ASSESSING THE DEVELOPMENT OF LOCALIZED DAMAGE IN

billowycookieΠολεοδομικά Έργα

29 Νοε 2013 (πριν από 3 χρόνια και 9 μήνες)

246 εμφανίσεις



ASSESSING THE DEVELOPMENT OF LOCALIZED DAMAGE IN
CONCRETE UNDER COMPRESSIVE LOADING
USING ACOUSTIC EMISSION



A Thesis
Submitted to the Faculty
of
Purdue University
By
Sunil Puri

In Partial Fulfillment of the
Requirements for the Degree
of
Master of Science in Civil Engineering


May 2003





ii












To my parents, for their love and caring


iii
ACKNOWLEDGEMENTS


I am grateful to my advisor Dr. Jason Weiss for his constant support and advice
during the course of this research work. I also wish to thank Dr. Cohen and Dr. Grandt
for being part of my committee. The financial support from Purdue Research Foundation
and Center for Advanced Cement Based Materials is greatly acknowledged.
I express sincere thanks to Janet Lovell and Mark Baker for their assistance. I
would also like to acknowledge my fellow students in Materials Area of School of Civil
Engineering for their help and company. I am greatly thankful to my family for their
motivation and love.


iv

TABLE OF CONTENTS

Page
LIST OF TABLES...........................................................................................................viii
LIST OF FIGURES...........................................................................................................ix
ABSTRACT......................................................................................................................xii
CHAPTER 1: INTRODUCTION.....................................................................................13
1.1. Background................................................................................................................13
1.2. Research objectives....................................................................................................14
1.3. Scope of research.......................................................................................................15
1.4. Organization of work.................................................................................................15
CHAPTER 2: LITERATURE REVIEW..........................................................................17
2.1. Introduction................................................................................................................17
2.2. Concrete in compression............................................................................................17
2.2.1. Stress strain response.......................................................................................17
2.2.2. Localization in compression.............................................................................19
2.2.3. Snapback in compression.................................................................................20
2.2.4. Factors influencing the testing of concrete in compression.............................21
2.2.4.1. Concrete composition..............................................................................21
2.2.4.2. Shape and size of test specimen...............................................................21
2.2.4.3. Effect of boundary conditions..................................................................22
2.2.4.4. Testing control.........................................................................................24
2.2.4.5. Rate of loading.........................................................................................25
2.2.4.6. Stiffness of testing machine.....................................................................26
2.2.4.7. Allowable rotations of the loading platens before and during the
experiment...............................................................................................26
2.2.4.8. Gage length of the lineal variable differential transformer (LVDT).......27
2.3. Acoustic emission......................................................................................................27

v
2.3.1. Introduction......................................................................................................27
2.3.2. Crack analysis..................................................................................................27
2.3.3. Acoustic emission for studying the cracking phenomena................................28
2.3.4. Types of acoustic waves...................................................................................30
2.3.5. Acoustic emission equipment...........................................................................31
2.3.5.1. Sensors.....................................................................................................31
2.3.5.1.1. Angle of incidence............................................................................32
2.3.5.1.2. Frequency of wave............................................................................32
2.3.5.1.4. Available space on the surface of specimen......................................33
2.3.5.2. Coupling material.....................................................................................33
2.3.5.3. Preamplifier..............................................................................................33
2.3.5.4. Software...................................................................................................33
2.3.6. Terminology.....................................................................................................34
2.3.6.1. Felicity ratio.............................................................................................34
2.3.6.2. Kaiser effect.............................................................................................35
2.3.6.3. Triangulation............................................................................................36
2.3.7. Loss of acoustic energy....................................................................................37
2.3.8. Acoustic emission analysis procedure..............................................................38
2.4. Optical damage assessment........................................................................................40
2.4.1. Introduction......................................................................................................40
2.4.2. Image analysis..................................................................................................40
2.4.2.1. Damage stabilization................................................................................40
2.4.2.2. Sample preparation..................................................................................41
2.4.2.3. Image acquisition.....................................................................................41
2.4.2.4. Image processing.....................................................................................41
2.4.2.5. Damage quantification.............................................................................42
2.5. Research significance.................................................................................................42
CHAPTER 3: EXPERIMENTAL PROGRAM................................................................43
3.1. Introduction................................................................................................................43
3.2. Experimental approach..............................................................................................43

vi
3.2.1. Damage quantification and localization...........................................................43
3.2.2. Effect of length-to-diameter ratio.....................................................................43
3.2.3. Cycling loading and assessment of acoustic activity.......................................44
3.3. Application of research work.....................................................................................44
3.4. Selection of specimen size and geometry..................................................................44
3.5. Constituent materials.................................................................................................45
3.5.1. Cement.............................................................................................................45
3.5.2. Aggregate.........................................................................................................45
3.5.3. Fibers................................................................................................................45
3.5.4. Superplasticizer................................................................................................46
3.6. Mixture design...........................................................................................................46
3.7. Casting, curing, specimen preparation and inspection...............................................47
3.7.1. Casting..............................................................................................................47
3.7.2. Curing and demolding......................................................................................47
3.7.3. Specimen preparation.......................................................................................48
3.7.4. Specimen inspection.........................................................................................49
3.8. Testing equipment......................................................................................................49
3.8.1. Universal testing machine................................................................................49
3.8.2. Linear variable differential transformer...........................................................50
3.8.3. Ring frame........................................................................................................50
3.8.4. Teflon sheets....................................................................................................50
3.8.5. Acoustic emission system................................................................................51
3.9. Sensor locations and attachment mechanism.............................................................51
3.10. Testing plan..............................................................................................................53
3.11. Mode of testing........................................................................................................53
CHAPTER 4: EXPERIMENTAL RESULTS..................................................................56
4.1. Introduction................................................................................................................56
4.2. Experimental result categorization............................................................................56
4.2.1. Damage quantification and localization...........................................................56
4.2.1.1. Unloading stiffness and mechanical energy............................................57

vii
4.2.1.2. Acoustic emission measurement..............................................................62
4.2.1.3. Damage localization.................................................................................67
4.2.1.3.1. Location and size of CDZ.................................................................68
4.2.1.3.2. Effect of CDZ on stiffness degradation.............................................70
4.2.2. Effect of length to diameter ratio.....................................................................71
4.2.3. Cycling loading and acoustic activity..............................................................72
4.3. Categorization of stress-strain curve..........................................................................76
CHAPTER 5: OPTICAL DAMAGE ASSESSMENT.....................................................79
5.1. Introduction................................................................................................................79
5.2. Surface preparation....................................................................................................79
5.2.1. Specimen strengthening...................................................................................79
5.2.2. Sampling...........................................................................................................80
5.2.3. Grinding and polishing.....................................................................................81
5.3. Image analysis............................................................................................................81
CHAPTER 6: SUMMARY AND CONCLUSIONS........................................................86
6.1. Introduction................................................................................................................86
6.2. Summary....................................................................................................................86
6.3. Summary....................................................................................................................87
6.4. Recommendations for future......................................................................................88
LIST OF REFERENCES..................................................................................................89
APPENDICES..................................................................................................................98
APPENDIX A...................................................................................................................99
APPENDIX B.................................................................................................................101
APPENDIX C.................................................................................................................104


viii

LIST OF TABLES


Page
Table 2.1: Direct and indirect methods for crack analysis................................................28
Table 2.2: Types of acoustic waves..................................................................................31
Table 2.3: Categorization of events based on amplitude (Rao et al. 1999)......................39
Table 3.1: Aggregate gradation.........................................................................................46
Table 3.2: Mixture proportions.........................................................................................47
Table 3.3: Specimen characteristics..................................................................................49
Table 3.4: Testing plan.....................................................................................................54
Table 5.1: Color classification for processing image........................................................83
Table A2: Chemical properties of cement (Lonestar, Type I)........................................100
Table A3: Mechanical properties of cement (Lonestar, Type I)....................................100
Table A3: Properties of fiber..........................................................................................100



Table

ix

LIST OF FIGURES

Page

Figure 2.1: Typical curves for normal, medium and high strength concrete (Jansen, 1996)
.......................................................................................................................18
Figure 2.2: A conceptual illustration of the concept of localized damage in concrete
(Jansen 1996).................................................................................................20
Figure 2.3: Experimental stress-strain curves (a) Normal strength concrete (b) High
strength concrete (Jansen and Shah, 1997)....................................................22
Figure 2.4: Effect of boundary restraint on stress-strain curve for (a) low strength and (b)
high strength concrete (Kotsovos, 1983).......................................................24
Figure 2.5: Uniaxial compressive stress-strain curves for concrete for different loading
rates (Rusch, 1960)........................................................................................25
Figure 2.6: A typical acoustic wave..................................................................................30
Figure 2.7: Illustration of Kaiser effect.............................................................................35
Figure 2.8: A Typical AE results during the unload-reload process for the study of Kaiser
effect in mortar (Weng et al. 1992)...............................................................36
Figure 2.9: Triangulation technique for locating crack....................................................37
Figure 2.10: Relation of fracture energy with AE energy (Landis et al. 2000)................39
Figure 3.1: Grinding machine with attached guide...........................................................48
Figure 3.2: Ring apparatus used for testing......................................................................51
Figure 3.3: Locations of sensors.......................................................................................52
Figure 3.4: Experimental setup for compression testing..................................................55
Figure 4.1: Normalized stress-strain curves for specimens unloaded at different levels..57
Figure 4.2: Focal point: A reference for measurement of compliance.............................57
Figure 4.3: Typical stress-strain curve..............................................................................58
Figure 4.4: Normalized stress versus inelastic strain........................................................59
Figure 4.5: Calculation of fracture energy from individual tests and tested using the focal
point...............................................................................................................61
Figure

x
Figure 4.6: Application of focal point: Calculation of stiffness at different strain levels 62
Figure 4.7: Number of acoustic hits versus strain in specimen........................................63
Figure 4.8: Amplitude of acoustic events versus stress-strain response...........................63
Figure 4.9: Categorization of acoustic events on the basis of amplitude..........................64
Figure 4.10: Duration of acoustic events versus stress-strain response............................65
Figure 4.11: Categorization of acoustic events on the basis of duration..........................65
Figure 4.12: Application of focal point: Relation between acoustic energy and fracture
energy density................................................................................................66
Figure 4.13: Stiffness degradation and acoustic energy versus normalized strain...........67
Figure 4.14: Acoustic energy versus normalized strain....................................................68
Figure 4.16: Development of CDZ and its stiffness degradation.....................................70
Figure 4.17: Stress-strain response for different L/D ratios.............................................71
Figure 4.18: Loading pattern for cycling loading.............................................................72
Figure 4.19: Normalized stress versus normalized strain for cycling loading..................73
Figure 4.20: Focal point for cycling stress-strain response..............................................74
Figure 4.21: Stiffness degradation of specimen tested in cyclic loading..........................74
Figure 4.22: Cumulative acoustic hits for cyclic loading.................................................75
Figure 4.23: Zone categorization for stress-strain response.............................................76
Figure 5.1: Partially damaged specimens after testing.....................................................80
Figure 5.2: Typical ring surface........................................................................................82
Figure 5.3: Different levels of damage on ring surfaces...................................................82
Figure 5.4: (a) Original image (b) Processed image........................................................83
Figure 5.5: Percentage areas for different damage levels.................................................84
Figure 5.6: Acoustic energy versus percentage total damaged area.................................85
Figure B1: Energy density curves (304 mm specimen unloaded at 80 % postpeak)......102
Figure B2: Energy density curves (304 mm specimen unloaded at 60 % postpeak)......102
Figure B3: Energy density curves (228 mm specimen unloaded at 40 % postpeak)......103
Figure C1: Different levels of 304 mm (12 inch) specimen studied for visual damage
assessment...................................................................................................105

xi
Figure C2: Damage at different levels (275 mm, 250 mm, 225 mm and 200 mm) of 304
mm specimen...............................................................................................106
Figure C3: Damage at different levels (175 mm, 150 mm, 125 mm and 100 mm) of 304
cm specimen................................................................................................107
Figure C4: Damage at different levels (75 mm, 50 mm, 25 mm and 0 mm) of 304 mm
specimen......................................................................................................108
Figure C5: Damage at different levels (200 mm, 175 mm 150 mm and 125 mm) of 228
mm specimen...............................................................................................109
Figure C6: Damage at different levels (100 mm, 75 mm, 50 mm and 25 mm) of 228 mm
specimen......................................................................................................110
Figure C7: Damage at particular level (0 mm) of 228 mm specimen............................111


xii

ABSTRACT



Puri, Sunil, M.S.C.E, Purdue University, August 2002. Assessing the Development of
Localized Damage in Concrete Under Compressive Loading Using Acoustic Emission.
Major Professor: Jason Weiss


The stress-strain behavior of concrete in compression is one of the critical
material parameters that are used in the design of concrete structures. While significant
research has been performed to describe the importance of the peak strength and initial
elastic modulus, further work is needed to explain how damage develops and influences
the stiffness degradation and energy dissipation of concrete. While some have considered
the stress-strain response to be a material property, recent work has illustrated that the
stress-strain relationship is dependent upon specimen geometry and size. This research
investigated the stress-strain behavior of concrete cylinders with different length to
diameter ratios under compressive loading. The presence of a region or band of localized
damage, its size and its role in the size-dependent stress-strain response was explored.
Elastic and inelastic strains have been characterized under load-unload testing conditions
to better understand the elastic and dissipated fracture energies at different load levels. In
addition to load and deformation measurements, acoustic emission has been used to
investigate the damage zone localization and its growth. Specifically, detailed analysis of
acoustic activity is presented in terms of amplitude, duration and acoustic energy. To
support the correlation between the acoustic activity and mechanical damage in concrete
at different load stages, damaged cylindrical specimens were sectioned, ground, polished
and imaged to inspect extent of cracking and hence the damage at different levels of
cylinders. Image analysis was applied using Image Pro Plus software to quantify cracking
at the different levels of specimen. Results of acoustic emission evaluation and image
analysis were compared to present clear picture of damage localization and its
development in concrete under compressive loading.

CHAPTER 1: INTRODUCTION



1.1. Background

The stress-strain behavior of concrete is a crucial factor that is used in the design
of concrete structures
.
Peak compressive stress and initial elastic modulus is applied to
define the Whitney’s stress block. (ACI-318)
Research has focused on relating the stress-strain behavior of concrete with the
ultimate compressive strength and size of specimens (Shah and Chandra, 1968, Jansen et
al. 1995, Popovics 1998) The majority of the investigations have modeled the concrete as
a material that undergoes uniform damage and many continuum damage approaches have
been used. Recent investigations have questioned the use of continuum damage
mechanics and have advocated the existence of localized damaged in concrete (Jansen,
1996, Van Mier, 1984, Palmquist and Jansen, 2001, Weiss et al. 2001). Though the
presence of region of local damage has gained acceptance (RILEM 148-SSC, 1997), the
size of compression damage zone and its role in the stress-strain response has not been
fully understood (Jansen, 1996). This thesis is intended to build on the fundamental
understanding of the behavior of damage zone in concrete in compression. The
compressive behavior of concrete has been investigated to develop an improved
understanding of crack growth, damage development, stiffness degradation, failure
mechanisms and the resulting stress-strain response.
A variety of experimental techniques have been used in the past to understand
failure damage mechanisms of concrete through crack patterns and sizes (Malhotra and
Carino, 1991).
A real-time non-destructive technique is needed to assess micro and macro level
damage in concrete and to specifically locate the damage that develops. Acoustic
Emission (AE) is one such technique that provides non-invasive information of the
material under testing. AE has been successfully used to understand the failure

14
mechanisms in metals and fiber reinforced composite materials (Clough and
McDonough, 1996, Hamstad et al. 1979, Hamstad et al. 1992, Nesvijski and Sarkis,
1999). AE has also been used to investigate the fracture processes in concrete (Landis et
al. 1991, Yoon et al. 2000, Ohtsu and Watanabe, 2001, Kim and Weiss, 2002). The
characteristics of acoustic waves (number of hits, amplitude and corresponding acoustic
energy) may provide a method to quantify fracture mechanism in concrete (Kim and
Weiss 2002, Yoon et al. 2000). Recent work has illustrated that energy measured by
acoustic sensors can be correlated to the mechanical fracture energy and damage in the
specimen (Landis et al. 2000).
While AE provides indirect information about damage processes, an optical
technique is needed to verify the damage localization and to endorse the effectiveness of
AE. Image Analysis has been found to be an effective tool to quantify cracking in
concrete (Qi et al. 2002, Ammouche et al. 2000).



1.2. Research objectives

The objectives of this research work are:
1) to correlate the stress level and damage development
2) to establish a correlation between the acoustic activity and damage development,
3) to determine whether AE can be used to quantify the size of compressive damage
zone (CDZ) in concrete in uni-axial compression,
4) to determine the size of damage zone and its development,
5) to determine how the damage zone influences the stiffness degradation in a
specimen,
6) to determine whether the damage zone size is independent of specimen size



15

1.3. Scope of research

Concrete is a quasi-brittle material. Under compressive loading, stresses cause
cracks to initiate and grow. At higher stress levels these microcracks coalesce resulting
into a region of dense cracks, which finally lead to failure. This failure of concrete in
compression is believed to occur as a result of localization of damage in a particular
region of the specimen, known as compression damage zone (CDZ) (Van Mier 1984,
Jansen 1996).
Both microscopic and macroscopic cracking contribute to the development of
damage which causes a release of energy. This energy release is associated with the
initiation of microcracks and their development to form macrocracks. Recently, several
studies have attempted to correlate the acoustic activity with the damage occurring in the
concrete (Landis et al. 1991, Yoon et al. 2000, Ohtsu and Watanabe 2001, Kim and
Weiss 2002). It should be noted that for improvements in the interpretation of the AE
results, the acoustic emission signals must be analyzed to understand the nature of their
source. Analysis of the quality of these AE events may make it possible to understand
initiation of microcracks, their propagation, their coalescence to form macroscopic cracks
and ultimately resulting into the compression damage zone (CDZ). The development of
CDZ will also be investigated.
Fracture energy dissipated during loading may also provide significant
information on the extent and rate of damage development. This information, in tandem
with AE observations, may be used explain how damage develops and influences
stiffness degradation.


1.4. Organization of work

The research presented in this thesis is divided into six chapters. The first chapter
provides background information on the research objectives and scope of the work. The
second chapter is a literature review that provides an indepth review of the existing state
of the knowledge. The third chapter begins with a brief introduction of the experimental
approach and follows with more details about the equipment and the experimental

16
procedures. The fourth chapter presents the results of experimental testing, analysis and
interpretation of these results. The fifth chapter contains the details about the optical
damage assessment of tested specimens using image analysis and compares this analysis
with the results of AE. Finally, the sixth chapter provides a concise summary of the work
and suggestions for further work.


17


CHAPTER 2: LITERATURE REVIEW



2.1. Introduction

This chapter provides the background for thesis. The first section of this chapter
provides basic information about the behavior of concrete in compression, the
significance of the stress-strain response, a physical explanation for the prepeak and
postpeak stress-strain behavior of concrete, and the current state of work done on the
concept of compression damage zone.
The second section of this chapter explains the theory behind using acoustic
emission technique for damage assessment in concrete.
Finally, the third section of this chapter explains the basic concepts of the image
analysis and presents its application in quantifying damage in concrete.


2.2. Concrete in compression



2.2.1. Stress strain response
Compressive strength is one of the most widely used mechanical properties of the
concrete. The study of stress-strain behavior of concrete has long been a challenging
topic of research for civil engineers. Complete compressive stress versus strain curves
(including the postpeak portion) for normal and high strength concrete have been
experimentally established by a number of researchers (Wang et al. 1978, Shah et al.
1981, Glavind and Stang, 1991, Taerwe, 1992, Han and Walraven, 1993, Jansen and
Shah, 1993, Hsu and Hsu, 1994, Jansen and Shah, 1994, Jansen et al. 1995, and Choi et
al. 1996). Typical examples of these curves are shown in the Figure 2.1. It can be noticed
that for all strength levels the stress-strain response remains linear at low stress levels but

18
at higher stresses (just before peak stress), the stress-strain curve becomes non-linear. It
can also be observed that for low and normal strength concretes, the strain at peak is
lower than that for higher strength concrete.
Figure 2.1 also shows that the postpeak part of curve becomes steeper with the
increasing strength levels. Steepness in postpeak behavior signifies an increase in the
brittleness of the concrete. This behavior of concrete has been explained using the energy
concepts. It has been speculated that during the compressive loading, the work done
against specimen results into storage of strain energy in the specimen. A part of this
stored energy works against concrete and dissipates with time as fracture energy, while
rest of the stored energy gets released. The proportion of strain energy dissipated as
fracture energy is directly related to the fracture surfaces that form during cracking.
In high strength concrete, large amount of strain energy gets stored due to
absence of micro level fracture surface formation. This large quantity of stored energy
leads to higher rate of propagation of cracks that leads to the direct formation of
macrocracks resulting into more brittle failure as compare to that in low strength
concretes.


Figure 2.1: Typical curves for normal, medium and high strength concrete (Jansen, 1996)


19
2.2.2. Localization in compression

The failure of concrete in compression is believed to occur due to localization of
damage in a particular part of specimen, which is termed as the compression damage
zone (CDZ). (Van Mier, 1984, Jansen, 1996) The remaining portion of concrete specimen
is known as the bulk zone.
To understand the behavior of stress-strain curve of concrete under compressive
loading, behavior mechanism of concrete can be idealized as follows. Initially, the stress-
strain curve responds linearly and the material behaves elastically (Figure 2.1). As the
load is increased, damage develops in different parts of specimen. As a result concrete
undergoes permanent damage resulting in the stress-strain response deviating from
linearity and energy being dissipated. Immediately before the peak load is reached, the
curve begins to flatten due to coalescence of cracking (Van Mier, 1984). It has been
speculated that localization of damage zone formation initiates at the peak stress (Torrenti
et al. 1991, Torrenti et al. 1993) or just prior to the peak stress (Shah and Sankar, 1987,
Comi et al. 1995). It, therefore, would seem reasonable to assume that the stress-strain
graph for bulk concrete and the eventual compression damage zone must be
approximately same up to the point of initiation of CDZ.
During the postpeak, bulk zone and CDZ behave differently from each other.
(Hsu et al. 1994, Shah and Jansen, 1993) A typical model for unloading of these parts has
been shown in Figure 2.2. It can be noticed that stress-stain curve of bulk concrete has
been shown to be unloading parallel to the initial (loading) elastic modulus. It means
elastic modulus of the material in the bulk zone is assumed to remain unchanged during
the process of loading and unloading. But for the compression damage zone, unloading
curve shows a stress-displacement response.







20













Figure 2.2: A conceptual illustration of the concept of localized damage in concrete
(Jansen 1996)


2.2.3. Snapback in compression
Snapback is defined as the work done by specimen against the testing machine. In
other words, we can explain that during the process of being compressed by testing
machine, specimen releases energy to counteract the work done by the machine.
Strain energy is stored in the specimen during the process of loading. When
damage occurs in the specimen in a localized region, energy is dissipated as cracking;
however as bulk zone unloads, the energy stored in it is released. This released energy
works against both the loading machine and the damage zone. The work done against the
loading machine leads to snapback. Extent of snapback can be calculated by subtracting
the net movement of end plates from the applied movement of the piston with an
assumption that loading machine remains uncompressed. More snapback has been
noticed in high strength concretes due to high amount of strain energy stored in
specimen, leading to a brittle behavior (Jansen 1996).

Post-
Localization

Applied
Stress
σ
ε
䉵Bk⁚潮e⁂敨慶=潲
σ
ε
䍄娠䉥桡癩潲
Pre-Localization
Pre-Localization
Post-
Localization

21

2.2.4. Factors influencing the testing of concrete in compression
According to RILEM Technical Committee (RILEM 148-SSC, 1997), the
following factors are the most significant influences in the testing of concrete in
compression.
1) Concrete composition.
2) Shape and size of test specimen
3) Effect of boundary conditions
4) Type of feedback signal
5) Rate of loading
6) Stiffness of testing machine
7) Allowable rotations of the loading platens before and during the experiment
8) Gage length of the Linear Variable Differential Transformer (LVDT)

2.2.4.1. Concrete composition
Concrete, being a composite material, attains its properties from the aggregate and
binder used to make it. Size, shape, gradation, stiffness and strength of aggregate (coarse
and fine) affect the fracture behavior of concrete (Shah and Chandra, 1968, Rao and
Prasad, 2002, Zhou et al. 1995, Tasdemir et al. 1996).
Composition of concrete directly affects the strength of concrete and hence the
fracture behavior. It can be noticed in Figure 2.1 that with the increase in strength,
postpeak curve becomes steeper which signifies more brittle failure.

2.2.4.2. Shape and size of test specimen
Cylinders and prisms are two specimen shapes, which are commonly used in most
of the research works (Choi et al. 1996, Shah and Sankar, 1987, Van Vliet and Van Mier,
1995, Gobi and Ferrara, 1995). It has been found that tri-axial state of stresses develops
in cube due to friction at the ends resulting in a strength value that is typically 25%
greater than the strength of the same concrete tested in cylindrical form (Van Mier 1997).

22
The size of the specimen also has prominent effect on the compressive behavior
of concrete. Different size specimens have been tested in compression in various research
studies. (Jansen and Shah, 1997, Sankar and Shah, 1987, Lee and Willam, 1997, Choi et
al. 1996). Typically, the size of the specimen is represented in the terms of length to
diameter ratios. ASTM C39 has also discussed the specimen specifications for
compressive strength testing in terms of L/D ratio. Jansen and Shah (1997) conducted a
series of experiments using cylindrical specimens of different slenderness ratios (H/D =
2, 2.5, 3.5, 4.5, 5.5); where H is the length or height of the specimen and D is the
specimen’s diameter. The stress-strain responses for normal and high-strength concrete
this experimental series are shown in Figure 2.3. It can be observed in this figure that for
same strength, the prepeak response of all the specimens is nearly identical whereas the
postpeak response depends upon specimen length. The descending branch in postpeak
region becomes steeper with an increase in specimen’s length corresponding to an
increase in brittleness.

Figure 2.3: Experimental stress-strain curves (a) Normal strength concrete (b) High
strength concrete (Jansen and Shah, 1997)
It can also be observed that size-effect is more predominant in the high strength
concrete as compare to normal strength specimens. Size of CDZ is the governing factor
for controlling the postpeak behavior of different strength concretes.

23

2.2.4.3. Effect of boundary conditions
Boundary conditions can have a significant effect on the stress-strain response
especially on the postpeak portion. (Kotsovos, 1983, Van Mier, 1984, Vonk, 1992, Choi
et al. 1996, Choi, 1996). The shear stresses between loading platen and specimen are
caused by a mismatch in their lateral expansion due to the Poisson’s effect (Gerstle et al.
1978, Kotsovos, 1983, Van Mier, 1984). Lateral compressive stresses are exerted on the
specimen due to restriction on its expansion by restraint from the loading platens. These
lateral stresses, along with vertical compressive stresses create an area of triaxial state of
stresses at the ends of the specimen. This triaxial state of stresses hinders the formation of
microcracks, leads to typical cup-cone failure and may increase the measured
compressive strength of specimen.
Presence of these triaxial stresses can be controlled by using friction reducing
teflon layers and/or grease on the ends. RILEM Technical Committee (148-SSC, 1997)
has cautioned against the use of excessive grease as it can cause tensile splitting of the
specimen due to pressure exerted by grease filled in concrete pores. As a result of tensile
splitting, the apparent strength may decrease and the “true compressive” stress may not
be measured. But the use of teflon sheets has been found to provide compressive strength
independent of the slenderness of specimens (RILEM 148-SSC, 1997). Effect of different
boundary conditions can also be noticed in the Figure 2.4. The postpeak curves become
steeper i.e. more brittle with decrease in boundary restraint. It is evident here that failure
changes from shear mode under high boundary restraint conditions (Case 1 &2) to a
splitting mode under very low boundary restraint (Case 3,4 &5) (Figure 2.4a)


24

Figure 2.4: Effect of boundary restraint on stress-strain curve for (a) low strength and (b)
high strength concrete (Kotsovos, 1983)

2.2.4.4. Testing control
Feedback signal for testing a specimen in compression can generally be
categorized as open-loop and closed-loop control. In case of open-loop, loading is
provided by increasing load or displacement irrespective of the resultant effect on the
specimen. For closed-loop testing, measured values like stroke of the actuator, force, the
displacement in the specimen (vertical or lateral) or the crack opening measured directly
from the specimen can be used to control the test. The advantage of closed-loop system is
that the control of the test is not dependent on the force or on the displacement of the
machine and effectiveness of the testing can be checked for more than 2000 times per
second. Various research works have been performed using closed-loop by combination
of vertical, lateral or circumferential displacement as feedback signal for controlling the
test (Jansen 1995, Taerwe 1992, Glavind and Stang 1991, Dahl and Brincker 1989, Choi
1996, Weiss et al. 2001)
Open-loop testing can be further categorized as load-control and displacement
control testing. For load-control testing, load is applied and corresponding displacement
is measured. The load is increased until the material fails at peak load, therefore it is not
possible to capture stress-strain response during strain softening of concrete in the
postpeak region. In displacement control, displacement is increased at a constant rate and

25
resulting load is recorded. During compressive loading, the total displacement of the
piston of the testing machine is equal to the sum of the deformation of the testing
machine and the specimen under testing. Testing machine deforms elastically in
proportion to the applied load therefore during the postpeak region when load carrying
capacity of the specimen decreases drastically, the deformed machine can unload
uncontrollably to destabilize the postpeak response. Therefore, caution must be practiced
while selecting the displacement reading from the loading cell for controlling the test.
External LVDTs have been recommended to obtain platen-to-platen displacement to
obtain a stable displacement control (RILEM 148-SSC, 1997).

2.2.4.5. Rate of loading
The rate of loading influences the response of concrete in compression. (Rusch,
1960) Rate of displacement-control loading for concrete in compression generally varies
from 1 to 20 micrometer per second where 10 micrometer per second is equivalent to
0.0236 inch/min. Experimental work has been done to study the effect of rate of loading
on the stress-strain behavior (Figure 2.5).

Figure 2.5: Uniaxial compressive stress-strain curves for concrete for different loading
rates (Rusch, 1960)


26
It can be observed from Figure 2.5 that with the increase in the rate of loading, the
strength of the cylinder increased. It has been explained that with high rate of loading,
micro-cracks cannot choose the optimum path to develop and propagate in to aggregate
particles. Hence, this changes the mode of fracture from inter-granular to trans-granular.
Trans-granular fracture provides a brittle failure and gives deceptive higher strength of
concrete than original. This can be also be noticed in Figure 2.5 that increase in rate of
loading led to brittle response resulting into steeper postpeak response.

2.2.4.6. Stiffness of testing machine
The stiffness of the testing machine plays an important role in compressive
loading of specimens, especially for high strength concrete. High stiffness of testing
machine is needed to avoid snapback and hence to record accurate displacement data and
attain stable postpeak response. A high stiffness testing machine is expected not to
deform during loading for the validity of the assumption that displacement applied is
equal to displacement of specimen.
Stiffness of testing machine can be calculated by finding out the exact
displacement of the specimen as determined by the platen to platen displacement.
Deformation in machine can be obtained by subtracting measured platen-to-platen
displacement from the applied displacement of the piston. Machine with minimum
difference is generally assumed to be good for testing.


2.2.4.7. Allowable rotations of the loading platens before and during the experiment
Testing machine must have loading platens fixed against rotation during
experimentation. It helps to avoid any torsion in specimen during loading. RILEM
Technical Committee (RILEM 148-SSC, 1997) further advised not to rely on the
accuracy of flatness of ends of specimen. Therefore, to have complete contact of the
loading platen with the end surface of specimen before the test, transverse movement of
one of the platen is asked for.


27
2.2.4.8. Gage length of the lineal variable differential transformer (LVDT)
Linear variable differential transformer (LVDT) is an electronic device used to
measure displacements. The gage length of LVDT is defined as the length of the
specimen for which displacement is measured. RILEM Technical Committee (148-SSC
1997) recommended platen-to-platen distance as gage length; therefore this was used in
this investigation.

2.3. Acoustic emission



2.3.1. Introduction
This section starts with a concise comparison of the different non-destructive
techniques available for crack analysis and explains the reason behind selection of
acoustic emission for this research work. The basic aspects related to acoustic emission
(AE) measurement are explained and a brief introduction is provided about the
terminology and equipment used in AE technique.

2.3.2. Crack analysis
Different methods can be used for crack and damage analysis in concrete. These
methods are divided into two types; direct methods and indirect methods. Table 2.1
describes the different aspects of these methods and compares them for different
experimental techniques.


28
Table 2.1: Direct and indirect methods for crack analysis
Direct methods
Indirect methods
1

Cracks are located visually and are
analyzed based on these optical
observations.

Presence of cracks is located by the
interpretation and analysis of data
recorded in the form of voltage signals,
ultrasonic waves, X-ray/ infrared rays.
2 Petrography, optical scanning, visual
inspection and penetration method can
be examples of direct methods.
Ultrasonic pulse velocity, acoustic
emissions, x-rays diffraction represent
indirect methods.
3 Results come as crack width, crack
density, crack aspect ratio and crack
length.
Voltage amplitude, wave frequencies and
film density constitute the results.
4 The changes that occur during the
preparation of sample for testing or
during the unloading of the specimen
cannot be adjudged. Uncertainty exists
about the issue that what is being
observed is actually the same
microstructure that was present when
the sample was under loading.
Working with indirect methods provides
live results of what actually is happening
during the process of loading and
unloading.
5 Only a small section of the sample is
examined, which may not present
statistically a true picture of crack
population. Inaccessible regions
remain unexamined.
Complete section can be analyzed by
systematically providing the wave
recorders (sensors) or pulse sources at
proper locations of specimen. These
methods can also access hidden parts for
inspection and hence can ensure complete
damage evaluation.

6 As these techniques need thin sections
of particular size so new cracks and
other surface damages are inevitable
during the process of preparation of
specimens. Due to this aspect these
techniques cannot be declared fully
non-destructive and on some occasions
also termed as semi-destructive
techniques.
Thin sections are not essential for testing.
Specimens can directly be tested in the
same form as would be cast or build. No
outside damage to material during
inspection makes these techniques as
purely non-destructive techniques.


2.3.3. Acoustic emission for studying the cracking phenomena
AE is a technique where the elastic waves generated by the rapid release of
energy from the crack processes are measured with piezoelectric transducers. In AE

29
technology the material under stress emits acoustic waves, which are picked up by
transducers attached at the surface of the material. These acoustic signals get converted
into voltage signals by sensors and are amplified before sending them to a computer
system. With the help of specially designed software, data collected by computer can be
formatted into desired form of graphs based upon amplitude, energy and number of
events recorded.
The voltage signals captured during the process of testing provide information
about the happenings of the interior of specimen. Therefore, this characteristic of AE
technology can be used to detect a growing crack inside a material.
As acoustic emission data can be digitized and stored in a personal computer, it
provides permanent record of the test that can be used at any time for re-evaluation and
post processing analysis. The study of acoustic emission, being a real-time and non-
invasive technique, has offered an excellent means in the past for observing the cracking
process that occurs during the loading of structures or lab specimens (Ohtsu and
Watanabe, 2001, Yuyama et al. 1999, Niiseki and Satake, 1991, Yoon et al. 2000, Kim
and Weiss, 2002). Due to these advantages, AE has emerged as an important tool to
assess damage in concrete structures.
As no standard code specifications are available for selection of parameters of
acoustic wave, the characterization of acoustic waves differs based upon the personal
preferences and understanding of the wave mechanisms. Figure 2.6 shows a typical wave
recorded during AE testing. It can be seen that the amplitude is the maximum voltage
recorded for the particular event whereas duration of the event is the time period for
which the voltage of event remains higher than the threshold value. Acoustic energy is
defined in this study as the absolute value of area under the acoustic signal (Figure 2.6).
Peak amplitude, signal duration and acoustic energy (Figure 2.6) have typically been used
in previous studies as standard parameters (Rao et al. 1991, Weng et al. 1992, Rajachar et
al. 1999, Ohtsu and Watanabe, 2001, Yoon et al. 2000, Kim and Weiss, 2002).

30

Figure 2.6: A typical acoustic wave

A number of micro and macro processes contribute to the deformation and
deterioration of a concrete under stress and to the resulting series of emission events.
Since the information available from AE monitoring is indirect in the form of voltage
signals, it needs to be analyzed to understand the nature of source of these events and the
damage mechanisms. By analysis of these AE events, it may be possible to understand
crack initiation, their propagation, their coalescence into macrocracks and hence the
formation of CDZ.
2.3.4. Types of acoustic waves
Sound waves are categorized as longitudinal, transverse, raleigh, lamb and
standing waves. Lamb waves, being part of only thin sections, are not discussed in this
work. Standing waves, which form due to overlapping of other waves, are difficult to
quantify and characterize. Therefore in this work, events recorded by acoustic sensors are
categorized in three categories; longitudinal, transverse and raleigh waves. Table 2.2
distinguishes these waves from each other and briefly describes their characteristics on
the basis of their velocity and type of movement.


31

Table 2.2: Types of acoustic waves


2.3.5. Acoustic emission equipment
In this section, a brief explanation has been provided about the equipment
selection criterion used in AE technique and an overview has been given to optimize their
use for getting best results in AE measurement.

2.3.5.1. Sensors
As mentioned earlier, the AE technique is based upon the detection and
conversion of high frequency elastic waves to electrical signals. This is accomplished by
directly coupling piezoelectric transducers on the surface of the specimen under testing.
Longitudinal waves
Transverse waves
Raleigh waves
Wave particles move in the
direction of propagation of
wave.
Wave particles move
in the direction
perpendicular to
propagation of wave.
Wave particles move along
the surface of material. The
plane of particle motion can
be either parallel or
perpendicular to the surface.
Also named as compression or
P-waves.
Also named as shear
waves or S-waves
Also called as surface or R-
waves.
These types of waves have
speed greater than both
transverse and raleigh waves;
approximately 1.5 times the
velocity of transverse waves.
Speed is less than
longitudinal waves
but more than raleigh
waves.
Have least speed among these
three types of waves.
Particle motion, being
perpendicular to direction of
wave motion, can be easily
detected by sensors.
These waves are not
easy to detect.
Perpendicular component is
detected more often as
compare to parallel one.

32
Sensors are coupled to the specimens by means of a fluid or jelly and are secured with
tape, adhesive or rubber bands.
As the hardened concrete surface would not be uniform and smooth enough to
have complete contact with the surface of sensor, the coupling agent must also have the
characteristic to fill the pores on the surface of concrete and hence to facilitate the
movement of waves from specimen surface to sensors. Attention must be paid to have
minimum possible thickness of adhesive.
Sensors work on the principle of conversion of displacements into voltages. The
output of each piezoelectric sensor (during specimen loading) is amplified through a low-
noise pre-amplifier, filtered to remove any extraneous noise and furthered processed by
suitable electronic equipment.
The angle of incidence, frequency of wave, type of testing material, and the
availability of space on the surface of specimen are few of the factors, which can
influence the selection of sensors. The following section provides introductory
information on the importance of each of these factors.

2.3.5.1.1. Angle of incidence
Right angle is an ideal angle for the movement of waves from specimen surface to
sensor. But onsite testing defies this condition quite often and angle of incidence varies
from a right angle. The farther the angle of incidence is from a right angle, higher
dispersion of waves would be expected and more accurate and sensitive sensors would be
asked for capturing the acoustic waves.

2.3.5.1.2. Frequency of wave
The selection of sensors also depends on the estimated frequencies of the waves it
is expected to record. The increase in the sensitivity of sensor for different frequencies
increases the cost of sensors.

33
2.3.5.1.3. Type of material
Concrete is not an isotropic material. Anisotropy in concrete increases with
increase in the size of aggregate and with the difference of modulus of elasticity of
aggregate and the cement paste. Increase in anisotropy leads to more dispersion of waves
and hence asks for more accurate sensors.

2.3.5.1.4. Available space on the surface of specimen
The selection of sensors cannot be finalized without taking into consideration the
availability of space where the sensors could be attached to the surface of the specimen.
Typically ¾ inch -1 inch sensors are used in the lab
.


2.3.5.2. Coupling material
Coupling material used to attach sensors to the testing specimen also has a crucial
role in transfer of acoustic waves from specimen to sensors. Factors affecting the
selection of coupling material for sensors are surface of material, temperature of
surroundings, humidity, type of waves expected and sensors with which it is used.
Detailed information exists on how different coupling materials influence acoustic
emission, principles and instrumentation (Beattie, 1983).

2.3.5.3. Preamplifier
A preamplifier is used to amplify the voltage signals received from sensors.
Preamplifiers can be categorized on the basis of number of channels, storage capacity,
and the speed with which it processes the data. More information about the selection
criterion for sensors and preamplifiers can be obtained from Beattie (1983).

2.3.5.4. Software
Development of AE techniques has led to arrival of many types of software. To
have full advantages of the software, following aspects can be kept in mind while
selecting it. An acoustic analysis software

34
• should have the capability to analyze the waveform in time/frequency domain in
different windows.
• must investigate AE events by location algorithms so as to locate the origin of
events.
• should have possibility to import data files and to perform analysis.
• must be capable of exporting data as text files to MS-Word / MS-Excel and
graphics files as BMP or JPG.
• must have a fast data processing, high-resolution display and a good printing
capacity.
• must have user-friendly library functions (import / export of diagrams, locations,
filter etc.)

2.3.6. Terminology
AE technique has its own technical terms that describe the presentation of data.
This section covers a few of these terms.

2.3.6.1. Felicity ratio
The felicity ratio (FR) is defined as the ratio of the load at which AE events begin
on successive load cycle to the maximum load of the previous cycle (Equation 2.1).
Figure 2.7 represents a typical load-unload cycle. Cumulative acoustic activity profile has
also been plotted against time.
FR =
cycle previous on the load Maximum
cycle load successiveon begin events AEat which Load
=
M
S
Ρ
Ρ
(2.1)
Felicity ratio has been used to describe the failure of fiber-reinforced materials or as a
criterion to evaluate the degree of damage in structures (Choi et al. 1992, Lee and Lee
1999). The felicity ratio is also termed as Concrete Beam Integrity Ratio (CBI ratio) to
analyze the structural integrity of reinforced beams (Yuyama et al. 1999).

35














Figure 2.7: Illustration of Kaiser effect

2.3.6.2. Kaiser effect
During unload and reload cycling process, AE signals are not likely to be
observed until the newly applied load exceeds the maximum load of previous cycle. This
phenomenon is called Kaiser effect (i.e. felicity ratio of 1). Figure 2.7 shows an example
of the Kaiser effect. As the load does not change from first cycle to second cycle, AE
activity remains nil, but when in the third cycle, the load exceeds the load applied in
previous maximum load of second cycle, AE activity starts immediately.
A similar illustration of Kaiser effect has been presented in the experimental
results shown in Figure 2.8. It shows actual AE results as an event count rate (rate of
acoustic events) obtained during unload and reload process. In this figure, the dotted lines
represent the loading history.
Time
Load Exceeds the
Previous Level
Acoustic
Activity Starts
Load &
Cumulative
Acoustic Events
I
II
III
Load Cycles
AE Activity
P
m
P
s
Cycle I Cycle II Cycle III

36

Figure 2.8: A Typical AE results during the unload-reload process for the study of Kaiser
effect in mortar (Weng et al. 1992)

It can be seen that at the beginning of the test, very high AE count rate appeared
but when the load was held constant at a level of approximately 20 kN for a short period
of time, no AE signals were recorded. Unloading cycle shows few AE events. As the
specimen was reloaded to a higher load level (40 kN), AE was rare until the applied load
exceed the maximum load of the previous cycle i.e., 20kN. This process continues in
each cycle as the load is increased further from the load in the previous cycle. It can also
be noted in the later cycles that AE events appeared continuously even at a constant load
level. This means that the cracks within the specimens were growing (i.e., the specimen
was in an unstable condition), which signifies that Kaiser effect did not hold at that stage.

2.3.6.3. Triangulation
Triangulation is a technique used to locate the position of origin of acoustic
events and hence the cracks. Typical velocity of acoustic waves in a particular material
and the location of sensors in a particular coordinate system are used to determine the
location of the original event (Figure 2.9).

37
-4
-3
-2
-1
0
1
2
3
4
5
-4 -3 -2 -1 0 1 2 3 4 5

Figure 2.9: Triangulation technique for locating crack

For triangulation, one sensor is used as a trigger to activate other sensors. On
being activated, sensors pick up acoustic signals and capture time of arrival of events.
The distance of the crack from each sensor can be determined as a product of acoustic
velocity and captured time of arrival.

2.3.7. Loss of acoustic energy
Concrete constitutes different phases (i.e. matrix, aggregate, micro and macro
level cracks). The percentage of energy transmitted from one phase to another depends
upon the acoustic impedances of the each phase. Acoustic Impedance (Z) of the material
is defined as the product of the density of the material with the wave velocity (V) in that
phase.
A transient wave changes its mode many times from longitudinal to shear and/or
to surface waves and vice versa due to reflection and/or refraction during traveling in
different phases.
The percentage energy transmitted (E
t
) between two phases with acoustic
impedances Z
1
and Z
2
is given as
Crack

L
3

L
1
L
2
Sensor #2
Sensor #3
Sensor #1


38
2
21
21
)(
1004
ZZ
ZZ
E
t
+
×
=
(2.2)
and the percentage reflected energy (E
r
) is
100
)(
)(
2
21
21
×








+

=
ZZ
ZZ
E
r
(2.3)
Complete information about the acoustic impedances of the different materials
and the wave velocities in these materials can be obtained from literature (Beattie 1983,
ASTM E494-95).
Loss of energy of wave in a single phase is considered different for two and three
dimension problems. For two-dimension, it is proportional to the linear distance traveled
from the source; whereas for three-dimension, loss of energy is proportional to the square
of linear distance traveled from the source.


2.3.8. Acoustic emission analysis procedure
Analysis of AE signals can be broadly categorized as time domain and frequency
domain analysis. Due to convenience in handling high data rate, time domain analysis is a
more common approach. Time domain analysis usually consists of studying a number of
parametric plots; the simplest of these is the estimation of cumulative events and event
rate.
Work of Weng et al. (1992) on mortar gave the results that when the compressive
strength of the mortar specimens was increased, the number of AE events was found to
increase accordingly. An approximately linear relationship was observed between the
increase in strength and number of AE events.
Works by Landis and Whittaker (2000) found that the AE energy corresponding
to acoustic events recorded from the damage of wood is directly proportional to the
fracture energy dissipated during the loading. Results from the experiments indicated that
roughly one sixth of the fracture energy is measured as acoustic emission energy. This
relation is evident from the Figure 2.10.


39

Figure 2.10: Relation of fracture energy with AE energy (Landis et al. 2000).

Rao et al. (1999) proposed the idea of distinguishing the acoustic events on the
basis of their peak amplitude. Acoustic events were divided in to different categories
naming α, β, γ and δ on the basis of their peak amplitude. To explain the relation of
events with micro & macro cracking with damage the events were categorized as follows.

Table 2.3: Categorization of events based on amplitude (Rao et al. 1999)

Phase
Event Type
Peak Amplitude (dB)
Micro-crack Initiation
α
44-60dB
Micro-crack Extension
β
61-70dB
Macro-crack Initiation
γ
71-80dB
Macro-crack Extension
δ
81-100dB

Based upon the acoustic characteristics, the events may indicate the condition of
the component under test in terms of cumulative cracks, extent of damage or defect
growth, thereby providing information about the concrete specimen.



40
2.4. Optical damage assessment



2.4.1. Introduction
This section provides a brief introduction about image analysis. Information has
been provided about the procedure used to stabilize damaged specimens and the method
that could be used to prepare surface of concrete so that it can be captured as an image
effectively.


2.4.2. Image analysis
Image analysis is a technique in which images of surfaces are analyzed to
categorize the contents of image on the basis of color, shape and size. This study covers
two-dimensional image analysis. In two-dimensional image analysis, surfaces are
transformed into an image that consists of a plane of pixels. Pixels of different color and
intensity are categorized and are correlated with the characteristics of original surface.
Damage quantification in concrete using image analysis involves following steps:
1) damage stabilization 2) sample preparation 3) image acquisition 4) image processing
5) damage quantification. These steps have been explained in brief in the following
section.

2.4.2.1. Damage stabilization
Damaged specimen needs to be stabilized before it can be analyzed. Low-
viscosity epoxy is typically used to stabilize the damaged specimen. Epoxy impregnation
of the pore system serves two purposes: a) it fills the voids and, upon curing, supports the
microstructure and b) it enhances contrast between the pores, cracks, aggregate and
cementitous material. For analysis of microcracks, an epoxy of low viscosity is
necessary, but for the less permeable matrix in high strength concrete, an ultra-low
viscosity epoxy is needed to aid its rapid infiltration into the open pores and cracks in the
system.


41
2.4.2.2. Sample preparation
Sample preparation constitutes cutting, grinding and polishing of stabilized
specimen. After stabilization of the damaged specimen, it is cut and painted. It should be
noted that the process of cutting can also cause micro level damage to concrete. Methods
for solving the problem of introduction of cracks during sample preparation have been
explained by Horanin et al. (1996).
The surfaces of the cut samples need to be ground and polished before images can
be captured. Polishing removes the damage imparted by the sawing and grinding
operations. This stage involves use of a sequence of successively fine grits or diamond
pastes ranging from 6 µm to 0.25 µm. Smooth glass or a lap wheel covered with a
polishing cloth can also be used for polishing. When looked under light, a finely ground
surface looks dark and rough due to light scattering with blurring, whereas polished
surface appears bright and crisp almost like a mirror, with sharp edges and good
differentiation between components. The details of grinding and lapping methods have
been well described in the literature of Allman and Lawrence (1972) and John et al.
(1998).

2.4.2.3. Image acquisition
In this process, the prepared surface of the specimen is captured as image.
Different capturing devices like optical microscope, optical scanner, still or movie
camera can be used to capture the images. A constantly controlled light source is needed
to have uniform capturing process. A frame size is selected to have uniform resolution in
all the images. Depending upon the resolution of the image, captured images attain
particular number of pixels.

2.4.2.4. Image processing
This step may typically constitute thresholding, image cleaning, filtering and grid
masking. More information about these processes can be obtained from Qi et al. (2002)
and Ammouche et al. (2000).


42
2.4.2.5. Damage quantification
Processed image can be used to quantify damage. Depending upon the color,
intensity of the color, size and aspect ratio, different elements of the image can be studied
for micro and macro level cracking.

2.5. Research significance

The design of concrete structures typically utilizes the peak strength and initial
elastic stiffness to approximate the stress-strain response of concrete (i.e., Whitney’s
stress block) (ACI-318). While the stress-strain response is commonly thought of as a
material property, it has been shown that the stress-strain response is dependent on the
specimen size and geometry (Van Mier, 1981, Jansen and Shah, 1997). This research is
aimed at investigating the initiation of microcracks, their coalescence, development of the
CDZ and its effect on the overall stiffness of the material.
Acoustic measurements will be used to quantify the size and properties of the
CDZ. By quantifying the size of the CDZ, a composite modeling approach can be used to
quantify the role of specimen size and geometry on stress-strain behavior. In addition, the
relationship between damage and acoustic measurements can be related to indicate how
acoustic sensors may be used in structural health monitoring
.

43

CHAPTER 3: EXPERIMENTAL PROGRAM



3.1. Introduction

This chapter describes the experimental approach that has been followed in this
work. Details are provided about the constituent materials, testing equipment and process
used for preparing and testing the specimens.
Information has been provided to describe how data has been recorded by both
mechanical loading equipment and AE equipment.

3.2. Experimental approach

The experimental approach has been divided into three test series: 1) damage
quantification and location 2) the effect of length-to-diameter ratio and 3) damage
assessment for cycling loading. The following section describes the testing mechanism
and goal of each of these series.

3.2.1. Damage quantification and localization
Cylindrical specimens were tested in compression. Each specimen had
length/diameter ratio of 4. Specimens were unloaded at predetermined load levels [58%
prepeak, 78% prepeak, 98% postpeak, 80% postpeak, 60% postpeak, 38% postpeak]. The
goal of this series was to analyze the unloading stress-strain response at different load
levels.

3.2.2. Effect of length-to-diameter ratio
Cylindrical specimens with length/diameter (L/D) ratios of 1, 2, 3 and 4 were
tested in compression. The goal of the testing was to analyze the effect of variable L/D on
the postpeak stress-strain response and on the development of damage zone.

44

3.2.3. Cycling loading and assessment of acoustic activity
Once again specimens with a L/D of 4 were tested under compression. Specimens
were tested in load-unload-reload cycles with an increment of 10% of the peak load from
the maximum loading level of previous cycle. The objective behind this testing series was
to analyze the unloading stress-strain response for cycling loading, to assess the Kaiser
effect and felicity ratio, and to capture the initiation and development of damage zone.

3.3. Application of research work

Structural design utilizes peak compressive stress to define the Whitney’s stress
block. Knowledge about stage of initiation of the CDZ, its size, and development can
help to understand the overall stiffness degradation of concrete. Information about the
size of damage zone can also help to predict its effect for different size structures. Size-
independent factors can be selected to have their application on universally applicable
structural design for all sizes. Correlation of acoustic characteristics with mechanical
fracture processes can provide a tool to assess structural health monitoring non-
invasively.

3.4. Selection of specimen size and geometry

Cylindrical specimens were chosen for this work. Cylindrical specimens were
preferred since cubical specimens typically exhibit corner splitting and triaxial stresses
along ends. While ASTM C-39 recommends specimens with L/D=2, different L/D ratios,
lengths and diameters have been used in the previous research. Shah and Sankar (1987)
used 76.2 mm (3 inch) specimens for their work on internal cracking and strain softening
response under uniaxial compression. Jansen (1996) used three different size diameters;
76.2 mm (3 inch), 101.6 mm (4 inch) and 152 mm (5 inch). Round Robin testing
performed by the RILEM Technical Committee (148-SSC, 1992) typically used 100 mm
as dimension for cylinder or prism. However, this work chose to adopt 75 mm specimens
to permit greater testing control and to have more transmission of waves by reducing AE
losses.

45
Consulting previous works (Palmquist and Jansen 2001) and making an
assumption that the size of damage zone is double the size of the diameter, L/D ratios of
1, 2, 3, and 4 have been used in this work.

3.5. Constituent materials

This section describes the constituent materials used for manufacturing the
specimens.


3.5.1. Cement
Commercially available Type I portland cement supplied by Lonestar Industries
Inc. has been used for preparation of specimens. Details on the cement are provided in
Appendix A.

3.5.2. Aggregate
Locally available sand and coarse aggregate were sieved to obtain gradations
similar to those described in RILEM report (RILEM 148-SSC 1997). It should however
be noted that the gradations described in this thesis are based upon customary US sieve
sizes and not on European sizes as presented in RILEM 148-SSC. Table 3.1 provides the
details of the aggregate gradation recommended by RILEM 148-SSC and the gradation
used for this work.

3.5.3. Fibers
Polypropylene fibers (Structural fibers manufactured by W.R.Grace) were used to
increase the toughness of the specimen. Properties of fibers are provided in Appendix A.
The fibers were used as 0.75% of total volume and were added slowly during the process
of mixing to facilitate uniform distribution throughout the mixture.






46
Table 3.1: Aggregate gradation













The absorption coefficient for the sand used was 1.79% (ASTM C128-97) and for
the #8 aggregate was 1.31 % (ASTM C127-88).

3.5.4. Superplasticizer
Daracem-19 superplasticizer provided by W.R Grace Company was used.

Details
of superplasticizer are provided in Appendix A.


3.6. Mixture design

The concrete mixture used in this study is given in the Table 3.2.


Gradation recommended by
RILEM SSC-148 (1997)
Gradation as per
US standard sieve sizes
Sieve Openings
(mm)
% Passing
Sieve Openings
(mm)
% Passing
8 100 9.51 100.00
4 70.13 4.76 77.75
2 50.06 2.38 54.45
1 35.01 1.19 35.88
0.5 19.97 0.60 21.10
0.25 7.02 0.25 6.76
0.125 0 0.15 0

47
Table 3.2: Mixture proportions
Material
Weights (kg/m
3
)
Cement 367.4
Aggregate 1842.0
Water 183.7
Fiber (Polypropylene) 9.1
Super Plasticizer (Daracem-19) 3.9


3.7. Casting, curing, specimen preparation and inspection


3.7.1. Casting
In this study 30 cylindrical specimens were cast. Standard practices for making
concrete test specimens (ASTM C-192) were followed. The concrete was mixed in a pan
mixer. The specimens were cast in PVC cylindrical molds with a diameter of 76 mm (3
inch) and height of 406 mm (16 inch). One end of the molds was capped to form
cylinders. Concrete was placed in 100 mm (4 inch) lifts
.
Specimens were cast in the
batches of ten. A water-to-cement (w/c) ratio of 0.5 was used for all the mixtures. The
same aggregate and w/c were used for all the specimens. An internal needle vibrator was
used along with an external vibrating table to ensure proper consolidation of the mixtures.

3.7.2. Curing and demolding
Specimens were kept in the molds for the first 24 hours, with the top surface
covered with wet burlap and a plastic sheet. At an age of 24 hours the specimens were
removed from their forms. The top and bottom ends of the specimens (approximately 50
mm) were removed to attain the desired specimen length (302 mm) and to remove any
end-effects that may be present due to casting processes. Specimens were placed in a

48
moist environment (98 % relative humidity and 23°C) until the time of specimen
preparation and testing.

3.7.3. Specimen preparation
After 28 days, the specimens were cut to attain desired L/D ratio. The ends of the
specimens were ground using a lapping wheel to ensure that ends of the specimen were
flat. The lapping wheel was equipped with a vertical guide that kept the axis of the
specimen perpendicular to the end surface of the specimen. Figure 3.1 shows the lapping
wheel fitted with vertical guide. Grinding and polishing of ends were performed using
different size silicon carbide grit particles (#150, #180, #240 and #320 for grinding and
#400, #600, #800 and #1000 for polishing).


Figure 3.1: Grinding machine with attached guide


After grinding, the diameter and length of each specimen was measured at two
locations. The average values of characteristics of specimens from a set of mixture are
reported in Table 3.3. The weight of the specimens was also recorded.


Vertical
Guide
Lapping
Wheel


49
3.7.4. Specimen inspection
Ultrasonic Pulse Velocity test (UPV) was measured in accordance with ASTM
E494-95 on the specimens to verify the initial integrity of the specimens. UPV
measurements provided average speed of the sound, which was further used for crack
location determination. Table 3.3 provides results from the UPV test.


Table 3.3: Specimen characteristics
Specimen
ID
Average
Length (mm)
Average
Diameter (mm)
Velocity
(mm/ msec)
Weight
(gm)
Density
(gm/mm
3
)
1 315.0 76.01 4.999 3407.2 0.002384
2 313.7 76.10 4.971 3384.5 0.002372
3 307.3 75.95 4.894 3347.0 0.002404
4 304.8 76.33 4.815 3285.4 0.002356
5 307.3 76.37 4.973 3366.9 0.002392
6 315.0 76.26 4.984 3382.5 0.002351
7 304.8 75.87 4.916 3327.2 0.002415
8 304.8 76.40 4.932 3339.8 0.002390



Specimens with low density, surface anomalies or low velocity were discarded.
Specimens with confirmed integrity were placed back in the moist-curing room where
they were kept until the time of testing.

3.8. Testing equipment

The following section describes the equipment used for compressive testing and
to record mechanical and acoustic response.

3.8.1. Universal testing machine
A universal testing machine, manufactured by Machine Testing Services (MTS),
was used in this work. Eight-inch circular hardened steel platens were used for loading.
The bottom steel platen was fixed against rotation while top steel platen was allowed to
rotate. Acrylic (Lexan) screen was designed to protect the user from flying debris during
any uncontrolled failure of specimen. Load was recorded using a 490 kN (110 kip) load

50
cell. A computer was interfaced with the controller and the loading frame, which enabled
the tests to be programmed to obtain desired loading or displacement control. Data was
conditioned and recorded every second by the computer for the testing load cell and the
varying piston locations.


3.8.2. Linear variable differential transformer
Two linear variable differential transformer (LVDTs) displacement transducers
with gage length of ±5 mm were used to measure platen-to-platen displacement. The
LVDTs were calibrated before they were attached to the platens using the ring assembly
shown in figure 3.2. The LVDTs were interfaced with a computer for signal conditioning
and data acquisition.

3.8.3. Ring frame
Based upon RILEM Technical Committee’s (148-SSC, 1997) recommendations,
platen-to-platen distance was used as gage length for the testing. A ring apparatus was
designed to attach LVDTs to the platens as shown in figure 3.2. Reverse movement of
LVDTs was chosen to ensure their safety during uncontrolled damage of specimen. For
that, LVDTs were compressed before the start of testing and were controlled to open with
increase in compressive displacement.

3.8.4. Teflon sheets
Two sheets of teflon of 100 mm × 100mm (4 inch × 4 inch), each with thickness
of 0.254 mm were used on each end of the specimen. This was done to reduce friction
between the ends of the specimens and platens, thereby reducing triaxial stresses at the
ends of the specimens (Choi et al. 1996; Shah and Sankar 1987; RILEM 148-SSC 1997).
Figure 3.2 shows teflon sheets used at the ends.


51

Figure 3.2: Ring apparatus used for testing



3.8.5. Acoustic emission system
Acoustic emission system manufactured by Vallen System was used for capturing
acoustic activity. AMSY4 preamplifier was used to amplify the acoustic signals. Vallen’s
Visual AE software was used to analyze the data. More details about the acoustic
emission system and the concerned AE analysis software can be found from
http://www.vallen.de/index.html (2002).

3.9. Sensor locations and attachment mechanism

In addition to the mechanical measurements, four piezoelectric (375 MHz) wide
band sensors were used to record acoustic activity generated in the specimen. The sensors
were spaced along the length of the cylinder with the top and bottom sensors located 50
mm from the platens. The remaining sensors were spaced at 67.5 mm from one another.
Teflon
Sheets
LVDT


Acoustic
Sensor
Loading
Platen




Ring
Frame



52
The sensor surface was 20 mm in diameter. To attach the flat surface of sensors to the
surface of concrete cylinder, the surface of specimen was flattened at sensor locations by
using a grinding disk of 25 mm in diameter.






















Figure 3.3: Locations of sensors

The ground surfaces were thoroughly cleaned and vacuum grease was applied as
a couplant to both sensor surface and area of sensor location. Elastic bands were used to
apply pressure on sensors to maintain their contact with the specimen surface. Figure 3.3
shows the location of sensors in a 304.8 mm specimen.

120°
3

4

1

2

50.8mm
50.8mm
3@67.8mm
3in

53


3.10. Testing plan

Table 3.4 gives the detailed plan of testing carried out.

3.11. Mode of testing

At the time of testing (approximately 90 days), the specimens were removed from
the curing room and cleaned. The surface of the specimen was thoroughly towel dried.
Initially a load of 2.2 kN (0.5 kips) was applied to bring the platens (i.e., teflon sheets) in
contact with the ends of the specimens. Care was taken to ensure that the load did not
exceed 2.2 kN which could have generated damage in the specimen. In addition, sensors
were not used to record events in the preloading phase due to possible friction at the end
platens and the AE generated by seating. Once the pre-load was applied, the LVDTs were
adjusted to zero as the initial readings. Counters on the both the acoustic emission and
mechanical testing computers were synchronized.
Channel 1 (the sensor located at 117 mm from bottom) was used as the trigger
and complete waveforms were recorded from this sensor. A threshold of 50 dB was set
up to screen out environmental noise and the acoustic activity generated by the
equipment. Pencil lead break was used to ensure the working and the location of each
sensor. Separate acoustic files were saved for this testing procedure. Once testing started
for cyclic loading, messages were recorded in acoustic data for distinguishing each cycle
from other. Figure 3.2 and 3.4 shows in detail the experimental setup for the compression
testing.




Table 3.4: Testing plan
Objective
Experimental
Work
Mode of Achievement
Number of
Specimens
Additional Information
• To determine mechanical properties of
specimens
Calculation of f′c
& E
f′c = Load/ Area
E=Stress (f′c)/Strain

2
Displacement control
Rate = 200 µε/min
• To analyze unloading stress-strain
response.

• To establish a correlation between the AE
and damage mechanism in concrete

• To understand localization of damage and
to determine the size of compressive
damage zone (CDZ)
Unloading of
specimens at pre-
determined load
levels

Calculation of AE
characteristics

Calculation of
fracture energy

Analysis of unloading
stress-strain response

Analyzing peak
amplitude, signal
duration and acoustic
energy along increasing
strain levels.

Relation of fracture
energy and AE energy

Determining damage
location using
triangulation


6






L/D=4

Displacement control

Number of sensors = 4

Number of LVDTs = 2



• To determine the effect of L/D ratio on size