the spreadsheet program. Figure B.3
shows a typical temperature cycle obtained for an OPC mix at 0.26 w/b ratio from the
centre of a Im diameter column with the formwork being removed at 24 hours.
 Table B.l: Typical output from spreadsheet (heating)

TEMPERATURE DEVELOPMENT DURING P(RADIAL FLOW)

with heat transfer from surface (prog HEATS)
Typical Adiabatic heating curve OPC 0.26

D= 1 .OE-06 sq m/sec
h= 3.5 W sq m/d eg C
k= 2.4 W/m deg 0 Specific heat = 100 J
X = 0.111 metres
calc constant 2k/hx = 12.3552124

max time increment for stability 0.5 x"2/D = 1.71 hours

Form
time adiab Distance from surface (metres) work
temp 0.6105 0.4995 0.3885 0.2775 0.1665 0.0555 surface 0.019
hours deg 0 centre
0 0 0 0 0 0 0 0 0 0
1 4.5 4.5 4.5 4.5 4.5 4.5 4.5 4.5 4.5
2 7.5 7.5 7.5 7.5 7.5 7.5 7.5 7.5 7.2
3 8.8 8.8 8.8 8.8 8.8 8.8 8.8 8.7 8.1
4 10.7 10.7 10.7 10.7 10.7 10.7 10.7 10.6 9,6
5 12.3 12.3 12.3 12.3 12.3 12.3 12.2 12.1 10.8
6 14.5 14.5 14.5 14.5 14.5 14.5 14.4 14.2 12.6
7 17.1 17.1 17.1 17.1 17.1 17.0 16.9 16.7 14.7
8 19.6 19.6 19.6 19.6 19.6 19.5 19.3 19.1 16.7
9 24.0 24.0 24.0 24.0 23.9 23.9 23.6 23.4 20.5
10 34.5 34.5 34.5 34.5 34.4 34.3 34.0 33.8 30.2
11 38.8 38.8 38.8 38.8 38.7 38.5 38.2 37.9 33.3
12 41.4 41.3 41.3 41.3 41.3 41.1 40.7 40.2 34.7
13 43.9 43.8 43.8 43.8 43.7 43.5 43.0 42.5 36.0
14 46.3 46.2 46.2 46.2 46.0 45.8 45.2 44.6 37.4
15 46.7 46.5 46.6 46.5 46.4 46.0 45.3 44.7 36.8
16 48.2 48.0 48.0 48.0 47.8 47.4 46.6 45.9 37.5
17 48.6 48.3 48.3 48.3 48.1 47.6 46.7 46.0 37.1
18 49 48.6 48.7 48.6 48.3 47.8 46.9 46.1 36.9
19 49.2 48.7 48.8 48.7 48.4 47.8 46.8 46.1 36.5
20 49 48.4 48.5 48.4 48.1 47.4 46.4 45.6 35.8
21 49 48.3 48.4 48.3 47.9 47.2 46.1 45.3 35.4
22 49 48.1 48.2 48.1 47.7 47.0 45.9 45.1 35.0
23 49 48.0 48.1 48.0 47.6 46.8 45.6 44.8 34.7
24 49 47.8 47.9 47.8 47.4 46.6 45.4 44.6 34.5
 Table B.2: Typical output from spreadsheet (cooling)

TEMPERATURE DEVELOPMENT DURING HYD(RADIAL FLOW)

with heat transfer from surface (prog HEATS)
Typical Adiabatic coolin g curve OPC 0.26

D= 0.000001 sq m /sec
h= 10 W sq m/d eg C
k= 2.4 W/m deg C Specific heat = 100 J
x= 0.111 metres
calc constant 2k/hx = 4.324324

max time increment for stability 0.5 x"2/D = 1.71125 hours

time adiab Distance from surface (metres) Air

tem p 0.6105 0.4995 0.3885 0.2775 0.1665 0.0555 surface 0
hours deg 0 centre
0 0 0 0 0 0 0 0 0 0
24 49 47.8 47.9 47.8 47.4 46.6 45.4 44.6 0
25 49 47.6 47.8 47.6 47.2 46.4 45.2 36.7 0
26 49 47.5 47.6 47.5 47.0 46.2 39.9 32.4 0
27 49 47.3 47.4 47.3 46.8 44.2 36.6 29.7 0
28 49 47.1 47.2 47.1 46.0 42.2 34.0 27.6 0
29 49 46.6 47.1 46.6 44.8 40.3 31.9 25.9 0
30 49 45.9 46.5 45.9 43.6 38.6 30.1 24.5 0
31 49 45.0 , 45.8 45.0 42.3 36.9 28.6 23.2 0
32 49 43.9 i 44.8 43.9 40.9 35.4 27.2 22.1 0
33 49 42.7 ! 43.7 42.7 39.5 33.9 25.9 21.0 0
34 49 41.5 42.6 41.5 38.2 32.6 24.8 20.1 0
35 49 40.2 41.3 40.2 36.9 31.3 23.7 19.2 0
36 49 38.9 40.0 38.9 35.5 30.0 22.7 18.4 0
37 49 37.6 38.7 37.6 34.3 28.9 21.8 17.7 0
38 49 36.3 37.4 36.3 33.0 27.8 20.9 17.0 0

94 49 4.35 4.50 4.35 3.93 3.28 2.45 1.99 0

95 49 4.19 4.33 4.19 3.79 3.16 2.36 1.92 0
96 49 4.03 4.17 4.03 3.65 3.04 2.27 1.85 0
97 49 3.88 4.01 3.88 3.51 2.93 2.19 1.78 0
98 49 3.74 3.86 3.74 3.38 2.82 2.11 1.71 0
99 49 3.60 3.72 3.60 3.25 2.71 2.03 1.65 0
100 49 3.46 3.58 3.46 3.13 2.61 1.95 1.59 0
 60

50

u
0)
m
3 30
2

g
E 20
d)
H*
R em oval of formwork

10

0
0 20 40 60 80 1 0 0

Time ( hours )

Figure B.3 Typical temperature cycle for an OPC mix at 0.26 w/b ratio.
Calculated Temperature Cycles
 60

50 - OPC
- - 10%CSF
40 -

Q- 2 0 -

10 -

0 20 40 60 80 100 120
Time ( hours )

Figure Bl: Temperature cycles for OPC and 10%CSF mixes at

0.26 w/b ratio

60
OPC
50 - - - 20% PFA
---•40% PFA
40 -

^ 20 -

10 -

0 20 40 60 80 100 120

Time ( hours )

Figure B2: Temperature cycles for binary PFA mixes at 0.26 w/b
ratio
 60

OPC
50
30%GGBS
o 60%GGBS
40
0 )
2
3 30
2
&
E 20
CD

10

0
0 20 40 60 80 100 120 140
Time ( hours )

Figure B3: Temperature cycles for binary GGBS mixes at

0.26 w/b ratio

60
10%CSF
50
O
o
- - 40%PFA+10%CSF
0)
40
CO 50%PFA+10%CSF
"C
30
a
CL 20 S '
E
) 0
\-
10

0 20 40 60 80 100 120

Time ( hours )

Figure B4: Temperature cycles for ternary PFA/CSF mixes at

0.26 w/b ratio
 10%CSF
50 -
60%GGBS+10%CSF

40 - •70%GGBS+10%CSF
•c /— - - 80%GGBS+10%CSF
V

Q- 2 0 -

10 -

0 20 40 60 80 100 120
Time ( hours )

Figure B5: Temperature cycles for ternary GGBS/CSF mixes at

0.26 w/b ratio

10%CSF
50 -
- - 40%PFA+10%CSF

40 - 60%GGBS+10%CSF

S
i 30 - .N

10 -

0 20 40 60 80 100 120
Time { hours )

Figure B6: Temperature cycles for binary 10%CSF and Ternary

PFA/CSF and GGBS/CSF mixes at 0.30 w/b ratio
 50 - 10%CSF

40%PFA+10%CSF
40 -
60%GGBS+10%CSF
30 -

g- 2 0 -

10 -

0 20 40 60 80 100 120 140

Time ( hours )

Figure B7: Temperature cycles for binary 10%CSF and

Ternary PFA/CSF and GGBS/CSF mixes at 0.20 w/b ratio
Table 1 : C om pressive strength and density data for 100% O PC mix (at 0.26 w/b)
Age Stan dard C lired
Compressive strengths ( N/mm^ ) Density* Compressive strengths ( N/mm^ ) Density*
( d ays) Individual cube values Mean ( kg/m^ ) Individual cube values Mean ( kg/m^ )
1 50.7 57 57.3 55 2501 72 76 77 75 2500
3 65 58 63 62 2504 80.1 84.9 81 82 2505
7 81 86.5 87.5 85 2505 89.2 88.7 80.1 86 2510
28 105.5 107 99.5 104 2510 93.1 90.1 86.8 90 2511
90 109.2 118.3 117.5 115 2510 95.7 92 103.3 97 2505
180 119.1 128.6 115.3 121 2511 109 114 101 108 2508
* represent average density, determined from idividual cues or gross value.

Table 2 : Compressive strength and density data for 10% CSF mix (at 0.26 w/b)
Age
Compressive strengths ( N/mm^ ) Density Compressive strengths ( N/mm^ ) Density
( d ays) Individual cube values Mean ( kg/m^ ) Individual cube values Mean ( kg/m^ )
1 54 60 58.5 57.5 2490 82 87 89 86 2495
3 68 73.5 68.5 70 2492 104 98 101 101 2498
7 88.3 94 93.7 92 2505 103 109 100 104 2500
28 108.5 110 105.5 108 2510 102.8 111.3 103.9 106 2502
90 125 118 120 121 2495 109 121.1 114.9 115 2497
180 124.5 135 136.5 132 2487 120.3 112.7 121 118 2490

No significant differences are evident between the densities of the standard and TMC cubes
Table 3 : C om pressive strength and density data for 20% PFA mix (at 0.26 w/b)
Age .red . m m m im
Compressive strengths ( N/mm^ ) Density Compressive strengths ( N/mm^ )

( days ) Individual cube values Mean ( kg/m^ ) Individual cube values Mean
1 47 51 46 48 2495 62.5 69 69.5 67

3 57 53 52 54 2490 73 81.5 70.5 75

7 69.1 77.5 75.4 74 2492 83 86.5 78 82.5

28 102.1 88.2 82.7 91 2470 80 84.7 90.3 85

90 109.7 99 97.3 102 2475 95 88 96 93

180 108 117 105 110 2480 90 103.5 97.5 97

Table 4 : Compressive strength and density data for 40% PFA mix (at 0.26 vy/b)
A ge
Compressive strengths ( N/mm^ ) Density Compressive strengths ( N/mm^ )

( days ) Individual cube values Mean ( kg/m" ) Individual cube values Mean
1 17.5 20.1 16.4 18 2487 45 38 37 40
3 - - - - 2485 - - - -
7 57 53 55 55 2450 69.3 68 51.7 63

28 73 82 79 78 2452 70 79 73 74

90 82.5 94 96.5 91 2450 78.2 89.4 87.4 85

180 109.5 96.4 109.1 105 2456 85.7 100 96.3 94

Table 5 : C om pressive strength and density data for 30% GGBS mix (at 0.26 w/b)
Age m m ïiS i& K
Compressive strengths ( N/mm Density Compressive strengths ( N/mm^ )

( days ) Individual cube values Mean ( kg/m' : Individual cube values Mean
1 38 45 43 42 2475 62 71 68 67

3 51.5 59.1 57.4 56 2481 73 80 72 75

7 77 86 80 81 2483 90 78 79.5 82.5

28 97 110 108 105 2488 83 87.5 84.5 85

90 117 111 108 112 2477 97.5 95 86.5 93

180 109 123 119 117 2475 92.5 104.5 94 97

Table 6 : Compressive strength and density data for 60% GGBS mix (at 0.26 w/b)
Age
Compressive strengths ( N/mm^ ) Density Compressive strengths ( N/mm^ )
( days ) Individual cube values Mean ( kg/m^ ) Individual cube values Mean
1 14.2 16.5 14.3 15 2467 29.5 33 27.5 30
oo
7 48 55 53 52 2472 62 55.1 56.9 58

28 86 82 87 85 2475 69 68 76 71

90 101.5 95.5 94 97 2469 93 85 86 88

180 98 112 102 104 2471 87 101 97 95
 Ternary m ixes at 0.26 w/b ratio
Table 7 : Compressive strength and density data for 40% PFA + 10% CSF mix (0.26 w/b)

Age Standard Cured TMC '

Compressive strengths ( N/mm^ ) Density Compressive strengths ( N/mm^ )

( days ) Individual cube values Mean ( kg/m^ ) Individual cube values Mean
1 8.7 10.5 9.3 9.5 2447 40.5 38.7 36 38.4
7 63 59 61 61 2455 85 77 84 82

28 87 92 85 88 2457 95.2 93.5 100.8 96.5

90 115 103 112 110 2460 98.7 108.5 110.8 106

180 125 114.3 119.2 119.5 2470 112 118.5 107 112.5

Table 8 : Compressive strength and density data for 50% PFA + 10% CSF mix (0.26 w/b)

Age Stan dard C l

Compressive strengths ( N/mm^ ) Density Compressive strengths ( N/mm^ )
( days ) Individual cube values Mean ( kg/m=" ) Individual cube values Mean
1 8.5 8.7 6.8 8 2445 29.8 33.1 27.1 30
7 36.5 41.3 34.7 37.5 2454 80.3 76.1 77.6 78
28 73.5 82.5 80.4 78.8 2453 85.3 91 93.7 90
90 107 98 111.2 105.4 2455 108 96.1 97.4 100.5
180 109.2 117.5 119.8 115.5 2467 100.3 114 115.7 110
Table 9 : C om pressive strength and density data for 60% G GBS + 10%CSFmix (at 0.26 w/b)
Age mtm B e#
Compressive strengths ( N/mm^ ) Density Compressive strengths ( Wmm^ )

( days ) Individual cube values Mean ( kg/m^ ) Individual cube values Mean
1 6.7 8.2 7.6 7.5 2469 39 33 37.5 36.5
7 68 74.5 72 71.5 2474 81 73 72.8 75.6
28 89.7 97 95.3 94 2481 106.4 99 94.6 100
90 106.4 120 115.6 114 2476 101.8 108 106.7 105.5

180 110 122 119 117 2475 103 114 107 108

Table 10 : Compressive strength and density data for 70%GGBS + 10%CSFmix (at 0.26 w/b)

Age M l
Compressive strengths ( N/mm^ ) Density Compressive strengths ( N/mm^ )
( days ) Individual cube values Mean ( kg/m^ ) Individual cube values Mean
1 3.4 2.8 5.8 4 2467 19 24.5 22.5 22
7 52 47 49.5 49.5 2471 49 44 48 47
28 83 77 74 78 2473 57 58.2 64.8 60
90 97 89 98.7 94.9 2475 61 69 68 66
180 97 100 109 102 2468 68 81 79 76

Table 11 : Compressive strength and density data for 80%GGBS + 10%CSFmix (at 0.26 w/b)

Age
Compressive strengths ( N/mm^ ) Density Compressive strengths ( M/mm^ )
( d ays) Individual cube values Mean ( kg/m^ ) Individual cube values Mean
1 2 * 2 2375 14 17.5 16.5 16
7 42 37 35 38 2398 39.1 33.4 35.5 36
28 57 64 59 60 2441 49.3 52.1 57.6 53
90 83 92 86 87 2445 60 64.1 58.9 61
180 89 94 102 95 2450 72.1 67.3 70.6 70
* Demoulding problems
Table 12 : Com pressive strength and density data for 10%CSFmix (at 0.30 w/b)

Compressive strengths ( N/mm^ ) Density Compressive strengths ( N/mm^ )

( days ) Individual cube values Mean ( kg/m^ : Individual cube values Mean
1 48.9 52.1 49.9 50.3 2487 80.5 84.3 81.2 82
7 85.6 92 92.4 90 2492 99.8 101.5 101.7 101
28 103.9 107.3 112.8 108 2475 108 102 111 107

90 114.7 121.1 126.3 120.7 2485 111 114.7 110.3 112

180 118.9 129.8 132.3 127 2483 117.4 112.8 123.8 118

Table 13 : Compressive strength and density data for 40%PFA + 10%CSF mix (at 0.30 w/b)

Age
Compressive strengths f^N/mm^ Density Compressive strengths ( N/mm^ )
( days ) Individual cube values Mean ( kg/m^ ) Individual cube values Mean
1 8.9 9.8 11.3 10 2439 33.4 33.5 38.1 35
7 58 59.7 62.3 60 2442 75 78.5 83.5 79
28 92.5 90.7 86.8 90 2445 90.3 102 92.7 95

90 99.4 109.3 112.3 107 2443 107.1 98.4 106.5 104

180 109.9 117.2 120.9 116 2440 104 112.8 113.2 110

Table 14 : Compressive strength and density data for 60%GGBS + 10%CSF mix (at 0.30 w/b)

Age Standard Cured ;

Compressive strengths ( M/mm^ ) Density Compressive strengths ( N/mm^ )

( d ays) Individual cube values Mean ( kg/m^ ) Individual cube values Mean
1 6.3 7.8 9.9 8 2455 31.9 34.2 38.9 35

7 68 73 75 72 2457 69.5 77.1 75.4 74

28 88 95 93 92 2462 91.9 98.7 100.4 97

90 106 109 115 110 2460 103.5 109.5 99 104

180 109.2 118.7 117.1 115 2463 106 115 103 108
T a b le 15 : C om pressive strength and density data for 10%CSFmix (at 0.20 w/b)

Age
C o m p ressive strengths ( N /m m Density C o m p re s s iv e stren gth s ( N /m m

( days) Individual cube values Mean ( kg/m^ ) Individual cube v a lu e s Mean

1 59 64 57 60 2507 9 7 .5 106 1 0 2 .5 102
7 93 91 98 94 2510 10 4 111 109 108

28 108.1 10 9.5 1 1 6 .9 11 1.5 2497 11 7 109 10 7 111

90 1 2 9.3 138.1 1 3 7 .6 135 2515 1 1 2 .5 122 1 1 0 .5 11 5

180 129 1 3 9 .5 1 4 2 .5 137 2511 111 125 115 117

T a b le 16 : C o m p ressive strength and density data for 4 0 % P F A + 1 0 % C S F m ix (a t 0 .2 0 w /b )

Age
C o m p ressive strengths ( N /m m D ensity C o m p re s s iv e strengths ( N /m m )

( days ) Individual cube values Mean (kg/m': Individual cube v a lu e s Mean

1 11.2 12 .5 13 .8 12 .5 2457 40 4 0 .7 4 2 .3 41

7 72 74 70 72 2464 8 7 .5 83 8 4 .5 85

28 9 8 .5 10 6.7 9 7 .8 101 2465 108 101 109 10 6

90 11 6 .8 12 6.3 1 2 2 .9 122 2470 11 7 109 1 2 0 .5 1 1 5 .5

180 13 1 .5 120.1 1 2 3 .4 125 2473 10 9 1 2 4 .5 1 2 6 .5 120

T a b le 17 : C om p ressive strength and density data for 6 0 % G G B S + 1 0 % C S F mix (a t 0 .2 0 w /b )

Age Standard Cured . ‘v Tl

C o m p ressive strengths ( N/m m ^ ) Density C o m p re s s iv e strengths ( M/mm^ )

( days ) Individual cube values Mean ( k g /m ' ) Individual cube va lu e s Mean

1 24.1 2 5 .2 2 7 .2 2 5 .5 24 71 4 8 .2 4 5 .5 5 0 .3 48

7 88 8 2 .5 9 0 .5 87 2479 98.1 1 0 4 .8 9 5 .6 9 9 .5

28 1 0 4 .5 112 115 1 1 0 .5 2485 10 8 116 11 2 112

90 116 125 11 9 120 2483 11 3 120 115 116

180 1 1 7.9 12 9 .4 1 2 7 .7 125 2488 1 1 0 .7 1 2 4 .5 1 2 7 .8 121

 APPENDIX C

Dielectric Properties of High Strength Concrete

DIELECTRIC PROPERTIES OF HIGH STRENGTH CONCRETE

C.l Introduction

In the past 40 years in-situ and non destructive testing of concrete has gained wide
acceptance for evaluating existing concrete structures'^’ Since then there have been
many different techniques used for monitoring structures.

The continuing development of high strength concrete and more recently self
compacting concrete, has increased the demand for monitoring these concretes during
production, preferably non-destructively. One possible method to monitor the strength
during the early stages of hydration is the maturity method. This requires continuous
measurement of the concrete temperature and its applicability to high strength
concrete has been criticised^'^®\ Also many of the existing non-destructive
measurement techniques are not practical when used on site for routine testing, and
are mainly confined to laboratory based testing.

A novel non-destructive method has been recently developed at the Technical

University of Delft in the Netherlands. This is based on measuring the dielectric
properties of concrete namely, the permittivity and conductivity, and has been used in
the past to characterise soil behaviour^^’^l

The dielectric properties of concrete are dependent on the amount and state of water
in the pore system and this technique therefore monitors the progress of the hydration
process in concrete and to predict the mechanical properties, e.g. strength.

Hydronix Ltd (the UK distributors) made it possible for me to use this technique
during the final stages of my research programme to assess its potential for predicting
the strength development of high strength concrete. The theoretical aspects of the
dielectric measurements are initially discussed followed by the results of the
experimental investigation.
C.2 Theoretical Background
The dielectric properties are the electro-magnetic properties of a material. If we
consider a capacitor formed by two metal plates, an applied electric potential will
charge the plates. An electric field, E, between the plates will result in a force acting
on a point charge, Q. The electric field, E, and the resulting force acting on the charge
are vector quantities. The force vector, F, is related to the E-field vector, E, by

F = QE

From Coulomb’s law the force between two point charges, Qi and Q%, distance, d,
apart.
F = QiQ2 / 4 7 r£oerd^ r 1,2

Where eo = 8.854 x 10 is the permittivity of free space, &tthe dimensionless relative

permittivity of a material with respect to that of free space and the unit vector r
points from Qi to Q2 .
When a material is between the two plates of a capacitor, the forces acting on the
charged particles will orient them in the electric field, E. As a result they become
polarised. However, spontaneous fluctuations of the molecules tend to randomise this
alignment and eventually a dynamic equilibrium is established among the molecules
as a result of these two effects. The relative permittivity 8 r is a measure of the
completion between these two effects; i.e. a measure of polarisability.
The polarisation of permanent dipoles with and without the application of an external
E-field is shown in figure Cl.
b) c)

Figure C l Polarisation of dipoles with and without an E-field

i ta
(4

I
%
When the electric field is removed, the induced energy will be dissipated within a
certain time period. By applying an alternating field, energy will be stored and/or
absorbed depending on the frequency applied. The frequency dependence of the
polarisation process can be described by a complex representation of the relative
permittivity, Gr.
In this context, the relative permittivity will be referred to as permittivity, and denoted
by 8 . It is defined as:
8 = 8 - je

where the real part of the permittivity, 8 , is a measure of the total polarisability. For a
static E-field, é is usually referred to as a dielectric constant, which is usually higher
than of free space.
The imaginary part of the permittivity, e, represents the total energy absorption which
includes the dielectric loss, 8 d, and loss by ionic conduction.

8 = 8 d + o / 27T8of
where o is the ionic conductivity of the water in the pores and f the frequency of the
E-field applied.
A more practical way to represent this equation is in terms of the specific electrical
conductivity of the pore water, which can be defined as

Ow 27rf 8 w. 8 o

Where éw is the imaginary part of the permittivity of water. Ow includes the dielectric
losses; if these are negligible then Ow approximates to the specific ionic conductivity,
o, i.e. Ow~ o. Often the specific electrical conductivity of the water is also referred to
as the conductivity and is attributed to the movement of ions.

Note: The theoretical aspects of the dielectric properties have only been briefly
discussed, the reader is referred to Hasted^^^®^ and Granf^^’^ for a comprehensive
account.
C.3 Dielectric Properties

The dielectric properties, described in the previous section, can be used to characterise
materials and individual phases present in the material. The permittivity represents the
electrical polarisation of a material, whereas the conductivity represents the amount of
electrical current that propagates through it. The frequency of the electric field with
which the dielectric properties of concrete are measured is critical in the interpretation
of the results.

The changes in the dielectric properties of concrete result from the changes in the
hydrating cement paste and are mainly dependent on the amount and state of the
water. The cement paste can be considered as a three-phase system of:

• the unhydrated cement.

• the cement gel, which contains the hydration products and bound water.
• the capillary pores, which are filled with free water and air.

Unhydrated cement has a dielectric behaviour similar to a solid such as glass and
sand. It has a low permittivity of about 3-4 (dimensionless units) and a very low
conductivity, 10'^ mS/cm (milli Siemens/cm). The water in the cement gel is
physically bound to the hydration products, therefore, more energy is required to
rotate the water molecule. At very low frequencies, ions in the pore solution will
dominate the measurement and at high frequencies molecules cannot respond to the
frequency. The water in the capillary pores can be regarded as free water, and its
dielectric behaviour is similar to salty water. The permittivity of salt water is about
80-83 whereas the permittivity of fresh cement paste is in the region of 10^ - 10^
and that of hardened cement paste is usually less than 10. The conductivity of
concrete can be in some cases zero for dry concrete and up to 300 mS/cm for fresh
concrete. In comparison, water has a conductivity between 200-500 mS/cm.
These values indicate that both the permittivity and the conductivity of concrete
cannot be attributed to the amount of water in the pore system alone and is also
dependent on its microstructure.
As mentioned earlier, the delectric properties are strongly dependent on the
measurement frequency. At high frequencies, the charge cannot follow the applied
frequency and dielectric losses will result in high conductivity Haddad has
shown at very high frequencies (>lGHz) the dielectric properties of water in concrete
carmot be measured. Similarly Al-Quadi et has shown that low frequencies
yield a high permittivity due to the concrete/electrode boundary.

Researchers at Delft University carried out tests at different levels of applied

frequencies to establish the optimum frequency for the practical applications and for
imderstanding the dielectric properties of concrete^“^\ For frequencies in the range of
lOMHz to 6 GHz, the permittivity of a concrete mix was measured using a network
analyser. The results illustrated in figure C. 2 showed that at frequencies greater than
100 MHz there is hardly any change in the permittivity whereas at low frequencies
there is an increase after the setting period. Based on these results a frequency range
between 10 and 100 MHz gave the required dielectric measurements. The optimum
frequency was found to be 20 MHz and this was chosen in the design of the concrete
sensor.

C.4 The Dielectric Strength Sensor (CONSENSOR)

To measure the dielectric properties a sensor was developed by IMAG-DLO Institute

for Agricultural and Environmental Engineering in The Netherlands. This is a chip of
4 x 4 mm which generates an electrical field at 20 MHz between two stainless steel
rods, 3 cm long and 1 cm in diameter, spaced 2 cm apart on a plastic plate. These rods
act as electrodes and are cast into the concrete cube or structure, as shown in figure
C.3. The sensor is controlled by a computer which also logs and stores the measured
values of permittivity and conductivity. The chip also measures the temperature of the
concrete.

This prototype apparatus has been successfully used in the laboratory and on site and
since its initial manufacture several modifications have been made to make it more
practical. A recent version of the apparatus is shown in figure C.4. One of the major
improvements is that the new device can be reused, thus making it cost effective.
Figure C.3 : Prototype dielectric sensor

Figure C.4: Recent version of CONSENSOR device

 Tvm e Qi)

10
2 OVIH2

, r* :
m

W 20

.(226-)

frotn ne;tV/OÎ^
figure C.1
An improvement to the analysis software has also been made to make it more user
friendly.

• Relationship between dielectric properties and strength

Both the dielectric properties and strength development are directly related to the
microstructure of the cement paste. To relate the dielectric properties to strength,
Beek et aP^^ investigated three normal strength concrete mixes, shown in Table C.l.
The permittivity and conductivity were monitored during hydration and are shown in
figures C.5 and C.6 . They related the dielectric properties to the degree of hydration
concept^*and have arrived at the following conclusions. The initial increase in
permittivity is a result of the formation of the hydration products, after having reached
a maximum, the permittivity decreases as the degree of hydration increases. It can be
seen that the conductivity only increases immediately after mixing then starts to
decrease. This is caused by the reduction in the amount of capillary water and by the
loss of connectivity of the capillary pores.

Table C.l: Concrete mix proportions Beek^^^^)

Mix No. Cement w/c ratio Cement gravel sand 28-day
type content strength
(kg/m^) (kg/m^) (N/mm^)
(kg/m^)
1 OPC 0.45 372 1044 787 65
2 OPC 0.5 349 1044 787 58
3 OPC 0.6 320 1044 787 44
 Perm ittivity (-)
45
40
35
30
25
20
15 CEM I 52.5 wcr 0.45
10 CEM I 52.5 wcr 0.5
5 CEM 1 52.5 wcr 0.6

0 20 40 60 80 100 120 140 160

Time (hours)

Figure C.5: Relationship between permittivity and time^^^^l

Conductivity (mS/cm)
3

2.5
CEM I 52.5 wcr 0.45
2 CEM I 52.5 wcr 0.5
■— CEM I 52.5 wcr 0.6
1.5

0.5

0 20 40 60 80 100 120 140 160

Time (hours)

Figure C.6: Relationship between conductivity and time^^^’^

The relationship between the dielectric properties and strength has to be determined
on a concrete cube in the laboratory. After this relationship has been established for a
particular concrete mix the strength can then be determined on site. Beek^^^^^ et al cast
150mm cubes and these were tested for compressive strength at 1,2,3, 6 and 28 days.
These cubes were sealed cured in the same conditions as the cube that was used for
the dielectric measurements. Figures C.l and C.S show the permittivity/conductivity
strength relationships. The relationship between permittivity and strength is only valid
after the maximum value of permittivity is reached, whereas the conductivity/strength
relationship represents the early stages of hydration i.e.first 24 hours. It can also be
seen that this relationship is independent of the water/cement ratio. Based on these
results Beek^^^^) concluded that both relationships can be used for detemining strength.
However, the conductivity/strength relationship is preferred and forms the basis of the
monitoring system.

• Practical application

As part of the evaluation of CONSENSOR, a visit to The Netherlands took place to

carry out site trials during the construction of the Lekbrug bridge in Vianen. Sensors
were placed at various locations on the bridge deck. The strength was determined by
using three methods:
• Maturity method
• Compressive cube tests
• The CONSENSOR device

The results of the comparison between these methods are shown in figure C.9. The
strengths obtained from the CONSENSOR are in close agreement with the average
cube results.
 70

60

50 E C E M I 5 2 .5 w cr 0 .5
^ C E M I 5 2 .5 w cr 0 .4 5
40 X C E M I 5 2 .5 w cr 0 .6

30

y = -2.5399X + 115.99
20

10

0
20 25 30 35 40 45 50

Figure C.7: Relationship between permittivity and strength^^^’\

60

03
C E M I 5 2 .5 w cr 0 .6
I 40 ACEM I 5 2 .5 w cr 0 .5
xz
cn X C E M I 5 2 .5 w cr 0 .4 5
c
2
w

-80.354X f- 74.727
= 0.9^4___
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

C onductivity (m S /c m )

Figure C.8: Relationship between conductivity and strengtW^^^\

 s t r e n g t h [N /m n f]

dielectnc
average compressive strength cubes
X compressive strength cube
Dutch maturity method

40
Tim e [h o u r s]

Figure C.9: Strength development of cubes determined by compression test, maturity

method and with the CONSENSOR^^^^\

C.4 Experimental Programme

• Initial tests

Initial tests were carried out to assess the use of the technique on concrete with UK
materials, some of the normal strength mixes, as shown in Table 9.1, that were tested
by researchers at Delft were replicated. The water/cement ratios were 0.5 and 0.6.
From each mix, ten 150x150x150 mm cubes were cast, one cube had the dielectric
sensor placed in it and the remaining nine cubes were cured under standard BS1881
conditions. These were then crushed at 1,7 and 28 days for determination of their
compressive strength.

• Tests on high strength concrete

Several high strength concrete mixes were selected from previous chapters together
with a self-compacting concrete mix. Details of these mixes were given in Table 4.1.
c .6 Results and Discussion

C.6.1 Change of Dielectric properties with time

Figures C.IO (a-b) show the permittivity and conductivity traces for the two normal
strength concrete mixes respectively. The permittivity increases with time during
hydration during the initial period of hydration, up to 15-30 hours, and then remains
constant before starting to decrease. The conductivity reaches a maximum before the
maximum permittivity is reached, and then decreases rapidly. These results are in
close agreement to those obtained by Beek^^^^^ ,(Figures. 9.5 and 9.6).

As mentioned earlier, these observations can be attributed to the amount of water that
is present in the pore space. The dielectric properties are known to be determined by
the ionic water that fills the pore spaces. Unlike the unhydrated and hydrated cement
the pore water has both high permittivity and high conductivity. As the hydration
continues, the ionic water is consumed and a dense matrix of CSH gel forms and
reduces the dielectric properties.

The results of the low water cement ratio concretes are shown in figures C.11-C.16.
The results obtained show the characteristic traces for both permittivity and
conductivity, the main feature of the results is that the conductivity and permittivity
values are both lower than those obtained at a high water/cement ratio. This is
expected due to the reduced amount of water in these mixes. An interesting feature is
however seen in the self compacting concrete mix, the initial conductivity is 4.5
mS/cm which is abnormally high compared to the other mixes tested. This could be
attributed to the high cement paste content in the mix together with the effect of the
type of superplasticiser used. The effects of these are not clearly understood and
require further investigation.

In contrast, the ggbs/csf mix is showing low conductivity and permittivity values, this
can be ascribed to the lower activity of the ions in the pore water and a lower cement
content in the mix. A similar low value was also found at research carried out at Delft
on ggbs mixes.
 2.5 -

0.5 w/c
> 0.6 w/c

?O
O

0 20 40 60 80 100 120 140 160

(a) Time (hours)

35 -

>, 25 -

E 20 -

10 -
0.5 w/c
0.6 w/c

0 20 40 60 80 100 120 140 160

Time (hours)

Figure C.IO: Dielectric properties - (a) Conductivity

(b) Permittivity
 2.5

E
o
w
1 .5

>
3 1
TD
C
o
O
0 .5

20 40 60 80 100 120 140 160

Time (hours)

30

25 4

20

I
1
<D
CL

5 4

0 20 40 60 80 100 120 140 160

Time (hours)
Figure C .ll: Dielectric properties - (a) Conductivity (b)
Permittivity for OPC mix at 0.26 w/c ratio
 2.5

2
E
(/)
& 1.5
>
3 1
■CD
O
o
0.5

0 20 40 60 80 100 120 140 160

Time (hours)

25

20

15
■I

CL
10

0 20 40 60 80 100 120 140 160

Time (hours)

Figure C.12: Dielectric properties - (a) Conductivity (b)

Permittivity for 10% CSF mix at 0.26 w/b ratio
 4.5 -

o 3.5

•5 2.5

0.5

0 20 40 60 80 100 120 140 160

Time (hours)

40 1

35 -

25
1 20

0 20 40 60 80 100 120 140 160

Time (hours)
Figure C.13: Dielectric properties - (a) Conductivity (b)
Permittivity for SCC mix at 0.30 w/b ratio
 3

2.5
I 2
E
1.5
o
3
"O
c
1
o
Ü
0.5

0 20 40 60 80 100 120 140 160

Time (hours)

40

25 -

0 20 40 60 80 100 120 140 160

Time (hours)
Figure C.14: Dielectric properties - (a) Conductivity (b)
Permittivity for 40% PFA+10% CSF mix at 0.26
w/b ratio.
 0.9
^ 0.8
I 0.7
Ü5
^ 0.6
•f 0.5
1 0.4
"O

0.2

0 20 40 60 80 100 120 140 160

Time (hours)

20
18
16
14
> 12

E 10
8
6
4
2
0
0 20 40 60 80 100 120 140 160
Time (hours)
Figure C.15: Dielectric properties - (a) Conductivity (b)Permittivity
for 60% GGBS+10% CSF mix at 0.26 w/b ratio
C.6.2 Relationship between dielectric properties and strength

Figure C.16 shows the correlation between conductivity and strength for the normal
strength mixes tested in the present study. The correlation is similar to that obtained
in figure 9.9. Indeed a unique relationship is found for both of the initial water/cement
ratios. This implies that it is sufficient to determine this relationship at one
water/cement ratio and apply it over a range of water/cement ratios.

Figure C.17 shows the same relationship at low water/cement ratios of 0.26 and for
the SCC mix at 0.30.
A strong correlation is evident, but it is interesting to note that even at low water
/binder ratios the correlation is independent of the initial water/cement ratio. These
findings are certainly promising and highlight the advantages of the dielectric
technique compared to existing monitoring systems, such as the maturity method,
where the strength-maturity relationship has to be determined for each concrete mix.

Figure C.18 shows the effectiveness of the correlation at 0.26 w/b ratio in a binary
blended cement of 10%CSF and two ternary blends of 40% PFA+10% CSF and
60%GGBS+10%CSF. Although a similar correlation is found between strength and
conductivity for both CSF and PFA/CSF mixes the GGBS/CSF has a unique effect on
the correlation line. This maybe linked with the low conductivities that were found for
this mix together with the use of a low cement content.

This investigation has shown the practical advantages of using the dielectric
measuring device on site, it is currently in its developing stage and only a few
prototype versions exist. Further research is required to understand the dielectric
properties of concrete, particularly in the use of blended cements.
 120

100

E
E
z
£O)
c
E
■ 0 . 6 w /c
A 0.5 w /c

0 0.2 0.4 0.6 0.8 1 1.2 1.4

Conductivty (mS/cm)

Figure C.16: Relationship between conductivity and strength for

OPC mix at 0.6 and 0.5 w/c ratio

1 2 0

1 0 0

E
E
z

£
S 40
OD

♦ SCC 0.3 w /c
A 0.26 w /c

0 0.5 1 1.5
Conductivity (mS/cm)

Figure C.17: Relationship between conductivity and strength

for OPC mix at 0.26 and SCC mix at 0.3 w/c ratio
 180

160
■ csf
140 A pfa/csf
1 2 0 ♦ ggbs/csf
1 0 0

05
C
80
2
W 60
40
20
0
0 0.5 1 1.5 2
Conductivity (mS/cm)

Figure C.18: Relationship between conductivity and strength

for 10% CSF, 40%PFA+10%CSF and
60%GGBS+10%CSF mixes at 0.26 w/b ratio
As a result of the present investigation, researchers at Delft are planning to investigate
the effects of blended cements and temperature on the dielectric properties of low
water/cement ratio concretes.

After completion of the tests in the current investigation, it was found that researchers
at TU-Delft are currently developing a new version of the apparatus with a hand held
data logger , which would eliminate carrying a portable computer on site. Preliminary
tests performed on various construction sites in Holland, have also shown that there is
a need for a smaller casing unit for the electrodes in order to reduce cutting
unnecessary large diameter holes in the formwork.

C.6 Conclusions

The dielectric properties can be used for evaluating the strength development in
hardening concrete. The technique gives a strong correlation between compressive
strength and conductivity for mixes with Portland cement and water/cement ratios
between 0.26 to 0.60.

Similar results were found with binary and ternary blended cements, however there is
a need for further research into the dielectric properties of such mixes.

The dielectric monitoring system is in its development stage and the results clearly
show the practical significance of the device.