Citation
Condition assessment and durability study of the Green Mountain Dam parapet wall

Material Information

Title:
Condition assessment and durability study of the Green Mountain Dam parapet wall performance of different cement types and implementation of the ultrasonic pulse velocity method
Creator:
Mohamed, Osama Ahmed
Publication Date:
Language:
English
Physical Description:
xiii, 193 leaves : illustrations ; 29 cm

Subjects

Subjects / Keywords:
Concrete -- Analysis ( lcsh )
Cement -- Analysis ( lcsh )
Ultrasonic testing ( lcsh )
Cement -- Analysis ( fast )
Concrete -- Analysis ( fast )
Ultrasonic testing ( fast )
Green Mountain Dam (Colo.) ( lcsh )
Colorado -- Green Mountain Dam ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 192-193).
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Osama Ahmed Mohamed.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
40275409 ( OCLC )
ocm40275409
Classification:
LD1190.E53 1998m .M64 ( lcc )

Full Text
CONDITION ASSESSMENT AND DURABILITY STUDY OF THE GREEN
MOUNTAIN DAM PARAPET WALL:
PERFORMANCE OF DIFFERENT CEMENT TYPES AND IMPLEMENTATION
OF THE ULTRASONIC PULSE VELOCITY METHOD
by
Osama Ahmed Mohamed
B.Sc., University of Khartoum, 1992
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Civil Engineering
1998
---------.Ji


This thesis for the Master of Science
degree by
Osama Ahmed Mohamed
has been approved
by
Kevin L. Rens
1218


Mohamed, Osama Ahmed (M.S., Civil Engineering)
Condition Assessment and Durability Study of the Green Mountain Dam Parapet
Wall: Performance of Different Cement Types and Implementation of the Ultrasonic
Pulse Velocity Method
Thesis directed by Assistant Professor Kevin L. Rens
ABSTRACT
This thesis has two main objectives. The first is to find the effect of aging on
the performance of concrete. The second is to find the correlations between the
ultrasonic pulse velocity (UPV), compressive strength, density, cement type, and
specimen size.
Age is one of the most important factors that influence the performance of
concrete structures. Initial concrete properties change with time and many forms of
concrete deterioration are also affected by aging. This results in probable overall
degradation of a concrete structure.
All aspects of this thesis are based on testing and inspection of the 53-year old
concrete from a parapet wall. The wall is located atop Green Mountain Dam (GMD),
near Kremmling, CO.
Several cement constituents and degree of fineness are investigated for their
effect on concrete performance over time. In particular, the effects of tricalcium
silicate, dicalcium silicate, and degree of fineness on the long time strength of
in


concrete is discussed. The effect of the amount of equivalent alkalis on the concrete
deterioration due to alkali-aggregate reaction is also investigated with respect to
certain critical amounts of equivalent alkalis. The amounts of equivalent alkalis are
discussed for the effect of their reaction with certain types of minerals present on
aggregates. Finally, the effect of time on the deterioration caused by freeze-thaw of
concrete is discussed and the resulting influence on concrete compressive strength is
emphasized.
The ultrasonic pulse velocity method is widely used to assess concrete
condition and to determine concrete properties. This study investigates the
relationship between the UPV and the uniaxial compressive strength of concrete at a
certain age and the effect of cement type on the correlation. The study also
investigates the effect of sample size and concrete density on UPV. It correlates the
UPV through cores determined in the laboratory to the UPV through the parapet wall
panels.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
Signed
Kevin L. Rens
IV


ACKNOWLEDGEMENT
I would like to express my profound appreciation to Dr. Kevin L. Rens and
Dr. Judith J. Stalnaker for their support, guidance, and encouragement throughout the
study.
Thanks are also extended to Dr. John Mays for accepting to be in my thesis
committee. Your distinguished teaching style has always attracted me to your
classes.
The National Science Foundation is sincerely thanked for funding the study. I
would like to thank the United States Bureau of Reclamation for participating in cost
sharing, providing background materials, and performing substantial amounts of the
testing related to the study.


CONTENTS
Chapter
1. Introduction...........................................................1
2. Background of the Study................................................3
2.1 Initiation and Objectives of the Study..............................3
2.2 Construction of the Parapet Wall....................................5
3. Overview of Current Testing...........................................13
3.1 Visual Inspection of the Parapet Wall..............................13
3.2 Mechanical T esting................................................15
3.3 Ultrasonic Testing.................................................17
3.3.1 Test Equipment and Setting....................................... 18
3.3.2 Testing Principle..................................................19
3.3.3 Procedure..........................................................20
4. Effect of Chemical and Physical Characteristics of Cement on Long
Time Compressive Strength.............................................27
4.1 Introduction..................................................... 27
4.2 Objective..........................................................27
4.3 Measurement of the Degree of Fineness..............................28
4.3.1 Air Permeability vs. Wagner Turbidimeter Methods..................31
4.4 Theoretical Background on Cement Chemistry and Degree of Fineness .31
VI


4.5 Presentation and Analysis of Test Results...........................33
4.6 Summary of Conclusions..............................................43
5. Time Effect of Alkali-aggregate Reaction on the Performance of
Concrete..............................................................46
5.1 Introduction........................................................46
5.2 Objective...........................................................47
5.3 Effect of Alkali-aggregate Reaction in Contrast with the Other
Chemical and Physical Characteristics of Cement....................47
5.4 Testing Related to the Chapter......................................47
5.4.1 Preview of Petrographic Analysis....................................48
5.4.2 Preview of Visual Inspection........................................49
5.4.3 Preview of Mechanical Testing.......................................49
5.5 Theoretical Background............................................. 50
5.5.1 Mechanism of the Alkali-aggregate Reaction..........................51
5.6 Presentation and Analysis of Test Results...........................52
5.6.1 Petrographic Examination and Alkali Reactivity of Aggregate.........53
5.6.2 Visual Inspection...................................................59
5.6.3 Mechanical Testing (compressive strength)...........................68
5.7 Summary of Conclusions..............................................74
6. Factors Affecting Resistance of Concrete to Freeze-thaw Damage.........79
6.1 Introduction....................................................... 79
6.2 Objective...........................................................80
Vll


6.3 Testing Related to this Chapter.....................................80
6.3.1 Overview of Freeze-thaw Testing................................... 80
6.3.2 Overview of Visual Inspection.......................................80
6.3.3 Overview of Compressive Strength....................................82
6.4 Theoretical Background..............................................82
6.5 Presentation and Analysis of Test Results...........................84
6.5.1 Environmental Conditions at the Parapet Wall Location...............84
6.5.2 Freeze-thaw Testing.................................................86
6.5.3 Visual Inspection...................................................87
6.5.4 Compressive Strength................................................89
6.6 Summary of Conclusions..............................................91
7. Factors Affecting Velocity of Ultrasonic Pulses through Concrete and
Correlation to Concrete Properties.....................................93
7.1 Introduction...................................................... 93
7.2 Objective...........................................................94
7.3 Theoretical Background..............................................94
7.3.1 Wave Types..........................................................95
7.3.2 Couplants...........................................................96
7.3.3 Ultrasonic Transducers..............................................96
7.3.4 Velocity of Sound Waves and Material Properties.....................97
7.4 Presentation and Analysis of Test Results...........................98
viii


7.4.1 Effect of Sample Size on Velocity of Ultrasonic Pulses
through Concrete...................................................99
7.4.2 Effect of Direction of Compression Wave
(longitudinal or lateral directions with respect to test specimen).102
7.4.3 Effect of Concrete Density on Velocity of Ultrasonic Pulses........104
7.4.4 Relationship between Compressive Strength of Concrete and UPV......106
7.4.5 Effect of Cement Type on Correlation of UPV and Compressive
Strength........................................................109
7.5 Summary of Conclusions..........................................113
8. Summary, Conclusions, and Recommendations for Further Work..........133
8.1 Effects of Tricalcium Silicate, Dicalcium Silicate, and Degree of
Fineness of Cement on Short and Long Term Compressive Strength
of Concrete.....................................................133
8.2 Effects of Equivalent Alkalis on Long Term Performance..........134
8.3 Effects of Cement Type, Air-entrainment, and Air-entrainment Method
on Resistance of Concrete to Freeze-thaw Damage.................135
8.4 Factors Affecting UPV through Concrete and Correlation to
Compressive Strength............................................136
8.5 Recommendations for Further Work................................138
Appendix
A. Definitions and Terminology........................................139
B. Visual Inspection..................................................143
Bibliography...........................................................192
ix


FIGURES
Figure
2.1 General View of the Parapet Wall................................4
2.2 Typical Parapet Wall Panel..................................... 9
2.3 Section through the Parapet Wall...............................10
3.1 Direct Measurement of Travel Time..............................21
4.1 Relationship Between C3S% and the 28-Day Compressive Strength of
Field Concrete................................................36
4.2 Relationship between C2S% and the 28-Day Compressive Strength of
Field Concrete.............................................. 37
4.3 Strength Development through Time for Laboratory Mixed Concrete.41
5.1 Expansion of Selected Materials from the GMD Aggregate with
High and Low Equivalent Alkalis...............................57
5.2 Panel Number 24-4..............................................65
5.3 Panel Number 43-1..............................................67
5.4 Panel Number 43A-1........................................... 69
5.5 Variation of Compressive Strength with Time for Cement


Types I,II, IV.....................................................72
6.1 Climatological Data at the GMD Location............................85
7.1 Variation of UPV along Two Lines of Measurement for Wall Panels
(figures 1 through 14)...........................................126
7.2 Correlation of Wall UPV and Core UPV..............................100
7.3 Relationship between UPV through Cores and Concrete Density.105
7.4 Relationship between UPV through Wall and Concrete Density........107
7.5 UPV through Wall Panels, through Cores, and Compressive Strength ..108
7.6 Relationship between UPV and Compressive Strength for Cement
Type I............................................................110
7.7 Relationship between UPV and Compressive Strength for Cement
Type III.........................................................Ill
7.8 Relationship between UPV and Compressive Strength for Cement
Type IV..........................................................112
XI


TABLES
Table
3.1 Dimensions and Concrete Density of Cores Obtained from the
GMD Parapet Wall.................................................24
4.1 Degrees of Fineness Obtained Using the Air Permeability Method
and the Wagner Turbidimeter......................................29
4.2 Degree of Fineness, W/C Ratio, 28-Day Compressive Strength,
C3S, and C2S for Different Cement
Types............................................................34
4.3 28-Day Compressive Strength VS. Tricalcium Silicate..............35
4.4 Compressive Strength of Laboratory Cured Concrete Cylinders at
Different Ages...................................................38
4.5 Strength Development for Different Cement Types Representing
the Five ASTM Cement Types ... ..................................40
5.1 Petrographic Analysis of Aggregates..............................54
5.2 Chemical Analysis of Cements.....................................56
5.3 Percent Expansion of Parapet Wall Aggregate With Various Cement
Types............................................................58
5.4 Visual Inspection of Selected Panels.............................60
XU


5.5 Compressive Strength of Concrete at Different Ages..............71
6.1 Summary of Freeze-thaw Tests....................................81
6.2 Compressive Strength of Selected Panels.........................83
7.1 Field Sonic Testing............................................116
7.2 Sonic Testing of Cores Obtained from the GMD...................131
XU1


1. Introduction
Concrete properties are sensitive to time and may vary substantially through the
life of the concrete structure. The deviation of concrete properties from their initial
design values may affect the performance of the structure significantly. Generally,
concrete gains strength beyond the initial design value. The gain of strength through
time provides the designer with an additional factor of safety. The variation of
compressive strength with time is a function of several parameters such as the water
content and cement type. The type of cement, characterized by its chemical and
physical characteristics, influences the amount and nature of change in concrete
properties through time. As time may improve the concrete properties, it may also
affect these properties adversely. Deterioration of concrete (such as internal
microcracking) may be aggravated with time resulting in loss of strength and
degradation of other properties including modulus of elasticity. Knowledge of the
physical and chemical characteristics of construction materials will help anticipate the
future performance of concrete structures. The performance of existing concrete
structures can be evaluated using different techniques such as coring and mechanical
testing, and nondestructive testing techniques. A widely used method of
nondestructive testing is the ultrasonic pulse velocity (UPV) method. The
interpretation and utmost implementation of the test results obtained using this
l


method is still not fully explored. Further research needs to be conducted at academic
institutions and industry to help introduce this method as one of the most easy to use
and cost effective inspection methods.
The need to determine the current properties of existing infrastructure and
assess its condition arises from the fact that most of the nations infrastructure is
aging. This infrastructure is still in service and its safety, serviceability, and
adequacy of structural performance needs to be determined.
While traditional structural analysis and design research is important to build
tomorrows infrastructure, the research related to assessment of existing infrastructure
should be given special consideration. The need for assessment of existing
infrastructure stems from the fact that most of this infrastructure is aging and still in
service.
This thesis investigates two aspects related to assessing the quality of existing
concrete infrastructure:
The first aspect is the effect of cement chemical and physical characteristics
on the long time strength and durability of concrete structures and the effect
of environmental conditions on strength and durability of different concrete
types.
The second aspect is the application of the ultrasonic pulse velocity method
to determine concrete properties. Correlations pertinent to application of this
method to assess concrete quality and properties are developed.
2


2. Background of the Study
The following sections summarize the history of an earlier study and its relation
to the current study. The details presented in the following section are found in the
Douglass report (Douglass et. al., 1947) and the materials laboratory report No. C-
224,1943. Some of the testing results presented in the subsequent chapters are also
found in the Douglass report. ^
2.1 Initiation and Objectives of the Study
The original study was initiated in 1940 by the Portland Cement Association
(PCA). At that time, the objective of the study was to investigate the service
characteristics of different cement types. The 28 cement types under investigation
were used in the construction of different concrete structures at various locations in
the United States. These structures are exposed to different climatic conditions. The
same cements were used in adjacent panels of highway slabs located in New York,
South Carolina, and Missouri. The cements were also placed in concrete piles
exposed to tidewater on the coasts of Florida, Massachusetts, California as well as
fresh water in New York. No attempt was made to obtain the findings of the studies
conducted in other states. In Colorado, the cements were used in the construction of
the parapet wall located at the top of the GMD. Figure 2.1 shows a general view of
the parapet wall. By using these cements in actual structures, their performance
3


Figure 2.1. General View of the Parapet Wall


under actual service conditions can observed. The properties of concrete measured in
the laboratory can then be related to behavior under real service conditions.
The United States Bureau of Reclamation (USBR) agreed to use the 28 cements in the
construction of the parapet wall of the GMD. The cements were tested thoroughly
prior to their use in the construction of the wall to determine their chemical
composition and physical characteristics. Eight of the cements were ASTM (see
Appendix A) type I, six were type II, three were type III, four were type IV, one was
type V. In addition, six were also treated with air entraining agent to determine the
effect of air entrainment on concrete behavior. The aggregates used in the
construction of the wall were petrographically examined to determine their mineral
and geological composition.
2.2 Construction of the Parapet Wall
The Portland Cement Association provided 24 of the 28 cements considered in
this study and used in the construction of the parapet wall. Three of the 24 cements
were air-entrained in the mixer during construction of the wall. These panels in
which the method of air-entrainment was applied have the letterT in their
identification number and will be referred to later in this thesis. A local cement
company provided the remaining 4 cement types making a total of 28 cement types.
The aggregates used in the construction of the wall were mined from the railroad
pit located at Kremmling, Colorado. The same aggregate was used in all construction
work on the dam site. The maximum aggregate size used in the wall is 38 mm.
5


Water for construction is obtained from two sources. The first source was the
contractors domestic supply obtained from springs close to the site location.
The second source was the reservoir itself to help supplement the supply in the
mixing plant.
The concrete was patched and mixed in one plant. The aggregate was patched in
one weighing hopper equipped with dial scales. From the weighing hopper, the
patched aggregate was dumped into another hopper directly above the back of the
concrete mixer. The cement was weighed on a platform scale in a cone cart and
introduced into a hopper on the back of the mixer in a separate chute. Some cement
types were treated with a Vinsol Resin air-entraining agent. Where the Vinsol Resin
was added, the standard solution was poured into a cement and aggregate charge
while it was in the hopper. The mixing water was not weighed and was discharged
manually from a calibrated tank. The mixer was charged by introducing one-half of
the required water first, and then adding the aggregate, cement, and the remaining
half amount of mixing water simultaneously. As the last of the aggregate entered the
mixer and the gate to the hopper was closed; an automatic timer was started, mixer
controls locked, and mixing continued for one and one-half minutes at 15 revolutions
per minute.
Each panel, which is 2.74 meter long encompasses a volume of 1.3 cubic meters
of concrete. The concrete was mixed in two batches of 0.65 cubic meters each. The
6


mixer was not washed between batches. The mixer was clean at the start of the days
run, and was primed with a grout mix of sand, cement, and water.
The cement scales were cleaned and checked prior to the start of the test program.
The regular USBR mixing plant inspectors had previously checked the aggregate
scales.
The base of the parapet wall was placed on thoroughly compacted earth fill.
Round 16 mm diameter dowels at 0.61 meters centers were left protruding 0.46
meters at the horizontal construction joint. The dowels were cleaned and construction
joint sandblasted before setting the panel forms.
Wooden plywood construction forms were used on the downstream face with
one-half inch absorptive form lining. On the upstream face, the forms consisted of 6
mm plyboard. The so called She-bolts were used to hold the forms in place as
well as the usual amount of external bracing for rigidity. The panel forms were
checked for line and grade by the USBR survey crew. An open construction joint 6
mm wide was constructed at every other panel.
Concrete was unloaded from the trucks by hand shoveling into an open chute and
pushed into the forms, where it was compacted into 46 cm lifts, by an immersion type
vibrator. An average of about 25 minutes was required in placing the concrete in
each panel. In order to ensure that the proper cement went into each panel, each truck
driver was required to carry a card identifying each load.
7


The forms were stripped from the panels the following day or about 18 hours after
placing. The surface voids were filled, and care was exercised in using the same
cement as was used in the panel. The upstream face of the panels contained many
entrapped air voids while the downstream face, where celotex was used, was
practically free of such voids.
The casting of the parapet wall, which consisted of 104 panels, was completed in
May, 1943. This total number of wall panels permitted each of the 28 cement types
to be used 3 or 4 times. The lower part of the wall below the ground level was
constructed with one cement type, designated type 0, (see Appendix A) while the
upper part of the wall was constructed with different cement types. Each of the 28
cement types belonged to one (or more) of the current ASTM cement types. Each
cement is identified by a two-digit number (or two digits plus a letter). The first digit
refers to the ASTM type (e.g. group 1,2,3,4, or 5), while the second digit is a
subclassification within a particular ASTM type. The subclassifications recognize the
differences in chemical constituents and physical characteristics within ASTM
cement types. For example, the cement designation 43-4 means that this cement
belongs to ASTM group IV. The second digit, 3, is used to identify this cement type
among the type IV cements used in this study. The last digit, 4, means that this
cement type is placed in panel number 4 in a set of 4 panels. Each panel was 1.14
meters deep by 2.74 meters long (see Figure 2.2 and 2.3). The thickness of the wall
panels on the top part was about 0.46 meters.
8




Figure 2.3. Section through the Parapet Wall
0.457 m
<---------------->

0.914m
10


Numerous tests were performed prior to and in 3 years after the wall was constructed.
Some of the tests were performed on the concrete materials while others were
performed on the concrete cylinders and mortar bars. Cylinder size varied depending
on test type, standards, and specifications.
Tests on construction materials consisted of tests on cements and tests on
aggregates. Tests on different cement types were conducted to determine the
chemical constituents and physical properties. Some of these chemical constituents
and physical properties are studied in detail in this study in terms of their effect on
durability and strength of concrete. Some of these chemical constituents studied
further in this research include tricalcium silicate, dicalcium silicate, and equivalent
alkalis. The degree of fineness of cement is also studied in detail. The aggregates
used in the construction of the parapet wall were examined petrographically to
determine their mineral composition. The aggregate was examined 18 months after
the powerhouse was completed. The aggregate was examined due to the appearance
of some cracks in the powerhouse suggesting possible expansion due to alkali-
aggregate reaction. This current study verifies the reactivity of the GMD aggregates.
Tests on concrete cylinders and mortar bars were conducted to determine the
initial properties of the parapet wall concretes. Some of the properties determined
include sonic modulus, compressive strength, volume change, freeze-thaw resistance,
and sulfate resistance. Detailed test results are included in Douglass report (Douglass
et. al., 1947). Determination of the initial condition of the parapet wall concrete and
11


properties is essential for comparison with future condition and properties of the same
concrete (characterized by cement type). Variation in condition and in properties is
studied to provide guidelines for assessment of aging concrete infrastructure.
The study initiated in 1940 by the Portland Cement Association in collaboration
with the USBR is revisited again in 1997-1998. The finalization of the study is
funded by the National Science Foundation with cost share contributions from the
USBR. The research team included members from the University of Colorado at
Denver and the USBR.
12


3. Overview of Current Testing
As outlined in Chapter 2, initial testing was performed between 1943 and 1946
and is documented in the Douglass report (Douglass et. al., 1947). This current study
was initiated in 1997 and the majority of the test results obtained are included in this
thesis. The tests performed in summer of 1997 included: visual examination of the
parapet wall panels, mechanical testing of cores procured from the wall, sonic testing
of the wall panels and the cores obtained from the wall, and petrographic examination
of cores. The petrographic examination results of the cores are not included in this
thesis since they were not made available by the time this thesis was prepared.
3.1 Visual Inspection of the Parapet Wall
The parapet wall was visually inspected prior to the mechanical and non-
destructive ultrasonic tests. This augmented the visual inspection records which were
performed between 1943 and 1946 which are provided in the Douglass report
(Douglass, et. al., 1947). Each of the 28 cements under consideration was used in
three or four panels and it was considered unnecessary to inspect all the 104 panels.
Two or three panels per cement type were selected for the visual inspection.
Complete records of the visual inspection results for the selected panels are included
in Appendix B. The objective of the visual inspection is to record all signs of
deterioration in the form of cracking, color change, efflorescence, spalls, popouts, etc.
13


Visual inspection observations were recorded in forms specially prepared for this
purpose. The forms included sketches of the upstream, top, and downstream faces of
the parapet wall, which facilitates location of the cracks within the panel face. The
environmental conditions such as snow, wind, clouds, etc. which affected the
visibility during inspection were also recorded. On the bottom of the inspection form,
a section is allocated for comments that could not be recorded in the main body of the
form. Cracks were identified in terms of pattern, width, and location within the panel.
The inspection form used in this project is included at the end of Appendix B. The
identification of cracks, color changes, and efflorescence staining helps diagnose the
deterioration causes, however it is not a simple process. For instance, alkali-silica-
reaction (ASR) gel was found infilling cracks that are caused by other reasons such
cracking caused by freeze-thaw (Non-Structural Cracks in Concrete, 1992). Since
visual inspection is the least expensive method of non-destructive testing, there is a
need to develop skills required to inspect structures and further identify causes of
distress visually. Identification of causes of cracking is not an easy task. A single
type of cracking can be caused by numerous reasons and ruling out of the causes is
not a straightforward process. However, further testing, such as petrographic
examinations, can provide more evidence to the causes of cracking deterioration.
Visual inspection results were analyzed and related to mechanical test results which
will be shown in Chapters 4 through 7. It was observed that the cracking pattern of
panels describes the degree of exposure to environmental and stress conditions. The
14


upper parts of most panels were more deteriorated than the lower parts due to
exposure to alternate sun and snow conditions. In the 1997 inspection, D-cracking
(see Appendix A) was observed along the vertical joints due to entrapped moisture
between panels or due to expansion of panels that causes longitudinal stresses in
panels. Due to the above reasons, many panels experienced cracking in a parabolic
pattern (vertex of parabola pointing upwards and opening downwards) with more
deterioration in the upper and vertical joint zones. It was also observed that the
upstream face of the panel experienced some sort of pitting of various degrees of
severity (see Appendix B for definition of pitting).
In many cases, the extent of cracking can be roughly visualized by looking at the
exterior appearance and width of the crack. When the parapet wall was inspected in
the summer of 1997, the degree of deterioration within most of the panels was
indicated by the exterior appearance of the wall panel. Mechanical testing which is
discussed in the following chapters provided stronger evidence on the degree of
deterioration within each wall panel interior.
3.2 Mechanical Testing
Drills and other supplies for procurement of cores were provided by the USBR.
On site training for coring was provided by a technician from the USBR. Cores were
procured from the wall panels in the rate of 3 to 5 cores per panel depending on
testing requirements and panel condition. Each core obtained was labeled to identify
its panel number, location within the panel, size, and any noticeable damage. Each
15


core obtained is logged in a special form that defines the core and panel numbers,
length, and location within the panel. The cores were wrapped in plastic to maintain
moisture content and transported to the USBR laboratories. At the USBR
laboratories, the cores were stored in a humidity controlled room prior to testing. The
targeted core size is 152 x 304 mm (6 x 12 inches) as recommended in ASTM
standards. Some panels were highly deteriorated and obtaining the cores was a
difficult process. This resulted in some cases in cores that were shorter than the
required 304 mm of length. Short cores required the application of the ASTM
correction factors for cores measured in compression or dynamically. The dynamic
testing of cores is not discussed in this thesis. Care was taken during cutting and
capping of cores to ensure that ASTM C42 requirements regarding length to
diameter ratio (1/d), were not violated. The maximum and minimum length to
diameter ratio permitted by ASTM C42 are 2.1 and 1.0, respectively. The ASTM
correction factor is to account for the fact that short cores tend to exhibit higher
strength than longer cores for the same type of concrete. Tables of compressive
strength included in this thesis have the ASTM correction factors applied to them. In
the testing laboratories, all cores were cut, measured, and densities were calculated.
The length of a core used to determine (1/d) ratio is the average of three
measurements while the diameter is the average of two measurements. Measured
dimensions of cores are shown in Table 3.1 which is included at the end of this
chapter. Cores were capped to ensure even surfaces during testing and were strain-
16


gauged. All strain-gauges were tested during preparation of cores for mechanical
testing. The following mechanical tests were performed:
1) Compressive strength
2) Split tension
3) Dynamic loading
4) Non-linear loading
All selected panels were tested for compressive strength while only few of the
panels were tested for split tension and dynamic loading. The intended objective of
dynamic loading is to compare the mechanical dynamic modulus to the sonic
dynamic modulus to investigate possible relationship. However, the final number of
tested cores was insufficient to deduce a relationship and no such comparison was
possible. The non-linear loading was intended to compare the behavior of air-
entrained concrete to non-air-entrained concrete under unload-reload testing.
Observations revealed inconsistent behavior. This thesis emphasizes compressive
strength of concrete as the highly important concrete property and relates it to UPV.
3.3 Ultrasonic Testing
The non-destructive testing technique used in this thesis is the UPV method.
Selected wall panels were tested insitu using this technique. Cores procured from the
wall panels were also tested in the laboratory prior to mechanical testing. The
objectives of using this method are to:
17


1) compare UPV through wall panels to UPY through cores. The ultimate goal
is to develop a tentative relationship between the two UPVs through a
continuous structure (such as the wall) and 152 x 304 mm concrete cylinders
or cores. The relationship will reflect the correlation between UPV through
cores (or cylinders) and a continuous concrete structure (the parapet wall in
this case). This will help determine the quality of concrete in existing
structures from that of the cores obtained.
2) investigate the factors affecting the relationship between compressive strength
and other parameters such as concrete density, cement type, etc.
3.3.1 Test Equipment and Setting
The equipment used for testing the wall and the cores is the V-Meter MK II
which has been designed to be portable and simple to operate. Similar device was
used during testing of the parapet wall. As a measure of the compatibility of the two
devices, some wall panels were tested used both equipment and produced similar
results. The equipment consisted of:
1) the V-Meter MK II
2) two 54 KHz transducers
3) carrying case
4) V-Meter MK II manual
5) couplant
6) AC/DC charger unit
18


The transducers used were low frequency 54KHz transducers (suitable for concrete)
that send P waves (compression type wave). Such frequency is considered low
when compared to the frequencies used with metals.
Setting of the V-Meter is simple. The High/Low gain switch compensates for
distance by increasing or decreasing the signal strength presented to the electronics.
During testing of the parapet wall, the High/Low gain switch was set on low gain
except at locations where concrete was so deteriorated that signals became too weak.
In the laboratory testing of cores, the High/Low gain switch was set to low gain.
The pulse level switch also has the High/Low options. The pulse level should be set
to low pulses at short distances and high pulses at long distances. The low pulses
option was convenient for both the parapet wall (see figures 2.2 and 2.3) and cores.
3.3.2 Testing Principle
Different factors affect the velocity of ultrasonic pulses travelling through
concrete. The most important factors include the density and elastic properties.
When ultrasonic pulses are sent through a metal, the homogeneity of the metal allows
the pulses to travel in a certain path without echoing unless an obstruction (flaw) is
encountered. Such non-homogeneity areas cause the pulse to echo back to a
receiving transducer. Consequently, the time needed for the pulses to travel from a
surface to a flaw located within the homogeneous material and back to the surface
enables the flaw to be located from the velocity, distance, and time relationship.
19


Concrete is a highly heterogeneous material that causes much echoing to happen
while the ultrasonic pulses travel through the material. Therefore, detection of flaws
within concrete is highly complicated.
3.3.3 Procedure
In this study, the ultrasonic pulse velocity method has been used to determine the
quality of concrete. The velocity of ultrasonic pulses travelling through concrete was
measured using the V- Meter previously described. The travel time of ultrasonic
pulses is measured to calculate the velocity by a simple velocity, distance, and time
relationship. The ultrasonic pulses are sent through a transmitting transducer and
received in a receiving transducer and the instrument measures the transmit time.
There are three main ways of arranging the transducers to measure transmit time; the
direct, indirect, and semi-direct methods. In this study, the direct method has been
used in which transducers are held in opposite direction, one in each side of the tested
specimen, at right angles to the surface of the material as shown in Figure 3.1. To
reduce inaccuracies caused by the presence of air voids between the sending and
receiving transducers and the tested material, a couplant material is applied between
the transducers and the tested material. Although much better than air voids, the
couplant material itself is a material that is different from the concrete material in
terms of transmit time and results in some incidental inaccuracies. It was noticed
during testing that higher amounts of couplant material slightly increase the travel
20


Figure 3.1. Direct Measurement of Travel Time
21


time. In laboratory testing of cores, the amount of couplant was reduced to minimize
errors. The couplant used in the field was petroleum vaseline.
The transmit time through selected panels of the parapet wall was measured to
determine their sonic elastic properties. The following panels were tested 14-1,16 -
4,16B 4,18 1,21 1,21 -1,24 4,31 -1, 34 4, 34B 4,41 1,42 4,42B
- 4,43 1,43A -1,51-4. The panel designations mentioned characterize the
ASTM cement type used as shown in Appendix A.
The cores obtained from the above panels were tested in the laboratory for sonic
elastic properties prior to mechanical testing. The couplant material used in the
laboratory was honey which is easily cleaned off to prepare the core for mechanical
testing (capping of cores).
Since sonic elastic modulus is a function of the density of the material, the density
of concrete cores was measured in the laboratory prior to mechanical testing. The
density was calculated by obtaining the dry and submerged weights of the particular
core. The dry weight of cores used in calculations is the average of 3 readings to
improve the accuracy. The submerged weight measurement was not repeated due to
possible errors caused by absorption of water by the core. The temperature of water
used to determine the submerged weight of the core was recorded. It is used later to
determine the density of water which is required to calculate the density of concrete.
Values of the measured densities are shown in Table 3.1. Generally, concretes that
experienced more extreme deterioration ended up with lower densities than
22


comparable concretes. The average density of cores obtained from panel 24-4 is the
least of all the panels made with type I cement. It will be seen in the following
chapters that this panel is one of the worst in overall conditions. Also, as expected,
the average density of air-entrained panels is lesser than comparable panels. This is
obvious if we compare panel 16B-4 to 16-4, 34B-4 to 34-4, and 42B-4 to 42-4. The
difference is more pronounced between panels 16B-4 and 16-4 than with other
panels.
23


Table 3.1. Dimensions and Concrete Density of Cores Obtained From the GMD Parapet Wall.
Concrete Designation Core No. Measurement of Length Average Measurement of Diameter Average Measurement of Wt. Density
(inches) (inches) (inc hes) (inches) (I bs) (pcf)
Length 1 Length 2 Length 3 Diameter 1 Diameter 2 Dry Submerged
14-1 1 9.81 9.75 9.75 9.77 5.94 5.94 5.94 23.13 13.18 145.06
2 10.69 10.63 10.63 10.65 5.94 5.94 5.94 25.36 14.53 146.12
3 10.19 10.25 10.25 10.23 5.94 5.94 5.94 24.34 13.97 146.46
4 7.5 7.5 7.5 7.5 5.81 5.88 5.84 16.97 9.53 142.33
5 7.63 7.63 7.56 7.6 5.81 5.88 5.84 17.43 9.88 144.1
16-4 1 9.5 9.56 9.44 9.5 5.75 5.81 5.78 21.45 12.37 147.41
2 9.81 9.88 9.81 9.83 5.75 5.81 5.78 22.13 12.75 147.22
3 9.25 9.13 9.25 9.21 5.75 5.81 5.78 20.89 12.01 146.79
4 10.81 10.81 10.88 10.83 5.75 5.81 5.78 24.37 14.02 146.93
16B-4 1 10.75 10.69 10.63 10.69 5.75 5.81 5.78 23.51 13.23 142.71
2 10 9.94 10.06 10 5.75 5.81 5.78 22.06 12.37 142.06
3 10.13 10.06 10.19 10.13 5.75 5.81 5.78 24.41 12.65 129.52
4 9.63 9.63 9.63 9.63 5.81 5.75 5.78 21.19 11.9 142.33
18-1 2 11.94 11.94 12 11.96 5.81 5.81 5.81 27.24 15.59 145.9
4 9.38 9.38 9.44 9.4 5.88 5.81 5.84 21.42 12.21 145.13
5 10.88 10.81 10.88 10.84 5.88 5.81 5.84 24.71 14.06 144.78
21-1 1 12.94 12.88 12.88 12.9 5.94 5.94 5.94 30.6 17.56 146.43
2 11.75 11.75 11.69 11.73 5.94 5.94 5.94 27.55 15.72 145.32
3 12.06 12 12 12.02 5.94 5.88 5.91 28.43 16.23 145.41
4 12 12.06 11.94 12 5.88 5.94 5.91 28.51 16.4 146.91
5 12 11.94 11.94 11.96 5.94 5.94 5.94 28.43 16.28 146.01


Table 3.1. (Continued)
Concrete Designation Core No. Measurement of Length Average Measurement of Diameter Average Measurement of Wt. Density
(inches) (inches) (inches) (inches) (lb?) (pcf)
Length 1 Length 2 Length 3 Diameter 1 Diameter 2 Dry Submerged
24-4 2 6.5 6.44 6.5 6.48 5.94 5.94 5.94 14.98 8.46 143.37
3 8.94 8.94 8.94 8.94 5.81 5.81 5.81 5.81 11.18 144.28
31-1 1 11.94 11.94 11.94 11.94 5.94 5.94 5.94 28.3 16.32 147.4
3 12.06 12.06 11.94 12.02 5.94 5.88 5.91 28.39 16.38 147.5
34-4 1 10.81 10.75 10.75 10.77 5.88 5.84 5.84
2 10.69 10.75 10.75 10.73 5.88 5.88 5.88 - - -
3 12.16 12.19 12.12 12.16 5.81 5.81 5.81 27.26 15.38 143.18
4 10.81 10.88 10.84 10.84 5.91 5.91 5.91 25.29 14.26 143.07
5 11.88 10.9 11.94 11.9 5.88 5.91 5.91 - - -
34B-4 1 10.5 10.44 10.5 10.48 5.88 5.88 5.88 26.34 14.8 142.43
2 O 11.81 11.75 11.72 11.76 5.81 5.81 5.81 - - -
O 4 11.75 11.81 11.75 11.77 5.81 5.88 5.84 _ _
5 11.91 11.97 12 11.96 5.81 5.81 5.81 26.85 15.13 142.96
41-1 1 12 12 12 12 5.81 5.81 5.81 27.76 16.04 147.8
2 8.81 8.88 8.75 8.81 5.88 5.81 5.84 20.45 11.71 146
3 11.88 11.94 12 11.94 5.81 5.81 5.81 27.53 15.85 147.08
4 9 9.12 9.25 9.12 5.88 5.81 5.84 21.07 12.12 146.9
5 11.44 11.38 11.44 11.42 5.82 5.82 5.82 26.3 15.13 146.92


Table 3.1. (Continued)
Concrete Designation Core No. Measurement of Length Average Measurement of Diameter Average Measurement of Wt. Density
(inches) (inches) (inches) (inches) (I DS) (pcf)
Length 1 Length 2 Length 3 Diameter 1 Diameter 2 Dry Submerged
42-4 1 9 9 9 9 5.75 5.81 5.78 20.06 11.35 143.71
2 9.5 9.56 9.56 9.54 5.75 5.81 5.78 21.24 12.04 144.06
3 7.88 7.81 7.81 7.83 5.75 5.81 5.78 17.44 9.84 143.19
4 10.13 10.19 10.13 10.15 5.75 5.81 5.78 22.81 13.03 145.54
42B-4 1 10.75 10.69 10.75 10.73 5.94 5.94 5.94 25.08 14.12 142.79
2 10 10 10 10 5.81 5.81 5.78 22.15 12.55 143.98
3 10.06 10.06 10 10.04 5.94 5.94 5.94 23.49 13.29 143.85
4 9.75 9.75 9.75 9.75 5.75 5.81 5.78 21.73 12.32 144.1
43-1 1 6.88 5.94 6.91 6.6 5.88 5.88 5.88 15.9 9.01 144
3 6.31 6.25 6.38 6.3 5.94 5.88 5.91 14.53 8.23 143.9
43A-1 1 11.06 1.06 11.06 11.06 5.94 5.94 5.94 28.9 16.76 148.55
2 12 11.94 11.94 11.96 5.94 5.94 5.94 28.81 16.73 148.82
3 12.06 12.13 12.13 12.1 5.88 5.88 5.88 26.6 15.36 147.67
51-4 1 11.25 11.31 11.25 11.27 5.94 5.88 5.91 . .
2 8.88 8.94 8.81 8.88 5.94 5.88 5.91 - - -
3 10.97 10.81 11 10.93 5.91 5.94 5.92 26.13 15 146.5
4 12.06 12 12.09 12.05 5.78 5.78 5.78 27.68 15.89 146.5
5 10.94 10.81 10.88 10.88 5.88 5.94 5.91 -
Note: The da shes n this tab e implies hat the df insities fc ar these cor ss were not obta ined.


4. Effect of Chemical and Physical Characteristics of Cement on Long Time
Compressive Strength
4.1 Introduction
The short and long term compressive strength of concrete varies with different
parameters such as cement type. The chemical and physical characteristics of a
specific cement define its classification among the five main ASTM types and hence
the different uses of the different ASTM types. In order to study the effect of a
certain physical characteristic or chemical constituent of cement on short and long
time compressive strength of concrete, other parameters need to be fixed. In this case
study, all concretes were placed under the same environmental conditions on the
parapet wall panels of the Green Mountain Dam. The same mix proportions,
aggregate type and grading, mixing and placing techniques, curing methods, and
curing duration were applied to all panels.
4.2 Objective
The objective of this chapter is to investigate the effect of two chemical
constituents and one physical characteristic of cement on the long time compressive
strength of concrete. The two chemical compounds are tricalcium silicate
(3Ca0.Si02) and dicalcium silicate (2Ca0.Si02). The physical characteristic in
consideration is the degree of fineness of cement. The cements in consideration vary
27


in the amounts of the above chemical compounds as well as the physical
characteristic. Even though the 28 different cement types have been tested, only
selected groups can be used in this section. This is to insure that either the chemical
compound or the degree of fineness can be fixed to investigate the other. Variables
not related to cement type that affect long and short time compressive strength are
fixed as mentioned above.
4.3 Measurement of the Degree of Fineness
Two methods have been used to determine the degree of fineness for the purpose
of comparison. One is the wagner turbidimeter method and the other is the air
permeability method. The chemical constituents and physical characteristics of
cements were determined in 1940s at the Denver Laboratories of the USBR. The two
methods are still in use today and are included in the ASTM standards to set cement
classification limits on the degree of fineness. Table 4.1 shows the degrees of
fineness of the cements under consideration measured using these two methods. The
average ratio, p, of the degree of fineness measured using the air permeability method
to the degree of fineness measured using the wagner turbidimeter method is 1.897.
This ratio is obtained using Equation 4.1.
Z*
(4-'>
x : ratio of degrees of fineness measured using the two methods
given in table 4.1.
28


Table 4.1. Degrees of Fineness Obtained Using the Air
Permeability Method and the Wagner Turbidimeter.
I Cement Number Air Permeability sq.m/kg Wagner sq.m/kg Air Permeability/Wagner (ratio)
11 350.5 174 2.01
11T 362 184.5 1.96
12 309 169 1.83
12T 306 162.5 1.88
13 324.5 163.5 1.98
14 334 156 2.14
15 325.5 177.5 1.83
16 324.5 166.5 1.95
16T 309.5 163.5 1.89
17 315 161 1.96
18 329 170.5 1.93
18T 371.5 192.5 1.93
21 285.5 154 1.85
21T 290 159 1.82
22 300.5 168 1.79
23 297 ' 184.5 1.61
24 370 176 2.1
25 318.5 182.5 1.75
31 517.5 281.3 1.84
33 503 257 1.96
33T 463 228 2.03
34 371 187 1.98
41 388 189 2.05
42 318 194.5 1.63
43 377.5 191.5 1.97
43A 354 186.5 1.9
51 337.5 195 1.73
4500 296.5 163.5 1.81
29


n : number of ratios for all cements.
Standard deviation ,S, of the degree of fineness ratios is 0.127 and is given by:
A variability of 6.7% in average degrees of fineness measured using the two methods
mentioned above can be considered reasonable, i.e. the average ratio of 1.897 can
reasonably describe the relationship between degree of fineness measured using the
Air Permeability method and the Wagner Turbidimeter method.
Another measure of the strength of the relationship between the two methods is
the coefficient of correlation. The coefficient of correlation between the degrees of
fineness obtained using the two methods is 0.92. This coefficient reflects a
reasonably strong relationship between the values obtained using the two methods.
The measured levels of correlation and variability of degrees of fineness obtained
using the two methods permits the use of either one for this section. The degrees of
fineness used in the subsequent sections in this chapter are obtained using the air
permeability method.

(4.2)
The coefficient of variation ,V, is 6.7% and is given by:
(4.3)
30


4.3.1 Air Permeability vs. Wagner Turbidimeter Methods
As illustrated in the previous section, the two methods provide different standards
to measure the degree of fineness of cement. However, results obtained using these
two methods are related by a specifc factor. In ASTM Cl50, the difference between
the two methods is recognized by providing different guidelines for cement
classification for each method. The minimum degree of fineness that characterizes a
specific ASTM type is given for each of the two methods. When the air permeability
method is used, the minimum degree of fineness for ASTM types I, II, IV, and V is
280 m2/kg. When the Turbidimeter test is used the minimum degree of fineness for
the cement types above is 160 m2/kg. It is not stated explicitly in ASTM that there is
an adopted ratio for the degree of fineness obtained using the two methods. However,
the ratio of the two minimum values specified is 280/160 =1.75. This is to be
compared to the ratio obtained in this study of approximately which is 1.9. The ratio
of degrees of fineness obtained in this study is 8.6% higher than the implied ASTM
ratio.
4.4 Theoretical Background on Cement Chemistry and Degree of Fineness
The most important hydration product is calcium silicate hydrate (C3S2H3) which
occupies one-half to two-thirds of the volume of hydrated cement paste. Calcium
silicate hydrate eventually determines the behavior of concrete and the overall
properties of concrete. Calcium silicate hydrate is produced by the hydration of
dicalcium silicate and tricalcium silicate according to the following equations:
31


2C3S + 6H -----------> C3S2H3 + 3CH (4.4)
Tricalcium Water
silicate
Calcium silicate Calcium
hydrate hydroxide
2C2S + 4H -----------> C3S2H3 + CH
(4.5)
Dicalcium Water
Silicate
Calcium silicate
hydrate
Calcium
hydroxide
Tricalcium silicate and dicalcium silicate are the most important chemical
compounds in terms of their contribution to the strength of concrete since their
hydration results in the formation of C3S2H3. Tricalcium silicate is more reactive
than dicalcium silicate and contributes more to the strength of concrete at early stages
of concrete development. The contribution of dicalcium silicate develops with time
and is more prominent in the long term. It has been reported that in the long term,
concrete made with cement containing higher 3Ca0.Si02 develops lower strength
than similar concrete made with higher 3Ca0.Si02 (Mindess and Young, 1981; Lea
and Desch, 1956).
The degree of fineness of cement is also of prime importance. The greater the
specific surface area of the cement the higher the rate of strength development. In
fact, increased surface area of the cement increases the area in direct contact with the
mixing water, which results in a faster hydration rate. This is contrasted by relatively
larger cement particles that delay the penetration of water to the core of the cement
grain which in turn slows the hydration rate. Therefore, the degree of fineness affects
32


the performance of tricalcium and dicalcium silicates in that unless the cement is
sufficiently fine, the reaction of the chemical compounds takes longer and the rate of
strength development becomes slower. In other words, the lesser the degree of
fineness the lesser the degree of hydration at a certain age. However, as long as the
cement satisfies the standard for the degree of fineness, changing the degree of
fineness alone will not substantially affect the strength of concrete (Mindess and
Young, 1981; Lea and Desch, 1956; McMillan and Tyler, 1948). ASTM Cl50
provides minimum values of degree of fineness specific to each cement type.
4.5 Presentation and Analysis of Test Results
Table 4.2 shows the 28-day compressive strength of 152x304 mm cylinders made
with field concrete (Douglas et. al., 1947). Also shown in the table are selected
chemical and physical properties of different cements used. To determine the effects
of the amounts of tricalcium and dicalcium silicates one can fix other parameters that
contribute to the strength. The cements selected in Table 4.3 are very similar in their
degree of fineness and water-cement ratios. Figure 4.1 illustrates graphically the data
in Table 4.3 and shows a strong linear relationship between percentage of tricalcium
silicate and the 28-day compressive strength. This is indicated by the high
coefficient of correlation. Figure 4.2 illustrates the relationship between percentage
dicalcium silicate and the 28-day compressive strength. This figure illustrates that the
higher the amount of dicalcium silicate the lower the 28-day compressive strength.
Table 4.4 illustrates the variation of compressive strength with time starting with the
33


Table 4.2. Degree of fineness, W/C Ratio, 28-Day Compressive Strength, C3S, and C2S
for Different Cement Types.
PCA Cement type Degree of Fineness (sq.m/kg) w/c ratio 28-Day Strength (Mpa) C3S (percent) C2S (percent)
11 350.5 0.52 35.44 54.5 17.9
12 309 0.53 33.44 45.2 27.7
12T 306 0.52 31.23 46.3 26.8
13 324.5 0.55 28.2 50.3 25.7
14 334 0.53 33.3 44.7 29.1
15 325.5 0.55 35.5 67 7.7
16 324.5 0.54 31.72 53.5 20.3
16T 309.5 0.51 30.75 51.4 22.3
17 315 0.54 32.75 51.8 22
18 329 0.55 31.44 47.1 25.2
21 285.5 0.54 25.86 40.6 37.1
21T 290 0.49 25.78 39.1 38.9
22 300.5 0.53 28.75 42.2 31.6
23 297 0.53 30.68 52.9 21.5
24 370 0.51 33.23 40.9 28.5
25 318.5 0.53 27.79 37.1 36.6
31 517.5 0.58 32.13 52 19.4
33 503 0.57 31.37 56.9 14.6
34 371 0.57 31.3 64.3 10.2
41 388 0.52 23.1 20.3 50.5
42 318 0.55 17.44 24.2 57.4
43 377.5 0.52 22.41 24.9 48.2
43A 354 0.52 22.89 28.2 50.8
51 337.5 0.53 26.75 40.8 39.4
Aggregate Grading :
(1) Sand :
The fine aggregate (sand) part of the aggregate varies between 39% and 41%
Fine aggregate grading; Pan #100 #50 #30 #16 #8
3.2% 10% 28.8% 30.5% 14.6% 12.9%
Fineness Modulus (F.M.) = 2.82
(2) Gravel:
Grading of aggregate; #4 3/8" 3/8" 3/4" 3/4"- 1-1/2"
15.8% 36.5% 47.7%
Each value of compressive strength is the average of 12 determinations of 152.4X304.8
mm cylinders taken from the field during the construction of the parapet wall of the
Green Mountain Dam.
34


Table 4.3. 28-Day Compressive Strength vs. Tricalcium Silicate
P.C.A Degree of Fineness w/c 28-Day strength C3S C2S
Cement type (sq.m/kg) ratio (N/sq.mm) (percent) (percent)
13 324.5 0.55 28.2 50.3 25.7
15 325.5 0.55 35.5 67 7.7
16 324.5 0.54 31.72 53.5 20.3
17 315 0.54 32.75 51.8 22
18 329 0.55 31.44 47.1 25.2
42 318 0.55 17.44 24.2 57.4
w/c ratio Degree of fineness
- Average 0.547 322.75 sq.m/kg
- Standard deviation 0.005164 5.2
- Coefficient of variation 0.94% 1.61%
35


Figure 4.1: Relationship Between C3S% and the 28-Day Compressive Strength of Field
Concrete.
36


Figure 4.2: Relationship between C2S% and the 28 -Day Compressive Strength of
Field Concrete
&
37


Table 4.4. Compressive Strength of Laboratory Cured Concrete at Different Ages
P.C.A Cement type 7- Day 28 Day 90 Day 1-Year | 2 Year C3S C2S D. Fineness (sq. m/kg)
[N/sq.mm (percent)
11 22.27 28.75 33.51 36.47 35.51 54.4 17.9 350.5
12 20.96 29.99 37.92 38.27 38.82 45.2 27.7 309
12T 22.68 31.23 36.13 38.82 38.82 46.3 26.8 306
13 16.96 27.1 33.1 35.85 39.23 50.3 25.7 324.5
14 20.69 29.23 37.1 38.85 39.85 44.7 29.1 334
15 20.2 31.37 34.89 36.27 35.85 67 7.7 325.5
16 19.03 30.06 36.75 37.51 36.41 53.5 20.3 324.5
16T 21.58 28.13 30.75 35.37 36.13 51.4 22.3 309.5
17 23.86 35.1 39.37 41.16 40.96 51.8 22 315
18 21.17 29.51 36.34 35.3 37.78 47.1 25.2 329
21 17.51 30.27 38.89 44.61 47.02 40.6 37.1 285.5
21T 17.17 26.2 29.53 38.27 38.75 39.1 38.9 290
2 16.75 29.51 36.41 37.23 39.85 42.2 31.6 300.5
23 19.37 29.51 35.37 41.78 41.92 52.9 21.5 297
24 20.89 32.13 39.51 41.09 44.06 40.9 28.5 370
25 14.69 28.27 38.89 42.61 44.75 37.1 36.6 318.5
31 27.24 34.06 37.23 38.13 39.16 52 19.4 517.5
33 28.41 35.03 35.99 37.47 37.58 56.9 14.6 503
34 25.44 32.96 35.72 38.89 38.41 64.3 10.2 371
41 9.79 27.03 40.34 44.2 44.33 20.3 50.5 388
42 7.45 20.41 36.06 41.23 41.99 24.2 57.4 318
43 11.72 27.03 38.47 43.68 45.02 24.9 48.2 377.5
43A 10.69 26.06 38.54 45.64 44.27 28.2 50.8 354
51 16.48 28.48 36.34 42.2 44.13 40.8 39.4 337.5
Aggregate Grading:
(1) Fine aggregate: Pan #100 #50 #30 #16 #8
3.2% 10% 28.8% 30.5% 14.6% 12.9%
(2) Gravel: #4 3/8" 3/8" 3/4 3/4" 1-1/2"
15.8% 36.5 47.7%
38


7-day compressive strength and ending with the 2-year compressive strength. All
concrete mixes were made with an average w/c ratio of 0.53. The coefficient of
variation of w/c ratios is about 3.9%. Aggregate grading is shown in Table 4.2. The
amount of sand in all mixes is about 40% of the amount of aggregate. Cylinders were
cast and fog cured at 70 degrees Fahrenheit in the Denver laboratories of the U.S.
Bureau of Reclamation.
Table 4.5 illustrates the time variation of compressive strength for concrete
cylinders selected from Table 4.4. Table 4.5 represents each of the current ASTM
cement types and includes the amounts of 3CaO.SiC>2 and 2CaO.SiC>2 as well as the
degrees of fineness. Figure 4.3 is a graphical representation of the data in Table 4.4.
For the selected cement types plotted in Figure 4.3, the following observations can be
made:
(1) At age seven days, the compressive strength of the five cement types varied
between 10.69 N/mm^ and 27.24 N/mm^ (maximum value approximately 155%
greater than the minimum). At age 2 years, the variation narrowed down in the
range 35.85 N/mm^ to 44.75 N/mm^ (maximum value approximately 25%
greater than the minimum). This same phenomenon is notable in Table 4.4: At
age 7 days, the maximum compressive strength (28.4 N/mm^) developed by type
33 cement (See Appendix A), was approximately 280% greater than the minimum
strength (7.45 N/mm^), developed by type 42. At age 2 years, the maximum
39


Table 4.5. Strength Development for Different Cement Types Representing the Five Cement Types
Cement type
Age #15 #25 #31 #43A #51
(ASTM type 1) (ASTM type 2) (ASTM type 3) (ASTM type 4 (ASTM type 5)
Compressive 7 Day 20.2 14.69 27.24 10.69 16.48
Strength 28 Day 31.37 28.27 34.06 26.06 28.48
(N/sq.mm) 90 Day 34.89 38.89 37.12 38.54 36.34
1 Year 36.27 42.61 38.13 45.64 42.2
2 Year 35.85 44.75 39.16 44.27 44.13
Chemical & D.F (sq.m/kg) 325.5 318.5 517.5 354 337.5
Physical C3S % 67 37.1 52 28.2 40.8
Properties C2S %$ 7.7 36.6 19.4 50.8 39.4


I
Figure 4.3 : Strength Development through Time for Laboratory Mixed Concrete.
Cement Types Shown Represent the Five Main ASTM Types.
50 -I STRENG TH DEVELOPM t ENT OVER TIM E FOR DIFFER ENT CEMENT TY PES
An . * * m i m ^ *
X
5 35 2 111 I Type 15 Type 25 A Type 31 Type 43A tA Type 51
<0 SH 25 - 1
<75 V) lu on . sj
K 20 Q. 1 15 f. . 1
o in . w X
5 -

7-Day 28 Day AG 90 -Day e!of concre 1 Year TE 2 Year
41


compressive strength (47.02 N/mm^) developed by type 21, was only 32%
greater than the minimum (35.5 N/mm^), developed by type 11.
(2) Type 15 cement started out with a reasonably good strength at age 7 days (20.2
N/mm^), however, it ended up one of the lowest in strength (35.85 N/mm^) at age
2 years. Table 4.4 shows that this cement type gained about 26% additional
strength in average between age 28 days and age 2 years.
(3) Type 25 cement (an ASTM type 2) started at a lower rate of strength development
(the second lowest in Figure 4.3) at age 7 days, and ended up one of the highest at
age 2 years. The same observation is confirmed in Table 4.4 where the average
compressive strength for ASTM type 2 cement improved from the second lowest
to one of the highest between age 7 days and 2 years. During this period, this
cement type gained about 45% in average compressive strength (from 29.32
N/mm^ to 42.73 N/mm^).
(4) Type 31 cement had the highest rate of strength development up to 28 days after
which it started to exhibit a low rate of strength gain. Table 4.4 indicates that the
average strength for ASTM type 3 cement increased between age 28 days and age
2 years by only 13% (from 34 to 38.47 N/mm^).
(5) Type 43A cement started with the least compressive strength of all cement types
at age 7 days. At age 2 years, type 43 A was one of the highest in compressive
strength. This observation is confirmed in Table 4.4. The most important
42


observation about ASTM type 4 is that between age 28 days and 2 years, it gained
about 75% more compressive strength on average (25.13 N/mrn^ to 43.9 N/mm^).
(6) Type 51 cement also started at a relatively low strength (16.48 N/mm^) at age 7
days and ended one of the highest (44.13 N/mm^) at age 2 years. Table 4.4
indicates that this cement type gained 55% more strength on average between age
7 days and age 2 years.
4.6 Summary of Conclusions
(1) Effect of cement type on concrete compressive strength is greatly reduced in the
long term. Approximately 280% difference in compressive strength between the
maximum and the minimum strengths at 7 days is reduced to 32% at age 2 years.
(2) ASTM type 4 cement that has been used in the construction of much of the
nations essential infrastructure gained about 75% in average compressive
strength between age 28 days and 2 years.
(3) ASTM type 4 cement gained on average about 14% more than type 3,15.7%
more than type 1, and no significant increase over types 2 and 5. The improved
long term performance of type 4 is due to the relatively high percentage C2S as
shown in Table 4.4.
(4) When the cement meets the recommended standards for degree of fineness, the
amount of C3S becomes an important factor governing the short term strength up
to the 28 day strength. This effect extends to some age between 28 days and 90
43


days. The higher the amount of C3S the higher the short term compressive
strength. Since the sum of C3S and C2S percent is between 70% and 80% by
weight of cement, the higher the percentage of C2S, the lower the short term
compressive strength. A higher percentage of C2S happens at the expense of
C3S, and hence adversely affects the short term strength (see Figures 4.1 and 4.2).
(5) The higher the degree of fineness of cement, the higher the short term strength.
More appropriately, the combination of a high degree of fineness and a high
percentage of C3S produces higher short term strength.
This is evident when comparing cement type 15 to cement type 33. Type 15 has a
moderate degree of fineness (325.5 m^/kg) and the highest percent of C3S (67%)
of all cement types in Table 4.4; type33 has a relatively high degree of fineness
(503 m^/kg) and percent of C3S (56.9%). Type 33 cement developed 40% more
strength than type 15 in 7 days.
(6) When the cement meets the standards for the degree of fineness, the higher the
amount of C2S the higher the long term strength. A very high degree of fineness
does not substantially affect the long term strength. The reason being that as long
as the cement is sufficiently fine to keep the hydration reaction active, strength
development will continue in the presence of sufficient amount of water and
reasonable environmental conditions. This is the situation with type 4 cements
that did not need to have a very high degree of fineness to develop the observed
44


relatively high long term strength. Similar to the discussion in (3) above, in
general, the higher the amount of C3S the lower the long term strength. High
amounts of C3S happens at the expense of the amount of C2S.
45


5. Time Effect of Alkali-Aggregate Reaction on the Performance of Concrete
The case study parapet wall, which was constructed in the 1940s, consists of 104
panels. Each panel is approximately 2.74m long by 1.22m tall by 0.41m thick as
shown in Figure 2.2a and Figure 2.2b. Data presented in this chapter are obtained
from testing of the cements, aggregate, and concretes used in the construction of the
parapet wall at the Green Mountain Dam.
5.1 Introduction
The alkali-aggregate reaction phenomenon discussed in this chapter is the cause
of many failures of concrete structures built during the late 1920s to early 1940s.
These failures were the result of overall cracking throughout the concrete manifested
at the surface as extensive map cracking or pattern cracking. The map or pattern
cracking is frequently accompanied by gel exuding or weeping from cracks or surface
popouts and spalling. Such problems were confined mostly to specific regions of the
country, including Kansas, Nebraska, Alabama, and Georgia (Mindess and Young,
1981).
In the United Kingdom (UK), alkali-aggregate reaction was not identified until
1976. A number of cases have been reported in the Southwest of England and the
Trent Valley. In these reported cases where alkali-aggregate reaction has caused
cracking, persistent dampness had been observed along the edges of cracks together
with discoloration and signs of expansion of concrete. In these reported
46


cases, no spalling has been reported. In the UK, the risk of cracking due to alkali-
aggregate reaction is minimized by limiting the alkali content to 3 Kg/m3.
(Non-Structural Cracks in Concrete, 1992).
5.2 Objective
The primary objective of this chapter is to determine the long time effect of the
alkali-aggregate reaction on the durability and strength of concrete. The concrete
parapet wall which is now more than 50 years represents the best example of the such
effect. Other objectives include the determination of the potential for reactivity with
alkalis of some mineral types.
5.3 Effect of Alkali-aggregate Reaction in Contrast with the Other Chemical and
Physical Characteristics of Cement
The effect of variation in physical and chemical characteristics of cements on
early compressive strength was discussed in Chapter 4. In this chapter, the variation
of compressive strength of parapet wall panels in the first two years and the recent 53
year old compressive strengths are related to the amounts of equivalent alkalis. Also
discussed in this chapter is effect of alkali-aggregate reaction on the general trends of
the different concrete types observed in Chapter 4.
5.4 Testing Related to the Chapter
The objective of this chapter is to investigate the development and long time
effect of alkali-aggregate reaction and its impact on concrete properties, in particular
compressive strength. This chapter also discusses some suspected aggregate
47


constituents with regard to their contribution to alkali-aggregate reaction. The
conclusions derived reflect the current condition of the concrete after more than 50
years. The following sections briefly introduce tests conducted in 1943,1946, and
1997.
5.4.1 Preview of Petrographic Analysis
Natural sand and gravel aggregate were used in the construction of the parapet
wall. They were obtained from a railroad deposit at Kremmling, Colorado (Douglass
et.al., 1947). Petrographic examination of the aggregate determines its composition
and identifies its mineral constituents. It is important to determine the amount and
type of reactive silica in the aggregate used in construction if the cement used
contains harmful amounts of equivalent alkalis. Some types of silica are more
reactive than the others and the presence of a pessimum quantity of (see Appendix A)
reactive silica type can possibly cause initiation of alkali-aggregate reaction. A 1940s
petrographic examination revealed the mineral constituents of the aggregate, the
amounts of these minerals, and their geologic descriptions. As will be discussed later,
these aggregates were suspected in regard to their reactivity with alkalis after the wall
was constructed. Therefore, included in this chapter are laboratory test results
performed on some of the minerals present in the aggregates used in the construction
of the wall. The purpose of the testing is to examine the reactivity of these
components with alkalis present in cement.
48


5.4.2 Preview of Visual Inspection
Visual inspection is one of the most important and effective methods to observe
signs of distress (Pepper, 1988). The wall was visually inspected several times
between 1943 and 1946. The results of visual inspection in the first three years of the
age of concrete show pattern cracking in panels such as 43-1 can be attributed to
alkali-aggregate reaction. Most panels remained in good condition. The wall was
reportedly inspected two times, after 10 and 25 years, between 1946 and 1997.
Appendix B includes visual inspection results of the panel inspected in 1997 together
with the inspection results of 1943 and 1946 for the same panels.
5.4.3 Preview of Mechanical Testing
Compressive strength of concrete is an important design property that varies with
time. Most concretes, regardless of cement type, gain more strength with time.
Alkali-aggregate reaction results in disintegration within the internal structure of
concrete induced by cracking of aggregate and cement matrix. In order to observe the
variation of concrete compressive strength through time, similar concretes are tested
at different periods of time. Each concrete type (characterized by cement type) was
made with the same aggregate type and mix proportions and exposed to
approximately similar curing and environmental conditions. The cements used varied
in chemical and physical characteristics including variations in alkali content. The
test cylinders (152.4x304.8mm) were made representing each concrete type and
tested at different time intervals within the first two years after the construction of the
49


wall was completed. In June 1997 cores were procured from the parapet wall of the
same size as the cylinders tested within the first two years. Test results that show the
variation in compressive strength within the first two years (obtained from Douglass,
1943) in addition to the compressive strength of cores after more than 50 years are
presented and analyzed in this chapter.
5.5 Theoretical Background
Alkali-aggregate reaction is a chemical reaction that takes place between alkalis
contained in cement paste and certain reactive forms of minerals within the aggregate.
Therefore, deterioration due to alkali-aggregate reaction is anticipated when the
cement used contains high amount of equivalent alkalis and the aggregate used
contains some form of reactive silica. ASTM defines a low-alkali cement as having
equivalent alkalis less than 0.6% (ASTM Standard C150,1996). This limit on
equivalent alkalis is specified when cement is to be used with aggregate that may be
harmfully reactive. Experiments have shown that expansions do not occur in cements
with equivalent alkalis below 0.6% (Mindess and Young, 1981).
Both type of reactive silica and amount of reactive silica influence amount of
deterioration. Different types of silica have different degrees of reactivity. Opal is
the most reactive form of natural silica. Other forms of reactive component such as
quartz and silica glass have also been identified as being reactive. The percentage of
reactive silica in aggregate is also important. The amount of reactive silica, known as
the pessimum percentage, that causes the maximum expansion of the concrete,
50


depends on the type of reactive silica in the aggregate used. Most of the failures that
occurred in the 1930s and 1940s occurred within 1 to 10 years, probably involving
opaline rocks. In some structures, however, severe deterioration did not occur until
after 15 to 20 years, which suggests that less reactive silicas were involved (Mindess
and Young, 1981).
5.5.1 Mechanism of the Alkali-aggregate Reaction
High alkali-cement can increase the solubility of amorphous silica and the rate at
which it dissolves. This initial hydrolysis opens up the structure and allows pore
fluid, which is essentially sodium or potassium hydroxide, to further hydrolyze the
silica fraction to form an alkali-silica gel. The formation of alkali-silica gel is given
by (Mindess and Young, 1981):
S + N (K) H ---------------> N (K) S H (5.1)
Aggregate pore fluid alkali-silica gel
The alkali-silica gel formed in Equation (5.1) results in weakening of the aggregate
particle. However, the gel has the ability to absorb considerable amounts of water,
which is accompanied by a volume expansion. The resulting expansion induces
internal stresses. Depending on the extent of the reaction, the resulting stresses may
cause cracking of the cement paste and the aggregate that has been weakened by the
absorbed water. Further absorption of water after significant cracking has occurred
changes the formed solid gel into a solution. The aggregates may then be surrounded
with calcium-alkali-silica gel formed by a reaction with calcium hydroxide present in
51


the cement paste. Calcium hydroxide is present in the cement paste as a byproduct of
hydration of dicalcium silicate and tricalium silicate.
The alkali-aggregate reaction described above occurs only at locations of reactive
silica particles. The alkali metal ions present in a solution runs to the various
locations of reactive silica particles where they react with these particles to form the
disruptive alkali-silica-gel. The amount of alkali-silica-gel increases with the number
of locations of reactive silica particles. When the number of reactive silica particles
is large compared to the available alkali silica ions, reaction at a location of a reactive
silica particle will not complete. Therefore, any amount of reactive silica beyond the
Pessimum Percentage will not react and will have no adverse effect on the integrity
of concrete i.e. percentage of reactive silica higher than the pessimum percentage will
not cause further expansion of the concrete (Mindess & Young, 1981).
The size of reactive silica particle influences the consequences of the alkali-
aggregate reaction. A smaller particle size results in expansion due to absorption of
water at that location but it will not be sufficient to cause stresses leading to cracking
of concrete.
5.6 Presentation and Analysis of Test Results
The following sections analyze in detail the petrographic examination results,
visual inspection, and mechanical testing results.
52


5.6.1 Petrographic Examination and Alkali Reactivity of Aggregate
During the 1940s petrographic study, the aggregates used in the construction of
the parapet wall were examined with hand lenses and segregated into consistent
lithological groups. A detailed geologic description is included in the Douglass
report (Douglass et. al., 1947), and is summarized in Table 5.1. Some of the groups
were studied in detail by means of thin sections and a petrographic microscope. The
coarse sands were examined under wide field microscope; the fine sands were studied
in detail by means of grain mounts under the petrographic microscope. No
segregation of the constituents of the sand was attempted by the petrographer. The
results (given as percentage of rock type) of the petrographic analyses of different
size ranges are shown in Table 5.1. The size ranges covered are typical to the size
ranges of aggregate used in the construction of the wall. The segregations designated
as diorites and diorite porphyries, andesites, thyolites, tuffs, and chalcedonic rocks are
lithologically similar to rock types which previous field and laboratory experience has
shown to be susceptible to attack by alkalis in cement (Douglass et. al., 1947). Table
5.1 shows that tuffs and chalcedonic rocks comprise a very small proportion of the
gravel, but the other suspected rock types are significantly abundant. Therefore, the
segregations made during the petrographic analyses could be crushed down and used
as aggregates in a mortar bar testing program with high-alkali and low-alkali cements.
The segregation designated as andesites contain two types of andesites, one closely
related to the albitized diorite porphyries but containing a smaller proportion of
53


Table 5.1. Petrographic Analysis of Aggregate. Percentage by weight. (Douglass et. al., 1947)
Rock type 1-1/2" 3/4" 3/4" 3/8" 3/8" No. 4 Description of Rock types
Granitic Rock 36 36.8 28.4 Slightly to moderate weathered, fine to coarse grained granites.
Gneisses & schists 39.4 42.6 48.7 Slightly to moderately weathered, fine to coarse grained, some quartzites.
Diorites and Diorite porphyries 10 6.8 13.7 Slightly to moderately weathered, fine-grained to aphanitic massive to porphyritic.
Rhylites 1.1 1.5 0.9 Slightly to moderately weathered, aphanitic, gray to buff, massive to porphyritic.
Andesites 10.2 8 4.4 Slightly to moderately weathered, aphanitic, gray to buff, porphyritic.
Basalts 2.1 2.7 1 Slightly to moderately weathered, aphanitic, black, dense to vesicular.
Tuffs 0.3 0.2 0.3 Soft, porous, massive, light-colored, acidic.
Sandstoness 0.8 1.2 1.6 Hard, quartzose, massive, fine to medium grained.
Shales - - 0.8 soft, platy, dark buff, silty shales.
Coal 0.1 - - Soft, friable, black, massive.
Chalcedonic rocks - 0.2 0.2 Very hard, massive, silicified volcanics.


phenocrysts and being considerably finer in grain. This type may be classified as
andesite porthyries. The second type of andesite is gray to buff in color. In order to
test alkali reactivity, the expansion potential of aggregates, segregations of andesites,
diorites and diorite porphyries, and rhyolites were made during the examination of the
coarse aggregate.
These minerals were crushed to make sand from which, together with crushed
quartz, bars (2.5x2.5x25 cm) were fabricated with high alkali (#2742) and low alkali
(#2735) cements (see Table 5.2 for chemical constituents of these cements). The
segregations were used in various percentages with the crushed quartz in order to
make sure that they were included in the pessimum quantity. The bars were cured
at 100 Fahrenheit in sealed containers. As portrayed in Figure 5.1, all of these
segregated materials expanded with high alkali cement, in all percentages. A
pessimum quantity was indicated with diorite and rhyolite. As shown in Figure 5.1,
rhyolite was not tested for reaction with low-alkali cement as there was a limited
amount after it has been tested for reaction with high-alkali-cements. However
neither andesite nor diorite and diorite porphyries expanded with low-alkali cement.
Further testing for reactivity of the wall aggregate was performed on 5x5x25 cm
bars. Table 5.3 shows the expansion test results. Three different combinations were
used, 1) sand with natural grading, 2) gravel crushed to sand sizes and used in straight
grading, and 3) a 37% sand with 63% crushed gravel mixture. All three aggregate
combinations were used with five different cement types, namely, high-alkali
55


Table 5.2. Chemical Analysis of Cements (Douglass et.al., 1947)
Lab. No. Cement Type Si02 AI2O3 Fe2Q3 CaO MgO Na2Q K2O Equivalent Alkalis
percent
2742 II 21.62 4.86 4.75 62.71 1.58 1.3 0.12 1.38
2735 II 22 4.52 3.81 64.78 1.3 0.04 0.14 0.13
2754 II 20.86 6.47 5.39 63.21 1.28 0.22 0.67 0.66
MS V 21.84. 4.42 4.41. 63.65 1 0.28 0.64 0.7
4500 II 21.65 5.12 5.05 64.24 0.75 0.13 0.5 0.43
- Alkali reactivity of the wail aggregates.
- Cement types used to examine the reaction of aggregate segregations.


Figure 5.1. Expansion of Selected Materials from the GMD Aggregate with High
and Low Equivalent Alkalis (from Dougalss et. al., 1947)
57


Table 5.3. Percent Epansion of Parapet Wall Aggregate with Various Cement Types. (Douglass et. al., 1947)
Agg. Cement Lab. No. Na20+ 0.658K2O Age
Day Months
7 1 2 3 4 5 6 8 12 18 24 30 36 42
2742 1.42 0 0.013 0.02 0.044 0.069 0.097 0.121 0.144 0.183 0.238 0.288 0.312 0.341 0.359
2735 0.18 0 0 -0 -0 -0 0 0 -0 0.001 0.006 0.018 0.011 0.017 0.011
Sand 2754 0.89 0 -0 0.004 0.005 0.007 0.012 0.014 0.014 0.025 0.046 0.038 0.061 0.059 0.066
MS 0.92 0 -0.01 -0 -0 0.001 0.003 0.003 0.002 0.003 0.007 0.001 0.015 0.012 0.014
4500 0.63 0 -0.01 -0 -0 0.002 0.003 0.003 0.002 0.002 0.004 0.001 0.011 0.01 0.011
2742 1.42 0 0.004 0.017 0.028 0.047 0.07 0.094 0.117 0.148 0.184 0.19 0.22 0.221 0.221
Gravel 2735 0.18 0 -0.01 -0.01 -0.02 -0.01 -0.01 -0.01 -0.01 -0.04 -0.01 0.005 0.003 0.004 0.002
1-1/2" 2754 0.89 0 -0.01 -0.01 -0.01 0 0 0.004 -0 -0.03 0.008 0.023 0.028 0.042 0.048
max. MS 0.92 0 -0.01 -0.01 -0 -0 -0 -0 -0.01 -0.01 -0 0.005 0.008 0.014 0.014
4500 0.63 0 -0.01 -0.01 -0 -0 -0.01 -0.01 -0.01 -0.01 -0.01 0.009 0.013 0.018 0.015
37% 2742 1.42 0 0.011 -0.01 0.022 0.033 0.048 0.065 0.077 0.119 0.172 0.197 0.242 0.276 0.279
sand 2735 0.18 0 -0.01 -0.03 -0.01 -0.01 -0.01 -0.01 -0.01 -0.01 0.001 0.009 0.009 0.012 0.006
+ 2754 0.89 0 -0 0.004 0.005 0.01 0.009 0.012 0.014 0.01 0.021 0.032 0.031 0.038 0.041
63% MS 0.92 0 -0 -0 -0 0.003 0.002 0.002 0.001 -0 0.007 0.008 0.002 0.007 0.01
gravel 4500 0.63 0 0.001 0.004 0.005 0.008 0.006 0.006 0.005 0.001 0.012 0.02 0.013 0.017 0.014
5x5x25 cm bars, 1:2 Mortar, Sealed, Moist Cured at 100 Degrees Fahreheit.


(#2742), low-alkali (#2735), medium high-alkali (#2754), a sulfate resisting type of
high-alkali content, and a reasonably low alkali (#4500).
Table 5.3 shows that the sand with high-alkali cement had expanded 0.121% at 6
months and the expansion reached 0.359% at 42 months and the length increase was
continuing. The sand with medium-high-alkali cement expanded by 0.066% at 42
months and the expansion was also continuing. The other two aggregate
combinations showed the same trend in expansion but were less significant than the
sand.
The above tests on the wall aggregate, performed in 1947, led to the tentative
conclusion that these aggregates were reactive. The impact of their reactivity is
studied in conjunction with the visual inspection results as well as the time variation
in compressive strength as shown in the following sections. It should be noted that
the above indications have resulted from accelerated tests only, that is, when the
specimens have been cured at elevated temperature (100 Fahrenheit) under sealed
moist storage. An additional experiment was performed with the 28 cements in
which 7.5x7.5x38 cm bars were stored in moist cabinets held at 70 Fahrenheit for a
period of 36 months. No significant expansions were registered with any of the 28
cements.
5.6.2 Visual Inspection
Table 5.4 shows three of the visual inspection results of 11 selected panels. The
visual inspections of more panels are included in Appendix B. The inspections were
59


Table 5.4. Visual Inspection of Selected Panels (1943 and 1946 Inspections from Douglass et. al., 1947)
Panel No. 1943 Inspection 1946 Inspection 1997 Inspection
14-1 Nothing to report. Horizontal crack eight inches long at conduit level*, north end. Considerable tight branch- cracking; rather coarse spacing. Noticeable pattern cracking covering most area which decreases in width and intensity from top to bottom. The crack at conduit level now extends full panel width. Maximum measured crack width is 0.229 mm located at south end of crack at conduit level.
16-4 Entire area thinly covered with white granular deposit. Horizontal crack at conduit over full length except south quarter; maximum opening 0.051 mm. Moderate amount of tight branch cracking. Some barely visible cracks appeared in upper half of the panel. Maximum crack width is 0.076 mm. Horizontal crack at conduit level runs across the full width.
18-1 Nothing to report. No change. Some efflorescence on upper half at limited locations. Some delamination has started at limited locations.
21-1 Scar or spall 25.4 mm diam- meter 25.4 cm from north end. Slight raveling at edges of blockout in pilaster; no comer cracks. Horizontal crack 30.5 cm long at conduit on south end. Blockout has short crack from bottom south corner, and a horizontal crack 15.24 cm from top extending through chamfer. Rather severe raveling at top and bottom of blockout. Minor crazing at limited locations. Horizontal crack at conduit level is barely visible (1.07 m long). Slight, barely visible cracks at limited locations.
24-4 Nothing to report except very small deposit of carbonate just below corner crack. Horizontal crack at conduit over north two- thirds of panel; maximum opening 0.051 mm mm. Carbonate deposit has disappeared. Extensive wide pattern cracking covers entire area, maximum crack width has exceeded 15.2 mm. Panel is fully covered with orange discoloration and many cracks are covered with efflorescence.
* Electrical conduit that provides power to the wall for lighting and runs horizontally across the wall panels. The conduit is located approximately
2 feet from top of the wall.


Table S.4. (Continued)
Panel No. 1943 Inspection 1946 Inspection 1997 Inspection
34-4 Nothing to report. Slight branch cracking barely visible. visible map cracks over most of the area accompanied with discoloration and/or exudation.
41-1 Nothing to report. Very small amount of branch cracking. No change except for conspicuous pattern cracks at left triangular zone. Max. crack width is 0.178 mm.
42-1 Numerous small spots and streaks of light colored deposit showing no grain structure at 30X magnification. Branch cracking in top 15.2 cm over entire area. Light colored deposit still evident. A small spall at top near south end. Same as previous inspection. Cracking in top 15.2 cm. is barely visible.
43-1 Nothing to report. Coarse pattern conspicuous over entire area; average spacing about 15.2 cm, maximum opening 0.076 mm. Horizontal crack at conduit full length, maximum 0.076 mm. Extensive coarse pattern cracking over entire area, maximum measured crack width increased to 15.2 mm. Horizontal crack extends along the full length of the panel. Extensive orange discoloration over entire area Some efflorescence visible along cracks.
43A-1 Appearance is mottled by large light and dark areas; no visible deposit. Horizontal crack at conduit 20.32 cm long, north end. Horizontal crack at conduit level is now 8.89 cm long with light exudation along crack. Some crazing in upper left triangular zone with maximum crack width of 0.229 mm.
51-4 Faintly mottled by lighter spots about 25.4 mm diameter. No change. No change except the appearance of horizontal crack at conduit level, approximately 1.22 m long; maximum crack width is 1.016 mm.


performed in October, 1943; June, 1946; and June, 1997. The panels were selected
on the basis of the alkali content to cover three ranges, namely, low, medium-high,
and high alkali cements. Low-alkali cement contains equivalent alkalis less than
0.6% as described in ASTM Cl 50. For the purpose of this study, cements that
contain total equivalent alkalis greater than or equal to 0.6% and less than or equal to
1.0% are classified as medium-high-alkali cements. Cements containing equivalent
alkalis greater than 1.0% are considered high alkali cements. For most panels,
remarkable changes took place on the downstream face of the panel, therefore only
changes in the downstream face of the wall are reported here. The visual inspection
results of selected panels are discussed in the following sections. The designations
for these panels as well as the visual inspection terminology are shown in Appendix
A.
Panel number 14-1: This panel was constructed using cement type 14 that can be
classified as ASTM type I cement. Cement type 14 contained medium-high
equivalent alkalis (0.96%). The development of alkali-aggregate reaction is well
demonstrated with this cement type. The wall was first inspected in October, 1943
(i.e. five months after the construction of the wall was completed). There was no
visible sign of alkali-aggregate reaction and this panel was in excellent condition. In
1946 signs of distress were first observed but the overall condition of the wall was
relatively good. In 1997, the reaction caused significant expansion manifested with
conspicuous (see Appendix A, number b.6) pattern cracking covering most area.
62


Panel number 16-4: This panel was constructed using cement type 16 that can be
classified as ASTM type I cement. Cement type 16 contained low equivalent alkalis
(0.57%). No significant distress was observed in the three inspections shown in
Table 5.4.
Panel number 18-1: This panel was constructed using cement type 18 that can be
classified as ASTM type I cement. The cement contained low equivalent alkalis
(0.28%). As anticipated for a low alkali-cement, no signs of distress in the first three
years of the age of concrete were observed. In 1997, the amount of distress was still
minimal and could probably be due to other causes of distress such as freeze-thaw or
other environmental conditions. No pattern cracking was observed in any of the three
inspections documented in Table 5.4.
Panel number 21-1: This panel was constructed with cement type 21 that can be
classified as type ASTM type II cement. Cement type 21 contained a low-equivalent
alkalis amount of 0.58%. Although some signs of distress had been documented over
the prior 53 years, no pattern cracking or discoloration has been reported as shown in
Table 5.4.
Panel number 24-4: This panel was constructed using cement type 24 that can be
classified as an ASTM type II cement. Cement type 24 contained a medium-high
equivalent alkalis amount of 0.94%. No significant signs of distress were observed
within the first three years of the wall age. In 1997 this panel became one of the
worst in overall condition. Significant amounts of silica gel were produced by the
63


reaction and caused serious cracking that exceeded 0,6 inches. Pattern cracking,
discoloration, and efflorescence represent the most apparent signs of distress as
shown in Figure 5.2. Although the amount of equivalent alkalis in this cement type
was slightly less than that present in type 14-1, panel 24-4 was much more distressed
than panel 14-1.
Panel number 34-4: This panel was constructed using type 34 cement that can be
classified as ASTM type 3 cement. Cement type 34 contained a low amount of
equivalent alkalis 0.52%. The cement type showed minimal signs of distress within
the first three years of the age of concrete. Although the overall condition of this
panel was much better than panels 14-1 and 24-4, this panel was distressed in the
form of pattern cracking and some discoloration. The reason for distress is probably
alkali-aggregate reaction. If cause of distress is truly alkali-aggregate reaction then it
could be inferred that even with low alkali-cement cements the likelihood for distress
due to alkali-aggregate reaction exists. The variation in conditions of exposure to
environmental conditions greatly affects the reaction for example, sun exposure.
The relative orientation of the panels results in variation in the duration and intensity
of exposure to different environmental factors. The stoichiometry of the reaction is
greatly affected by these variations (e.g., variation in temperature from one panel to
another).
Panel number 41-1: This panel was constructed using cement type 41 that can be
classified as ASTM type IV. This cement type contained medium-high equivalent
64


Figure 5.2. Panel Number 24-4 (Notice efflorescence, pattern cracking, and change of color)


alkalis (0.9%). This panel remained in generally good condition during all three
inspections. The 1997 inspections showed conspicuous cracking limited to the upper
left triangular zone. The overall appearance of this panel does not show substantial
distress due to alkali-aggregate reaction. This medium-high alkali cement behaved
well compared to the other medium high alkali cements.
Panel number 42-1: This panel was made with ASTM type IV cement. This
cement contained low equivalent alkalis (0.38%). Additional branch cracking had
been appeared between 1943 and 1946. The overall condition of the panel remained
good and stable since the 1946 inspection.
Panel number 43-1: This panel was constructed using cement type 43 that can be
classified as ASTM type IV cement. This cement contained a high-equivalent alkali
amount of 1.14%. The panel was in excellent condition according to the October,
1943 inspection. Similar to the cement types containing different amounts of
equivalent alkalis, the five month period was insufficient to produce any visible signs
of distress externally (internal distress would probably start earlier than external
distress). This is the only cement type in our selected panels that showed early signs
of distress at the age of three years. Pattern cracking was visible over entire area in
the 1946 inspection. In 1997, this panel became one of the worst in overall condition
as shown in Figure 5.3. Deterioration is manifested with extensive pattern cracking,
orange discoloration, and efflorescence. The maximum measured crack width of
66


Figure 5.3. Panel Number 43-1 (Notice efflorescence, pattern cracking, and change of color)


0.6 inches is also significant. Such a large amount of equivalent alkalis causes early
distress of concrete.
Panel number 43A-1: This panel was constructed using cement type 43A that
can be classified as ASTM type IV cement. Cement type 43A contained low
equivalent alkalis (0.44%). The general chemical and physical characteristics of this
cement type are similar to type 43. The main difference between type 43 and type
43A cements is the amount of equivalent alkalis. Type 43 cement is a high-alkali
cement while type 43A cement is a low-alkali cement. Type 43A behaved much
better than type 43. Some distress was observed for type 43 A but the overall
condition was good. No pattern cracking or discoloration was reported through the
53-year life span of the panel. One can see difference in the overall behavior by
comparing Figure 5.3 to Figure 5.4.
Panel number 51-4: This panel was constructed using cement type 51 that can be
classified as ASTM type V cement. Cement type 51 contained low equivalent alkalis
(0.26%). This cement type behaved very well. No evidence of distress due to alkali-
aggregate reaction and the overall condition was good throughout the life of the wall.
5.6.3 Mechanical Testing (compressive strength)
In June 1997, cores were obtained from the 11 selected panels previously visually
inspected. Five cores were obtained from most of the selected panels and three or
four cores from other panels. The test types include uniaxial compression, dynamic
loading, and split tension. In this chapter, emphasis will be on the variation of
68


Figure 5.4. Panel Number 43A-1 (No visible pattern cracking, discoloration, or efflorescence).


compressive strength through time and the effect alkali-aggregate reaction on
deterioration of compressive strength. The deterioration in compressive strength
caused by alkali-aggregate reaction is due to cracking of aggregate and the
surrounding cement paste matrix.
The variation in compressive strength in the first two years of wall age was
observed by testing cylinders made and cured in the laboratory. Table 5.5 shows the
variation in compressive strength of the 11 selected cement types through time. The
variation in chemical and/or physical characteristics within each ASTM group is
indicated by subclassifications within the group as shown in Appendix A.
Cement type I: Panel number 14-1 was built with a medium high-alkali cement.
The exterior appearance was affected by alkali aggregate reaction as discussed in the
visual inspection section. But the strength of this panel increased continuously with
time as shown in Table 5.5. The alkali-aggregate reaction did not result in internal
disintegration sufficient to affect the long time strength. All ASTM type 1 cements
(type 14, type 16 and type 18) increased in strength continuously over the past 53
years. Cement types 16 and 18 contained low equivalent alkalis. However, it is
worthy of note that panel 14-1 was the lowest in strength among panels constructed
with type 1 cements. This is probably due to some internal disintegration caused by
alkali-aggregate reaction. Cement type 16 contained equivalent alkalis very close to
the ASTM limit for low equivalent alkalis of 0.6%. Cement type 16 behaved very
well as far as appearance and strength. Figure 5.5 shows that for cement type I all
70


Table 5.5. Compressive Strength of Concrete at Different Ages. (Age 7-days to age 2 years, from
Douglass et.al., 1947)
P.C.A Cement type Na2 + 0.658 K2O (percent) 7 Day (N/sq.mm) 28 Day (N/sq.mm) 90 Day (N/sq.mm) 1-Year (N/sq.mm) 2 Year (N/sq.mm) 53 Year (N/sq.mm)
14 0.96% 20.69 29.23 37.1 38.85 39.85 44.25
16 0.57% 19.03 30.06 36.75 37.51 36.41 47.82
18 0.28% 21.17 29.51 36.34 35.3 37.78 46.18
21 0.58% 17.51 30.27 38.89 44.61 47.02 48.21
24 0.94% 20.89 32.13 39.51 41.09 44.06 28.14
34 0.52% 25.44 32.96 35.72 38.89 38.41 35.12
41 0.90% 9.79 27.03 40.34 44.2 44.33 49.67
42 0.38% 7.45 20.41 36.06 41.23 41.99 41.62
43 1.14% 11.72 27.03 38.47 43.68 45.02 22.97
43A 0.44% 10.69 26.06 38.54 45.64 44.27 44.09
51 0.26% 16.48 28.48 36.34 42.2 44.13 42.21


Figure 5.5. Variation of Compressive Strength with Time for Cement Types I, II, & IV
VARIATION OF COMPRESSIVE STRENGTH FOR TYPE I CEMENTS
VARIATION OF COMPRESSIVE STRENGTH FOR TYPE II CEMENTS
60 -i
7-Day 28-Day 90 -Day 1-Year 2-Year 53-Year
TIME
i im{nofi<Â¥.
Equivalent
alkalis)
16 (0.57%
Equivalent
alkalis)
I 118(0.28%
Equivalent
alkalis)
--------Poly. (14
(0,96%
Equivalent
alkalis))
........Poly. (16
(0.57%
________-I
21 (0.58%
Equivalent
alkalis)
I 194(0 04%
Equivalent
alkalis)
---------Poly. (24
(0.94%
Equivalent
alkalis))
.......Poly.
(21(0.58%
Equivalent
alkalis))
VARIATION OF COMPRESSIVE STRENGTH FOR TYPE IV CEMENT
Day Day Year Year
TIME
i iai( o.9%
Equivalent
alkalis)
lmma&AO (0.38%
Equivalent
alkalis)
I. 143(1.14%
Equivalent
alkalis)
-Poly. (43
(1.14%
Equivalent
alkalis))
Poly. (42
72


panels gained strength continuously since 1943 till the time of this study. It is worthy
of note that this gain of strength includes cement type 14 which is medium-high alkali
cement according to the classification of this study. Cement type 14 is a high-alkali
cement base on ASTM classification.
Cement type II: Panel 24-4 gained strength continuously over the first two years
of its age. Due to alkali-aggregate reaction, this panel lost more than 36% of its
compressive strength between age 2 years and age 53 years. This is to be compared
to type 14 cement that contained approximately the same amount of equivalent alkalis
but was not significantly affected by alkali-aggregate reaction. Cement type 21
contained equivalent alkalis of 0.58% which is very close to the ASTM limit of 0.6%.
This cement type behaved very well as far as strength gain. Figure 5.5 shows the gain
of strength for cement type 21 and loss of strength for cement type 24 at age 53 years.
Cement type III: Panel number 34-4 gained strength in the first year of the age of
concrete and slightly dropped in strength at age 2 years and age 53 years. Cement
type 34 contained low equivalent alkalis. The slight reduction in compressive
strength (8.6%) between age 2 years and age 53 years is expected for ASTM type III
cements (high early strength).
Cement type IV: Panel number 43-1 gained strength continuously over the first
two years of its age. Nonetheless, this panel suffered the greatest deterioration in
compressive strength between age 2 years and age 53 years. The reduction in
strength was close to 50%. Cement type 43 contained high equivalent alkalis
73


(1.14%). The control to cement type 43 is cement 43A that is very similar to type 43
in chemical composition except that cement type 43 A contained low equivalent
alkalis (0.44%). Cement type 43 A gained strength through the first two years of its
age and maintained a reasonably high strength at age 53 years. The other cement
types that belong to ASTM type IV (cement type 41 and cement type 42) either
gained strength throughout the 53 years of their ages or remained close to the two
years compressive strength. Cement type 41 contained medium-high equivalent
alkalis of 0.9%. Nonetheless, this cement type remained in good condition and
gained high strength at age 53 years. This is compared to type 24 that also contained
medium high equivalent alkalis but was significantly distressed between age 2 years
and age 53 years. Figure 5.5 shows the continuous gain of strength for cement type
41 and cement type 42 and the loss of strength for cement type 43.
Cement type V: Panel 51-4 gained strength continuously over the first two years
and experienced a slight reduction in compressive strength at age 53 years. Cement
type 51 contained low-equivalent alkalis of 0.26% and therefore was not affected by
alkali-aggregate reaction.
5.7 Summary of Conclusions
Outlined below are the main findings of this chapter. These findings focus on the
types of reactive minerals, the long time effect on strength and appearance, and the
reaction of the minerals with different alkali contents.
74


1) Aggregate that contains Ryholites, Andesite, Diorite, and Diorite Porphyries can
react with cements that contain significant amounts of alkalis and this results in
significant loss of strength. Figure 5.1 shows that these materials expanded with
high alkali cements when mixed with quartz in various percentages. A pessimum
quantity was shown with Diorite and Diorite Porphyries. Reduction in
compressive strength of some wall panels has also been observed due to the
reaction of aggregate containing the above materials and high alkali cements.
Table 5.1 shows the amounts of these materials present in aggregates that caused
substantial reduction in long time strength.
2) A combination of reactive aggregates with high alkali cements results in
significant loss of strength in the long term. No visible signs of alkali-aggregate
reaction are observed within the first five months of the age of concrete. This
means that regardless of the amount of equivalent alkalis, a five-month period is
not sufficient for the alkali-aggregate reaction to cause visible signs of distress
externally. Only one cement type (cement type 43) contained high equivalent
alkalis based on the classification of this chapter. Deterioration in the form of
extensive pattern cracking alone can be seen in two years and may not be
accompanied at the two-year-mark by significant loss of strength as indicated by
panel number 43-1. However, a combination of efflorescence, orange
discoloration, and extensive pattern cracking will accompany significant loss of
strength (up to 50%) in the long term (53 years). Cement type 43 is an example
75


of high alkali cement (1,14% equivalent alkalis) that lost approximately 50% of
strength between age 2 years and age 53 years.
3) A combination of reactive aggregates with medium-high-alkali cements produced
a variety of results in this study. Three cement types contained medium-high-
alkalis (0.6 % to 1.0% equivalent alkalis), namely, type 14, type 24, and type 41
as shown in Table 5.5. These cements belonged to ASTM classification types I,
II, and IV. None of the three medium-high alkali cements experienced strength
reduction in the first two years of the wall age. Cement types 14 and 41 gained
strength continuously since 1943. Cement type 41 (0.9% equivalent alkalis)
ended up with the highest strength of all cements shown in Table 5.5 at the 1997
study. On the other hand, cement type 24 experienced a significant reduction in
strength (about 36%) at age 53 years. Visible signs of deterioration including
efflorescence, discoloration, and extensive pattern cracking had been observed in
panel 24-4 as shown in Table 5.4. It is worthy of note that all panels were
constructed with the same mix proportions and aggregate types. There is a
notable conflict among cements containing medium-high equivalent alkalis. Two
out of the three panels made with medium-high alkali cements gained strength
continuously since 1943 to the time of the 1997 study.
4) A combination of reactive aggregates with low-alkali cements did not result in
reduction of compressive strength. All low-alkali cements gained strength
continuously since 1943 as indicated in Table 5.5. Even the cements that
76


contained equivalent alkalis very close to the ASTM limit of 0.6% (type 16 and
type 21), gained strength continuously since 1943. The aggregate constituents in
(1) above do not produce deterioration in strength with cements containing
equivalent alkalis less than but very close to the ASTM limit.
5) The Douglass report includes a detailed visual inspection of all the wall panels.
The report which was prepared in March, 1947 after the last visual inspection
considers the overall condition of the wall as excellent except for some panels like
43-1. Medium-high alkali cements, such as type 24, did not show signs of
distress due to alkali-aggregate reaction in the first three years of the wall age.
High-alkali cements, such as type 43, showed early signs of distress at age 3
years. However these two cement types showed significant signs of visible
distress after 53 years. As long as the conditions are favorable for an alkali-
aggregate reaction to continue, loss of strength in the long term should be
anticipated. The harmful pessimum quantity of the aggregate should be
considered in conjunction with the time required for the alkali-aggregate reaction
to form sufficient amount of alkali-silica gel to cause significant expansion of the
concrete. The time required for the reaction to form harmful amounts of
expansive alkali-silica gel depends on the type of reactive silica, moisture content,
and environmental conditions. The petrographic and expansion tests outlined
previously showed that the parapet wall aggregates were reactive under
accelerated conditions. For most panels, no significant distress was observed
77


within the first three years of the age of concrete. In the 1997 inspection, some of
the panels that contained high equivalent alkalis suffered substantial cracking
(greater than 0.6 inches in panel 24-4). Finally, alkali-aggregate reaction is
manifested by efflorescence and change of concrete color to orange. In general
the panels constructed with low-alkali cements remained in relatively good
condition.
78


6. Factors Affecting Resistance of Concrete to Freeze-Thaw Damage
6.1 Introduction
The variation in cement type greatly affects the performance of concrete in terms
of durability and strength. Concretes made with different cement types are known to
behave differently under similar conditions such as environmental conditions, mix
proportions, etc. This chapter investigates the influence of various factors on freeze-
thaw resistance of concrete.
Alternate freezing and thawing can seriously affect the integrity of concrete
structures. It is therefore essential to study the factors that can contribute to freeze-
thaw damage. Degradation of concrete due to freeze-thaw generally continues
through time as concrete is subjected to continuous freeze-thaw cycles. Freeze-thaw
can affect the appearance of concrete structures due to the development of surface
cracks of small width. The development of these small cracks into larger cracks is a
function of many factors such as the stress condition of the structural member.
Freeze-thaw damage can cause substantial loss of strength as will be demonstrated in
this chapter. After substantial loss of strength below design values, a structure will
have to rely on other means to resist loading such as factors of safety. The
development of compressive strength in the first two years of the wall age was
presented in Chapter 4. The effect of time is revisited in this chapter to present its
effect of on the performance of concrete after more than 50 years.
79


6.2 Objective
The objective of this chapter is to present the effect of various factors such as air-
entraining of concrete and cement type on susceptibility of concrete to freeze-thaw
damage. The long time effect of freeze-thaw on compressive strength is presented
and related to factors affecting freeze-thaw action.
6.3 Testing Related to this Chapter
6.3.1 Overview of Freeze-thaw Testing
To determine the effect of cement type on freeze-thaw resistance of concrete,
standardized tests were performed on concretes made with different cement types in
thel940s. Table 6.1 shows the freeze-thaw tests performed on different concretes,
characterized by cement type. The freeze-thaw tests shown in Table 6.1 were
performed on 15.2 x 7.6 cm concrete cylinders. The objective of the test is to
determine the number of freeze-thaw cycles endured by each cement type.
6.3.2 Overview of Visual Inspection
The wall was visually inspected six times between 1943 and 1946. It has been
reported that the wall was also inspected at age 10 years and 25 years. In this chapter,
only the first and sixth inspections (between 1943 and 1946) are reported together
with the current inspection. The writers of Douglass report consider the overall
condition of the parapet wall as excellent after the sixth inspection (Douglass et. al.,
1946). No significant signs of distress due to freeze-thaw or due to other causes of
distress were evident in the 1946 inspection. The 1997
80


Table 6.1. Summary of Freeze-thaw Tests.
Number of Cycles Endured by 7.6x15.2 cm Cylinders at Different
Percents of Expansion (Douglass et.al., 1947)
Cement type Set Number (1)* Expansion Set Number (2) Expansion t*
0.30% 0.60% 0.30% 0.60%
11 81 113 22 32
12 84 119 27 39
13 52 92 14 20
14 64 93 27 35
15 34 67 16 24
16 44 78 23 32
17 50 80 29 34
18 18 28 21 30
0 46 90 58 72
21 64 87 56 67
22 50 80 28 37
23 83 157 25 35
24 44 87 15 21
25 62 88 25 34
31 29 46 12 18
33 21 36 17 24
34 50 91 5 12
41 55 86 19 25
42 57 115 33 42
43 83 115 48 67
43A 140 198 57 68
51 66 164 22 32
12T 400 400 230 -
16T 0.05% at 570 - 0.11% at 300 -
16B 0.18% at 500 - 260 -
21T 0.2% at 520 - 0.25% at 300 -
34B 0.07% at 550 - 230 -
42B 0.06% at 550 - 0.07% at 300 -
* Fog cured for 90 days before freeze-thaw testing
** Fog cured for 14 days and placed on laboratory roof for 1 year before
freeze-thaw testing________________________________________________
81


inspection revealed substantial signs of distress in various panels due to freeze-thaw.
Table 5.4 shows the visual inspection results of selected cement types. Appendix B
includes the complete record of the panels inspected in the current study.
6.3.3 Overview of Compressive Strength
Compressive strength is one of the most important design parameters for concrete
structures. Compressive strength is influenced by freeze-thaw in terms of its effect on
the integrity of the internal concrete structure. Table 6.2 shows the development of
compressive strength through time of different concrete types. The effect of freeze-
thaw on compressive strength of some of the wall panels is apparent and will be
discussed in detail later.
6.4 Theoretical Background
Air entraining of concrete is considered one of the most important methods of
enhancing the resistance of concrete to freeze-thaw damage. Many factors influence
the effective air-entraining of concrete such as the spacing factor (see Appendix A)
and the volume of entrained air. A small spacing factor will enhance the resistance of
concrete to frost action (Mindess and Young, 1981). The resistance of concrete to
freeze-thaw damage is considerably enhanced by air-entrainment in the range of 2%
to 6% (Water Resources Manual, 1988).
Different proposals have been introduced to explain the mechanism of damage
due to freeze-thaw. Mindess and Young (1981) discuss the mechanism of freeze-
thaw which is summarized below.
82


Table 6.2. Compressive Strength of Selected Panels, N/sq.mm
Cement Designation 7 -day 28 day 90 day 6 month 1 year 2 years 53 years 1997
16 19.03 30.06 36.75 37.58 37.51 36.41 47.64
16B 18.06 27.1 32.75 33.16 33.44 34.68 42
18 21.17 29.51 36.33 37.44 35.3 37.78 45.81
21 17.5 30.27 38.89 42.68 44.61 47.02 48.84
31 27.24 34.06 37.23 37.23 38.13 39.16 21.72
34 25.44 32.96 35.72 37.51 38.89 38.41 35.68
34B 22.27 28.62 32.2 34.2 34.82 34.68 41.54
42 7.45 20.41 36.06 38.75 41.23 42 41.33
42B 9.17 20.48 32.4 35.72 35.58 38.27 44.34
43A 10.69 26.06 38.54 40 45.64 44.27 44.5
51 16.48 28.48 36.34 36.89 42.2 44.13 43.69


The transformation of water from the liquid state to solid state is accompanied by
volume increase. Therefore, the transformation of water present in the pores of the
cement paste into ice will result in expansion of concrete. Some researchers consider
the main dilation in concrete to be the result of hydraulic pressure caused by
compression of residual water. The expansion of the ice previously mentioned causes
the water that did not transform into ice to experience pressure due to the decrease in
the available volume of pores. The capillaries containing the ice and unfrozen water
will then expand if the pressure on water is not relieved. This pressure will actually
occur in many capillaries resulting in a greater combined effect on the cement paste.
The cement paste can fail due to this stress state if its tensile strength is exceeded.
Air-entraining of concrete is therefore beneficial to resist damage caused by freeze-
thaw action by providing a short escape for the compressed water. The short escape
is determined by the spacing factor (Mindess and Young, 1981).
6.5 Presentation and Analysis of Test Results
The following sections Outline the testing performed at different periods to
investigate the parameters affecting freeze-thaw damage as well as its effect.
6.5.1 Environmental Conditions at the Parapet Wall Location
The environmental conditions at the GMD were severe in terms of freeze-thaw.
Figure 6.1 shows the monthly variation in temperature, monthly precipitation, and
reservoir water surface elevation in the first 3 years of the age of the wall. It is likely
that these conditions remained representative of the environmental conditions of the
84


85


GMD up until 1997. The monthly variation in temperature reflects the severity of
environmental conditions at the GMD location and is likely to produce deterioration
due to freeze-thaw or frost attack.
6.5.2 Freeze-thaw Testing
To determine the freeze-thaw resistance of concretes made with different cement
types, a standardized test was performed after the construction of the wall was
completed. The concrete samples tested were 7.6x15.2 cm cylinders with maximum
aggregate size of 19mm. Two sets of samples were prepared each representing the
cements under consideration. The first set of cores was fog cured at 70 for 90 days
then subjected to freeze-thaw tests. The second set was fog cured for 14 days then
placed on the laboratory roof at the USBR for one year prior to freeze thaw testing.
The test cylinders were equipped with gage points at their ends for the purpose of
observing the expansions. Table 6.1 shows the number of freeze-thaw cycles endured
by each concrete type to a specified failure point. A predetermined failure point is
characterized by a specific percentage expansion. Air-entrained concretes endured
more freeze-thaw cycles before the specified expansion limits were reached. As
shown in Table 6.1, the test on air-entrained concretes was not completed to the
predetermined failure point since their superior performance was well established by
the very expansion at high number of freeze-thaw cycles. The cements identified by
the letter T were air-entrained by intergrinding with Vinsol Resin air-entraining
agent. The cements identified by the letter B were treated by adding the Vinsol
86


Resin in solution form to the concrete (see appendix A for definitions and
terminology). To compare the difference between the air-entrainment methods in
terms of resistance to freeze-thaw cycles, the performance of cement type 16 can be
evaluated. Cement type 16T expanded by 0.05% after 570 cycles of freeze-thaw,
where as cement type 16B expanded by 0.18% after only 500 cycles of freeze-thaw.
This means that the cement air-entrained in solution form expanded 2.6 times more
than the cement air-entrained by intergrinding the air-entraining agent in a fewer
number of freeze-thaw cycles.
The average number of cycles endured by cement types II, IV, and V is more than
the average number of cycles endured by cement types I and III for all degrees of
expansion. When exposed to environment conditions after 14 days then subjected to
freeze-thaw tests, cement type III endured the least number of freeze-thaw cycles of
all cement types. That is to say cement type III is suspected as to performing poorly
with respect to its resistance to freeze-thaw effects.
6.5.3 Visual Inspection
Deterioration observed through visual inspection is detailed in Table 5.4, with a
complete listing in Appendix B. Observed deterioration is discussed below as are the
results of the two sets of freeze-thaw tests performed in the 1940s.
Panels 16-4 and 16B-4: Deterioration is limited to barely visible cracking and
the horizontal crack caused by the presence of the power conduit. No pattern
cracking is observed. The overall appearance of panel 16B-4 is better than panel 16-
87