Bineshian Hossein 2014

Failure and post-failure aspects of mechanical response of
concrete structures to compression and tension
By
HOSSEIN BINESHIAN
BSc, MSc
This thesis is presented for the degree

of Master of Philosphy
of
The University of Western Australia
School of Civil, Environmental and Mining Engineering
2014
DECLARATION
This thesis does not contain work that I have published, nor work under review for
publication.
Hossein Bineshian
2
DEDICATION
To my love,
Leila jaan
3
ABASTRACT
The mechanical response of concrete to tension and compression including pre- and
post-peak portions are examined in this thesis. Pre-peak normally consists of three parts;
plastic, elastic, and brittle that can be obtained by any load-control testing machine.
Failure is occurred when a network of major cracks is formed at peak stress while
failure criterion describes failure behaviour of concrete under triaxial loading. Post-peak
is a mixture of material and structural responses of concrete after failure and unlike the
pre-peak is significantly influenced by geometry of specimen and stiffness of loading
frame. Stress–deformation curve is used in displacement-control mode. Strain softening
is a post-peak highly localized phenomenon caused by coalescence of micro-cracks that
forms a damage zone after peak stress while load-carrying capacity is diminished.
In total 102 uniaxial/triaxial compression/tension tests are conducted on 102 concrete

cylindrical 150 × 300 mm specimens in accord with ASTM standards. Major fracture
pattern of cone/shear and cone for concrete specimens under uniaxial and triaxial
compression tests are recognized respectively. An empirical classification; Concrete
Quality Designation (CQD), is introduced in this research that classifies concrete to six
classes, based on the UCS. It can be applicable in practical purposes.
A new strength criterion is developed for concrete based on the mathematical function
of Bineshian criterion (2000) having three parameters; however, new constant values
are determined for different CQDs. It can provide linear and nonlinear envelopes
capable of describing triaxial actual data for both compressive and tensile quadrants and
provides credible estimations of these values similar to those determined in laboratory
with correlation and accordance coefficients adjacent to 1 and 0 respectively.
Concrete Damage Plasticity (CDP) constitutive model is used for a CQD-G class plain
concrete in Abaqus. Suitable constitutive parameters for the behavioural model are
determined using experimental data, analytical methods, and other sources while proper
results obtained from numerical simulations using FEA. Findings in this sector of the
research are applicable in numerical modelling and simulations of failure and post-
failure responses of concrete and especially in design of high strength concrete (CQD-G
class) structures and concrete-like brittle materials like rocks.
4
TABLE OF CONTENTS
DECLARATION 2
DEDICATION 3
ABSTRACT 4
LIST OF FIGURES 7
LIST OF STMBOLS 10
LIST OF TABLES 14
ACKNOWLEDGEMENTS 17
CHAPTER 1:
INTRODUCTION
1.1. Preface…….……………………..………………………………………….18
1.2. Objectives…………………………………………………………………...20
1.3. Methodology…….…………..……………………………………………....22
1.4. Thesis Structure………………………………………………………….…23
CHAPTER 2:
MECHANICAL BEHAVIOUR OF CONCRETE
2.1. Introduction...……………………………………………………………....25
2.2. Pre-peak Response and Failure…..……………………………………….27
2.3. Post-peak Response………………………………………………………...30
2.3.1. Stiffness……………………………………………………....30
2.3.2. Geometry and Boundary Restraint…………………………...32
2.3.3. Strain Localization…………………………………………...35
2.3.4. Damage Plasticity……..……………………………………...36
2.4. Conclusions………………………………………………………………….41
CHAPTER 3:
EXPERIMENTAL AND EMPIRICAL WORKS
3.1. Introduction.………………………………………………………………..43
3.2. Plain Concrete Preparation………………………………………………..44
3.2.1. Mix Proportion and Design………………………………….44
3.2.2. Casting and Capping………………………………………...45
3.3. Strain Measurements……………………………………………………....46
3.4. Testing Plan………………………………………………………………....47
3.5. Test Results………………………………………………………………....49
3.6. Concrete Quality Designation……………………………………………...52
3.7. Conclusions………………………………………………………………….54
5
CHAPTER 4:
NEW STRENGTH CRITERION FOR CONCRETE
4.1. Introduction.………………………………………………………………..56
4.2. Bineshian’s Strength Criterion (2000) for Intact Rocks and Coals….….57
4.3. Development of Bineshian Strength Criterion for Concrete…………….65
4.3.1. Introduction…………………………………………………..65
4.3.2. Performance………………………………………………….67
4.3.3. Parameters of the Criterion…...……………………………...79
4.4. Conclusions………………………………………………………………….81
CHAPTER 5:
CONCRETE DAMAGE PLASTICITY MODEL
5.1. Introduction.………………………………………………………………..83
5.2. CDP Theory………………………………………………………………...84
5.2.1. Mechanical Response………………………………………...84
5.2.2. Uniaxial Cyclic Response…………………………………....87
5.2.3. Multiaxial Response………………………………………….90
5.2.4. Post-failure Stress-Strain Relation…………………………...90
5.2.5. Fracture Energy Cracking Criterion………………………….92
5.2.6. Response to Compressive Load……………………..…….....94
5.2.7. Yield Surface……………………………………………..…..95
5.2.8. Plastic Flow Potential……………….………………..……....96
5.3. Visualization of Crack Directions………………………………..………..97
5.4. Output…………………………………………………………….……..…..97
5.5. Identification of Constitutive Parameters……………………………..….98
5.6. 3D FEA using CDP………………………………………………………..103
5.7. Conclusions………………………………………………………………...113
CHAPTER 6:
CONCLUSIONS
6.1. Observations……………………………………………………………….114
6.2. Future Research…………………………………………………………...116
REFERENCES 118
6
LIST OF FIGURES
Figure 2.1. Complete stress-strain diagram for concrete consisting pre-peak and
post-peak responses.
Figure 2.2. Load-control and displacement-control modes in compression testing.
Complete 𝜎𝜎 − 𝜀𝜀 curve only in displacement-control mode is obtained.
Figure 2.3. Failure process on concrete under uniaxial compression including flaw
initiation and finally coalescence of network of cracks causing failure.
Figure 2.4. Failure envelops for actual triaxial data (Johnston, 1985).
Figure 2.5. Stability of machine-specimen system considering stiffness of machine,
and geometry of specimen (Hudson et al., 1972).
Figure 2.6. Typical force–displacement diagrams of large/small structure (Jansen
and Shah, 1997; Nemecek and Bittnar, 2004; Gamino et al., 2004).
Figure 2.7. Softening curves for specimens with different height (van Mier, 1986).
Figure 2.8. Stress–strain curves obtained from cylinder tests on high strength
concrete using high friction steel platens (van Mier et al., 1997).
Figure 2.9. 𝜎𝜎 − 𝜀𝜀 curves for normal strength concrete specimens of different
geometry loaded between rigid high friction steel platens (Gobbi and
Ferrara, 1995; van Mier et al., 1997).
Figure 2.10. Influence of end friction for concrete specimens under uniaxial
compression (van Vliet and van Mier, 1996).
Figure 2.11. Bazant’s model fitting on 𝜎𝜎 − 𝜀𝜀 data of different slenderness ratio
(Bazant, 1989 based on van Mier’s (1984) test results).
Figure 2.12. Mechanical response of concrete to uniaxial loading in compression.
Figure 2.13. Mechanical response of concrete to uniaxial loading in tension.
Figure 2.14. Definition of the cracking strain 𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 used for the definition of tension
stiffening data for CDP model.
Figure 3.1. Cylindrical concrete mold used in this research.
Figure 3.2. 𝜎𝜎 - 𝜀𝜀 curve for concrete specimens under uniaxial compression.
Figure 3.3. Sketches of types of concrete fracture under uniaxial compression
(redrawing from ASTM C39/C39M-04).
Figure 3.4. Baldwin loading frame used for uniaxial/triaxial compression testing.
Figure 4.1. Fitting the proposed strength criterion to the data of Johnston (1985) on
Westerly Granite.
7
Figure 4.2. Accordance coefficient for strength criteria with pairs of tensile data.
Figure 4.3. Observed value against predicted value by proposed strength criterion for
concrete specimens with CQD-VP and UCS<15 MPa.
concrete specimens with CQD-P and UCS≈15 MPa.
concrete specimens with CQD-P and UCS≈20 MPa.
concrete specimens with CQD-F and UCS≈30 MPa.
concrete specimens with CQD-F and UCS≈40 MPa.
concrete specimens with CQD-G and UCS≈55 MPa.
concrete specimens with CQD-G and UCS≈70 MPa.
concrete specimens with CQD-VG and UCS≈100 MPa.
concrete specimens with CQD-VG and UCS≈130 MPa.
concrete specimens with CQD-E and UCS>130 MPa.
𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 is the equivalent compressive plastic strain, 𝑑𝑑𝑐𝑐 is the compressive
damage variable, 𝐸𝐸0 is the initial elastic stiffness, 𝜎𝜎𝑐𝑐0 is the value of
initial yield in compression, and 𝜎𝜎𝑐𝑐𝑐𝑐 is the ultimate uniaxial stress.
~𝑝𝑝𝑝𝑝
Figure 5.2. Mechanical response of concrete to uniaxial loading in tension. 𝜀𝜀𝑡𝑡 is
the equivalent tensile plastic strain, 𝑑𝑑𝑡𝑡 is the tensile damage variable,
𝐸𝐸0 is the initial elastic stiffness, and 𝜎𝜎𝑡𝑡0 is value of initial yield in tension.
Figure 5.3. Effect of the compression stiffness recovery.
Figure 5.4. Definition of the cracking strain 𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 used for the definition of tension
~𝑝𝑝𝑝𝑝
stiffening data for CDP model. 𝜀𝜀𝑡𝑡 is the equivalent tensile plastic
strain, 𝑑𝑑𝑡𝑡 is the tensile damage variable, 𝐸𝐸0 is the initial elastic stiffness,
and 𝜎𝜎𝑡𝑡0 is the value of initial yield in tension.
Figure 5.5. Post-failure stress-displacement curve.
8
Figure 5.6. Post-failure stress-fracture energy curve.
Figure 5.7. Definition of the compressive inelastic (crushing) strain 𝜀𝜀𝑐𝑐~𝑖𝑖𝑖𝑖 used for the
definition of compression hardening data. 𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 is the equivalent
compressive plastic strain, 𝑑𝑑𝑐𝑐 is the compressive damage variable, 𝐸𝐸0 is
the initial elastic stiffness, 𝜎𝜎𝑐𝑐0 is the value of initial yield in compression,
and 𝜎𝜎𝑐𝑐𝑐𝑐 is the ultimate uniaxial stress.
Figure 5.8. Yield surface in plane stress corresponding to different values of the ratio
of the second stress invariant (𝐾𝐾𝑐𝑐 ) on the tensile meridian. CM and TM
indicate compressive and tensile meridians respectively.
Figure 5.9. Uniaxial tensile test conducted on concrete class CQD-G.
Figure 5.10. Uniaxial compressive test conducted on concrete class CQD-G.
Figure 5.11. Yield surface in plane stress. The values of 𝜎𝜎𝑏𝑏0 and 𝜎𝜎𝑐𝑐0 can be extracted
from this curve. 𝛼𝛼 is a material constant, 𝜎𝜎𝑐𝑐0 is the uniaxial compressive
strength, 𝜎𝜎𝑡𝑡0 is the uniaxial tensile strength, 𝑝𝑝̅ is the effective hydrostatic
pressure, and 𝑞𝑞� is the Mises equivalent effective stress.
Figure 5.12. Concrete specimen instance modeled in Abaqus.
Figure 5.13. Assembled parts as an instance (the whole concrete specimen) in Abaqus.
Figure 5.14. Loading system’s amplitudes applied on the model in Abaqus.
Figure 5.15. Loads configuration for the concrete specimen model in Abaqus.
Figure 5.16. Mesh generated for the concrete specimen model in Abaqus.
Figure 5.17. Strain energy density.
Figure 5.18. Scalar stiffness degradation.
Figure 5.19. Mises stress component.
Figure 5.20. Damage.
Figure 5.21. Max principal strain component.
Figure 5.22. Logarithmic max principal strain component.
Figure 5.23. Equivalent plastic strain.
Figure 5.24. Damage dissipation energy density.
Figure 5.25. Plastic dissipation energy density.
Figure 5.26. Displacement.
Figure 5.27. Magnitude of reaction forces.
Figure 5.28. Fracture pattern obtained from Abaqus FEA using constitutive
parameters identified in this research for CDP model for CQD-G
concrete specimen in comparison to actual fracture pattern obtained from
test for concrete specimens categorized in CQD-G.
9
LIST OF SYMBOLS AND ABBREVIATIONS
𝛽𝛽 A factor in Bineshian criterion that defines RUCS in terms of AUCS

𝜓𝜓2 Accordance coefficient
ACI American Concrete Institute
ASTM American Society for Testing and Materials
AUCS Apparent Unconfined Compressive Strength in Bineshian criterion
𝐶𝐶 Apparent Unconfined Compressive Strength in Bineshian criterion
bp Bedding planes
𝐹𝐹 CDP yield function
𝑅𝑅 2 Coefficients of determination
CAE Complete Abaqus Environment; Computer Aided Engineering
𝑑𝑑𝑐𝑐 Compressive damage variable
DAMAGEC Compressive damage variable, 𝑑𝑑𝑐𝑐
CDZ Compressive Damage Zone
PEEQ Compressive equivalent plastic strain, 𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝
CM Compressive meridian
𝑤𝑤𝑐𝑐 Compressive weight factor
CDP Concrete Damaged Plasticity
𝜆𝜆 Concrete/rock/coal constant in Bineshian strength criterion
𝜁𝜁 Concrete/rock/coal constant in Bineshian strength criterion
CQD Concrete Quality Designation
𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 Cracking strain
𝛼𝛼 Dimensionless material constant
𝛾𝛾 Dimensionless material constant
U Displacement
𝜎𝜎�𝑐𝑐 Effective compressive cohesion stress
𝑝𝑝̅ Effective hydrostatic pressure
𝜎𝜎�𝑡𝑡 Effective tensile cohesion stress
𝜎𝜎𝑐𝑐 ′ Effective unconfined compressive strength
ALLDMD Energy dissipated in the whole or partial model by damage
DMENER Energy dissipated per unit volume by damage
EDMDDEN Energy dissipated per unit volume in the element by damage
10
~pl
εc Equivalent compressive plastic strain
𝜀𝜀𝑐𝑐~̇ 𝑝𝑝𝑝𝑝 Equivalent compressive plastic strain rates
~𝑝𝑝𝑝𝑝
𝜀𝜀𝑡𝑡 Equivalent tensile plastic strain
~̇ 𝑝𝑝𝑝𝑝
𝜀𝜀𝑡𝑡 Equivalent tensile plastic strain rate
E Excellent quality concrete class in CQD classification
F Fair quality concrete class in CQD classification
FEA Finite Element Analysis
𝐺𝐺𝑓𝑓 Fracture energy
𝜑𝜑 Friction angle in degree
𝑠𝑠𝑐𝑐 Function of the compressive stress state
𝑠𝑠𝑡𝑡 Function of the tensile stress state
GF Gauge Factor
GB General-purpose Blended cement type in accordance with AS 3972
GP General-purpose Portland cement type in accordance with AS 3972
G Good quality concrete class in CQD classification
HE High Early strength cement type in accordance with AS 3972
𝜀𝜀𝑐𝑐~𝑖𝑖𝑖𝑖 Inelastic or crushing strain
𝐸𝐸0 Initial or undamaged elastic stiffness of concrete
𝐷𝐷0𝑒𝑒𝑒𝑒 Initial or undamaged elasticity matrix
𝑆𝑆 Internal cohesive strength
ISRM International Society for Rock Mechanics
LE Logarithmic maximum principal strain component
〈. 〉 Macauley bracket
RF Magnitude of reaction forces
mc Main cleats
𝜎𝜎1 ′ Major effective principal stress
𝜎𝜎1 Major principal stress at failure
𝜎𝜎�𝑚𝑚𝑚𝑚𝑚𝑚 Maximum principal effective stress
PE Maximum principal strain component
𝜎𝜎3 ′ Minor effective principal stress
𝜎𝜎3 Minor principal stress at failure
𝑞𝑞� Mises equivalent effective stress
S Mises stress component
𝑟𝑟(𝜎𝜎�) Multiaxial stress weight factor
11
𝜎𝜎𝑛𝑛 Normal stress at failure
𝜎𝜎1𝑒𝑒𝑒𝑒𝑒𝑒 Observed value of 𝜎𝜎1
𝑓𝑓𝑐𝑐 Peak stress
PENER Plastic dissipation energy density
PLT Point Load Test
P Poor quality concrete class in CQD classification
𝑓𝑓𝑖𝑖 Predefined field variables
𝜎𝜎�𝚤𝚤 Principal stress components
𝐾𝐾𝑐𝑐 Ratio of the second stress invariant on the tensile meridian
RUCS Regulated Unconfined Compressive Strength in Bineshian criterion
𝜎𝜎𝑐𝑐 Regulated Unconfined Compressive Strength in Bineshian criterion
𝑑𝑑 Scalar stiffness degradation variable
𝜏𝜏 Shear stress at failure
𝐵𝐵 Slope of Mohr-Coulomb linear envelope
𝜆𝜆 Slope of the complete force–displacement curve
𝑙𝑙0 Specimen length
SDEG Stiffness degradation variable, 𝑑𝑑
𝐾𝐾 Stiffness of loading machine
𝜀𝜀 Strain
𝜎𝜎1𝑐𝑐𝑐𝑐𝑐𝑐 Strength value calculated by failure criterion
𝜎𝜎 Stress
𝜃𝜃 Temperature
DAMAGET Tensile damage variable, 𝑑𝑑𝑡𝑡
𝑑𝑑𝑡𝑡 Tensile damage variable; post-peak tensile softening
~𝑝𝑝𝑝𝑝
PEEQT Tensile equivalent plastic strain, 𝜀𝜀𝑡𝑡
TM Tensile meridian
𝑤𝑤𝑡𝑡 Tensile weight factor
ELSE The recoverable part of the energy in the element
ALLSE The recoverable part of the energy in the whole (partial) model
SENER The recoverable part of the energy per unit volume
ESEDEN The recoverable part of the energy per unit volume in the element
TML Tokyo Sokki Kenkyujo
ELDMD Total energy dissipated in the element by damage
ALLIE Total strain energy
12
𝜎𝜎𝑡𝑡𝑡𝑡 Ultimate tensile stress
UCS Uniaxial Compressive Strength
𝑓𝑓𝑐𝑐′ Uniaxial compressive strength of concrete at 28 days
𝐶𝐶 Uniaxial compressive strength in Mohr-Coulomb criterion
𝜎𝜎𝑡𝑡 Uniaxial tensile strength
UTS Uniaxial tensile strength
𝑟𝑟 ∗ (𝜎𝜎11 ) Unit step function
𝜎𝜎𝑐𝑐0 Value of initial yield
𝜎𝜎𝑡𝑡0 Value of tensile failure stress
VG Very Good quality concrete class in CQD classification
VP Very Poor quality concrete class in CQD classification
W/C Water-cement ratio
13
LIST OF TABLES
Table 2.1. Failure criteria.

Table 3.1. Concrete mix proportion and design.
Table 3.2. Strain gauges characteristics.
Table 3.3. Triaxial testing results for concrete specimens – Batch # 1.
Table 3.13. Legend for Concrete Quality Designation introduced in this research.
Table 4.1. Evaluation of correlation between criteria and triaxial data for intact
rocks and coals.
Table 4.2. Accordance of strength criteria with actual triaxial data for different
types of limestone.
Table 4.3. Comparison of uniaxial tensile strengths calculated by the criteria with
uniaxial tensile strengths obtained from Brazilian testing on coal by
Hobbs (Sheorey, 1997; Bineshian, 2000).
Table 4.4. Parameters for Bineshian criterion applicable to intact igneous rocks.
Table 4.5. Parameters for Bineshian criterion applicable to intact sedimentary rocks.
Table 4.6. Parameters for Bineshian criterion applicable to intact metamorphic
rocks.
Table 4.7. Parameters for Bineshian criterion applicable to coals.
Table 4.8. Evaluation of correlation of the proposed criterion and triaxial data for a
variety of concrete.
Table 4.9. Accordance of the proposed criterion with actual triaxial data for
different types of concrete.
14
Table 4.10. Comparison between UCS obtained by uniaxial compression test, AUCS
(𝐶𝐶) measured using Mohr-Coulomb criterion, and RUCS (𝜎𝜎𝑐𝑐 ) measured
using proposed criterion.
Table 4.11. Comparison between UTS obtained by splitting tensile strength test and
𝜎𝜎𝑡𝑡 measured using proposed criterion.
Table 4.12. Difference between observed values by experiments and calculated
values by proposed strength criterion for concrete with CQD-VP,
UCS<15 MPa.
values by proposed strength criterion for concrete with CQD-P, UCS=15
MPa.
values by proposed strength criterion for concrete with CQD-P, UCS=20
MPa.
values by proposed strength criterion for concrete with CQD-F, UCS=30
MPa.
values by proposed strength criterion for concrete with CQD-F, UCS=40
MPa.
values by proposed strength criterion for concrete with CQD-G, UCS=55
MPa.
values by proposed strength criterion for concrete with CQD-G, UCS=70
MPa.
values by proposed strength criterion for concrete with CQD-VG,
UCS=100 MPa.
values by proposed strength criterion for concrete with CQD-VG,
UCS=130 MPa.
values by proposed strength criterion for concrete with CQD-E,
UCS>130 MPa.
15
Table 4.22. Constant values for proposed criterion’s parameters.
Table 5.1. Constitutive material parameters for concrete class CQD-G, to be used in
Concrete Damaged Plasticity (CDP) model.
Table 5.2. Concrete compression hardening.
Table 5.3. Concrete compression damage.
Table 5.4. Concrete tension stiffening.
Table 5.5. Concrete tension damage.
16
ACKNOWLEDGEMENT
I would like to express my gratitude to my supervisors Prof. Arcady V. Dyskin from

UWA’s School of Civil and Resource Engineering and Prof. Elena Pasternak from
UWA’s School of Mechanical and Chemical Engineering for the useful comments,
remarks, and engagement through the learning process of this master thesis.
Furthermore I would like to thank my loved ones, who have supported me throughout
entire process, both by keeping me harmonious and helping me putting pieces together.
I will be grateful forever for your love.
17
CHAPTER 1:
INTRODUCTION
1.1. Preface
Mechanical response of concrete for both failure and post-failure stages is the subject of
this research. Mechanical behaviour of concrete consists of two main stages: pre-peak,
and post-peak. Pre-peak stage of mechanical response of heterogeneous and
approximately isotropic materials such as concrete and some rocks in stress–strain co-
ordinates contains elastic, elastoplastic, and brittle regions until the specimen fails at
peak stress. Peak stress is the ultimate strength of concrete. Post-peak stage of
mechanical response of aforesaid materials can be seen in four different circumstances
of post-failure regime; strain hardening, ductile behaviour, strain softening, and brittle
behaviour (Cook, 1965; Hobbs, 1971; Gobbi and Ferrara, 1995; Geel, 1998; Crowder
and Coulson, 2006; Jaeger et al., 2007).
In the pre-peak stage, a very short initial non-linear behaviour is seen that relates to
closing process of existing voids and fissures; however it is not a plastic behaviour as it
is reversible. Major mechanical response in pre-peak stage is a linear part of stress–
strain curve followed by a short brittle, ductile, or elastoplastic stage until the stress
reached the ultimate strength. In triaxial loading, increasing the strength with increasing
the stress that is accompanied by plastic deformation is called work hardening, which
mechanical behaviour of the material in this state depends upon the confining pressure.
Increasing the deformation in almost constant level of stress is called ductile behaviour
that can be seen in ductile concrete under uniaxial compression. Strain softening of
concrete is the decline of stress at increasing strain. Brittle or very brittle behaviour
happens at exactly ultimate stress, which causes a violent failure that no more strain can
be measured after the peak stress (Nishihara, 1957; Rummel and Fairhurst, 1970;
Sangha, 1972; Vonk, 1992; Choi et al., 1996; van Mier et al., 1997; Torrenti, 1986;
Shang and Song, 2006; Carpinteri and Brighenti, 2010).
Study of both failure and post-failure responses of concrete by experiments can be

possible using load control testing machine for pre-peak stage and displacement control
loading machine for post-failure stage. Stress–strain curve should be used for pre-peak
18
region and stress–displacement curve should be used for post-peak region (Hudson et al.,
1972; Hustrulid and Robinson, 1972; Sangha and Dhir, 1972; Timco and Frederking,
1984; Chuan and Xiaohe, 1990; Yamaguchi and Chen, 1990; van Vliet and van Mier,
1996; van Mier and Ulfkjaer, 2000; Watanabe et al., 2004).
Factors influencing the post-peak mechanical response of concrete are very important;
however study of their effects is not the main goal of this research. These factors are the
relation between the stiffness of the loading machine and post-peak stiffness of concrete,
composition and geometry of specimen, boundary condition in loading process, loading
rate, allowable rotations of the loading platens, gauge length, and, the type of the
feedback signal of the data logger. Some of these factors affect the pre-peak behaviour;
however the effect of stiffness and geometry in pre-peak response is not significant. All
these factors should be taken into account in laboratory testing for studying concrete’s
mechanical response to compression and tension (Kesler, 1959; Symon, 1970; Bordia,
1971; Dhir et al., 1972; Gonnerman, 1975; Labuz and Biolzi, 1991; Jansen and Shah,
1997; Gamino et al., 2004; Nemecek and Bittnar, 2004; del Viso et al., 2008).
Mechanical response of concrete to compression and tension in triaxial loading

condition can be defined in terms of major and minor principal stresses at failure, which
can be described by failure/strength envelopes. Failure criteria are mathematical
functions that provide failure/strength envelopes for both compressive and tensile
quadrants in triaxial loading. These mathematical functions can be obtained by
analytical investigations – these types of criteria are called theoretical failure criteria, or
can be obtained by analyzing experimental data – these are called empirical criteria.
Failure criteria can be used to predict failure of concrete (Coulomb, 1776; Mohr, 1900;
Griffith, 1921; Balmer, 1952; Murrell, 1963; Fairhurst, 1964; Bieniawski, 1974; Hoek
and Brown, 1980; Ramamurthy et al., 1985; Bineshian, 2000).
Analytical and numerical methods are widely used to obtain post failure behaviour of
concrete and therefore some constitutive models are developed by researchers to
simulate concrete’s post-peak mechanical response to compression and tension. Most
applicable constitutive models are the Series Copping Model, Concrete Damage Zone
(CDZ) model, and Concrete Damaged Plasticity (CDP) model. Finite element analysis
can be used in simulations for abovementioned failure criteria and constitutive models
for both failure and post-failure stages of mechanical response of concrete to tensile and
19
compressive loading. Abaqus FEA is one of the most applicable software in this field
for concrete (Hillerborg et al., 1976; Hillerborg, 1988; Bazant, 1989; Lubliner et al.,
1989; Markeset, 1995; Lee and Fenves, 1998; Tang et al., 2000; Abaqus Theory Manual,
2002; Abaqus/CAE User’s Manual, 2011).
This thesis presents results obtained from the study of mechanical response of concrete
in tensile and compressive loading for both pre-peak and post-peak stages of complete
stress–strain curve of concrete. It aims to develop a failure criterion to model failure
behaviour of concrete for both tensile and compressive quadrants using analytical and
empirical procedures, simulating post-failure behaviour of concrete by numerical
models using finite elements methods to propose new constitutive parameters for CDP
model to be applicable in design of concrete structures, and suggestion of an empirical
concrete classification to be used in practical purposes.
The proposed failure criterion together with the proposed constitutive parameters for
Concrete Damaged Plasticity (CDP) model, and the suggested empirical classification
entitled “Concrete Quality Designation (CQD)” in terms of uniaxial compressive
strength are the major outcomes of the present dissertation that can be used in analytical,
numerical, and practical works in the field of concrete mechanics and technology.
1.2. Objectives
The main aims of the present research in the field of mechanical response of concrete
are:
• Developing a new comprehensive strength criterion for different types of

concrete applicable for estimating the strength for both tensile and compressive
quadrants,
• Providing practical constant values for the parameters of the proposed strength
criterion,
• Developing a new empirical concrete quality classification in terms of uniaxial

compressive strength to be applicable in practice, and finally
20
• Proposing new constitutive parameters for a Concrete Damaged Plasticity model
using numerical, analytical, and experimental procedures to simulate failure and
post-failure responses of concrete to be applicable in design of concrete
structures.
To obtain the goals, following jobs including study and laboratory experiments are
conducted:
• A comprehensive literature review of mechanical properties and behaviour of

concrete, rocks, coals, and soils, especially on the failure/strength criteria,
constitutive models, post-peak strain softening, concrete technology, and
numerical modelling and simulation of fracture and failure of concrete,
• Concrete batching, composition and mix design, and preparation in accordance

with ASTM C192/C192M-05, ASTM C143-78, and ACI 211.1,
• Concrete specimens manufacturing and ends preparation in accordance with

ASTM C617-10,
• Uniaxial compression tests on prepared concrete specimens in accordance with

ASTM C39/C39M-04,
• Splitting tensile strength tests on prepared concrete specimens in accordance

with ASTM C496/C496M-11,
• Brazilian tests on prepared concrete and rock specimens in accordance with

ASTM D3967-08,
• Triaxial compression tests on prepared concrete specimens at desired confining

pressures in accordance with ASTM C801-98,
• Mathematical study, practice and modelling of the proposed strength criterion’s

function,
21
• Regression analysis on obtained triaxial data pairs from uniaxial/triaxial
compression and splitting/Brazilian tensile tests,
• Calculations for suggested constant values for proposed strength criterion’s

parameters,
• Preparing the suggested empirical classification for different concrete types in

terms of their uniaxial compressive strength based on practical experience of
author, and
• Finite element study to find out suitable constitutive model and its parameters to
simulate prepared and tested concrete specimens’ failure and softening
behaviour.
Conventional strain measurements using foil strain gauges are used for uniaxial
compression tests.
The next section provides the research plan’s sectors briefly, which made this research
possible.
1.3. Methodology
Research plan for present study contains four main sectors that have been conducted
during Feb 2010 to July 2011:
• Pure study sector including literature review,
• Analytical and statistical investigations,
• Experimental sector including laboratory studies and field experience, and
• Numerical modelling and simulations using FEA software.
22
1.4. Thesis Structure
The thesis is organized as follows:
• Chapter 1 (the current chapter) is an introduction to the whole thesis containing

objectives and the methodology applied in the research plan.
• Chapter 2 contains literature review conducted in this research including two

sections; pre-peak response and post-peak response of concrete to tension and
compression. An overview of failure criteria covering both tensile and
compressive quadrants with linear and non-linear envelopes is presented. Effects
of stiffness of specimen and loading system, and influence of geometry
including size, shape, and slenderness ratios and boundary restraint of loading
platens and specimen’s ends on strain softening are presented in this chapter as
well. A study on strain localization phenomenon as a structural response is also
considered in this chapter. Chapter 2 finally presents damaged plasticity
considering three different appropriate constitutive models for damage
behaviour of concrete.
• Main parts of experimental and empirical investigation conducted in this

research are presented in Chapter 3. It includes concrete batch’s composition and
mix design, specimen manufacturing and ends preparation, strain measurements,
test plan, and test results. All specimens are prepared and tested in this research
in accordance with American Society for Testing and Material standards.
Concrete Quality Designation in abbreviation form of CQD as an applicable
empirical classification for concrete in terms of uniaxial compressive strength is
introduced in Chapter 3.
• New strength criterion for concrete is introduced in Chapter 4. Its performance

in prediction of tensile and compressive strength using statistical methods is
verified in this chapter. Parameters of proposed strength criterion for different
types of concrete from very poor to excellent quality concrete are calculated and
suggested in this chapter as a guide to its constant values to be used in practical
purposes.
23
• Chapter 5 presents the numerical study conducted in this research to simulate
failure and post-peak response of concrete by Concrete Damaged Plasticity
(CDP) constitutive model using Abaqus FEA. The suggested constitutive
parameters for CDP behavioral model and the methodology followed by
interpretation of results of numerical simulation conducted using Abaqus are
provided in this chapter too.
• Conclusions from all sectors of this research including experimental, empirical,

analytical, and numerical works conducted in this research are presented in
Chapter 6. This chapter also contains a section entitled “Future Research”,
which includes recommendations for future study and complementary research
on subjects similar to present dissertation. This section mainly contains
recommendations on improvements to the empirical classification called
Concrete Quality Designation that is introduced in this research by practical
experience, further experimental works to obtain more precise constant values
for the proposed strength criterion, further investigations on transition limit from
brittle to ductile behaviour of concrete types, numerical crack propagation study
of failure and strain softening of concrete, and finally calculation of Concrete
Damaged Plasticity modeling parameters for all classes of concrete that can be
done using Abaqus FEA.
24
CHAPTER 2:
MECHANICAL BEHAVIOUR OF CONCRETE
2.1. Introduction
Relationship between stress and strain for concrete under compression or tension is
called mechanical behaviour of concrete (Kiendle and Maldari, 1938). A complete
stress–strain curve of concrete under uniaxial compression (Figure 2.1) can be split to
two main portions; pre-peak region and post-peak region. The first complete
compressive stress–strain curve is obtained by Kiendle and Maldari (1938) and Brace
(1964) for concrete and rocks respectively.
Pre-peak Post-peak
Response Response
Ductile Behaviour
𝜎𝜎
Brittle
Behaviour
𝜀𝜀
Figure 2.1. Complete stress–strain diagram for concrete consisting pre-peak and post-peak responses.
Pre-peak region of complete stress–strain curve of concrete is related to the mechanical

response of concrete before failure (ultimate/peak stress), which normally consists of
three parts; plastic, elastic, and brittle. This part of curve is also called ascending
branch, strength zone, and unfailing regime. Typically, the initial short length non-linear
part does not occur and only the elastic and brittle parts can be observed. The initial
linear part of pre-peak region of complete 𝜎𝜎 − 𝜀𝜀 response of concrete is seen until the
stress level is reached to about 35% of ultimate uniaxial stress. Afterward until the peak
loading, non-linear behaviour is observed. The reason for the non-linear behaviour is
growth and possible coalescence of micro-cracks at the boundary of matrix-aggregates
of concrete (see Section 2.2). Failure occurs when a network of major cracks is formed.
25
At this state the peak stress is reached while micro-cracks are merged and a critical flaw
in the matrix is propagated. The pre-peak region can be obtained by any load-control
testing machine (Figure 2.2). In load-control machines, a constant rate of load is applied
to the specimen and if the load exceeds the peak level, catastrophic failure occurs such
that its violent sound can be heard.
Load-Control
Force
Displacement-Control
Displacement
Figure 2.2. Load-control and displacement-control modes in compression testing. Complete 𝜎𝜎 − 𝜀𝜀 curve
is obtained only in displacement-control mode.
The post-peak stage is the mechanical response of concrete after failure. In the complete
stress–strain curve it can be seen after the peak stress and is called descending branch,
damage zone, and failing regime. The post-peak behaviour of concrete is directly
affected by the factors such as loading frame stiffness in relation to the concrete
specimen post-peak stiffness, the strain rate applied by the loading frame, and the
specimen geometry.
The descending branch of 𝜎𝜎 − 𝜀𝜀 curve often manifests itself in strain localization. Post-
peak region can be obtained by a displacement-control testing machine, which applies
small increments of displacement to concrete specimen (Figure 2.2). Strain softening
occurs when micro-cracks, which begin forming during the pre-peak portion of the
stress–strain curve, coalesce to form a zone of damage (fracture, fault, shear band or
rupture zone) that weakening the concrete so its load-carrying capacity is diminished
and therefore additional deformation of the zone of damage weakens it further and
continued softening occurs (Jansen and Shah, 1997).
26
2.2. Pre-peak Response and Failure
Solid line in Figure 2.3 represents the pre-peak response for a concrete specimen under
uniaxial compressive loading. A load-control mode could be used to record the pre-peak
response of concrete; however, a displacement-control loading machine should be used
to obtain a complete 𝜎𝜎 − 𝜀𝜀 curve. As can be seen in Figure 2.3 the micro-cracks and
fissures are coalesced at the boundary of the aggregates and cement. It happens at 35%
of peak stress. When the load is increased to 50% of peak stress, cracks are developed
and propagate toward the matrix. By increasing the load level to about 75% of peak
stress, cracks are merged and major flaws are occurred. At the peak stress a network of
major flaws are occurred that cause failure. In this state, the specimen is not able to
carry load anymore and in displacement-control mode its strain softening can be
recorded.
Intact 35% 50% 75% 100%

Concrete Peak Stress Peak Stress Peak Stress Peak Stress
Ultimate Stress
(Failure happens at this level of stress)
%75 Ultimate Stress

𝜎𝜎
%50 Ultimate Stress
%35 Ultimate Stress
𝜀𝜀
Figure 2.3. Failure process on concrete under uniaxial compression including flaw initiation and finally
coalescence of network of cracks causing failure (Bineshian, 2000).
27
When the specimen is under triaxial compression a similar mechanical behaviour can be
seen in 𝜎𝜎 − 𝜀𝜀 co-ordinates; however, study of failure behaviour in major and minor
principal stresses co-ordinates 𝜎𝜎1 − 𝜎𝜎3, or in shear strength and normal stress co-
ordinates 𝜏𝜏 − 𝜎𝜎𝑛𝑛 , are more applicable.
Under triaxial loading the peak stress is replaced with failure criteria, which are surfaces
in 𝜎𝜎1 and 𝜎𝜎3 or 𝜏𝜏 and 𝜎𝜎𝑛𝑛 spaces. The first failure criterion was developed by Coulomb in
1776 (Jaeger et al., 2007). It was based on Coulomb’s (1776) friction hypothesis. After
that many functional dependences were tried to find the most accurate model to estimate
the failure behaviour. Figure 2.4 shows an example of failure envelope obtained by two
failure criterion using actual triaxial data.
210
Bineshian Criterion (2000)
𝝈𝝈𝒄𝒄+𝝀𝝀𝝈𝝈𝟑𝟑 Non-linear Envelope

𝝈𝝈𝟏𝟏=𝝈𝝈𝟑𝟑+𝝈𝝈𝒄𝒄 ( )
𝝈𝝈𝒄𝒄+𝜻𝜻𝝈𝝈𝟑𝟑
Linear Envelope
σ1 (MPa)
140
Tensile Compressive
Quadrant Quadrant
70
Bineshian Criterion R^2 = 0.996
Mohr-Coulomb R^2 = 0.893
0
-15 -10 -5 0 5 10 15 20
σ3 (MPa)
Figure 2.4. Failure envelops for actual triaxial data (Source of triaxial data is Johnston, 1985).
Table 2.1 shows a brief history of failure/strength criteria. Most of them are developed
for rocks, coals, and soils, but can be used for concrete as well. Some of these criteria
provide linear strength envelope and some of them provide non-linear envelope for
triaxial test data. Some of them are applicable for compression only and some of them
can be applied for both types of stress (compression and tension). It also should be
noted that some of these criteria have limitations that restrict their comprehensiveness.
28
Table 2.1. Failure criteria.
Failure Criteria Mathematical Function Failure Envelope Stress Quadrants
Linear Non-linear Tensile Compressive
𝜎𝜎1 − 𝜎𝜎3 𝜎𝜎1 + 𝜎𝜎3
Coulomb (1776) = �𝐶𝐶 cot 𝜑𝜑 + � sin 𝜑𝜑    
2 2
Mohr-Coulomb (1900) 𝜎𝜎1 = 𝐶𝐶 + 𝐵𝐵𝜎𝜎3    
Griffith (1921) (𝜎𝜎1 − 𝜎𝜎3 )2 = 8𝜎𝜎𝑡𝑡 (𝜎𝜎1 + 𝜎𝜎3 )    
𝜎𝜎3 𝑏𝑏
Balmer (1952) 𝜎𝜎1 = 𝜎𝜎𝑐𝑐 (1 + )    
𝜎𝜎𝑡𝑡
1 1
Price (1960) 𝜎𝜎1 = (𝜎𝜎 − 𝐶𝐶)𝐵𝐵    
𝐴𝐴 3
Fairhurst (1964) (𝜎𝜎1 − 𝜎𝜎3 )2 = 𝑎𝑎 + 𝑏𝑏(𝜎𝜎1 + 𝜎𝜎3 )    
Hobbs (1964) 𝜎𝜎1 = 𝜎𝜎𝑐𝑐 + 𝜎𝜎3 + 𝐵𝐵𝜎𝜎3 𝛼𝛼    
Murrell (1965) 𝜎𝜎1 = 𝜎𝜎𝑐𝑐 + 𝐵𝐵𝜎𝜎3 𝛼𝛼    
Mogi (1966) 𝜎𝜎1 − 𝜎𝜎3 = 𝜎𝜎𝑐𝑐 + 𝐵𝐵𝜎𝜎3 𝛼𝛼    
𝜏𝜏 − 𝐶𝐶 𝜎𝜎𝑛𝑛
Hoek (1968) = 𝐷𝐷( )𝛼𝛼    
𝜎𝜎𝑐𝑐 𝜎𝜎𝑐𝑐
Hobbs (1970) 𝜏𝜏 = 𝐶𝐶 + 𝑘𝑘𝜎𝜎𝑛𝑛 𝛼𝛼    
Hobbs (1971) 𝜎𝜎1 = 𝜎𝜎𝑐𝑐 + 𝑐𝑐𝜎𝜎3 𝑑𝑑    
Franklin (1971) 𝜎𝜎1 − 𝜎𝜎3 = 𝜎𝜎𝑐𝑐 1−𝐵𝐵 (𝜎𝜎1 − 𝜎𝜎3 )2    
𝜎𝜎1 − 𝜎𝜎3 𝜎𝜎3

Ohnaka (1973) = 1 + 𝑏𝑏( )𝛼𝛼    
𝜎𝜎1 𝜎𝜎3 𝛼𝛼
Bieniawski (1974) = 1 + 𝐵𝐵 � �    
𝜎𝜎1 − 𝜎𝜎3 𝜎𝜎1 + 𝜎𝜎3
Panek (1979) = 𝑎𝑎 + 𝑑𝑑( )    
2 2
Brook (1979) 𝜏𝜏⁄𝜎𝜎𝑐𝑐 = 𝐵𝐵(𝜎𝜎𝑛𝑛 ⁄𝜎𝜎𝑐𝑐 )𝛼𝛼    
0.5
Hoek-Brown (1980) 𝜎𝜎1′ 𝜎𝜎3′ 𝜎𝜎3′    
= ′ + �𝑚𝑚 ′ + 𝑆𝑆�
′
𝜎𝜎𝑐𝑐 𝜎𝜎𝑐𝑐 𝜎𝜎𝑐𝑐
𝐵𝐵
𝜎𝜎1′ 𝑀𝑀 𝜎𝜎3′
Johnston-Chiu (1984) = �1 + �    
′
𝜎𝜎𝑐𝑐 𝐵𝐵 𝜎𝜎𝑐𝑐′
𝜎𝜎1 − 𝜎𝜎3 𝜎𝜎𝑐𝑐 𝛼𝛼
Ramamurthy et al. (1985) = 𝐵𝐵 � �    
𝜎𝜎3 𝜎𝜎3
Kwasniewski (1987) 𝜎𝜎1 = 𝜎𝜎𝑐𝑐 + 𝑑𝑑𝜎𝜎3    
𝑏𝑏
Sheorey et al (1989) 𝜎𝜎1 = 𝜎𝜎𝑐𝑐 �1 − (𝜎𝜎3 ⁄𝜎𝜎𝑡𝑡 )�    
𝐵𝐵
Yoshida et al. (1990) 𝜎𝜎1 = 𝜎𝜎3 + 𝐴𝐴𝜎𝜎𝑐𝑐 �𝑆𝑆 + (𝜎𝜎3 ⁄𝜎𝜎𝑐𝑐 )�    
0.5
Carter et al. (1991) mσc σ3 + σc 2    
σ1 = σ3 + � �
1 + w(σ3 ⁄σc )
𝜎𝜎𝑐𝑐 + 𝜆𝜆𝜎𝜎3
Bineshian (2000) 𝜎𝜎1 = 𝜎𝜎3 + 𝜎𝜎𝑐𝑐 � �    
𝜎𝜎𝑐𝑐 + 𝜁𝜁𝜎𝜎3
All other parameters presented in this table are criteria’s parameters that can be determined by regression analysis.
𝜎𝜎1 and 𝜎𝜎3 are the major and minor principal stresses at failure and 𝜎𝜎1 ′ and 𝜎𝜎3 ′ are the major and minor effective principal stresses.
𝜏𝜏 and 𝜎𝜎𝑛𝑛 are shear and normal stresses at failure.
 Applicable
 Not applicable
29
2.3. Post-peak Response
Post-peak region of complete 𝜎𝜎 − 𝜀𝜀 curve represents the softening behaviour (Crowder

and Coulson, 2006), which is a mixture of material and structural behaviour. The main
parameters affecting the compressive softening of concrete are (van Mier et al., 1997):
• Stiffness of the testing machine,

• Stiffness and the composition of concrete,
• Geometry of the specimen,
• Boundary condition between the loading platens and the specimen,
• Loading rate,
• Allowable rotations of the loading platens,
• The strain gauge length, and
• Type of the feedback signal.
2.3.1. Stiffness
Stiffness of the loading machine has a significant effect on the stress–strain curve of
concrete specimen. By increasing the stiffness of testing machine, reducing the violence
of fracture of the specimen is possible. Behaviour of concrete-like material in the post-
failure region depends upon the relative stiffness of the loading system and specimen. If
the stiffness of the loading system is lower than that of the specimen, the specimen will
fail violently, instead if the stiffness of the loading system is greater than that of the
specimen, failure will occur in a quasi-stable manner with the decreasing of the
specimen strength in the damage regime more or less linearly with increasing
displacement. A stiff loading machine can be a hydraulic, hybrid, or a very fast response
servo controlled loading machine. The condition for controlled fracture is determined by
the ability of the hydraulic loading system to unload rapidly. Failure during the
compression test with flexible loading frame is violent but fracture occurs during the
test with the stiff loading frame is almost non-violent. Almost all the energy stored by
the stiff machine was absorbed in non-elastic deformation and fracturing of specimen
(Cook, 1965; Rummel and Fairhurst, 1970; Salamon, 1970; Hustrulid and Robinson,
1972).
30
Stability of a compression test of a brittle specimen depends on the relation between the
stiffness of the loading machine (𝐾𝐾) and the slope of the complete force–displacement
curve of the specimen (𝜆𝜆; It is positive in unfailed regime and negative in the failing
regime). If 𝐾𝐾 + 𝜆𝜆 > 0 therefore complete 𝜎𝜎 − 𝜀𝜀 curve of the specimen in compression
test can be recorded (Salamon, 1970).
In other words, the complete process of specimen collapse can be controlled if the
longitudinal stiffness of the testing machine including load frame, actuator, load cell,
compression platens etc. is always greater than the absolute value of the slope of the
complete force–axial displacement curve for the specimen during the failure process.
The presence of any elastic member in a testing system reduces the effective stiffness of
the system, which is always less than the stiffness of any single component, Figure 2.5
(Hudson et al., 1972; Timco and Frederking, 1984; Schubert and Blumel, 2006).
Uncontrolled Failure
Force
Displacement
Controlled Failure
Force
Displacement
Figure 2.5. Stability of machine–specimen system considering stiffness of machine, and geometry of
specimen (Hudson et al., 1972).
A testing machine can also be stiffened by reducing the height of load column
component and increasing the modulus of elements and platens area (Chuan and
Xiaohe, 1990; Abdullah, 2006).
31
2.3.2. Geometry and Boundary Restraint
While there is no influence of the specimen shape on the ascending branch, considerable
effect on the descending branch is seen (Hustrulid and Robinson, 1972).
Strain softening behaviour is highly affected by the specimen geometry including size,
shape and slenderness ratio and boundary condition. Higher strength will obtain by
increasing the end friction and decreasing the slenderness ratio of the specimen.
Slenderness ratio of the specimen has a significant effect (increasing the brittle
behaviour with increasing the slenderness ratio) on the descending branch of complete
𝜎𝜎 − 𝜀𝜀 curve (van Mier and Vonk, 1991; Vonk, 1992; van Vliet and van Mier, 1996;
Jansen and Shah, 1997; van Mier, 1998; Gamino et al., 2004; del Viso et al., 2008).
Nemecek and Bittnar (2004) stated that there is no significant effect of specimen size on
ascending branch of stress–strain curve and nominal strength of materials like concrete,
while other researchers believe that a strong increase of uniaxial compressive strength is
observed when the specimen’s size decreases, or in other words at a constant diameter,
the more slender the specimen is, the lower the strength value is. The difference is
straightforward; decreasing the specimen height results in an increase of ductility, i.e. a
decreasing slope in descending branch of the stress–strain curve; however, in pre-peak,
the curves are almost identical, independent of specimen size (Hudson et al., 1972;
Sangha and Dhir, 1972; van Mier, 1986; van Mier and Vonk, 1991; Gobbi and Ferrara,
1995; Jansen and Shah, 1997; Pellegrino et al., 1997; van Mier et al., 1997; Gamino et
al., 2004).
It also should be noted that slenderness ratio 2 commonly used for the testing of
concrete cylinders. As the maximum size of the aggregate increases, lower ratios of
specimen diameter/maximum size of aggregate are necessary in order to obtain
maximum strength for a given concrete mix (Nishihara, 1957; Kesler, 1959; Symons,
1970; Dhir et al., 1972; Sangha, 1972; Dhir and Sangha, 1973; Gonnerman, 1975).
Pre-peak portion of 𝜎𝜎 − 𝜀𝜀 curves is independent of specimen slenderness ratio when

low friction loading platens were used meanwhile elastic energy stored during pre-peak
phase of loading increases proportionally to the specimen length (van Mier and Vonk,
32
1991; Jansen and Shah, 1997; van Mier et al., 1997; Nemecek and Bittnar, 2004; del
Viso et al., 2008).
Softening curves become less steep as the slenderness ratio decreased (Hustrulid and
Robinson, 1972; van Mier, 1984; Torrenti, 1986; Hillerborg, 1988; Yamaguchi and
Chen, 1990; van Mier and Vonk, 1991; Rokugo and Koyanagi, 1992; Vonk, 1992; Choi
et al., 1996; Jansen and Shah, 1997; del Viso et al., 2008).
There is considerable size effect on the descending branch in terms of the energy
release, Figure 2.6. Meanwhile strain softening’s stress–displacement curves calculated
from displacement of loading platens are almost identical and therefore independent of
the slenderness ratio (Sangha and Dhir, 1972; van Mier, 1984; Yamaguchi and Chen,
1990; van Mier and Vonk, 1991; van Mier and Ulfkjaer, 2000; del Viso et al., 2008).
Force
Small Structure
Large Structure
Displacement
Figure 2.6. Typical force–displacement diagrams of large/small structure (Jansen and Shah, 1997;
Gamino et al., 2004; Nemecek and Bittnar, 2004).
Softening branch is a structural property, depending on both the specimen configuration

and the applied loading as the differences in post-peak response for specimens with
different axial dimensions almost disappear completely and in fact, localized failure in
uniaxial compression is demonstrated. The size effect is applicable for both uniaxial and
triaxial tests, Figure 2.7 (van Mier, 1986).
In pre-peak regime, 𝜎𝜎 − 𝜀𝜀 diagrams should be used, whereas in post-peak regime where

localization of deformation appears, stress–deformation diagram should be used in the
spirit of fracture mechanics of concrete (van Mier, 1998).
33
1
Stress / Peak Stress
L = 50 mm
L = 100 mm
L = 200 mm
0
0 Displacement (mm) 0.8
Figure 2.7. Softening curves for specimens with different height (van Mier, 1986).
Size effect can be found in geometrically similar structures of different sizes. Specimen
geometry and boundary conditions will affect the strain softening behaviour; however,
post-peak fracture energy is essentially independent of the specimen length and this
independence confirms the occurrence of localization. For all loading systems a strong
increase of post-peak ductility was found with decreasing specimen slenderness ratio,
Figure 2.8 and 2.9 (Bordia, 1971; van Mier and Vonk, 1991; Vonk, 1992; Gobbi and
Ferrara, 1995; Jansen and Shah, 1997; van Mier et al., 1997; Geel, 1998; van Mier,
1998; Tang et al., 2000; van Mier and Ulfkjaer, 2000; Nemecek and Bittnar, 2004;
Watanabe et al., 2004; Schubert and Blumel, 2006; del Viso et al., 2008).
150
100
𝜎𝜎 (MPa)
L/D = 0.5
50
L/D = 1
L/D = 2
0
0 5 𝜀𝜀 (%) 10 15
Figure 2.8. Stress–strain curves obtained from cylinder tests on high strength concrete using high friction
steel platens (van Mier et al., 1997).
34
80
L/D = 0.5
𝜎𝜎 (MPa)
40
L/D = 1
L/D = 2
0
0 4 𝜀𝜀 (%) 8 12
Figure 2.9. 𝜎𝜎 − 𝜀𝜀 curves for normal strength concrete specimens of different geometry loaded between
rigid high friction steel platens (Gobbi and Ferrara, 1995; van Mier et al., 1997).
2.3.3. Strain Localization
During triaxial testing of brittle materials, numerous cracks were developed near the
specimen surface and propagated toward the middle of the specimen, then a non-
uniform micro-fracture field is generalized within the specimen and eventually leads to
the typical shear band that is generated, what is happen now is strain localization
(Yukutake, 1989; Abdulla and Kiousis, 1997). Due to this localization, fractured areas
are more deformed while un-fractured portions recover their deformations (Tang et al.,
2000).
The softening is found to be a highly localized phenomenon (van Mier, 1984; Markeset
and Hillerborg, 1995; van Mier and Ulfkjaer, 2000). This behaviour is called the strain
localization and is considered as being a structural response rather than a material
characteristic.
Because of localization, the post-peak portion of stress–strain curve is also dependent

on specimen size (as discussed in previous section). Most experimental data indicate
that localization occurs before or at maximum load, therefore strain softening of
concrete in uniaxial compression tests, appears after localization of damage (Vonk,
1992; van Vliet and van Mier, 1996; Jansen and Shah, 1997; Gamino et al., 2004), so
that the post-peak response is due to the deformation of a structural element,
consequently if the deformation has localized before or at the peak load, the post-peak
35
behaviour is not an essential property of the material, but rather a typical structural
response (Labuz and Biolzi, 1991; van Mier and Ulfkjaer, 2000; Bisby and Take, 2009).
Localization occurs for all boundary conditions, but the way in which it takes place
strongly depends on the amount of friction along the specimen–platen interface (van
Mier, 1984; Labuz and Biolzi, 1991).
With increasing the slenderness ratio the localization zone decreases in size, the cons
shaped inner parts can fully develop and the size of the specimen parts that are not
affected by the horizontal confinement (near the boundaries) increases. In the post-peak
regime, the concrete cannot be regarded as a continuum anymore, Figure 2.10 (van Vliet
and van Mier, 1996).
L = 50 mm L = 50 mm L = 100 mm L = 100 mm L = 200 mm L = 200 mm

High friction end Low friction end High friction end Low friction end High friction end Low friction end
Figure 2.10. Influence of end friction for concrete specimens under uniaxial compression (van Vliet and
van Mier, 1996).
2.3.4. Damage Plasticity
Different models for compressive softening and strain localization are developed
(Hillerborg, 1979; Bazant, 1989; Lubliner et al., 1989; Hillerborg, 1990; Markeset,
1995; Lee and Fenves, 1998; Bazant and Novak, 2000; Bazant, 2002).
Bazant (1989) classified post-peak behaviour into two response paths: one is the strain
softening localizing into a certain zone within the specimen and another one is the
unloading path that the rest of the specimen undergoes. He showed the entire specimen
response as a combination of localization and unloading response paths. He stated that
localization at strain softening in the model must begin right at the peak-stress state.
Figure 2.11 shows very close data fit for the Bazant’s model.
36
1
Series Coupling Model
L = 50 mm
L = 100 mm
Normalized Stress L= 200 mm
L = 400 mm
0
0 Post-Peak Overal Mean Strain 0.01
Figure 2.11. Bazant’s model fitting on 𝜎𝜎 − 𝜀𝜀 data of different slenderness ratio (Bazant, 1989 based on
van Mier’s (1984) test results).
Markeset and Hillerborg (1995) proposed a model – the Compressive Damage Zone
(CDZ) – that takes both the localized shear deformation and the deformation due to
splitting cracks into account. According to the CDZ model, the steepness of the
descending branch of a formal 𝜎𝜎 − 𝜀𝜀 curve will increase with increasing specimen
length and slenderness ratio.
Another model that is widely applicable in numerical simulations for concrete failure
behaviour is the Concrete Damaged Plasticity (CDP) constitutive model. The finite
element CDP model (e.g. in Abaqus) provides a general capability for modelling
concrete in all types of structures. It uses concepts of isotropic damage elasticity in
combination with isotropic tensile and compressive plasticity to represent the inelastic
behaviour of concrete. It can be used for plain concrete that is subject of present study.
The CDP consists of a combination of non-associated multi-hardening plasticity and

scalar (isotropic) damaged elasticity to describe the irreversible damage that occurs
during the fracturing process. It can be used in conjunction with a visco-plastic
regularization of the constitutive equations in Abaqus/Standard to improve the
convergence rate in the softening regime; however, it requires that the elastic behaviour
of the material be isotropic and linear. This section is mainly extracted from Hillerborg
et al. (1976), Lubliner et al. (1989), Lee and Fenves (1998), Abaqus theory Manual
(2002), and Abaqus/CAE User’s Manual (2011).
37
The CDP model is a continuum, plasticity-based, damage model for concrete. It
assumes that the main two failure mechanisms are tensile cracking and compressive
crushing of the concrete material. The evolution of the yield or failure surface is
~𝑝𝑝𝑝𝑝
controlled by two hardening variables, 𝜀𝜀𝑡𝑡 and 𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 , linked to failure mechanisms
under tension and compression loading. The model assumes that the uniaxial tensile and
compressive response of concrete is characterized by damaged plasticity (Figures 2.12
and 2.13).
Under uniaxial tension the 𝜎𝜎 − 𝜀𝜀 response follows a linear elastic relationship until the
value of the failure stress, 𝜎𝜎𝑡𝑡0 , is reached. The failure stress corresponds to the onset of
micro cracking in the concrete material. Beyond the failure stress, the formation of
micro-cracks is represented macroscopically with a softening 𝜎𝜎 − 𝜀𝜀 response, which
induces strain localization in the concrete structure.
Under uniaxial compression the response is linear until the value of initial yield, 𝜎𝜎𝑐𝑐0 . In
the plastic regime the response is typically characterized by stress hardening followed
by strain softening beyond the ultimate stress, 𝜎𝜎𝑡𝑡𝑡𝑡 . This representation, although
somewhat simplified, captures the main features of the response of concrete. It is
assumed that the uniaxial 𝜎𝜎 − 𝜀𝜀 curves can be converted into stress versus plastic-strain
curves. Thus,
~𝑝𝑝𝑝𝑝 ~̇ 𝑝𝑝𝑝𝑝
𝜎𝜎𝑡𝑡 = 𝜎𝜎𝑡𝑡 (𝜀𝜀𝑡𝑡 , 𝜀𝜀𝑡𝑡 , 𝜃𝜃, 𝑓𝑓𝑖𝑖 ) (2.1)
𝜎𝜎𝑐𝑐 = 𝜎𝜎𝑐𝑐 (𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 , 𝜀𝜀𝑐𝑐~̇ 𝑝𝑝𝑝𝑝 , 𝜃𝜃, 𝑓𝑓𝑖𝑖 ) (2.2)
~𝑝𝑝𝑝𝑝
where the subscripts 𝑡𝑡 and 𝑐𝑐 refer to tension and compression, respectively; 𝜀𝜀𝑡𝑡 and
𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 are the equivalent tensile and compressive plastic strains, 𝜀𝜀𝑡𝑡
and 𝜀𝜀𝑐𝑐~̇ 𝑝𝑝𝑝𝑝 are the
equivalent plastic strain rates, 𝜃𝜃 is the temperature, and 𝑓𝑓𝑖𝑖 is other predefined field
variables (𝑖𝑖 = 1,2, … ).
As shown in Figures 2.12 and 2.13, when the concrete specimen is unloaded from any
point on the strain softening branch of the 𝜎𝜎 − 𝜀𝜀 curves, the unloading response is
weakened: the elastic stiffness of the material appears to be damaged or degraded. The
degradation of the elastic stiffness is characterized by two damage variables, 𝑑𝑑𝑡𝑡 and 𝑑𝑑𝑐𝑐 :
38
~𝑝𝑝𝑝𝑝
𝑑𝑑𝑡𝑡 = 𝑑𝑑𝑡𝑡 (𝜀𝜀𝑡𝑡 , 𝜃𝜃, 𝑓𝑓𝑖𝑖 ) (2.3)
𝑑𝑑𝑐𝑐 = 𝑑𝑑𝑐𝑐 (𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 , 𝜃𝜃, 𝑓𝑓𝑖𝑖 ) (2.4)
where 𝑑𝑑𝑡𝑡 and 𝑑𝑑𝑐𝑐 , are functions of the plastic strains, temperature, and field variables.
The damage variables can take values from zero, representing the undamaged material,
to one, which represents total loss of strength (0 ≤ 𝑑𝑑𝑡𝑡 ≤ 1 and 0 ≤ 𝑑𝑑𝑐𝑐 ≤ 1).
The 𝜎𝜎 − 𝜀𝜀 relations under uniaxial tension and compression loading are, respectively:
~𝑝𝑝𝑝𝑝
𝜎𝜎𝑡𝑡 = (1 − 𝑑𝑑𝑡𝑡 )𝐸𝐸0 �𝜀𝜀𝑡𝑡 − 𝜀𝜀𝑡𝑡 � (2.5)
𝜎𝜎𝑐𝑐 = (1 − 𝑑𝑑𝑐𝑐 )𝐸𝐸0 �𝜀𝜀𝑐𝑐 − 𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 � (2.6)
where 𝐸𝐸0 is the initial or undamaged elastic stiffness of the material. Effective tensile
and compressive cohesion stresses can be defined as below (The effective cohesion
stresses determine the size of the yield or failure surface):
𝜎𝜎 ~𝑝𝑝𝑝𝑝
𝜎𝜎�𝑡𝑡 = (1−𝑑𝑑𝑡𝑡 ) = 𝐸𝐸0 �𝜀𝜀𝑡𝑡 − 𝜀𝜀𝑡𝑡 � (2.7)
𝑡𝑡
𝜎𝜎
𝜎𝜎�𝑐𝑐 = (1−𝑑𝑑𝑐𝑐 ) = 𝐸𝐸0 �𝜀𝜀𝑐𝑐 − 𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 � (2.8)
𝑐𝑐
𝜎𝜎𝑐𝑐 0 𝜎𝜎𝑐𝑐 u
𝜎𝜎𝑐𝑐
E0
(1−𝑑𝑑𝑐𝑐)E0
𝜀𝜀𝑐𝑐~𝑝𝑝𝑙𝑙 𝜀𝜀𝑐𝑐0𝑙𝑙 𝜀𝜀𝑐𝑐
39
𝜎𝜎t 0
𝜎𝜎t
E0
(1−𝑑𝑑t)E0
𝜀𝜀t~𝑝𝑝𝑙𝑙 𝜀𝜀t0𝑙𝑙 𝜀𝜀t
Figure 2.13. Mechanical response of concrete to uniaxial loading in tension.
Post-failure stress can be defined as a function of cracking strain, 𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 , which it is

defined as the total strain minus the elastic strain related to the undamaged material
(Figure 2.14):
𝑒𝑒𝑒𝑒
𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 = 𝜀𝜀𝑡𝑡 − 𝜀𝜀0𝑡𝑡 (2.9)
𝑒𝑒𝑒𝑒 𝜎𝜎
𝜀𝜀0𝑡𝑡 = 𝐸𝐸𝑡𝑡 (2.10)
0
𝜎𝜎t 0
𝜎𝜎t
E0
E0
(1−𝑑𝑑t)E0

𝜀𝜀t ~ck 𝜀𝜀t e𝑙𝑙
Figure 2.14. Definition of the cracking strain 𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 used for the definition of tension stiffening data for
CDP model.
40
The choice of the tension stiffening parameters is important since, generally, higher
tension stiffening makes it easier to obtain numerical solutions (Abaqus/CAE User’s
Manual, 2011). Too little tension stiffening will cause the local cracking failure in the
concrete to introduce temporarily unstable behaviour in the overall response of the
model. Few practical designs exhibit such behaviour, so that the presence of this type of
response in the analysis model usually indicates that the tension stiffening is
unreasonably low.
2.4. Conclusions
Mechanical behaviour of concrete is the relationship between 𝜎𝜎 and 𝜀𝜀 when concrete

specimen is under compression or tension, which consists of two regions; pre-peak and
post-peak. Pre-peak region can be obtained by any load-control testing machine in load-
control mode – 𝜎𝜎 − 𝜀𝜀 diagrams should be used – while post-peak region can be
recorded in displacement-control mode using displacement-control testing machine –
stress–deformation diagrams should be used.
Failure process of concrete is started by coalescence of micro-cracks that are occurred at

35% of peak stress after initial linear part of 𝜎𝜎 − 𝜀𝜀 curve. By increasing the load to
50%, cracks are developed and propagate toward the matrix. At about 75% of peak
stress, cracks are merged and major flaws are occurred. A network of major flaws is
occurred at peak stress that causes the failure. The specimen in this condition is not able
to carry load anymore. Strain softening occurs after or exactly at peak stress, which is a
highly localized phenomenon that is mixture of material and structural behaviour. Strain
localization is actually happened in the descending branch of complete 𝜎𝜎 − 𝜀𝜀 curve of
concrete.
There are some constitutive models that are used to describe post-peak softening of
concrete; Series Coupling Model, Compressive Damage Zone, and Concrete Damaged
Plasticity.
The main parameters affecting the strain softening of concrete are the relative stiffness
of the loading system and specimen, the concrete composition, the size and the
slenderness ratio of the specimen, boundary condition between the steel loading platens
41
and the specimen, loading and strain rate, allowable rotations of the loading platens, the
strain gauge length, and the type of the feedback signals that used for recording the data
using the data logger.
All aforesaid parameters are considered in the experimental sector of the study of failure
and post-failure of plain concrete in this research. Slenderness ratio considered in this
research is 2, which is based on the suggested value by researchers. Further explanation
regarding the experimental works conducted in this research is presented in Chapter 3.
Study of theoretical and empirical failure criteria for concrete clarified that the current
criteria contain deficiencies in providing strength envelope for both stress quadrants,
linear and nonlinear surfaces of failure, and specially comprehensiveness to follow the
trend of triaxial data at failure that cause improper estimation of strength for both
compression and tension. Considering these deficiencies assisted the author to develop a
comprehensive strength criterion for different types of concrete covering both linearity
and nonlinearity of failure behaviour in both tensile and compressive quadrants, which
is presented in Chapter 4.
CDP is selected for further study in this research. Its constitutive parameters are
determined in the numerical sector of this research. A concrete specimen with
slenderness ratio equals 2 using the determined constitutive parameters for CDP is
simulated by Abaqus FEA, which its results are presented in Chapter 5.
42
CHAPTER 3:
EXPERIMENTAL AND EMPIRICAL WORKS
3.1. Introduction
All experimental works planned for this research are conducted in the concrete
laboratory of the University of Western Australia. Previous concrete experience of the
author together with the current laboratory experiments is presented in this chapter. The
main experimental works on concrete conducted in this research is as follows:
• Batch proportion and mix design,

• Batch production,
• Specimens’ casting,
• Specimens’ curing,
• Specimens’ ends cutting,
• Specimens’ ends grinding,
• Specimens’ ends capping,
• Surface treatment for strain gauges,
• Pre-coating for strain gauges,
• Bonding the strain gauges to the specimens,
• Strain measurements,
• Uniaxial compression testing,
• Triaxial compression testing, and
• Splitting tensile strength testing.
Stress–strain curves of the uniaxial compression tests on concrete specimens and their
mechanical behavior in the failure regime are assessed together with their fracture
patterns. Failure envelopes in the 𝜎𝜎3 – 𝜎𝜎1 co-ordinates are prepared for the specimens
under triaxial compression loading as well. Results of more than 15 years experience of
author in classification of the concrete types based on their uniaxial compressive
strength at 28 days under the name of Concrete Quality Designation is presented too.
The standards used in this research in preparations of the specimens and in testing
methods are also presented.
43
3.2. Plain Concrete Preparation
Plain concrete is selected in this research as the modeling material to study the failure
and post-failure of concrete. Totally 102 specimens are manufactured from plain
concrete batches prepared in laboratory.
ACI 211.1 and ASTM C192/C192M-05 are used in preparing the batches and
specimens.
Procedure for concrete batch’s proportion, mix design, and required specimens
preparation is discussed in details in following sub-sections.
3.2.1. Mix Proportion and Design
Batch mixtures in this research are designed based on binder types and percentages,
aggregate type, size, and gradation, compressive strength for 28 days age, and W/C
ratio. The W/C proportioning method based on ACI 211.1 is used in this research.
Table 3.1. Concrete mix proportion and design.

Max.
Cement
Batch UCS Mix Cement Grain Grain Slump*
W/C Content
(#) (MPa) Ratio Type Size Type (mm)
(Kg/m3)
(mm)
1 < 15 1:1.5:0 > 0.80 350 GP 0.5 Fine Sand 80
2 15 – 20 1:1:0 0.70 400 GP 2.5 Garnet 60
3 20 – 40 1:1:1.5 0.45 450 GP 5 Sandstone 45
4 40 – 70 1:1.5:3 0.40 500 GB 10 Limestone 40
5 70 – 130 1:1.25:2.5 0.35 550 HE 20 Granite 35
6 > 130 1:0.75:1.5 < 0.35 650 HE 25 Granite 30
UCS is uniaxial compressive strength at 28 days (𝑓𝑓𝑐𝑐′ ).

GP is General-purpose Portland cement type (AS 3972).
GB is General-purpose Blended cement type (AS 3972).
HE is High Early strength cement type (AS 3972).
*
Slump test is conducted in accordance with ACI ASTM C143-78.
44
3.2.2. Casting and Capping
Concrete batches in this research are casted in cylindrical 150 × 300 mm steel molds.
Specimen preparation is conducted according to ASTM C192/C192M-05. Figure 3.1
shows the mold used in this research.
Curing process of the specimens is done in a vibration-free humid room for first 48
hours and then they have been immersed and kept in water tank in 23 ± 2°C to keep
their moisture and temperature in standard condition for 28 days.
According to ASTM C617-10 the specimens’ ends are prepared based on the following
procedure:
• Removing the additional length of the specimens using vertical diamond disk
cutter.
• Grinding the ends of the specimens. An electrical spin grinding disk is used for
this purpose using an anti-abrasive high impact Mangalloy cylindrical sleeve.
• Capping the specimens’ ends using sulfur mortar, capping plates, and alignment
device.
Between the completion of capping and the time of testing all the specimens are kept in
a moist condition.
Figure 3.1. Cylindrical concrete mold used in this research.
45
3.3. Strain Measurements
Measuring the strain in uniaxial compression testing in this research is performed using
strain gauges. For this purpose single element TML Polyester Foil strain gauges with
pre-attached lead wire are used. Table 3.2 shows technical details of the strain gauges.
One strain gauge is installed on each concrete specimen in the axial direction.
Installation of the strain gauge on each specimen is done in a precise procedure
containing surface treatment of the area that the strain gauge should be placed on, pre-
coating the installation area, bonding the strain gauge on the pre-coated area, and finally
masking the strain gauge to protect against any damages. Recording the strain during
the uniaxial compression testing is conducted using a data logger with quarter bridge
circuit. Recorded data is in Volts and then converted to mm using the Gauge Factor
(GF). Finally the displacement across the gauge is divided by the gauge length to
calculate the strain. A sample of the measured strain is presented in Figure 3.2.
Table 3.2. Strain gauges characteristics.

Strain Resistance Gauge Length Gauge Width Backing Length Backing
GF
Gauge (Ω) (mm) (mm) (mm) Width (mm)
PFL-30-11 120 30 2.30 40 7 2.14
PFL-20-11 120 20 1.20 28 6 2.14
PFL-10-11 120 10 0.90 18 6 2.14
GF is Gauge Factor.
75
50
UCS = 71 MPa
𝜎𝜎 (MPa)
UCS = 57 MPa
UCS = 40 MPa
25 UCS = 31 MPa
UCS = 21 MPa
UCS = 15 MPa
UCS = 7 MPa
0
0 0.002 0.004
𝜀𝜀 (%)
Figure 3.2. 𝜎𝜎 – 𝜀𝜀 curve for concrete specimens under uniaxial compression.
46
3.4. Testing Plan
Three testing methods are considered to be conducted on cylindrical concrete specimens

in this research; uniaxial compression, triaxial compression, and splitting tensile
strength testing in accordance with ASTM C39/C39M-04, ASTM C801-98, and ASTM
C496/C496M-11.
Uniaxial compression testing is considered to determine the UCS values and to study
stress–strain behavior of concrete specimens for ascending and descending branches of
𝜎𝜎 – 𝜀𝜀 curves in displacement control mode.
Uniaxial compression test method consists of applying a compressive axial load to cast
concrete cylinders until failure occurs as follows according to ASTM C39/C39M-04:
• Cleaning the specimens’ ends and loading platens.
• Centering the specimen on the lower platen and aligning the axis of the
specimen with the center of thrust of the spherically seated upper platen.
• Adjusting the load after bringing the upper platen on the upper end of the
specimen to obtain uniform seating of the specimen.
• Applying the load at a loading rate of 5.49 to 13.73 N/s until the concrete
specimen fails.
According to ASTM C39/C39M-04, concrete fails in uniaxial compression producing

the fracture patterns as sketched in Figure 3.3.
Cone Cone & Split Cone & Shear Shear Columnar

Figure 3.3. Sketches of types of concrete fracture under uniaxial compression (redrawing from ASTM
C39/C39M-04).
47
Triaxial compression testing is considered to cover the requirements to study
failure/strength envelopes in 𝜎𝜎3 – 𝜎𝜎1 co-ordinates and to evaluate the new strength
criterion proposed in this research for different types of concrete.
All triaxial compression testing in this research are conducted in accordance with
ASTM C801-98 standard.
Baldwin loading frame is used to run the uniaxial and triaxial compression testing
(Figure 3.4).
Figure 3.4. Baldwin loading frame used for uniaxial/triaxial compression testing.
Splitting tensile strength testing is used in this research to determine uniaxial tensile
strength (UTS) of concrete specimens. UTS data pairs would assist to study concrete’s
mechanical behavior in tension.
All splitting tensile testing in this research are conducted in accordance with ASTM
C496/C496M-11.
48
3.5. Test Results
Results of uniaxial tension/compression and triaxial compression testing are presented

as triaxial testing results for concrete Batch # 1 to 10 as ten data groups in Tables 3.3 to
3.12.
Each data group contains at least 5 data pairs of (𝜎𝜎3 , 𝜎𝜎1 ). Data groups presented in
Tables 3.3 to 3.12 contains 67 qualified data pairs; negative signs represent the tensile
data pairs. These qualified data pairs are selected from 102 data pairs obtained from
testing after applying eight qualification criteria (see Section 4.3.2). Amongst the
qualification criteria considered in this study, Mogi’s transition limit (1966a; 1966b)
from brittle to ductile strength (Equation (3.1)) is also applied to data groups but this
condition can be a subject for further study that is not in present research plan.
𝜎𝜎1 > 4.40 × 𝜎𝜎3 (3.1)
Concrete Batch 𝜎𝜎3 (MPa) 𝜎𝜎1 (MPa)
-0.69 0.00
0.00 7.00
3.50 36.10
7.00 46.90
#1
14.00 68.00
21.00 88.23
28.00 107.00
35.00 113.01
𝜎𝜎3 is confining pressure (minor principal stress) at failure.
𝜎𝜎1 is axial stress (major principal stress) at failure.
-0.70 0.00
0.00 15.00
1.50 26.00
#2 3.00 39.00
5.00 44.00
7.00 60.02
9.00 64.35
49
-0.90 0.00
0.00 21.00
3.50 61.00
#3
7.00 75.00
21.00 122.00
35.00 146.10
-1.23 0.00
0.00 31.00
#4 11.00 87.62
21.00 127.00
35.00 164.10
-2.00 0.00
0.00 39.87
5.00 77.00
10.00 99.95
15.00 119.10
20.00 140.00
#5
25.00 152.04
30.00 167.00
35.00 182.88
40.00 200.03
45.00 209.79
50.00 226.00
-4.90 0.00
0.00 57.00
1.00 79.98
#6
11.00 131.00
21.00 180.01
24.00 210.05
50
-7.00 0.00
0.00 71.00
#7 2.00 106.02
8.00 169.90
31.00 221.45
-6.12 0.00
0.00 101.14
5.00 148.32
#8 10.00 185.42
15.00 219.05
30.00 281.18
50.00 370.00
-8.90 0.00
0.00 131.98
1.00 150.00
#9
10.00 190.07
21.00 220.00
30.00 290.00
-9.00 0.00
0.00 190.10
# 10 35.00 429.00
70.00 460.00
110.00 471.39
51
3.6. Concrete Quality Designation
Concrete Quality Designation or CQD is an empirical classification that qualifies

concrete types to six different descriptive classes based on uniaxial compressive
strength obtained from uniaxial compression testing on concrete specimens.
The CQD classifies concrete to six descriptive classes based on uniaxial compressive
strength as below:
• VP (Very Poor): Lightweight concrete with uniaxial compressive strength lower

than 15 MPa
• P (Poor): Routine concrete with uniaxial compressive strength between 15 and
20 MPa
• F (Fair): Normal strength concrete with uniaxial compressive strength between
20 and 40 MPa
• G (Good): High strength concrete with uniaxial compressive strength between
40 and 70 MPa
• VG (Very Good): High performance concrete with uniaxial compressive
strength between 70 and 130 MPa
• E (Excellent): Ultra high performance concrete with uniaxial compressive
strength higher than 130 MPa
A conclusion of more than 15 years practical and research experience in the field of
concrete technology shows that the CQD presents a helpful practical classification.
The CQD also can be used together with the proposed strength criterion introduced in
this research for concrete (see Section 4.3.1) to determine the suitable parameters for the
criterion. Equations (3.2 and 3.3) briefly present the strength criterion proposed in this
research (see Section 4.3):
𝜎𝜎 +𝜆𝜆𝜎𝜎
𝜎𝜎1 = 𝜎𝜎3 + 𝜎𝜎𝑐𝑐 �𝜎𝜎𝑐𝑐 +𝜁𝜁𝜎𝜎3 � (3.2)
𝑐𝑐 3
𝜎𝜎𝑐𝑐 = 𝛽𝛽𝛽𝛽 (3.3)
52
In above equations, 𝜎𝜎1 and 𝜎𝜎3 are the major and minor principal stresses at failure, 𝜆𝜆, 𝜁𝜁,
and 𝛽𝛽 are criterion’s parameters, 𝜎𝜎𝑐𝑐 is the Regulated Unconfined Compressive Strength
(RUCS), 𝐶𝐶 is the Apparent Unconfined Compressive Strength (AUCS).
The foundation of the CQD for concrete is based on an insight from RQD for rocks.
Table 3.13 provides a legend for the CQD introduced in this research including
description for each class.
It should be noted that the CQD classes are not limited to the examples that are
provided in the description column of Table 3.13.
Table 3.13. Legend for Concrete Quality Designation introduced in this research.
UCS
CQD Description
(MPa)
Lightweight, lean, fill, and plastic concrete, plastic mortar, and loose
VP < 15
cement grouts
P 15 – 20 Concrete for routine uses, low strength mortar, and hard cement grouts
Normal strength concrete, mass concrete, concrete used in large civil

F 20 – 40 projects like pavements, slabs, caissons, and footings, and normal strength
mortar
High strength concrete and concrete used for Specific building elements
G 40 – 70
like beams, foundations, and building columns
VG 70 – 130 High performance concrete and concrete used for special rigid structures
E > 130 Ultra high performance concrete
CQD represents Concrete Quality Designation.

UCS is Uniaxial Compressive Strength for cylindrical concrete specimens at 28 days.
VP means Very Poor quality concrete.
P means Poor quality concrete.
F means Fair quality concrete.
G means Good quality concrete.
VG means Very Good quality concrete.
E means Excellent quality concrete.
53
3.7. Conclusions
Different types of concrete mixtures are prepared in this research. 102 cylindrical
concrete specimens with 150 × 300 mm size are prepared to study the failure and post-
failure behaviour of different types of concrete. Concrete batches proportion, design,
and production, specimens’ manufacture, curing, ends preparation, and capping are
conducted according to ACI 211.1, ASTM C192/C192M-05, ASTM C143-78, and
ASTM C617-10.
TML PFL-10-11-1L, PFL-20-11-1L, and PFL-30-11-1L 120 Ω single element polyester

foil strain gauges are used as conventional strain measurements method. One strain
gauge installed on each concrete specimen parallel to longitudinal axis of the specimen.
Displacement control mode is set up for uniaxial compression testing using ASTM
C39/C39M-04 to obtain UCS and study the fracture pattern and descending branch of
stress–strain curve. Triaxial compression testing is conducted according to ASTM
C801-98 to obtain the failure envelopes of concrete specimens in 𝜎𝜎3 – 𝜎𝜎1 co-ordinates.
Splitting tensile strength testing is used as uniaxial tension tests to obtain UTS to study
tensile behavior of concrete types in accordance with ASTM C496/C496M-11. Baldwin
tension/compression loading frame is used to perform the uniaxial/triaxial compression
tests.
Totally 10 uniaxial compression tests, 10 splitting tensile strength tests, and 82 triaxial
compression tests are conducted. Therefore 102 data pairs of (𝜎𝜎3 , 𝜎𝜎1 ) are obtained.
Amongst these data pairs 67 data pairs in 10 data groups are qualified to be used in
study of failure and post-failure of concrete. Mogi’s transition limit from brittle to
ductile is applied on data groups.
The major fracture patterns in accordance with the ASTM for concrete specimens under
uniaxial compression in this research are recognized as the cone and shear types. The
major fracture pattern for concrete specimens under triaxial compression is recognized
as the cone type.
A new classification for concrete quality is introduced based on an empirical study on

concrete types’ uniaxial compression strengths. This classification is called CQD
54
representing Concrete Quality Designation. The CQD classifies concrete types to Very
Poor (VP), Poor (P), Fair (F), Good (G), Very Good (VG), and Excellent (E).
Laboratory results presented in this chapter together with the concrete classification
proposed in this chapter is used in Chapters 4 and 5 as the requirements to perform the
analytical and numerical sectors of the research.
55
CHAPTER 4:
NEW STRENGTH CRITERION FOR CONCRETE
4.1. Introduction
A strength criterion is a mathematical relationship between the stress components at

failure called the failure envelope. They have the following requirements to improve
their applicability.
• It should provide accurate and precise estimations for both compression and
tension regime of stresses similar to those determined in laboratory.
• It should provide acceptable correlation with triaxial data.
• It should provide regulated and credible estimation as real values for

Uniaxial/Unconfined Compressive Strength (UCS) and Uniaxial Tensile
Strength (UTS) according to natural trend of triaxial data pairs.
• Its functional representation should be comprehensive and flexible based on the

actual triaxial data and preferably be provided by a simple mathematical
equation.
Bineshian’s (2000) strength criterion for intact rocks and coals meets all
abovementioned criteria properly. This criterion is further developed in this project to
cover different types of concrete from very poor to excellent quality concrete (see the
CQD presented in Section 3.6).
In this chapter after a brief presentation of Bineshian strength criterion (2000) for rocks
and coals, its systematic generalisation for all types of concrete from the aspect of
uniaxial compressive strength based on the concrete quality designation will be
introduced.
56
4.2. Bineshian’s (2000) Strength Criterion for Intact Rocks and Coals
Bineshian’s (2000) strength criterion is a compressive-tensile criterion that has been

obtained after assessing different mathematical functions and analyzing the results of
more than 8,700 fittings conducted on about 1,743 qualified data pairs (𝜎𝜎3 , 𝜎𝜎1 ) consist
of 250 data groups; 32 types of rocks and coals (15 igneous rock types, 6 sedimentary
rock types, 7 metamorphic rock types and 4 coal types) from triaxial compression tests
out of 2,149 data pairs (392 data groups; 34 types of rocks and coals) that collected
from literatures and conducted by author (Schwartz, 1964; Horino and Ellikson, 1970;
Ouyang and Elsworth, 1991; Vutukuri and Hossaini, 1992; Hossaini, 1993; Sheorey,
1997; Mahab Ghodss Consulting Engineers Co., 1999, 2003; Bineshian, 2000, 2006,
2008; Bineshian and Bineshian, 2011a, 2011b).
The qualification criteria for selecting suitable data pairs are presented in Section 4.5.
All laboratory tests cited were carried out according to the methods suggested by the
International Society for Rock Mechanics (ISRM) (Brown 1981).
The mathematical representation of Bineshian’s strength criterion reads:
𝑐𝑐 3
Here, 𝜎𝜎1 and 𝜎𝜎3 are the major and minor principal stresses at failure, 𝜆𝜆 and 𝜁𝜁 are the
rock constants, 𝜎𝜎𝑐𝑐 is the Regulated Unconfined Compressive Strength (RUCS) that is
obtained by multiplying two different parameters (𝛽𝛽 and 𝐶𝐶), where 𝐶𝐶 is the Apparent
Unconfined Compressive Strength (AUCS), which is determined as the major principal
stress at the point where the Mohr-Coulomb envelope in the 𝜎𝜎3 − 𝜎𝜎1 co-ordinates
intersects the 𝜎𝜎1 axis, and 𝛽𝛽 is a factor that defines RUCS in terms of AUCS obtained
from regression analysis.
The proposed criterion (4.1) is based on a simple mathematical function, which can
reproduce both linear and nonlinear envelopes on the triaxial test data in accordance
with the input data pairs (𝜎𝜎3 , 𝜎𝜎1 ) to cover both linear and nonlinear envelopes for intact
rocks and coals.
57
Unlike many empirical criteria from 1952 to the present (i.e. criteria of Ballmer, 1952;
Murrell, 1963; Fairhurst, 1964; Hobbs, 1964; Murrell, 1965; Mogi, 1966; Hoek, 1968;
Hobbs, 1970; Franklin, 1971; Hobbs, 1971; Bieniawski, 1974; Hoek-Brown, 1980;
Johnston-Chiu, 1984; Ramamurthy et al., 1985; Hoek-Brown, 1988; Ramamurthy et al.,
1988; Sheorey et al., 1989; Yoshida et al., 1990; Carter et al., 1991; etc.), this criterion
is not expressed through a power function which often limits the applicability of the
criterion (Bineshian, 2000; Bineshian and Bineshian, 2011a, 2011b).
The criterion is a compressive-tensile strength criterion, which fits well the compressive
and tensile strength data. It is an empirical-theoretical strength criterion, because RUCS
is calculated using Mohr-Coulomb criterion (a theoretical strength criterion) and the
function has been formed by empirical study.
The constants of Bineshian’s criterion can be determined easily by performing the

regression analysis of the data group/s containing the triaxial compression and tension
test data.
The proposed criterion shows a highly satisfactory correlation with triaxial data and
absolutely abides by the natural trend of data. Estimation of strength by this criterion is
very tending to certainty in comparison to the other famous criteria.
Coefficient of determination and accordance coefficient have been calculated to assess

correlation and accordance of the criteria with data and to evaluate their applicability.
The coefficient of determination quantifies how well the strength criterion fits with
actual triaxial data while the accordance coefficient quantifies how well the strength
criterion defines the actual data.
Table 4.1 shows the results of calculating the coefficients of determination for 250 data
groups including 1,743 triaxial tests on 32 types of rocks and coals’ specimens. It shows
correlation between the proposed criterion and actual triaxial data better than other
criteria.
Figure 4.1 shows fitting the selected criteria to the data of Johnston (1985) for Westerly
Granite. Compressive criteria because of the existence of tensile data pairs cannot be
regressed on Johnston’s data group.
58
Table 4.1. Evaluation of correlation between criteria and triaxial data for intact rocks and coals.
𝑅𝑅2
Strength Criteria
= 1.00 ≥ 0.98 ≥ 0.96 ≥ 0.94
Bineshian 47% 87% 93% 98%
Bieniawski 22% 53% 67% 78%
Hoek-Brown 09% 44% 63% 73%
Ramamurthy et al 11% 40% 59% 66%
Mohr-Coulomb 2% 22% 43% 58%

R2 is Coefficient of Determination. The closer the value of 𝑅𝑅 2 approaches to 1, the better the fitting between the criterion and the
experimental data is.
300
Bineshian Criterion R^2 = 0.996

225
σ1 (MPa)
150
𝝈𝝈𝒄𝒄 +𝝀𝝀𝝈𝝈𝟑𝟑
𝝈𝝈𝟏𝟏 = 𝝈𝝈𝟑𝟑 + 𝝈𝝈𝒄𝒄
𝝈𝝈𝒄𝒄 +𝜻𝜻𝝈𝝈𝟑𝟑
𝜆𝜆 = 14.00
75 𝜁𝜁 = 7.08
𝛽𝛽 = 1.25
𝐶𝐶 = 100.93 MPa
𝜎𝜎𝐶𝐶 = 126.29 MPa
𝜎𝜎𝑡𝑡 = - 9.34 MPa
0
-15 -10 -5 0 5 10 15 20
σ3 (MPa)
Figure 4.1. Fitting the proposed strength criterion to the data of Johnston (1985) on Westerly Granite.
Table 4.2 shows the compatibility and accordance of the criterion in comparison with
other criteria for 12 data groups of Chamshir Limestone (Bineshian 2000). From Table
4.2, the proposed criterion shows the smallest regression fitness errors; hence it has a
satisfactory compatibility with natural trend of triaxial data, and therefore can present
better estimation of triaxial strengths in comparison to others. In 42% of the cases,
Bineshian’s criterion has an accordance coefficient lower than 0.01, which is indicative
of satisfactory compatibility. This criterion has better compatibility that those applicable
to the other criteria.
59
Table 4.2. Accordance of strength criteria with actual triaxial data for different types of limestone.
𝜓𝜓 2
Strength Criteria
≤ 0.001 ≤ 0.005 ≤ 0.010 ≤ 0.050 ≤ 0.150
Bineshian 9% 25% 42% 75% 92%
Bieniawski 0% 9% 17% 75% 83%
Hoek-Brown 0% 0% 0% 67% 83%
Ramamurthy et al 0% 0% 17% 50% 83%
Mohr-Coulomb 0% 0% 0% 67% 75%

𝜓𝜓 2 is Accordance Coefficient. The smaller the value of 𝜓𝜓 2 is, the better the strength criterion accords with the triaxial test data.
The accordance is considered good when 𝜓𝜓 2 = 0.
Table 4.3 shows the capacity of Bineshian’s criterion to define tensile strength 𝜎𝜎𝑡𝑡 by the
reference to 22 data groups of coals tested by Hobbs (Sheorey, 1997). It is indicated
that, in 82% and 100% of the comparisons, Bineshian’s criterion provided estimates of
𝜎𝜎𝑡𝑡 better than those estimated with Hoek-Brown and Mohr-Coulomb criteria,
respectively. Bineshian’s criterion estimated 𝜎𝜎𝑡𝑡 approximately 1.2 times the values
determined by UTS testing whereas Hoek-Brown and Mohr-Coulomb criteria estimated
𝜎𝜎𝑡𝑡 , 2.2 and 5.1 times the applicable values, respectively (Bineshian, 2011b) for the 22
data groups of coals. These results are confirmed by the accordance coefficients
determined for the criteria in tension, as shown in Figure 4.2. Furthermore, following
equations can be extracted from the examinations conducted on the abilities of the
strength criteria to estimate the uniaxial tensile strength:
𝜎𝜎𝑡𝑡 𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵ℎ𝑖𝑖𝑖𝑖𝑖𝑖 = 1.20 × 𝑈𝑈𝑈𝑈𝑈𝑈 (4.2)
𝜎𝜎𝑡𝑡 𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻−𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 = 2.20 × 𝑈𝑈𝑈𝑈𝑈𝑈 (4.3)
𝜎𝜎𝑡𝑡 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶−𝑀𝑀𝑀𝑀ℎ𝑟𝑟 = 5.20 × 𝑈𝑈𝑈𝑈𝑈𝑈 (4.4)
Equations (4.2) – (4.4) prove that the estimation of tensile strength by Hoek-Brown and
Mohr-Coulomb criteria are not accurate and reliable. On the contrary, the Bineshian
criterion estimation of the tensile strength has an acceptable accuracy. In a similar study
regarding the compressive stresses, the following equations can be extracted. Equation
(4.5) is applicable for igneous rocks, Equation (4.6) is applicable for metamorphic
rocks, Equation (4.7) is applicable for sedimentary rocks, and Equation (4.8) is
applicable for all types of coals.
60
𝜎𝜎𝑐𝑐 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 = 𝐶𝐶 (4.5)
𝜎𝜎𝑐𝑐 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 = 𝐶𝐶 (4.6)
𝜎𝜎𝑐𝑐 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 = 0.90 × 𝐶𝐶 (4.7)
𝜎𝜎𝑐𝑐 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 = 0.80 × 𝐶𝐶 (4.8)
Table 4.3. Comparison of uniaxial tensile strengths calculated by the criteria with uniaxial tensile
strengths obtained from Brazilian testing on coal by Hobbs (Sheorey, 1997; Bineshian, 2000).
𝜎𝜎𝑡𝑡
(MPa)
Data Group UTS (MPa)
Bineshian Hoek-Brown Mohr-Coulomb
criterion criterion criterion
1 −0.83 −1.06 −0.90 −2.85
2 −0.76 −1.09 −1.89 −5.17
3 −3.41 −3.67 −9.99 −14.55
4 −1.79 −2.08 −2.99 −6.46
5 −2.07 −2.39 −3.54 −7.39
6 −2.45 −2.69 −3.51 −7.16
7 −0.69 −0.70 −1.10 −1.88
8 −2.69 −3.57 −2.10 −4.79
9 −0.69 −1.33 −0.64 −2.96
10 −0.55 −1.18 −0.38 −2.75
11 −0.62 −0.66 −3.49 −5.64
12 −0.76 −0.79 −3.18 −7.00
13 −1.58 −1.75 −0.96 −8.01
14 −0.89 −1.22 −1.46 −3.78
15 −1.1 −1.71 −2.96 −7.28
16 −1.1 −1.18 −2.60 −5.94
17 −1.1 −1.54 −3.08 −6.94
18 −0.69 −0.90 −1.33 −2.99
19 −0.55 −0.95 −1.35 −3.90
20 −0.76 −0.88 −3.29 −5.87
21 −1.1 −1.64 −2.11 −5.38
22 -0.89 -1.15 -1.86 -4.75
𝜎𝜎𝑡𝑡 is uniaxial tensile strength calculated by strength criteria.
UTS is uniaxial tensile strength obtained from Brazilian test.
61
Strength Criteria
Bineshian
0.26
Hoek-Brown
6.54
Coulomb-Mohr
40.64
0 10 20 30 40 50
Accordance Coefficient
The parameters of the criterion can be obtained by performing the required triaxial tests
and fitting the criterion to the obtained data pairs or selecting suitable suggested values
for the parameters from Tables 4.4 – 4.7.
The nonlinear regression method using Nelder and Mead technique (Netter et al., 1988)
is used to determine the parameters of the Bineshian strength criterion. The first method
(performing the required tests and fitting the criterion to the actual data) is the most
accurate, because it is designed especially for the certain rock/coal; however the second
method is also quite applicable.
The suggested parameters are calculated based on the comprehensive study on an

extensive range of data groups for different rock types to provide applicable, credible,
and accurate constant values (Bineshian, 2000).
Summarily, the average rounded values for the parameters of the proposed strength
criterion vary as below:
10 < 𝜆𝜆 < 14.50
2.50 < 𝜁𝜁 < 4.50
0.80 < 𝛽𝛽 < 1
62
Table 4.4. Parameters for Bineshian criterion applicable to intact igneous rocks.
Average Suggested Values

Igneous Rocks
λ ζ β
Andesite 12.49 8.03 1.03
Basalt 8.98 7.95 1.50
Diabase 13.49 4.16 1.17
Diorite 7.05 3.38 1.06
Dunite 7.93 0.51 0.57
Gabbro 29.22 10.40 0.82
Granite 12.19 2.72 0.89
Granodiorite 10.02 1.84 1.10
Lamprophyre 9.21 4.70 1.25
Peridotite 4.79 0.45 0.79
Quartzdiorite 10.24 2.79 1.03
Rhyolite 8.40 3.71 1.41
Tuff 6.51 0.55 0.95
Average Rounded Parameters 11 4 1

λ is Bineshian criterion’s parameter.
ζ is Bineshian criterion’s parameter.
β is Bineshian criterion’s parameter.
Table 4.5. Parameters for Bineshian criterion applicable to intact sedimentary rocks.

Sedimentary Rocks
λ ζ β
Dolomite 12.84 4.01 1.04
Limestone 6.95 2.24 0.84
Sandstone 9.46 2.11 0.87
Shale 6.55 4.07 0.95
Siltstone 14.97 4.96 0.83
Average Rounded Parameters 10 3.50 0.90

63
Table 4.6. Parameters for Bineshian criterion applicable to intact metamorphic rocks.

Metamorphic Rocks
λ ζ β
Eclogite 7.01 0.34 0.55
Gneiss 16.91 6.58 0.92
Marble 4.91 0.87 0.95
Quartzite 11.82 3.22 1.17
Schist 13.47 2.43 1.11
Slate 29.02 14.40 1.34
Average Rounded Parameters 14 4.50 1

Table 4.7. Parameters for Bineshian criterion applicable to coals.

Coals
λ ζ β
Loading, ⊥bpmc 17.75 2.93 0.90
Loading,bpmc 12.58 1.74 0.68
Loading,bp⊥mc 14.91 2.28 0.76
Coal* 13.31 2.42 0.80
Average Rounded Parameters 14.50 2.50 0.80

bp represents bedding planes.
mc represents main cleats.
* bp and mc for this type is not considered.
The constant values for the Bineshian strength criterion presented in Tables 4.4 – 4.7
can be used for prediction of compressive and tensile strength of different types of
intact rocks and coals.
For precise strength calculations, nonlinear regression of the criterion on the actual
triaxial data obtained from laboratory tests is recommended.
64
4.3. Development of Bineshian Strength Criterion for Concrete
4.3.1. Introduction
A new strength criterion is developed in this research for different types of concrete,
based on the Bineshian (2000) strength criterion. Its mathematical representation is the
same equation as Equation (4.1) but its parameters are redefined for concrete types
(Equations (4.9) – (4.11)). The concrete types referred to in this research are classified
based on the compressive strength (see Chapter 3 for the new empirical classification
introduced in this research under the name of Concrete Quality Designation in
abbreviation form of CQD). The new strength criterion developed for concrete in this
research reads:
𝑐𝑐 3
2𝑆𝑆 cos 𝜑𝜑
𝐶𝐶 = 1−sin 𝜑𝜑
(4.11)
where, 𝜎𝜎1 and 𝜎𝜎3 are the major and minor principal stresses at failure in MPa, 𝜆𝜆 and 𝜁𝜁
are dimensionless concrete constants, 𝜎𝜎𝑐𝑐 is the Regulated Unconfined Compressive
Strength (RUCS) in MPa, 𝐶𝐶 is the Apparent Unconfined Compressive Strength (AUCS)
in MPa obtained from linear Mohr-Coulomb failure envelope, 𝛽𝛽 is a dimensionless
factor that defines RUCS in terms of AUCS, 𝑆𝑆 is internal cohesive strength in MPa
obtained from the linear Mohr-Coulomb failure envelope, and 𝜑𝜑 is the internal friction
angle in degrees.
The parameter 𝐶𝐶 in Equations (4.10) – (4.12) is determined as the major principal stress
at the point where the linear Mohr-Coulomb failure envelope in the 𝜎𝜎3 − 𝜎𝜎1 system
intersects the 𝜎𝜎1 axis. It can be calculated using Equation (4.11); 𝐶𝐶 is originally
extracted from Mohr-Coulomb criterion (Equation (4.12)).
𝜎𝜎1 = 𝐶𝐶 + 𝐵𝐵𝐵𝐵3 (4.12)
65
1+sin 𝜑𝜑
𝐵𝐵 = 1−sin 𝜑𝜑 (4.13)
𝐵𝐵 is the slope of the Mohr-Coulomb linear failure envelope.
Parameter 𝐶𝐶 is not necessarily the unconfined compressive strength of the concrete,

because if more than one uniaxial compressive test is carried out on the samples, then
several probably different values of the UCS will be obtained. When only one uniaxial
compression test or the point load test (PLT) has been carried out, its test result may not
be representative. Therefore it should be regulated based on the natural trend of triaxial
test data. We call this regulated value the RUCS.
The RUCS – the regulated value for UCS – means that the observed value of the
uniaxial compressive strength is corrected by the regulation factor (𝛽𝛽). Thus to calculate
the best regulation factor, 𝛽𝛽 should be added as a parameter to the strength criterion
(Equation (4.9)). Then the criterion should be fitted to the triaxial test data by the
regression analysis and thus 𝛽𝛽 that regulates the AUCS will be obtained. The AUCS in
this research is calculated by the Mohr-Coulomb criterion. Hence the uniaxial
compressive strength 𝜎𝜎𝑐𝑐 can be calculated in three steps as follows:
1. Calculating AUCS by fitting the Mohr-Coulomb criterion to the triaxial test data.
2. Calculating the regulation factor 𝛽𝛽 by fitting the criterion (Equation (4.9)) to
triaxial test data.
3. Calculating RUCS by multiplying 𝐶𝐶 by 𝛽𝛽.
The RUCS value is determined by setting 𝜎𝜎3 = 0 in Equation (4.9):
𝜎𝜎3 = 0 → 𝜎𝜎1 = 𝜎𝜎𝑐𝑐
Consequently,
By setting 𝜎𝜎1 = 0 in Equation (4.9), the value of 𝜎𝜎3 is equal to the uniaxial tensile
strength (𝜎𝜎𝑡𝑡 ):
66
𝜎𝜎1 = 0 → 𝜎𝜎3 = 𝜎𝜎𝑡𝑡
Therefore,
−𝜆𝜆−1+�𝜆𝜆2 +2𝜆𝜆−4𝜁𝜁−1
𝜎𝜎𝑡𝑡 = 𝛽𝛽𝛽𝛽 (4.15)
2𝜁𝜁
The validation of estimation of strength for both uniaxial compressive and tensile
strengths and performance of the proposed criterion is presented in Section 4.3.2.
4.3.2. Performance
To evaluate the performance of the criterion for different types of concrete, 102
cylindrical concrete specimens with different CQD are used. All these concrete
specimens are prepared under the same conditions from different batches and mix
designs for this research’s purposes. Uniaxial tensile tests, uniaxial compressive tests,
and triaxial compressive tests have been conducted on the abovementioned specimens
on their 28 days age (see Chapter 3). Finally 10 data groups including at least five data
pairs per group from 10 batches of different mixes are selected and totally 102 data
pairs of (𝜎𝜎3 , 𝜎𝜎1 ) representing six class of concrete based on CQD have been obtained.
Eight essential qualification criteria are applied to the triaxial data groups to exclude
unsuitable data from analyses. The triaxial data groups are assessed based on the
following qualification criteria:
• Data pairs must be in the form of major and minor principal stresses (𝜎𝜎3 , 𝜎𝜎1 ) at
failure.
• There have to be 5 or more (𝜎𝜎3 , 𝜎𝜎1 ) data pairs in each data group (Bineshian,
2000).
• All data groups must contain the value of UCS (Bineshian, 1999).
• All data pairs in each data group must contain the principal stress combination
above the Mogi’s transition limit (1966a; 1966b) from brittle to ductile strength
(Sheorey, 1997):
𝜎𝜎1 > 4.40 × 𝜎𝜎3 (4.16)
67
• Each data group must contain the results of triaxial compression tests, in which
the confining pressure is greater than zero (Bineshian, 2000; Bineshian and
Bineshian, 2011a).
• Each data group must contain a range of confining pressures with at least one
test at a confining pressure greater than a half of UCS (Hossaini, 1999).
• Data groups must be definable in terms of regressive shaped curve (Bineshian
and Bineshian, 2011b).
• Data pairs are considered invalid if 𝜎𝜎1 is higher than UCS (Bineshian, 2000).
Finally, after applying the qualification criteria, 10 data groups of triaxial compressive
tests including different types of concrete from very poor to excellent concrete quality
(CQD-VP to CQD-E; see Chapter 3) consisting of totally 67 data pairs – (𝜎𝜎3 , 𝜎𝜎1 ) – have
been qualified. Tables 4.8 and 4.9 show the quality of correlation of the proposed
criterion for concrete on triaxial data groups and accordance of the criterion with the
actual triaxial data respectively. Table 4.8 indicates excellent correlation between the
proposed criterion and actual triaxial data based on the calculated coefficient of
determination for 10 data groups of different types of concrete. Table 4.9 shows
satisfactory compatibility and accordance of the proposed criterion with natural trend of
triaxial data. Tables 4.8 and 4.9 prove that the estimation of triaxial strengths using the
proposed criterion is accurate and precise and therefore its estimation is credible.
Table 4.8. Evaluation of correlation of the proposed criterion and triaxial data for a variety of concrete.
𝑅𝑅2
Strength Criteria
= 1.00 ≥ 0.99 ≥ 0.98 ≥ 0.97
Proposed 40% 70% 90% 100%
Mohr-Coulomb 0% 0% 0% 10%
R2 is Coefficient of Determination. The closer the value of 𝑅𝑅 2 approaches to 1, the better the fitting between the criterion and the
experimental data is.
Table 4.9. Accordance of the proposed criterion with actual triaxial data for different types of concrete.
𝜓𝜓 2
Strength Criteria
≤0.001 ≤0.005 ≤0.010 ≤0.050 ≤0.150
Proposed 0% 30% 50% 100% 100%
Mohr-Coulomb 0% 0% 0% 10% 80%
𝜓𝜓 2 is Accordance Coefficient. The smaller the value of 𝜓𝜓 2 is, the better the strength criterion accords with the triaxial test data.
The accordance is considered good when 𝜓𝜓 2 = 0.
68
The performance of measurement of the uniaxial compressive and tensile strengths
using the proposed strength criterion is assessed as well.
Tables 4.10 and 4.11 show the quality of the measurements for 𝜎𝜎𝑐𝑐 and 𝜎𝜎𝑡𝑡 respectively
using the proposed strength criterion for different types of concrete. As can be seen
from Table 4.10, the proposed criterion estimates the uniaxial compressive strength, on
average as 0.91 times the 𝐶𝐶 value (Equation (4.17)).
𝜎𝜎𝑐𝑐 = 0.91 × 𝐶𝐶 (4.17)
In a similar way from Table 4.11, it can be concluded that the proposed criterion
estimates the uniaxial tensile strength, averagely 1.55 times the UTS values obtained by
splitting tensile strength test while Mohr-Coulomb criterion estimates uniaxial tensile
strength 4.35 times the UTS values (Equation (4.18)).
𝜎𝜎𝑡𝑡 = 1.55 × 𝑈𝑈𝑈𝑈𝑈𝑈 (4.18)
Table 4.10. Comparison between UCS obtained by uniaxial compression test, AUCS (𝐶𝐶) measured using
Mohr-Coulomb criterion, and RUCS (𝜎𝜎𝑐𝑐 ) measured using proposed criterion.
Concrete Type, UCS 𝐶𝐶 𝜎𝜎𝑐𝑐 𝜎𝜎𝑐𝑐
Based on CQD (MPa) (MPa) (MPa) 𝐶𝐶
VP, UCS < 15 MPa 7.00 16.10 7.28 0.45
P, UCS ≈ 15 MPa 15.00 13.23 11.52 0.87
P, UCS ≈ 20 MPa 21.00 30.11 18.67 0.62
F, UCS ≈ 30 MPa 31.00 25.18 20.40 0.81
F, UCS ≈ 40 MPa 39.87 46.53 30.06 0.65
G, UCS ≈ 55 MPa 57.00 53.12 58.96 1.11
G, UCS ≈ 70 MPa 71.00 76.97 89.29 1.16
VG, UCS ≈ 100 MPa 101.14 97.58 92.80 0.95
VG, UCS ≈ 130 MPa 131.98 107.44 130.00 1.21
E, UCS > 130 MPa 190.10 162.45 207.94 1.28
Average 0.91
UCS is Uniaxial Compressive Strength.
𝐶𝐶 is the Apparent Unconfined Compressive Strength (AUCS).
𝜎𝜎𝑐𝑐 is the Regulated Unconfined Compressive Strength (RUCS).
CQD is Concrete Quality Designation.
69
Table 4.11. Comparison between UTS obtained by splitting tensile strength test and 𝜎𝜎𝑡𝑡 measured using
proposed criterion.
Concrete Type, UTS 𝜎𝜎𝑡𝑡 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝜎𝜎𝑡𝑡 𝑀𝑀𝑀𝑀ℎ𝑟𝑟−𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝜎𝜎𝑡𝑡 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝜎𝜎𝑡𝑡 𝑀𝑀𝑀𝑀ℎ𝑟𝑟−𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶
Based on CQD (MPa) (MPa) (MPa) 𝑈𝑈𝑈𝑈𝑈𝑈 𝑈𝑈𝑈𝑈𝑈𝑈
VP, UCS < 15 MPa -0.69 -1.59 -5.14 2.30 7.45
P, UCS ≈ 15 MPa -0.70 -1.10 -2.11 1.57 3.01
P, UCS ≈ 20 MPa -0.90 -1.20 -8.09 1.33 8.99
F, UCS ≈ 30 MPa -1.23 -2.99 -5.83 2.43 4.74
F, UCS ≈ 40 MPa -2.00 -4.26 -12.02 2.13 6.01
G, UCS ≈ 55 MPa -4.90 -6.46 -8.16 1.32 1.67
G, UCS ≈ 70 MPa -7.00 -7.62 -14.25 1.09 2.04
VG, UCS ≈ 100 MPa -6.12 -7.54 -16.29 1.23 2.66
VG, UCS ≈ 130 MPa -8.90 -9.78 -16.92 1.10 1.90
E, UCS > 130 MPa -9.00 -9.33 -45.38 1.04 5.04
Average 1.55 4.35

UTS is Uniaxial Tensile Strength obtained by splitting tensile strength test.
𝜎𝜎𝑡𝑡 is Unconfined Tensile Strength.
Figures 4.3 – 4.12 show the fitting of the proposed criterion on ten actual triaxial data
groups of different concrete types in the 𝜎𝜎3 − 𝜎𝜎1 co-ordinate system. These figures
together with Tables 4.8 – 4.21 clearly show the quality of the strength prediction using
proposed criterion and the relative errors of the criterion.
Data groups used in Figures 4.3 – 4.12 are triaxial data obtained from uniaxial
compression, uniaxial tension, and triaxial compression tests conducted in this research.
Figure 4.3 represents triaxial mechanical behaviour of Very Poor concrete (VP). Figures
4.4 and 4.5 depict mechanical behaviour of Poor concrete (P). Figures 4.6 and 4.7 show
mechanical behaviour of Fair concrete (F) under triaxial compression. Figures 4.8 and
4.9 are related to mechanical behaviour of Good concrete (G). Figure 4.10 and 4.11
show how Very Good concrete (VG) behaves under compression and finally Figure
4.12 shows Excellent concrete’s (E) mechanical behaviour. The quality of concrete in
this research is tagged using the Concrete Quality Designation (CQD) based on UCS
values. The CQD is an empirical classification that categorizes the concrete quality to
six levels as; Very Poor (VP), Poor (P), Fair (F), Good (G), Very Good (VG), and
Excellent (E). The details of the CQD are presented in Chapter 3.
70
125
𝝈𝝈𝟏𝟏 = 𝝈𝝈𝟑𝟑 + 𝝈𝝈𝒄𝒄 𝝈𝝈𝒄𝒄 +𝜻𝜻𝝈𝝈𝟑𝟑
𝜆𝜆 = 7.95
𝜁𝜁 = 0.54
100 𝛽𝛽 = 0.45
𝐶𝐶 = 16.10 MPa
75
(MPa)
σ1
50
Proposed Criterion R^2 = 0.995
25 Mohr-Coulomb R^2 = 0.945
Observed Data
CQD = VP UCS < 15 MPa

0
-10 0 10 20 30 40
σ3 (MPa)
Figure 4.3. Observed value against predicted value by proposed strength criterion for concrete specimens
with CQD-VP and UCS < 15 MPa.
80
𝜆𝜆 = 12.79
𝜁𝜁 = 1.66
𝛽𝛽 = 0.87
60 𝐶𝐶 = 13.23 MPa
(MPa)
40
σ1

20
Observed Data
CQD = P UCS = 15 MPa

0
-3 0 3 6 9 12
σ3 (MPa)
with CQD-P and UCS ≈ 15 MPa.
71
180 𝝈𝝈𝒄𝒄 +𝝀𝝀𝝈𝝈𝟑𝟑
𝜆𝜆 = 17.76
𝜁𝜁 = 2.55
𝛽𝛽 = 0.62
135 𝐶𝐶 = 30.11 MPa
(MPa)
90
σ1

45
Observed Data
CQD = P UCS = 20 MPa

0
-10 0 10 20 30 40
σ3 (MPa)
with CQD-P and UCS ≈ 20 MPa.
200
𝜆𝜆 = 9.34
𝜁𝜁 = 1.01
𝛽𝛽 = 0.81
150 𝐶𝐶 = 25.18 MPa
(MPa)
100
σ1

50
Observed Data
CQD = F UCS = 30 MPa

0
-7 0 7 14 21 28 35
σ3 (MPa)
with CQD-F and UCS ≈ 30 MPa.
72
250
𝜆𝜆 = 9.27
𝜁𝜁 = 1.16
200 𝛽𝛽 = 0.65
𝐶𝐶 = 46.53 MPa
150
(MPa)
σ1
100
Observed Data
CQD = F UCS = 40 MPa

0
-14 -7 0 7 14 21 28 35 42 49
σ3 (MPa)
with CQD-F and UCS ≈ 40 MPa.
250
𝜆𝜆 = 10.67
𝜁𝜁 = 1.96
200 𝛽𝛽 = 1.11
𝐶𝐶 = 53.12 MPa
150
(MPa)
σ1
100
Observed Data
CQD = G UCS = 55 MPa

0
-10 0 10 20 30
σ3 (MPa)
with CQD-G and UCS ≈ 55 MPa.
73
300
𝜆𝜆 = 12.50
250 𝜁𝜁 = 4.08
𝛽𝛽 = 1.16
𝐶𝐶 = 76.97 MPa
200 𝜎𝜎𝑡𝑡 = - 7.62 MPa
(MPa)
150
σ1
100

50
Observed Data
CQD = G UCS = 70 MPa

0
-14 -7 0 7 14 21 28 35
σ3 (MPa)
with CQD-G and UCS ≈ 70 MPa.
400
𝜆𝜆 = 13.73
𝜁𝜁 = 2.80
𝛽𝛽 = 0.95
300 𝐶𝐶 = 97.58 MPa
(MPa)
200
σ1

100
Observed Data
CQD = VG UCS = 100 MPa

0
-20 -10 0 10 20 30 40 50
σ3 (MPa)
Figure 4.10. Observed value against predicted value by proposed strength criterion for concrete
specimens with CQD-VG and UCS ≈ 100 MPa.
74
375
𝜆𝜆 = 13.90
𝜁𝜁 = 5.60
300 𝛽𝛽 = 1.21
𝐶𝐶 = 107.44 MPa
𝜎𝜎𝐶𝐶 = 130 MPa
225
(MPa)
σ1
150
Observed Data
CQD = VG UCS = 130 MPa

0
-20 -10 0 10 20 30
σ3 (MPa)
specimens with CQD-VG and UCS ≈ 130 MPa.
740
𝜆𝜆 = 22.84
𝜁𝜁 = 11.10
𝛽𝛽 = 1.28
555 𝐶𝐶 = 162.45 MPa
(MPa)
370
σ1

185
Observed Data
CQD = E UCS > 130 MPa

0
-50 -25 0 25 50 75 100 125
σ3 (MPa)
specimens with CQD-E and UCS > 130 MPa.
75
Tables 4.12 – 4.21 show the absolute deference between the observed values for
strength and the values calculated using the proposed strength criterion for 10 data
groups relating to different types of concrete, Figures 4.3 – 4.12. In Tables 4.12 – 4.21,
σ 1exp represents the observed value or experimental value, σ 1cri represents the strength
value calculated by the strength criterion, and | σ 1exp - σ 1cri | shows the absolute values
for the difference between the observed and the calculated values for the strength.The
relative error related to each data pair is also presented in these tables. As can be seen in
90% of the cases, the average relative error is less than 10 percent. It means that the
strength predicted by the proposed criterion is close to the observed values.
Consequently the criterion shows satisfactory precision in estimating the triaxial
strength of concrete.
Table 4.12. Difference between observed values by experiments and calculated values by proposed
strength criterion for concrete with CQD-VP, UCS < 15 MPa.
σ 1exp (MPa) σ 1cri (MPa) | σ 1exp - σ 1cri | (MPa) Relative Errors (%)
0.00 1.20 1.20 NA
7.00 7.28 0.28 3.85
36.10 31.37 4.73 15.08
46.90 48.42 1.52 3.14
68.00 72.17 4.17 5.78
88.23 89.12 0.89 1.00
107.00 102.71 4.29 4.18
113.01 114.40 1.39 1.21
Average 2.31 4.89
σ 1exp is the observed value.
σ 1cri is the strength value calculated by the strength criterion.
strength criterion for concrete with CQD-P, UCS = 15 MPa.
0.00 2.15 2.15 NA
15.00 11.52 3.48 30.21
26.00 26.75 0.75 2.80
39.00 37.83 1.17 3.09
44.00 48.87 4.87 9.96
60.02 57.31 2.71 4.73
64.35 64.13 0.22 0.34
Average 2.19 8.52
76
strength criterion for concrete with CQD-P, UCS = 20 MPa.
0.00 2.16 2.16 NA
21.00 18.67 2.33 12.48
61.00 58.19 2.81 4.83
75.00 80.10 5.10 6.37
122.00 122.24 0.24 0.20
146.10 145.77 0.33 0.23
Average 2.16 4.82
strength criterion for concrete with CQD-F, UCS = 30 MPa.
0.00 8.26 8.26 NA
31.00 20.40 10.60 51.96
87.62 90.72 3.10 3.42
127.00 127.16 0.16 0.13
164.10 162.08 2.02 1.24
Average 4.83 14.19
strength criterion for concrete with CQD-F, UCS = 40 MPa.
0.00 10.48 10.48 NA
39.87 30.06 9.81 32.63
77.00 69.05 7.95 11.51
99.95 98.58 1.37 1.39
119.10 122.11 3.01 2.47
140.00 141.61 1.61 1.13
152.04 158.25 6.21 3.93
167.00 172.82 5.82 3.37
182.88 185.81 2.93 1.58
200.03 197.60 2.43 1.23
209.79 208.42 1.37 0.66
226.00 218.48 7.52 3.44
Average 5.04 5.76
77
strength criterion for concrete with CQD-G, UCS = 55 MPa.
0.00 3.08 3.08 NA
57.00 58.96 1.96 3.32
79.98 68.39 11.59 16.95
131.00 140.12 9.12 6.51
180.01 187.67 7.66 4.08
210.05 199.23 10.82 5.43
Average 7.37 7.26
strength criterion for concrete with CQD-G, UCS = 70 MPa.
0.00 1.29 1.29 NA
71.00 89.29 18.29 20.48
106.02 106.72 0.70 0.66
169.90 146.62 23.28 15.88
221.45 228.31 6.86 3.00
Average 10.08 10.01
strength criterion for concrete with CQD-VG, UCS = 100 MPa.
0.00 4.64 4.64 NA
101.14 92.80 8.34 8.99
101.14 92.80 8.34 8.99
148.32 145.29 3.03 2.09
185.42 186.77 1.35 0.72
219.05 220.67 1.62 0.73
281.18 294.91 13.73 4.66
370.00 360.65 9.35 2.59
Average 6.01 3.30
78
strength criterion for concrete with CQD-VG, UCS = 130 MPa.
0.00 1.30 1.30 NA
131.98 130.00 1.98 1.52
150.00 138.96 11.04 7.95
190.07 198.01 7.94 4.01
220.00 242.51 22.51 9.28
290.00 268.62 21.38 7.96
Average 11.03 6.14
strength criterion for concrete with CQD-E, UCS > 130 MPa.
0.00 0.01 0.01 NA
190.10 207.94 17.84 8.58
429.00 386.19 42.81 11.08
460.00 451.44 8.56 1.90
471.39 505.87 34.48 6.82
Average 20.74 6.82
4.3.3. Parameters of the Criterion
As previously discussed in Section 4.2, to use a strength criterion to estimate the

strength of the concrete, the parameters of the criterion should be determined first. The
determination of the strength criterion’s parameters can be accomplished by the
following two methods:
• Fitting the criterion on the actual triaxial data and thus calculating the
parameters using the regression analysis, or
• Using the suggested parameters for the certain criterion within the specific
condition recommended for the material.
79
The first method is adopted in this research as it is the most accurate method in
determining the criterion’s parameters. Of course in certain circumstances using the
suggested values for the parameters can be very helpful.
The Nelder and Mead nonlinear regression method (Netter et al., 1988) is used to
calculate the parameters of the proposed strength criterion for different types of
concrete based on empirical suggested concrete quality classification in this research
(CQD) and then the rounded suggested values are presented in Table 4.22 to be used in
practical purposes.
Suggested parameters in this research are calculated using 67 series of triaxial tests (we
called them triaxial data pairs) in ten test series (it is called data groups in this research)
on concrete specimens with different class of strengths based on the new proposed
qualifying classification.
Summarily after eliminating pick maximum and minimum values from the calculation,
the average rounded values of 𝜆𝜆 for concrete vary as follwoing:
9 < 𝜆𝜆 < 18
Similarly 𝜁𝜁 varies as:
1 < 𝜁𝜁 < 5.50
And finally 𝛽𝛽 varies as follows:
0.60 < 𝛽𝛽 < 1.20
The parameter values for proposed strength criterion presented in Tables 4.22 can be
used in application for the strength estimation for different types of concrete. For
precise strength calculations, nonlinear regression of the criterion on actual triaxial data
is recommended. It should be noted that in the case when one prefers to determine the
parameters using the non-linear regression on the actual triaxial data, at least one
uniaxial compression test, one splitting tensile test, and three triaxial compression tests
are required.
80
Table 4.22. Constant values for proposed criterion’s parameters.*
Concrete Type, UCS 𝐶𝐶 𝜎𝜎𝑐𝑐
𝛽𝛽 𝜆𝜆 𝜁𝜁 R2
Based on CQD (MPa) (MPa) (MPa)
VP, UCS < 15 MPa 7.00 16.10 7.28 0.45 7.95 0.54 0.995
P, UCS ≈ 15 MPa 15.00 13.23 11.52 0.87 12.79 1.66 0.995
P, UCS ≈ 20 MPa 21.00 30.11 18.67 0.62 17.76 2.55 0.997
F, UCS ≈ 30 MPa 31.00 25.18 20.40 0.81 9.34 1.01 0.989
F, UCS ≈ 40 MPa 39.87 46.53 30.06 0.65 9.27 1.16 0.992
G, UCS ≈ 55 MPa 57.00 53.12 58.96 1.11 10.67 1.96 0.987
G, UCS ≈ 70 MPa 71.00 76.97 89.29 1.16 12.50 4.08 0.968
VG, UCS ≈ 100 MPa 101.14 97.58 92.80 0.95 13.73 2.80 0.996
VG, UCS ≈ 130 MPa 131.98 107.44 130.00 1.21 13.90 5.60 0.976
E, UCS > 130 MPa 190.10 162.45 207.94 1.28 22.84 11.10 0.980
*
Parameters used for proposed strength criterion’s envelopes in Figure 4.3 to 4.12 are presented.
UCS is the Uniaxial Compressive Strength.
𝐶𝐶 is the Apparent Unconfined Compressive Strength (AUCS).
𝜎𝜎𝑐𝑐 is the Regulated Unconfined Compressive Strength (RUCS).
𝛽𝛽 is a factor that defines RUCS in terms of AUCS.
𝜆𝜆 is concrete constant.
𝜁𝜁 is concrete constant.
𝑅𝑅 2 is the coefficients of determination.
4.4. Conclusions
New strength criterion is developed for concrete based on Bineshian strength criterion
for rocks and coals. A comprehensive strength criterion should cover both linear and
nonlinear envelopes and should be applicable in both tension and compression
quadrants. The proposed criterion is based on this insight and satisfies the features that a
comprehensive criterion should enjoy.
The proposed strength criterion is defined in terms of three dimensionless parameters

(𝜆𝜆, 𝜁𝜁 and 𝛽𝛽). Parameters 𝜆𝜆 and 𝜁𝜁 represent the internal characteristics of concrete, 𝛽𝛽 is
an adjustment factor to 𝐶𝐶. It is a function of the unconfined compressive strength (𝜎𝜎𝑐𝑐 )
obtained by the Mohr-Coulomb criterion. It adjusts the value of 𝜎𝜎𝑐𝑐 to reflect the trend of
the triaxial test data.
81
𝜎𝜎𝑐𝑐 is a regulated value of 𝐶𝐶. It is called the regulated uniaxial compressive strength
(RUCS). Proposed criterion for concrete, estimates 𝜎𝜎𝑐𝑐 , 0.91 times the value of 𝐶𝐶.
𝐶𝐶 is called the apparent uniaxial compressive strength (AUCS), which is obtained from
the Mohr-Coulomb criterion.
The proposed criterion predicts the tensile strength close to UTS and more accurate than
others. The value of the uniaxial tensile strength of concrete 𝜎𝜎𝑡𝑡 is, according to the
proposed criterion, about 1.55 × 𝑈𝑈𝑈𝑈𝑈𝑈.
A rigorous program of comparing the estimated strengths to the actual test data has
suggested that the criterion can provide linear and nonlinear envelopes capable of
describing the majority of triaxial test data. The proposed criterion can be applied to
compressive and tensile strength data and provide estimates of these values that are
similar to those determined in laboratory.
The coefficient of determination for proposed strength criterion is in many cases close
to 1 that indicates a good correlation between the criterion and actual triaxial data for
concrete types.
The accordance coefficient for proposed criterion shows an excellent accordance

between triaxial data and the criterion. Thus the estimation of triaxial strength by the
proposed criterion is credible.
The parameters of the proposed strength criterion are also determined for different types
of concrete based on their CQD. The values of the parameters suggested in this research
can be used in practice with an acceptable accuracy.
82
CHAPTER 5:
CONCRETE DAMAGE PLASTICITY MODEL
5.1. Introduction
In addition to the experimental investigation on concrete mechanics, researchers using

analytical and numerical methods studied strain softening in compression and tension
and therefore different constitutive models for softening were developed (Hillerborg,
1979; Bazant, 1989; Lubliner et al., 1989; Hillerborg, 1990; Markeset, 1995; Lee and
Fenves, 1998; Bazant and Novak, 2000; Bazant, 2002).
Bazant (1989) classified post-peak behaviour into two response paths: one is the strain
softening localizing into a certain zone within the specimen and another one is the
unloading path that the rest of the specimen undergoes. He showed the entire specimen
response as a combination of localization and unloading response paths. He stated that
localization at strain softening in the model must begin right at the peak-stress state.
Markeset and Hillerborg (1995) proposed a model – the Compressive Damage Zone
(CDZ) – that takes both the localized shear deformation and the deformation due to
splitting cracks into account. According to the CDZ model, the steepness of the
descending branch of a formal 𝜎𝜎 − 𝜀𝜀 curve will increase with increasing specimen
length and slenderness ratio.
Another model that is widely applicable in numerical simulations for concrete failure
response is Concrete Damage Plasticity (CDP) model. The CDP model realized using
the finite element method (e.g. Abaqus) provides a means for modeling concrete in all
types of structures. It uses the concept of isotropic damage elasticity in combination
with isotropic tensile and compressive plasticity to represent the inelastic behaviour of
concrete. It can be used for plain concrete.
The CDP has been selected to simulate concrete’s strain softening in this research. The
parameters for the constitutive model are determined in the present research and proper
results obtained.
83
5.2. CDP Theory
Capabilities of CDP model realized in Abaqus are as follows:
• Modeling the concrete in all types of structures including beams, trusses, shells,
and solids, as well as all quasi-brittle materials
• Modeling the inelastic behaviour of concrete using concepts of isotropic damage

elasticity in combination with isotropic tensile and compressive plasticity,
• Modeling concrete reinforcement (rebar),
• Modeling the monotonic, cyclic, and/or dynamic loading under low confining
pressures,
• Describing the irreversible damage that occurs during the fracturing process
using the combination of non-associated multi-hardening plasticity and
scalar/isotropic damage elasticity,
• Defining the rate of straining, and
• Allowing the user to control the stiffness recovery effects during cyclic load
reversals.
However, it requires the elastic behavior of the material to be isotropic and linear.
5.2.1. Mechanical Response
The CDP model assumes two main failure mechanisms; tensile cracking, and
compressive crushing (Hillerborg et al., 1976; Lubliner et al., 1989; Lee and Fenves,
1998; Abaqus Theory Manual, 2002; Abaqus User’s Manual, 2011).
The evolution of the yield or failure surface is controlled by two hardening variables,
~𝑝𝑝𝑝𝑝
𝜀𝜀𝑡𝑡 and 𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 , linked to the above tensile and compressive failure mechanisms. The
84
model assumes that the uniaxial tensile and compressive response of concrete is
characterized by damage plasticity.
Figures 5.1 and 5.2 present the mechanical behaviour of plain concrete to uniaxial
loading in compression and tension respectively.
Under uniaxial compression the response is linear until the value of initial yield stress,
𝜎𝜎𝑐𝑐0 is reached. In the plastic regime the response is typically characterized by stress
hardening followed by strain softening beyond the strain that corresponds to the
ultimate (peak) stress, 𝜎𝜎𝑐𝑐𝑐𝑐 . This representation, although somewhat simplified, captures
the main features of the response of concrete.
𝜎𝜎𝑐𝑐
E0
(1−𝑑𝑑𝑐𝑐)E0
𝜀𝜀𝑐𝑐~𝑝𝑝𝑙𝑙 𝜀𝜀𝑐𝑐0𝑙𝑙 𝜀𝜀𝑐𝑐

~𝑝𝑝𝑝𝑝
Figure 5.1. Mechanical response of concrete to uniaxial loading in compression. 𝜀𝜀𝑐𝑐 is the equivalent
compressive plastic strains, 𝑑𝑑𝑐𝑐 is the compressive damage variable, 𝐸𝐸0 is the initial elastic
stiffness, 𝜎𝜎𝑐𝑐0 is the value of initial yield in compression, and 𝜎𝜎𝑐𝑐𝑐𝑐 is the ultimate uniaxial
stress.
Under the uniaxial tension the 𝜎𝜎 − 𝜀𝜀 response follows a linear elastic relationship until
the value of the failure stress, 𝜎𝜎𝑡𝑡0 , is reached.
The failure strain corresponds to the onset of micro cracking in the material. Beyond the
failure strain, the formation of micro-cracks is represented macroscopically with a
softening 𝜎𝜎 − 𝜀𝜀 response, which induces strain localization in the concrete.
85
𝜎𝜎t 0
𝜎𝜎t
E0
(1−𝑑𝑑t)E0

~𝑝𝑝𝑝𝑝
Figure 5.2. Mechanical response of concrete to uniaxial loading in tension. 𝜀𝜀𝑡𝑡 is the equivalent tensile
plastic strain, 𝑑𝑑𝑡𝑡 is the tensile damage variable, 𝐸𝐸0 is the initial elastic stiffness, and 𝜎𝜎𝑡𝑡0 is the
value of initial yield in tension.
Assuming that the uniaxial 𝜎𝜎 − 𝜀𝜀 curves can be converted into stress versus plastic-
strain curves, we will have (Abaqus Theory Manual, 2002),
~𝑝𝑝𝑝𝑝 ~̇ 𝑝𝑝𝑝𝑝
𝜎𝜎𝑡𝑡 = 𝜎𝜎𝑡𝑡 (𝜀𝜀𝑡𝑡 , 𝜀𝜀𝑡𝑡 , 𝜃𝜃, 𝑓𝑓𝑖𝑖 ) (5.1)
𝜎𝜎𝑐𝑐 = 𝜎𝜎𝑐𝑐 (𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 , 𝜀𝜀𝑐𝑐~̇ 𝑝𝑝𝑝𝑝 , 𝜃𝜃, 𝑓𝑓𝑖𝑖 ) (5.2)
~𝑝𝑝𝑝𝑝
where the subscripts 𝑡𝑡 and 𝑐𝑐 refer to tension and compression respectively; 𝜀𝜀𝑡𝑡 and
𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 are the equivalent tensile and compressive plastic strains, 𝜀𝜀𝑡𝑡
and 𝜀𝜀𝑐𝑐~̇ 𝑝𝑝𝑝𝑝 are the
equivalent plastic strain rates, 𝜃𝜃 is the temperature, and 𝑓𝑓𝑖𝑖 , 𝑖𝑖 = 1,2, … are other
predefined field variables.
As shown in Figures 5.1 and 5.2, when the concrete specimen is unloaded from any
point on the strain softening branch of the 𝜎𝜎 − 𝜀𝜀 curves, the unloading response is
weakened: the elastic stiffness of the already damaged or degraded material is lower.
The degradation of the elastic stiffness is characterized by two damage variables, 𝑑𝑑𝑡𝑡 and
𝑑𝑑𝑐𝑐 as below:
~𝑝𝑝𝑝𝑝
𝑑𝑑𝑡𝑡 = 𝑑𝑑𝑡𝑡 (𝜀𝜀𝑡𝑡 , 𝜃𝜃, 𝑓𝑓𝑖𝑖 ) (5.3)
86
𝑑𝑑𝑐𝑐 = 𝑑𝑑𝑐𝑐 (𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 , 𝜃𝜃, 𝑓𝑓𝑖𝑖 ) (5.4)
where 𝑑𝑑𝑡𝑡 and 𝑑𝑑𝑐𝑐 , are functions of the plastic strains, temperature, and field variables.
The damage variables can take values from zero, representing the undamaged material,
to one, which represents the total loss of strength (0 ≤ 𝑑𝑑𝑡𝑡 ≤ 1 and 0 ≤ 𝑑𝑑𝑐𝑐 ≤ 1).
The 𝜎𝜎 − 𝜀𝜀 relations under uniaxial tension and compression are, respectively:
~𝑝𝑝𝑝𝑝
𝜎𝜎𝑡𝑡 = (1 − 𝑑𝑑𝑡𝑡 )𝐸𝐸0 �𝜀𝜀𝑡𝑡 − 𝜀𝜀𝑡𝑡 � (5.5)
𝜎𝜎𝑐𝑐 = (1 − 𝑑𝑑𝑐𝑐 )𝐸𝐸0 �𝜀𝜀𝑐𝑐 − 𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 � (5.6)
where 𝐸𝐸0 is the initial elastic stiffness (stiffness of the undamaged material).
The effective tensile and compressive cohesion stresses can be defined as follows (The
effective cohesion stresses determine the size (geometric dimensions) of the yield or
failure surface):
𝜎𝜎 ~𝑝𝑝𝑝𝑝
𝜎𝜎�𝑡𝑡 = (1−𝑑𝑑𝑡𝑡 ) = 𝐸𝐸0 �𝜀𝜀𝑡𝑡 − 𝜀𝜀𝑡𝑡 � (5.7)
𝑡𝑡
𝜎𝜎
𝜎𝜎�𝑐𝑐 = (1−𝑑𝑑𝑐𝑐 ) = 𝐸𝐸0 �𝜀𝜀𝑐𝑐 − 𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 � (5.8)
𝑐𝑐
5.2.2. Uniaxial Cyclic Response
Under uniaxial cyclic loading conditions the degradation mechanisms are quite
complex, involving opening and closing of the previously formed micro-cracks, as well
as their interaction. Experimentally, it is observed that there is some recovery of the
elastic stiffness as the load changes sign during the cyclic test. The stiffness recovery
effect, also known as the “unilateral effect,” is an important aspect of the concrete
behavior under cyclic loading. The effect is usually more pronounced as the load
changes from tension to compression, causing tensile cracks to close, which results in
the recovery of the compressive stiffness.
87
The concrete damage plasticity model assumes that the reduction of the elastic modulus
is given in terms of a scalar degradation (damage) variable 𝑑𝑑 as:
𝐸𝐸 = (1 − 𝑑𝑑)𝐸𝐸0 (5.9)
where 𝐸𝐸0 is the initial or modulus of the undamaged material.
This expression holds both in the tensile (𝜎𝜎11 > 0) and the compressive (𝜎𝜎11 < 0)
phases of the cycle. The stiffness degradation variable, 𝑑𝑑, is a function of the stress state
and the uniaxial damage variables, 𝑑𝑑𝑡𝑡 and 𝑑𝑑𝑐𝑐 . For the uniaxial cyclic conditions it is
assumed that,
(1 − 𝑑𝑑) = (1 − 𝑠𝑠𝑡𝑡 𝑑𝑑𝑐𝑐 )(1 − 𝑠𝑠𝑐𝑐 𝑑𝑑𝑡𝑡 ) (5.10)
where 𝑠𝑠𝑡𝑡 and 𝑠𝑠𝑐𝑐 are functions of the stress state introduced to model stiffness recovery
effects associated with stress reversals. They are defined as follows
𝑠𝑠𝑡𝑡 = 1 − 𝑤𝑤𝑡𝑡 𝑟𝑟 ∗ (𝜎𝜎11 ) (5.11)
𝑠𝑠𝑐𝑐 = 1 − 𝑤𝑤𝑐𝑐 (1 − 𝑟𝑟 ∗ (𝜎𝜎11 )) (5.12)
where
0 ≤ 𝑤𝑤𝑡𝑡 ≤ 1
0 ≤ 𝑤𝑤𝑐𝑐 ≤ 1
1, 𝜎𝜎11 > 0
𝑟𝑟 ∗ (𝜎𝜎11 ) = 𝐻𝐻(𝜎𝜎11 ) = �
0, 𝜎𝜎11 < 0
The weight factors 𝑤𝑤𝑡𝑡 and 𝑤𝑤𝑐𝑐 , which are assumed to be material properties, control the
recovery of the tensile and compressive stiffness upon load reversal. To illustrate this,
consider the example in Figure 5.3, where the load changes from tension to
compression.
88
𝜎𝜎t 0
𝜎𝜎t
E0
(1−𝑑𝑑t)E0
𝑤𝑤𝑐𝑐 = 0
𝑤𝑤𝑐𝑐=1 𝜀𝜀t
Figure 5.3. Effect of the compression stiffness recovery. 𝑑𝑑𝑡𝑡 is the tensile damage variable, 𝐸𝐸0 is the initial
elastic stiffness, 𝜎𝜎𝑡𝑡0 is the value of initial yield in tension, and 𝑤𝑤𝑐𝑐 is the weight factor.
Assume that there was no previous compressive damage (crushing) in the material,
therefore,
𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 = 0
and
𝑑𝑑𝑐𝑐 = 0
Then,
(1 − 𝑑𝑑) = (1 − 𝑠𝑠𝑐𝑐 𝑑𝑑𝑡𝑡 ) = (1 − (1 − 𝑤𝑤𝑐𝑐 (1 − 𝑟𝑟 ∗ ))𝑑𝑑𝑡𝑡 ) (5.13)
In tension (𝜎𝜎11 > 0), 𝑟𝑟 ∗ =1, therefore,
𝑑𝑑 = 𝑑𝑑𝑡𝑡 as expected.
In compression (𝜎𝜎11 > 0), 𝑟𝑟 ∗ =0, and
𝑑𝑑 = (1 − 𝑤𝑤𝑐𝑐 )𝑑𝑑𝑡𝑡 (5.14)
89
If 𝑤𝑤𝑐𝑐 = 1,
then 𝑑𝑑 = 0, therefore, the material fully recovers the compressive stiffness (which in
this case is the initial stiffness, 𝐸𝐸 = 𝐸𝐸0 ).
If, on the other hand, 𝑤𝑤𝑐𝑐 = 0 , then 𝑑𝑑 = 𝑑𝑑𝑡𝑡 and there is no stiffness recovery.
Intermediate values of 𝑤𝑤𝑐𝑐 result in partial recovery of the stiffness.
5.2.3. Multiaxial Response
The stress–strain relations for the general three-dimensional multiaxial conditions are
given by the scalar damage elasticity equation (Kachanov, 1958, Rabotnov, 1969,
Kupfer et al., 1979, Abaqus Theory Manual, 2002, Jankowiak and Lodygowski, 2005):
𝜎𝜎 = (1 − 𝑑𝑑)𝐷𝐷0𝑒𝑒𝑒𝑒 : (𝜀𝜀 − 𝜀𝜀 𝑝𝑝𝑝𝑝 ) = 𝐷𝐷0𝑒𝑒𝑒𝑒 : (𝜀𝜀 − 𝜀𝜀 𝑝𝑝𝑝𝑝 ) (5.15)
where 𝐷𝐷0𝑒𝑒𝑒𝑒 is the initial (undamaged) elasticity matrix. The previous expression for the
scalar stiffness damage variable, 𝑑𝑑 is generalized to the multiaxial stress case by
replacing the unit step function 𝑟𝑟 ∗ (𝜎𝜎11 ) with a multiaxial stress weight factor, 𝑟𝑟(𝜎𝜎�),
defined as,
∑3𝑖𝑖=1〈𝜎𝜎�𝚤𝚤 〉
𝑟𝑟(𝜎𝜎�) = ∑3𝑖𝑖=1|𝜎𝜎�𝚤𝚤 |
(5.16)
where 𝜎𝜎�𝚤𝚤 is the principal stress components where 𝑖𝑖 = 1, 2, 3, and 0 ≤ 𝑟𝑟(𝜎𝜎�) ≤ 1. The
1
Macauley bracket 〈. 〉 is defined by 〈𝑥𝑥〉 = 2 (|𝑥𝑥| + 𝑥𝑥).
5.2.4. Post-failure Stress–Strain Relation
In the CDP model, the post-failure stress can be defined as a function of cracking strain,
𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 . The cracking strain is defined as the total strain minus the elastic strain
corresponding to the undamaged material:
𝑒𝑒𝑒𝑒
𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 = 𝜀𝜀𝑡𝑡 − 𝜀𝜀0𝑡𝑡 (5.17)
90
where;
𝜀𝜀0𝑡𝑡 = 𝐸𝐸𝑡𝑡 (5.18)
0
Figure 5.4 presents an illustration to define the cracking strain (𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 ) used for the
definition of tension stiffening.
𝜎𝜎t 0
𝜎𝜎t
E0
E0
(1−𝑑𝑑t)E0
𝜀𝜀t~𝑝𝑝𝑙𝑙 𝜀𝜀te𝑙𝑙 𝜀𝜀t

𝜀𝜀t~ck 𝜀𝜀te𝑙𝑙
Figure 5.4. Definition of the cracking strain 𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 used for the definition of tension stiffening data for CDP
~𝑝𝑝𝑝𝑝
model. 𝜀𝜀𝑡𝑡 is the equivalent tensile plastic strain, 𝑑𝑑𝑡𝑡 is the tensile damage variable, 𝐸𝐸0 is the
initial elastic stiffness, and 𝜎𝜎𝑡𝑡0 is the value of initial yield in tension.
The tension stiffening is controlled by the cracking strain, 𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 . The cracking strain
values can be converted to plastic strain values using below relationship,
~𝑝𝑝𝑝𝑝 𝑑𝑑 𝜎𝜎
𝜀𝜀𝑡𝑡 = 𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 − (1−𝑑𝑑𝑡𝑡 ) 𝐸𝐸𝑡𝑡 (5.19)
𝑡𝑡 0
~𝑝𝑝𝑝𝑝
In the absence of tensile damage 𝜀𝜀𝑡𝑡 = 𝜀𝜀𝑡𝑡~𝑐𝑐𝑐𝑐 . In the case of plain concrete, the
specification of a post-failure stress–strain relation introduces mesh sensitivity of the
results, in the sense that the finite element predictions do not converge to a unique
solution, as the mesh is refined because mesh refinement leads to narrower crack bands.
This problem typically occurs if the cracking failure concentrated only in the localized
regions and mesh refinement does not result in the formation of additional cracks. If the
cracking damage is distributed evenly (due to the presence of stabilizing elastic
91
material, as in the case of plate bending), the mesh sensitivity in the analysis is less of a
concern. The choice of the tension stiffening parameters is important since, generally,
more tension stiffening makes it easier to obtain numerical solutions. Too little tension
stiffening will cause the local cracking failure in the concrete to introduce temporarily
unstable behavior in the overall response of the model. The presence of this type of
response in the analysis model usually indicates that the tension stiffening is
unreasonably low.
5.2.5. Fracture Energy Cracking Criterion
In the case of plain concrete in significant regions of the model, the tension stiffening
approach described above will introduce the unwanted mesh sensitivity into the results.
However, it is generally accepted that Hillerborg et al.’s (1976) fracture energy proposal
is adequate to allay the concern for many practical purposes.
In the Linear Elastic Fracture Mechanics the energy required to produce and open a unit
area of crack, 𝐺𝐺𝑓𝑓 is considered as a material parameter. With this approach the brittle
behavior of concrete is characterized by a stress–displacement response rather than a
stress–strain response. Under tension a concrete specimen will crack across some
section. After it has been pulled apart sufficiently for most of the stress to be removed
(so that the undamaged elastic strain is small), its length will be determined primarily by
the opening at the crack (Hillerborg et al., 1976). The opening does not depend on the
specimen's length. This fracture energy cracking model can be invoked by specifying
the post-failure stress as a tabular function of cracking displacement, Figure 5.5.
𝜎𝜎t
utck
Figure 5.5. Post-failure stress–displacement curve.
92
Alternatively, the fracture energy, 𝐺𝐺𝑓𝑓 , can be specified directly as a material property; in
this case, define the failure stress, 𝜎𝜎𝑡𝑡0 , as a tabulated function of the associated fracture
energy. This model assumes a linear loss of strength after cracking, Figure 5.6.
𝜎𝜎t
𝜎𝜎t0
ut
Figure 5.6. Post-failure stress–fracture energy curve.
The cracking displacement at which complete loss of strength takes place therefore is
2𝐺𝐺𝑓𝑓
𝑢𝑢𝑡𝑡0 = (5.20)
𝜎𝜎𝑡𝑡0
Values of fracture energy 𝐺𝐺𝑓𝑓 range from 40 N/m for a typical construction concrete (it is
classified as CQD-P with a compressive strength of approximately 20 MPa; see Chapter
3), to 120 N/m for a high strength concrete (it is classified as CQD-G with a
compressive strength of approximately 40 MPa; see Chapter 3).
If the post-peak tensile softening damage, 𝑑𝑑𝑡𝑡 , is specified, the cracking displacement
values can be converted to “plastic” displacement values using below relationship:
𝑝𝑝𝑝𝑝 𝑑𝑑 𝜎𝜎𝑡𝑡 𝑙𝑙0

𝑢𝑢𝑡𝑡 = 𝑢𝑢𝑡𝑡𝑐𝑐𝑐𝑐 − (1−𝑑𝑑𝑡𝑡 ) (5.21)
𝑡𝑡 𝐸𝐸0
where the plain concrete’s specimen length, 𝑙𝑙0 , is assumed to be one unit length as
below:
𝑙𝑙0 = 1.
93
5.2.6. Response to Compressive Load
The stress–strain behavior of plain concrete in uniaxial compression can be defined

outside the elastic range. Compressive stress data are provided as a tabulated function of
inelastic (or crushing) strain, 𝜀𝜀𝑐𝑐~𝑖𝑖𝑖𝑖 , and, if desired, strain rate, temperature, and field
variables. Positive (absolute) values should be given for the compressive stress and
strain.
The stress–strain curve can be defined beyond the ultimate stress, into the strain-
softening regime. Hardening data are given in terms of an inelastic strain, 𝜀𝜀𝑐𝑐~𝑖𝑖𝑖𝑖 , instead
of plastic strain, 𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 . The compressive inelastic strain is defined as the total strain
𝑒𝑒𝑒𝑒
minus the elastic strain corresponding to the undamaged material, 𝜀𝜀𝑐𝑐~𝑖𝑖𝑖𝑖 = 𝜀𝜀𝑐𝑐 − 𝜀𝜀0𝑐𝑐 ,
where 𝜀𝜀0𝑐𝑐 = 𝐸𝐸𝑐𝑐 , as illustrated in Figure 5.7. The inelastic strain values can be converted
0
to plastic strain values as below:
𝑑𝑑 𝜎𝜎𝑐𝑐
𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 = 𝜀𝜀𝑐𝑐~𝑖𝑖𝑖𝑖 − (1−𝑑𝑑𝑐𝑐 (5.22)
𝑐𝑐 ) 𝐸𝐸0
In the absence of compressive damage 𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 = 𝜀𝜀𝑐𝑐~𝑖𝑖𝑖𝑖 .

𝜎𝜎𝑐𝑐
E0
(1−𝑑𝑑𝑐𝑐)E0 E0
𝜀𝜀𝑐𝑐~𝑝𝑝𝑙𝑙 𝜀𝜀𝑐𝑐e𝑙𝑙
𝜀𝜀𝑐𝑐 ~in 𝜀𝜀0ce𝑙𝑙 𝜀𝜀𝑐𝑐
Figure 5.7. Definition of the compressive inelastic (crushing) strain 𝜀𝜀𝑐𝑐~𝑖𝑖𝑖𝑖 used for the definition of
~𝑝𝑝𝑝𝑝
compression hardening data. 𝜀𝜀𝑐𝑐 is the equivalent compressive plastic strain, 𝑑𝑑𝑐𝑐 is the
compressive damage variable, 𝐸𝐸0 is the initial elastic stiffness, 𝜎𝜎𝑐𝑐0 is the value of initial yield
in compression, and 𝜎𝜎𝑐𝑐𝑐𝑐 is the ultimate uniaxial stress.
94
5.2.7. Yield Surface
The CDP model uses a yield condition based on the yield function proposed by Lubliner
et al. (1989) and incorporates the modifications proposed by Lee and Fenves (1998) to
account for different evolution of strength under tension and compression (Figure 5.8).
In terms of effective stresses the yield function takes below form (Kupfer et al., 1979;
Jankowiak and Lodygowski, 2005):
1
𝐹𝐹 = 1−𝛼𝛼 (𝑞𝑞� − 3𝛼𝛼𝑝𝑝̅ + 𝛽𝛽(𝜀𝜀 ~𝑝𝑝𝑝𝑝 )〈𝜎𝜎�𝑚𝑚𝑚𝑚𝑚𝑚 〉 − 𝛾𝛾〈−𝜎𝜎�𝑚𝑚𝑚𝑚𝑚𝑚 〉) − 𝜎𝜎�𝑐𝑐 (𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 ) = 0 (5.23)
𝜎𝜎 −𝜎𝜎 (𝜎𝜎 ⁄𝜎𝜎 )−1

𝛼𝛼 = 2𝜎𝜎𝑏𝑏0 −𝜎𝜎𝑐𝑐0 = 2(𝜎𝜎𝑏𝑏0 ⁄𝜎𝜎𝑐𝑐0 )−1 (5.24)
𝑏𝑏0 𝑐𝑐0 𝑏𝑏0 𝑐𝑐0
~𝑝𝑝𝑝𝑝
�𝑐𝑐 �𝜀𝜀𝑐𝑐
𝜎𝜎 �
𝛽𝛽(𝜀𝜀 ~𝑝𝑝𝑝𝑝 ) = ~𝑝𝑝𝑝𝑝 (1 − 𝛼𝛼) − (1 + 𝛼𝛼) (5.25)
�𝑡𝑡 �𝜀𝜀𝑡𝑡 �
𝜎𝜎
1
𝑝𝑝̅ = − 3 𝜎𝜎�: 𝐼𝐼 (5.26)
3
𝑞𝑞� = �2 𝑆𝑆̅: 𝑆𝑆̅ (5.27)
𝑆𝑆̅ = 𝑝𝑝̅ 𝐼𝐼 + 𝜎𝜎� (5.28)
𝜎𝜎� = 𝐷𝐷0𝑒𝑒𝑒𝑒 : (𝜀𝜀 − 𝜀𝜀 𝑝𝑝𝑝𝑝 ) (5.29)
where 𝛼𝛼, is a dimensionless material constant calculated based on Kupfer’s (Kupfer et

al., 1979) curve, 𝛾𝛾 is the shape of loading surface in the deviatoric plane (Jankowiak
and Lodygowski, 2005) that is a dimensionless material constant, 𝑝𝑝̅ is the effective
hydrostatic pressure, 𝑞𝑞� is the Mises equivalent effective stress, 𝜎𝜎� is the effective stress
tensor, 𝜎𝜎�𝑐𝑐 is the effective compressive cohesion stress, 𝜎𝜎�𝑡𝑡 is the effective tensile
cohesion stress, 𝜎𝜎�𝑚𝑚𝑚𝑚𝑚𝑚 is the maximum principal effective stress (algebraically maximum
eigenvalue of 𝜎𝜎�), 𝜎𝜎𝑏𝑏0 is the initial equibiaxial compressive yield stress, 𝜎𝜎𝑐𝑐0 is the initial
uniaxial compressive yield stress (the uniaxial compressive stress at failure), 𝑆𝑆̅ is the
deviatoric part of 𝜎𝜎�, 𝐷𝐷0𝑒𝑒𝑒𝑒 is the initial undamaged elasticity matrix, 𝜀𝜀 𝑝𝑝𝑝𝑝 is the plastic
strain, and 𝜀𝜀 is the stress tensor.
95
Kc=2/3 -S1
-S2
Kc=1
Tensile
Meridian
Compressive Meridian
-S3
Figure 5.8. Yield surface in plane stress corresponding to different values of the ratio of the second stress
invariant (𝐾𝐾𝑐𝑐 ) on the tensile meridian.
5.2.8. Plastic Flow Potential
Based on Drucker-Prager hyperbolic function, the flow potential for Concrete Damage
Plasticity (CDP) constitutive model is defined as below:
𝐺𝐺 = �(𝜎𝜎𝑐𝑐0 − 𝜖𝜖𝜎𝜎𝑡𝑡0 tan 𝜓𝜓)2 + 𝑞𝑞� 2 − 𝑝𝑝̅ tan 𝜓𝜓 − 𝜎𝜎 (5.30)
𝜎𝜎 = (1 − 𝑑𝑑)𝜎𝜎� (5.31)
where 𝜎𝜎𝑐𝑐0 is the uniaxial compressive stress at failure, 𝜎𝜎𝑡𝑡0 is the uniaxial tensile stress at
failure, 𝜖𝜖 is a parameter, referred to as the eccentricity, that defines the rate at which the
function approaches the asymptote; the flow potential tends to a straight line as the
eccentricity tends to zero, 𝜓𝜓 is the dilation angle measured in the 𝑝𝑝 − 𝑞𝑞 plane at high
confining pressure, 𝑝𝑝̅ is the effective hydrostatic pressure, 𝑞𝑞� is the Mises equivalent
effective stress, 𝜎𝜎 is the Cauchy stress, 𝑑𝑑 is the scalar stiffness degradation variable, and
𝜎𝜎� is the effective stress tensor.
96
5.3. Visualization of Crack Directions
Unlike concrete models based on the smeared crack approach, the concrete damage
plasticity model does not have the notion of cracks developing at the material
integration point. However, it is possible to introduce the concept of an effective crack
direction with the purpose of obtaining a graphical visualization of the cracking patterns
in the concrete structure. Different criteria can be adopted within the framework of
scalar-damage plasticity for the definition of the direction of cracking. Following
Lubliner et al. (1989), we can assume that cracking initiates at points where the tensile
~𝑝𝑝𝑝𝑝
equivalent plastic strain is greater than zero, 𝜀𝜀𝑡𝑡 > 0, and the maximum principal
plastic strain is positive. The direction of the vector normal to the crack plane is
assumed to be parallel to the direction of the maximum principal plastic strain.
5.4. Output
In addition to the standard output identifiers available in Abaqus, the following

variables relate specifically to material points in the concrete damage plasticity model:
• DAMAGEC, Compressive damage variable, 𝑑𝑑𝑐𝑐 .

• DAMAGET, Tensile damage variable, 𝑑𝑑𝑡𝑡 .
• PEEQ, Compressive equivalent plastic strain, 𝜀𝜀𝑐𝑐~𝑝𝑝𝑝𝑝 .
~𝑝𝑝𝑝𝑝
• PEEQT, Tensile equivalent plastic strain, 𝜀𝜀𝑡𝑡 .
• SDEG, Stiffness degradation variable, 𝑑𝑑.
• DMENER, Energy dissipated per unit volume by damage.
• ELDMD, Total energy dissipated in the element by damage.
• ALLDMD, Energy dissipated in the whole (or partial) model by damage. The
contribution from ALLDMD is included in the total strain energy ALLIE.
• EDMDDEN, Energy dissipated per unit volume in the element by damage.
• SENER, The recoverable part of the energy per unit volume.
• ELSE, The recoverable part of the energy in the element.
• ALLSE, The recoverable part of the energy in the whole (partial) model.
• ESEDEN, The recoverable part of the energy per unit volume in the element.
97
5.5. Identification of Constitutive Parameters
Based on CDP behavioral model, mechanical response of concrete for both failure and
post-failure regimes depends on following constitutive parameters and scalar values:
• 𝛾𝛾 is the shape of loading surface in the deviatoric plane.

𝜎𝜎𝑏𝑏0
• is the ratio of the biaxial and uniaxial compressive stresses at failure.
𝜎𝜎𝑐𝑐0
• 𝛼𝛼 is a dimensionless material constant.

• 𝜓𝜓 is the dilation angle measured in the 𝑝𝑝 − 𝑞𝑞 plane at high confining pressure.
• 𝜖𝜖 is a parameter related to the eccentricity.
• 𝐸𝐸 is the Young's modulus.
• 𝜈𝜈 is the Poisson’s ratio.
• 𝜎𝜎𝑐𝑐0 is the uniaxial compressive strength.
• 𝜎𝜎𝑡𝑡0 is the uniaxial tensile strength.
Uniaxial compression and tension, triaxial compression, and biaxial failure in plane
state of stress are required to determine the CDP constitutive model’s parameters. All
uniaxial/triaxial tests are conducted on CQD-G class concrete in this research (Chapter
3). The biaxial in plane state of stress, which is the Kupfer’s curve for concrete, is
extracted from Jankowiak and Lodygowski (2005) for similar class of concrete. Figure
5.9 and 5.10 show the uniaxial tensile and compressive tests respectively that are
conducted on concrete class CQD-G.
4
Stress (MPa)
0
0 0.0002 0.0004 0.0006
Strain (%)
Figure 5.9. Uniaxial tensile test conducted on concrete class CQD-G.
98
60
Stress (MPa) 40
20
0
0 0.001 0.002 0.003 0.004
Strain (%)
Figure 5.10. Uniaxial compressive test conducted on concrete class CQD-G.
Using the results of the aforementioned tests, the parameters, which determine the shape
of the flow potential surface in the deviatoric and meridian plane and the evolution rule
of the material parameters; the hardening and the softening rule in tension and
compression, can be identified (Jankowiak and Lodygowski, 2005).
Parameter 𝛾𝛾 describes the shape of the yield function. It is determined based on triaxial
compression tests on concrete. According to Lubliner et al. (1989), Jankowiak and
Lodygowski (2005), and the laboratory triaxial compression tests conducted in this
research, the value of 𝛾𝛾 can be as below:
𝛾𝛾 = 0.666
𝜎𝜎𝑏𝑏0
Parameter also describes the shape of the yield surface. It can be determined using
𝜎𝜎𝑐𝑐0
yield surface curve in plane stress. Kupfer’s curve is an example that can be fined in
Kupfer et al. (1979) and Jankowiak and Lodygowski (2005). Figure 5.11 shows the
yield surface curve.
Parameter 𝛼𝛼 depends on the ratio of the initial equibiaxial compressive yield stress (𝜎𝜎𝑏𝑏0)
and the initial uniaxial compressive yield stress; the uniaxial compressive stress at
failure (𝜎𝜎𝑐𝑐0 ). Therefore it simply can be calculated using Equation (5.24) when the ratio
𝜎𝜎𝑏𝑏0
of is determined:
𝜎𝜎𝑐𝑐0
99
𝜎𝜎𝑏𝑏0
= 1.12
𝜎𝜎𝑐𝑐0
α = 0.097
Parameters 𝜓𝜓 and 𝜖𝜖 describe the shape of the flow potential function. Jankowiak and
Lodygowski (2005) stated that these parameters could be obtained from best fitting in
the meridian plane (𝑝𝑝 − 𝑞𝑞 plane) of the flow potential surface to the experimental
results. They determined the constitutive parameters after a minimization of the error
procedure as:
𝜓𝜓 = 0.68 𝑟𝑟𝑟𝑟𝑟𝑟 = 39∘
𝜖𝜖 = 1.0
It is assumed that a range of concrete elasticity is 0.3𝜎𝜎𝑐𝑐0 in compression and 0.7𝜎𝜎𝑡𝑡0 in

tension, therefore in the elastic zone the Young's modulus and the Poisson’s ratio equal
as below (Jankowiak and Lodygowski, 2005):
𝐸𝐸 = 19.7 GPa
𝜈𝜈 = 0.19
The values of 𝜎𝜎𝑐𝑐0 and 𝜎𝜎𝑡𝑡0 can be obtained from uniaxial compressive and tensile tests
on concrete specimens. These values in this research can be calculated from the
proposed strength criterion (see Chapter 4), or obtained from the uniaxial
compression/tension tests:
𝜎𝜎𝑐𝑐0 = 54.77 MPa
𝜎𝜎𝑡𝑡0 = 3.62 MPa
Finally Table 5.1 summarizes the CDP’s constitutive material parameters for concrete
class CQD-G. Tables 5.2 – 5.5 show the hardening and softening rule and the evolution
of the scalar damage variable for compression and tension depend on the crushing or
cracking strains (the source of these data is Jankowiak and Lodygowski, 2005).
100
Uniaxial 𝜎𝜎�2
1 Tension
(𝑞𝑞� − 3𝛼𝛼𝑝𝑝̅ + 𝛽𝛽𝜎𝜎�2 ) = 𝜎𝜎𝑐𝑐0
1 − 𝛼𝛼
𝜎𝜎10
𝜎𝜎�1
Uniaxial
Compression
Biaxial
Tension
(𝜎𝜎𝑏𝑏0 , 𝜎𝜎𝑐𝑐0 ) 1
(𝑞𝑞� − 3𝛼𝛼𝑝𝑝̅ + 𝛽𝛽𝜎𝜎�1 ) = 𝜎𝜎𝑐𝑐0
1 − 𝛼𝛼
𝜎𝜎𝑐𝑐0
1
(𝑞𝑞� − 3𝛼𝛼𝑝𝑝̅) = 𝜎𝜎
Biaxial 1 − 𝛼𝛼 𝑐𝑐0
Compression
Figure 5.11. Yield surface in plane stress. The values of 𝜎𝜎𝑏𝑏0 and 𝜎𝜎𝑐𝑐0 can be extracted from this curve. 𝛼𝛼 is
a material constant, 𝜎𝜎𝑐𝑐0 is the uniaxial compressive strength, 𝜎𝜎𝑡𝑡0 is the uniaxial tensile
strength, 𝑝𝑝̅ is the effective hydrostatic pressure, and 𝑞𝑞� is Mises equivalent effective stress.
Table 5.1. Constitutive material parameters for concrete class CQD-G, to be used in Concrete Damage
Plasticity (CDP) model.
CDP Parameters for Concrete CQD-G Class Values
𝛾𝛾 (The shape of loading surface in the deviatoric plane) 0.666
𝜎𝜎𝑏𝑏0
(The ratio of the biaxial and uniaxial compressive stresses) 1.12
𝜎𝜎𝑐𝑐0
𝛼𝛼 (The material constant) 0.097
𝜓𝜓 (The dilation angle measured in the 𝑝𝑝 − 𝑞𝑞 plane) 39∘
𝜖𝜖 (The eccentricity parameter) 1.00
𝐸𝐸 (The Young's modulus) 19.70 GPa
𝜈𝜈 (The Poisson’s ratio) 0.19
𝜎𝜎𝑐𝑐0 (The uniaxial compressive strength) 54.77 MPa
𝜎𝜎𝑡𝑡0 (The uniaxial tensile strength) 3.62 MPa
101
Table 5.2. Concrete compression hardening.
Stress (MPa) Crushing Strain
15.0 0.0
20.197804 0.0000747307
30.000609 0.0000988479
40.303781 0.000154123
50.007692 0.000761538
40.236090 0.002557559
20.236090 0.005675431
5.257557 0.011733119
Table 5.3. Concrete compression damage.
Damage Crushing Strain
0.0 0.0
0.0 0.0000747307
0.0 0.0000988479
0.0 0.000154123
0.0 0.000761538
0.195402 0.002557559
0.596382 0.005675431
0.894865 0.011733119
Table 5.4. Concrete tension stiffening.
Stress (MPa) Crushing Strain
1.99893 0.0
2.842 0.00003333
1.86981 0.000160427
0.862723 0.000279763
0.226254 0.000684593
0.056576 0.00108673
Table 5.5. Concrete tension damage.
Damage Cracking Strain
0.0 0.0
0.0 0.00003333
0.406411 0.000160427
0.69638 0.69638
0.920389 0.000684593
0.980093 0.00108673
102
5.6. 3D FEA using CDP
Abaqus 6.10-1 has been used for the simulation of mechanical response of the
cylindrical concrete specimen (150 × 300 mm, see Chapter 3) for both failure and post-
failure behaviour using CDP constitutive model. The model has been created in two
parts to generate best possible mesh on parts and then assembled as a solid
homogeneous instance (see Figures 5.12 and 5.13). The material property and
constitutive parameters for the concrete studied here are defined based on Table 5.1.
Tables 5.2 – 5.5 have also been used for concrete compression hardening/damage, and
tension stiffness/damage parameters of the CDP.
Figure 5.12. Concrete specimen instance modeled in Abaqus.
Figure 5.13. Assembled parts as an instance (the whole concrete specimen) in Abaqus.
103
The full constraint contact behaviour between the parts through the whole instance is
considered using a tie constraint with standard surface-based contact, which provides a
way to bond surfaces together permanently. The contact property type is defined as
frictionless tangential behaviour. Loads have been applied in two steps; uniform static
distribution for gravity and dynamic considering full Newton solution technique on
whole specimen in ramp amplitude (see Figures 5.14 and 5.15).
(1, 1) (10, 1)
1
Amplitude
0
0 2 4 6 8 10
Time (Sec)
Figure 5.14. Loading system’s amplitudes applied on the model in Abaqus.
Figure 5.15. Loads configuration for the concrete specimen model in Abaqus.
The boundary conditions for both ends of the specimen are specified as
displacement/rotation (UR1, UR2, and UR3 free). U1=U2=0 and U3=-0.1 for upper end
with considered amplitude and U1=U2=U3=0 for lower end with instantaneous
amplitude.
104
Dependent hexagonal (Hex) element mesh on part has been generated with 34,062
elements of size of 2 mm. The bias for two parts with 2–4 mm size has been verified
with the minimum face corner angle of less than 10º and averagely 76.62º (see Figure
5.16). The Implicit-Dynamic FEA solution method was used. The results obtained from
the FEA are presented in Figures 5.17 – 5.27.
Figure 5.16. Mesh generated for the concrete specimen model in Abaqus.
Figure 5.17. Strain energy density.
105
Figure 5.18. Scalar stiffness degradation.
Figure 5.19. Mises stress component.
106
Figure 5.20. Damage.
Figure 5.21. Max principal strain component.
107
Figure 5.22. Logarithmic max principal strain component.
Figure 5.23. Equivalent plastic strain.
108
Figure 5.24. Damage dissipation energy density.
Figure 5.25. Plastic dissipation energy density.
109
Figure 5.26. Displacement.
Figure 5.27. Magnitude of reaction forces.
110
Different shape and size of meshes are considered in this research to generate for the
model. The results of FEA were different for different meshes. It proves the mesh
sensitivity for CDP constitutive model as stated in Section 5.2.4 and 5.2.5 under the
subjects of post-failure stress–strain relation and fracture energy cracking criterion; “In
cases of plain concrete, the specification of a post-failure stress–strain relation
introduces mesh sensitivity in the results”. Most suitable shape and size of mesh is
selected in this research as the optimized mesh shape and size that are shown in Figures
5.17 – 5.27. Mesh sensitivity analysis could be presented in this section but as it was not
included in the present research plan, therefore the study of mesh sensitivity for CDP
model is not provided here.
The shapes of cracking and fracture pattern obtained from these analyses are very close
to the shape of conjugate planes fractures observed during experimental works on
specimens under uniaxial compression (see Chapter 3).
Figure 5.28 shows the crack (or fracture) pattern occurred during the failure process in
the concrete specimen simulated using Abaqus FEA. Conjugate planes formed in this
figure represent the cone and the cone and shear fractures. According to ASTM
C39/C39M-04 (see Chapter 3, Section 3.4) these types of fracture should be happened
in concrete specimens under uniaxial compression.
As can be seen in Figure 5.28 fracture pattern obtained from the CDP model is
homological to the observations obtained in this research from laboratory experiments
on CQD-G class concrete specimens under uniaxial compression (see Chapter 3).
In presentation of fracture pattern in Figure 5.28 different graphical effects are used to
obtain best visibility of the conjugate planes formed in the failure process using the
constitutive parameters identified in this research for CDP model in Abaqus applicable
to high strength concrete.
The pattern obtained from the fracture study using CDP in comparison to the
experimental results obtained in laboratory (see Figure 5.28) proves the credibility of
the constitutive parameters identified and concluded in this research to be used in CDP
model in Abaqus for high strength concrete; CQD-G class.
111
Fracture pattern obtained from numerical simulation using Abaqus by CDP model.
The Cone facture obtained from uniaxial compression test. The Shear facture obtained from uniaxial compression test.
UCS ≈ 55 MPa UCS ≈ 70 MPa
Figure 5.28. Fracture pattern obtained from Abaqus FEA using constitutive parameters identified in this
research for CDP model for CQD-G concrete specimen in comparison to actual fracture
pattern obtained from test for concrete specimens categorized in CQD-G. Uniaxial
compression is applied on the specimen.
Finally based on the results obtained in this research, it can be concluded that the finite
element simulations and the identification of the CDP model’s constitutive parameters
conducted in this research for CQD-G class concrete using Abaqus is acceptable.
Therefore the parameters obtained in this research for CDP is trusted to be used in
design of high strength concrete (CQD-G class) structures and also for designing the
underground and surface structures in high strength brittle materials like rocks, which
show similar mechanical behaviour as high strength concrete.
112
5.7. Conclusions
As a brief conclusion following remarks describes well what studied in this sector of the
research:
• CDP constitutive behavioral model enables a proper definition for the failure
mechanisms in concrete under compression and tension.
• The results obtained from the numerical study proved that the CDP model can
be used to simulate the concrete responses to tension and compression; fracture
pattern study is an example of the compatibility of the result of FEA with
experiments.
• The proposed constitutive parameters that verified and determined in this

research for CDP can be used in the constitutive model applicable to numerical
simulations that can be used in design of high strength concrete (CQD-G) or any
brittle material.
113
CHAPTER 6:
CONCLUSIONS
6.1. Observations
Finally, the obtained outcomes of all sections of the conducted research including study
efforts, experiments, analytical works, and numerical simulations together with
empirical suggestions are presented in this chapter.
Following remarks in the field of mechanical response of concrete to compression and

tension are obtained from literatures:
• Mechanical behaviour of concrete is characterised by the relationship between

axial stress and axial strain obtained from a concrete specimen under uniaxial
compression or tension. The complete stress–strain curve of concrete under
uniaxial compression consists of two regions: the pre-peak and the post-peak.
• The pre-peak region is a mechanical response of concrete before failure, which

normally consists of three parts of plastic, elastic, and brittle behaviour, and can
be obtained by any testing machine that applies a constant rate of load to the
specimen (load control mode). The ascending branch of the stress–strain curve
of concrete and its nominal strength are not influenced significantly by the
specimen geometry and restraint condition. Failure occurs when a network of
major cracks is formed while peak stress is reached, micro-cracks are merged
and a critical flaw in the matrix is formed. Failure criteria describe failure
behaviour of concrete under triaxial loading for both compression and tension
quadrants.
• The post-peak region is a mixture of the material and structural (sample–loading

frame) response of concrete after failure. The complete stress–strain curve of
concrete is significantly influenced by the geometry of the specimen, stiffness of
the loading frame and the specimen, and the actual strain rate. A displacement-
control testing machine is used in this region, which applies small increments of
114
displacement to the specimen that should be able to unload rapidly. The strain
softening is an important phenomenon that happens when micro-cracks coalesce
to form a zone of damage after the strain associated with the peak stress is
reached. In this state, the concrete load-carrying capacity is diminished.
Softening in compression is a highly localized phenomenon, which begins right
at the peak stress.
• Stability of a compression test depends on the relation between the stiffness of

the loading machine and the slope of the complete force–displacement of
concrete specimen. If the stiffness of the loading system is greater than that of
the specimen (in the post-peak region), failure will occur in a quasi-stable
manner with the decreasing of the specimen strength in damage regime more or
less linearly with increasing displacement. If the stiffness of the loading system
is lower than that of the specimen, violent failure will be observed.
Totally 10 uniaxial compression tests, 10 splitting tensile strength/Brazilian tests, and

82 triaxial compression tests are conducted and 102 data pairs of (𝜎𝜎3 , 𝜎𝜎1 ) are obtained.
Amongst these, 67 data pairs in 10 data groups are qualified to be used in this research.
Major fracture pattern for concrete specimens in this research under uniaxial and triaxial
compression are recognized as the cone/shear and the cone respectively.
A new empirical classification entitled “CQD representing Concrete Quality

Designation” is proposed. CQD classifies concrete types to Very Poor (VP), Poor (P),
Fair (F), Good (G), Very Good (VG), and Excellent (E) based on their
uniaxial/unconfined compressive strength. It can be used in practice for designing
concrete structures.
A new strength criterion is developed for different types of concrete. New parameters
are defined for different classes of concrete. The suggested values of the parameters can
be used for practical purposes with an acceptable accuracy.
Constitutive models and numerical simulations are widely used in the field of strain
softening of concrete. Amongst the most applicable models developed for strain
softening; Series Coupling, Compressive Damage Zone, and Concrete Damage
Plasticity. CDP is selected in this research and according to the experimental data and
115
using Abaqus FEA its parameters are defined and proper results obtained for strain
softening of concrete. Findings in this sector of the research can be used in modelling
and simulations of failure and post-failure of concrete. The constitutive parameters
suggested in this research for CDP can be used in design of high strength concrete
(CQD-G) and brittle concrete-like structures.
6.2. Future Research
The CQD is concrete quality designation; an empirical classification for concrete

according to their uniaxial compressive strength. It is developed in this research based
on experimental and practical experience in the field of concrete mechanics and
technology. It is just beginning for CQD in its way to be an applicable classification for
concrete. It can be improved by further experiments and experience in this field.
Therefore as a future research, it is recommended that more practical study needs to be
done on this classification to make it a reliable and more applicable classification for
concrete. It is now a qualifying descriptive classification; however, it can be improved
by adding quantifying features to its classes.
The suggestion of suitable values of parameters for a strength criterion is an important

and applicable method that can save time and cost of practical and research projects. In
the present research the values for the parameters of the proposed strength criterion are
suggested for different classes of concrete based on the CQD using the obtained
laboratory test results. Performing an extensive range of uniaxial/triaxial compression
and splitting tensile strength tests on different classes of concrete is recommended as a
future study to determine more accurate values for the proposed strength criterion. In
this way the precision of the calculated values can be improved and therefore their
variation bound will be more limited. Consequently the strength estimation by the
proposed strength criterion using suggested values would be more accurate and more
precise.
Using the experimental data, the transition limit from brittle to ductile for concrete types
can be exactly defined. In this research the Mogi’s transition limit from brittle to ductile
for rocks is applied on concrete data groups, but for obtaining better results, an
116
investigation to explore the relation for concrete types is recommended, which requires
an extensive range of experimental data.
The parameters for the CDP constitutive model are obtained in this research for only
one class of concrete based on the CQD. Calculation of the CDP modelling parameters
for all classes of concrete as a future research is recommended, which its outcomes can
be a very helpful time and cost saving instruction to simulate failure and post-failure of
different classes of concrete to be used as a practical guide for numerical analysis in
application.
Numerical crack propagation study of failure and strain softening of concrete that can
be done using Abaqus FEA is another research subject that is recommended as a future
research.
117
REFERENCES
Abaqus Theory Manual, 2002, ‘ABAQUS 6.3’, Hibbitt Karlson & Sorensen, Inc.
Abaqus/CAE User’s Manual, 2011, ‘ABAQUS 6.10-1’, Dassault Systems Simulia
Corp., USA.
ASTM C39/C39M-04, 2004, ‘Standard Test Method for Compressive Strength of
Cylindrical Concrete Specimens’, ASTM International, West Conshohocken, PA,
DOI: 10.1520/C0039_C0039M-04, www.astm.org.
ASTM C143-78, 1978, ‘Standard Test Method for Slump of Portland Cement
Concrete’, ASTM International, West Conshohocken, PA, www.astm.org.
ASTM C192/C192M-05, 2005, ‘Standard Practice for Making and Curing Concrete
Test Specimens in the Laboratory’, ASTM International, West Conshohocken, PA,
ASTM C496/C496M-11, 2011, ‘Standard Test Method for Splitting Tensile Strength of
Cylindrical Concrete Specimens’, ASTM International, West Conshohocken, PA,
ASTM C617-10, 2010, ‘Standard Practice for Capping Cylindrical Concrete
Specimens’, ASTM International, West Conshohocken, PA, DOI: 10.1520/C0617-
10, www.astm.org.
ASTM C801-98, 1998, ‘Standard Test Method for Determining the Mechanical
Properties of Hardened Concrete Under Triaxial Loads’, ASTM International, West
Conshohocken, PA, DOI: 10.1520/C0801-98, www.astm.org.
ASTM D3967-08, 2008, ‘Standard Test Method for Splitting Tensile Strength of Intact
Rock Core Specimens’, ASTM International, West Conshohocken, PA, DOI:
10.1520/D3967-08, www.astm.org.
Abdulla, A. A. & Kiousis, P. D., 1997, ‘Behavior of cemented sands – I. Testing’,
International Journal for Numerical Methods in Geomechanics, Vol. 21, pp. 533 –
547.
Abdullah, R. A. B., 2006, ‘A study on stress-strain behaviour of Granite and Sandstone
using closed-circuit servo-controlled testing machine’, UTM, Master Degree
Thesis, pp. 22 – 24.
Balmer, G., 1952, ‘A general analytical solution for Mohr envelope’, American Society
of Testing Materials, 52: pp. 1260 – 1271.
118
Bazant, Z. P., 1989, 'Identification of strain-softening constitutive relation from uniaxial
tests by Series Coupling Model for localization', Cement and Concrete Research,
19: pp. 973 – 977.
Bazant, Z. P., 2002, ‘Size effect in tensile and compression fracture of concrete
structures: computational modeling and design’, Fracture Mechanics of Concrete
Structures, Proceedings FRAMCOS-3, AEDIFICATIO Publishers, Germany, pp.
1905 – 1922.
Bazant, Z. P., & Novak, D., 2000, ‘Energetic-statistical size effect in quasibrittle failure
at crack initiation’, ACI Materials Journal, No.97, M45.
Bieniawski, Z. T., 1974, ‘Estimating the strength of rock materials’, Journal of South
African Institute of Mining and Metallurgy, 74 (8): pp. 312 – 320.
Bineshian, H., 1999, ‘Failure mechanics of rock’, Engineering Department, Tarbiat
Modares University, Tehran, 124p.
Bineshian, H., 2000, ‘Verification of the applicability of rock failure criteria in relation
to the Limestone of IRAN and suggestion of a suitable failure criterion’, MSc
Thesis, Tarbiat Modares University, Tehran, Iran, 336p.
Bineshian, H., 2006, ‘Geomechanical properties of Towhid project’, 1st ed., Tehran:
Perlite Co.
Bineshian, H., 2008, ‘The results of triaxial compressive tests on injected materials of
Towhid project soil’, Tehran: Perlite Co.
Bineshian, H. & Bineshian, Z., 2011a, ‘Prediction of geomaterials’ compressive
strength on the basis of defining new values for Bieniawski failure criterion’s
parameters’, In: Proceedings of International Conference on Advances in
Geotechnical Engineering, Perth, pp. 227 – 234.
Bineshian, H. & Bineshian, Z., 2011b, ‘Modification to the Hoek-Brown strength
criterion for intact geomaterials based on new values for the parameter’, In:
Proceedings of the 13th International Conference of the International Association
for Computer Methods and Advances in Geomechanics, Melbourne: pp. 365 – 370.
Bisby, L. A. & Take, W. A., 2009, ‘Strain localization in FRP confined concrete: new
insights’, Proceeding of the Institution of Civil Engineers, Structure and Building.
162: pp. 301 – 309.
Bordia, S. K., 1971, ‘The effects of size and stress concentration on the dilatancy and
fracture of rock’, Int. J. Rock Mech. Min. Sci. Vol.8, Pergamon Press, pp. 629-640.
119
Brace, W. F., 1964, ‘Brittle fracture of rocks. In State of Stress in the Earth’s Crust’,
Proceedings of the International Conference, Santa Monica, Calif. Edited by W.R.
Judd. American Elsevier Publishing Company, New York, pp. 110 – 178.
Brook, N., 1979, ‘Model study of mine roadway deformation’, Trans. Inst. Min. Engrs.,
136, pp. 375 – 384.
Brown, E. T., 1981, ‘Rock characterization testing and monitoring’, ISRM suggested
methods, Pergamon Press.
Carpinteri, A., & Brighenti, R., 2010, ‘Fracture behaviour of plain and fiber-reinforced
concrete with different water content under mixed mode loading’, Materials &
Design, Vol. 31, Issue 4, pp. 2032 – 2042.
Carter, B.J., Duncan, E. J. S. & Lajtai, E. Z., 1991, ‘Fitting strength criteria to intact
rock’, Geotechn. Geol. Eng.
Choi, S., Thienel, K.C., Shah, S.P., 1996, ‘Strain softening of concrete in compression
under different end constraints’, Magazine of Concrete Research, 48 (175), pp. 103
– 115.
Chuan, T., & Xiaohe, X., 1990, ‘Stiffness machine characteristics and its application in
an earthquake model’, Engineering Fracture Mechanics, Vol. 35, No. 4/5, pp. 801 –
805.
Cook, N. G. W., 1965, ‘The failure of rock’, Int. J. Rock Mech. Min. Sci., Vol. 2, pp.
389 – 403.
Coulomb, C. A., 1776, ‘Essai sur une application des règles de maximis et minimis a
quelque problems de statique relatifs à l’architecture’, Mémoires de
Mathématique et de Physique, L’ Académie Royale des Sciences, Vol. 7, pp. 343
– 387.
Crowder, J. J., & Coulson, A. L., 2006, ‘The field-scale rock mechanics laboratory:
Estimation of post-peak parameters and behaviour of fractured rock masses’,
ARMA, USRM, Vol. 06.
del Viso, J. R., Carmona, J.R., & Ruiz, G., 2008, 'Shape and size effects on the
compressive strength of high-strength concrete', Cement and Concrete Research,
38: pp. 386 – 395.
Dhir, R. K. & Sangha, C. M., 1973, ‘Relationships between size, deformation and
strength for cylindrical specimens loaded in uniaxial compression’, Int. J. Rock
Mech. Min. Sci. & Geomech. Abstr.,Vol.10, Pergamon Press, pp. 699 – 711.
Dhir, R. K., Sangha, C. M. & Munday, J. G. L., 1972, ‘Influence of specimen size on
unconfined rock strength’, Colliery Guardian, 220, pp. 75 – 80.
120
Fairhurst, C., 1964, ‘On the validity of the Brazilian test for brittle materials’,
International Journal of Rock Mechanics and Mining Sciences and Geomechanics
Abstracts, 1 (4): pp. 535 – 546.
Franklin, J. A., 1971, ‘Triaxial strength of rock materials’, International Journal of Rock
Mechanics and Mining Sciences, 3 (2): pp. 86 – 98.
Gamino, A. L., Borges, J. U. A., & Bittencourt, T. N., 2004, ‘Size effect of concrete
under uniaxial and flexural compression’, Department of Structural and Foundation
Engineering, Brazil.
Geel, E. V., 1998, ‘Concrete behaviour in multiaxial compression – Experimental
research’, Eindhoven University of Technology, pp. 6 – 11 and 84 – 119.
Gobbi, M. E., & Ferrara, G. 1995, ‘Strain softening of concrete under compression’,
ENEL-CRIS, Milano, 250p.
Gonnerman, H. F., 1975, ‘Effect of size and shape of test specimens on compressive
strength of concrete’, In: Proc. ASTM, 2S, pp. 237 – 250.
Griffith, A. A., 1921, ‘The phenomena of rupture and flow in solids’, Philosophical
Transactions of the Royal Society of London, 221: pp. 163 – 198.
Hillerborg, A., 1988, ‘Fracture mechanics concepts applied to moment capacity and
rotational capacity of reinforced beams’, In: Proceedings, International Conference
on Fracture and Damage Mechanics of Concrete and Rock, Vienna, pp. 233 – 240.
Hillerborg, A., 1990, ‘Fracture mechanics concepts applied to moment capacity and
rotational capacity of reinforced concrete beams’, Engineering Fracture Mechanics,
Vol. 35, No. 1/2/3, pp. 233 – 240, Great Britain.
Hillerborg, A., 1979, ‘A discussion of the paper “studies of composites made by
impregnation of porous bodies. 2. Polymethyl methyl methacrylate in portland
cement systems by R.F. Feldman and J.J. Beaudoin’, Cement and Concrete
Research, Vol. 9, Issue 1, pp. 147-150.
Hillerborg, A., Modeer, M. & Petersson, P. E., 1976, ‘Analysis of Crack Formation and
Crack Growth in Concrete by Means of Fracture Mechanics and Finite
Elements’, Cement and Concrete Research, Vol. 6, pp. 773 – 782.
Hobbs, D. W., 1964, ‘The strength and the stress-strain characteristics of coal in triaxial
compression’, Journal of Geology, 72 (2): pp. 214 – 231.
Hobbs, D. W., 1970, ‘The strength and stress- strain characteristics of coal in
triaxial compression’, J. Geol., 72 (2), pp. 214 – 231.
Hobbs, D. W., 1971, ‘Strength of concrete under combined stress’, Cement and
Concrete Research, Vol. 1, pp. 41 – 56.
121
Hoek, E., 1968, ‘Rock mechanics in engineering practice’, London: Wiley.
Hoek, E. & Brown, E. T., 1980, ‘Empirical strength criterion for rock masses’, Journal
of Geotechnical Engineering Division, 106 (9): pp. 1013 – 1035.
Hoek, E. & Brown, E. T., 1988, ‘The Hoek-Brown failure criterion – a 1988 update’, In:
Proceedings of the 15th Canadian Rock Mechanics Symposium, Toronto: pp. 31 –
38.
Horino, F. G. & Ellikson, M. L., 1970, ‘A method for estimating the strength of rock
containing planes of weakness’, Washington, D.C: Bureau of Mines, US
Deptartment of Interior.
Hossaini, S. M. F., 1999, ‘Essential hints in selection of laboratorial data to evaluate
failure criteria’, Fanni Department Journal, 32 (1): pp. 89 – 94.
Hossaini, S. M. F., 1993, ‘Some aspects of the strength characteristics of intact and
jointed rocks’, PhD Thesis, Sydney: The University of New South Wales.
Hudson, J. A., Crouch, S. L., & Fairhurst, C., 1972, ‘Soft, stiff, and servo-controlled
testing machines: A review with reference to rock failure’, Engineering Geol., Vol.
6, pp. 155 – 189.
Hustrulid, W., & Robinson, F., 1972, ‘A simple stiff machine for testing rock in
compression’, Rock Mechanics, pp. 61 – 84.
Jaeger, J, C., Cook, N. G. W. & Zimmerman, R. W., 2007, ‘Fundamentals of rock
mechanics’, Victoria: Blackwell Publishing Ltd.
Jankowiak, T., & Lodygowski, T., 2005, ‘Identification of parameters of concrete
damage plasticity constitutive model’, Foundations of Civil and Environmental
Engineering, No.6, pp.54 – 69.
Jansen, D. C., & Shah, S. P., 1997, ‘Effect of length on compressive strain- softening of
concrete’, Journal of Engineering Mechanics, Vol. 123, No. 1, pp. 25 – 35.
Johnston. I. W., 1985, ‘Strength of intact geomechanical materials’, Journal of
Geotechnical Engineering Division, 111 (6): pp. 730 – 749.
Johnston, I. W. & Chiu, H. K., 1984, ‘Strength of weathered Melbourne mudstone’,
Journal of Geotechnical Engineering Division, 110 (G57): pp. 875 – 898.
Kachanov, L. M., 1958, ‘O vremeni razruenija v usloviach polzucesti’, Izv. Ak. Nauk
CCCP, Otd. Techn. Nauk, nr 8, pp. 26 – 31.
Kesler, C. E., 1959, ‘Effect of length to diameter ratio on compressive strength’, An
ASTM cooperative investigation. In: Proc. ASTM, 59, pp. 1216 – 1228.
122
Kiendle, O. G., & Maldari, J. A., 1938, ‘Comparison of physical properties of concretes
made of three varieties of course aggregates’, Unpublished Thesis, University of
Wisconsin, Madison Wisc.
Kupfer, H., Hilsdorf, H. K., & Rusch H., 1979, ‘Behavior of concrete under biaxial
stresses’, ACI Journal, 65, 8, pp. 656 – 666.
Kwasniewski, M., 1987, ‘A new linear criterion of brittle failure for rocks. Proc. 6th
Cong. Rock Mech., Montreal, pp. 1031 – 1038.
Labuz, J. F., & Biolzi, L., 1991, 'Class I vs. Class II stability: A demonstration of size
effect', Int. J. Rock Mech. Min. Sci. & Geomech. Abstr., 28 (2/3): pp. 199 – 205.
Lee, J., & Fenves, G. L., 1998, ‘Plastic-Damage Model for Cyclic Loading of Concrete
Structures’, Journal of Engineering Mechanics, Vol. 124, No.8, pp. 892 – 900.
Lubliner, J., Oliver, J., Oller, S., & Oñate, E., 1989, ‘A Plastic-Damage Model for
Concrete’, International Journal of Solids and Structures, Vol. 25, pp. 299 – 329.
Mahab Ghodss Consulting Engineering Co., 1999, ‘Rock mechanics report for Karun
III Dam – 2’, Izeh: Dam Department.
Mahab Ghodss Consulting Engineering Co., 2003, ‘Rock mechanics report for
Seimareh Dam – 1’, Seimareh: Dam Department.
Markeset, G., 1995, ‘A compressive softening model for concrete’, In: Proceedings of
Fracture Mechanics of Concrete Structures, FRAMCOS-2, pp. 435 – 443.
Markeset, G., & Hillerborg, A., 1995, 'Softening of concrete in compression –
localization and size effects', Cement and Concrete Research, 25 (4): pp. 702 – 708.
Mogi, K., 1966a, ‘Some precise measurements of fracture strength of rocks under
uniform compressive stress’, Rock Mechanics and Engineering Geology, 4 (1): pp.
43 – 55.
Mogi, K., 1966b, ‘Pressure dependence of rock strength and transition from brittle
fracture to ductile flow’, Bull. Earthquake Res. Inst., 44: pp. 215 – 232.
Mohr, O., 1900, ‘Welche Umstände bedingen die Elastizitätsgrenze und den Bruch
eines Materials?’, In: Zeitschrift des Vereins Deutscher Ingenieure. Nr. 24, S. 1524
– 1577.
Murrell, S. A. F., 1963, ‘A criterion for brittle fracture of rocks and concrete under
triaxial stress and the effect of pore pressure on the criterion’, Oxford: Pergamon
Press.
Murrell, S. A. F., 1965, ‘The effect of triaxial stress systems on the strength of rocks at
atmospheric temperatures’, Geophysical Journal, 10 (3): pp. 231 – 281.
123
Nemecek, J. & Bittnar, Z., 2004, ‘Experimental investigation and numerical simulation
of post-peak behaviour and size effect of reinforced concrete columns’, Material
and Structures, Springer Netherlands, Vol.37, No. 3, pp. 161 – 169.
Netter, J., Wasserman, W. & Whitmore, G. A., 1988, ‘Applied statistics’, Boston: Allyn
and Bacon Inc.
Nishihara, M., 1957, ‘Stress-strain-time relation of rocks’, Doshisha Engineering
Review, 8: pp. 32-55 and 85 – 115.
Ohnaka, M., 1973, ‘The quantitative effect of hydrostatic confining pressure on the
compressive strength of crystalline rocks’, Journal of the Physics of the Earth, 21
(2): pp. 125 – 140.
Ouyang, Z. & Elsworth, D., 1991, ‘A phenomenological failure criterion for brittle
rock’, Rock Mechanics and Rock Engineering, 24 (3): pp. 133 – 153.
Panek, L. A., 1979, ‘Estimating mine pillar strength from compressive tests’. Trans.
Soc. Min. Engrs., AIME, 268, 1749 – 1761.
Pellegrino, A., Sulem, J., & Barla, G., 1997, ‘The Effects of Slenderness and
Lubrication on the Uniaxial Behavior of a Soft Limestone’, Int. J. Rock Mech. Min.
Sci. Vol, 34, No. 2: pp. 333 – 340.
Price, N. J., 1960, ‘The initiation and development of asymmetrical buckle folds in non-
metamorphosed competent sediments’, Tectonophysics, Vol. 4, Issue 2, pp. 173 –
201.
Ramamurthy, T., Rao, G. V. & Rao, K. S., 1985, ‘A strength criterion for rocks’, In:
Proceedings of Indian geotechnical conference, Roorkee: pp. 59 – 64.
Ramamurthy, T., Rao, G. V. & Rao, K. S., 1988, ‘A non-linear-nonlinear strength
criterion for rocks’, In: Proceedings of the 5th Australia-New Zealand Conference
on Geomechanics, Sydney: Institution of Engineering, Australia: pp. 247 – 252.
Rabotnov, Y. N., 1969, ‘Creep problems in structural members’, North-Holland,
Amsterdam.
Rokugo, K. & Koyanagi, W., 1992, ‘Role of compressive fracture energy of concrete on
the failure behavior of reinforced concrete beams’, Application of Fracture
Mechanics to Reinforced Concrete (ed. A. Carpenteri), Elsevier Applied Science,
pp. 437 – 464.
Rummel, F., & Fairhurst, C., 1970, ‘Determination of the post-failure behavior of brittle
rock using a servo-controlled testing machine’, Rock Mechanics, Vol. 2, pp. 189 –
204.
124
Salamon, M. D. G., 1970, ‘Stability, instability and design of pillar workings’, Int. J.
Rock Mech. Min. Sci., Vol. 7, pp. 613 – 631.
Sangha, C. M., 1972, ‘Strength, deformation and fracture properties of rock and
concrete’, Ph.D. Thesis, University of Dundee.
Sangha, C. M. & Dhir, R. K., 1972, ‘Strength and complete stress-strain relationships
for concrete tested in uniaxial compression under different test conditions’, Mater.
Structures (RILEM), 5, pp. 361 – 370.
Schubert, W., & Blumel, M., 2006, ‘Innovative lab test procedures for determination of
post failure behaviour of rocks’, Graz University of Technology, Institute for Rock
Mechanics and Tunneling, pp. 1 – 16.
Schwartz, A. E., 1964, ‘Failure of rock in the triaxial shear test’, In: Proceedings of the
6th US Symposium on Rock Mechanics, Rolla: University of Missouri Rolla: pp. 64
– 109.
Shang, H. S., & Song, Y. P., 2006, ‘Experimental study of strength and deformation of
plain concrete under biaxial compression after freezing and thawing cycles’,
Cement and Concrete Research, Vol. 36, Issue 10, pp. 1857 – 1864.
Sheorey, P. R., 1997, ‘Empirical rock failure criteria’, Rotterdam: A. A. Balkema.
Sheorey, P. R., Biswas, A. K. & Choubey, V. D., 1989, ‘An empirical failure criterion
for rocks and jointed rock masses’, Engineering Geology, 26 (2): pp. 141 – 159.
Symons, I. F., 1970, ‘The effect of size and shape of specimen upon the unconfined
compressive strength of cement stabilized materials’, Magazine o f Concrete
Research, 22, pp. 45 – 50.
Tang, C. A., Tham, L. G., Lee, P. K. K., Tsui, Y., & Liu, H., 2000, 'Numerical studies
of the influence of microstructure on rock failure in uniaxial compression – Part II:
constraint, slenderness and size effect', International Journal of Rock Mechanics &
Mining Sciences 37: pp. 571 – 583.
Timco, G. W., & Frederking, R., 1984, ‘A procedure to account for machine stiffness in
uni-axial compression tests’, LAHR Ice Symposium Hamburg, pp. 39 – 47.
Torrenti, J. M., 1986, ‘Some remarks upon concrete softening’, Material and Structures,
Springer Netherlands, Vol. 19, No. 5, pp. 391 – 394.
van Mier. J. G. M., 1984, ‘Strain softening of concrete under multiaxial loading
conditions’. Thése de Docteur de Ingénieuer, Technische Hogeschool Eindhoven.
van Mier, J. G. M., 1986, ‘Multiaxial strain-softening of concrete Part I: Fracture’,
Material and Structures, Springer Netherlands, Vol. 19, No. 3, pp. 179 – 190.
125
van Mier, J. G. M., 1998, ‘Special issue on compressive failure of concrete and rock’,
Delft University of Technology, Faculty of Civil Engineering and Geo-Sciences,
pp. 123 – 125.
van Mier, J. G. M., Shah, S. P., & Arnaud, M., 1997, ‘Strain-softening of Concrete in
uniaxial compression’, Material and structures, Springer Netherlands, Vol. 30, No.
4, pp. 195 – 209.
van Mier, J. G. M., & Ulfkjaer, J. P., 2000, 'Round-Robin analysis of over-reinforced
concrete beams – Comparison of results', Materials and Structures, 33: pp. 381 –
390.
van Mier, J. G. M., & Vonk, R .A., 1991, 'Fracture of concrete under multiaxial stress –
recent developments', Materials and Structures, 24: pp. 61 – 65.
van Vliet, M. R. A., & van Mier, J. G. M., 1996, 'Experimental investigation of concrete
fracture under uniaxial compression', Mechanics of Cohesive-Frictional Materials,
1: pp. 115 – 127.
Vonk, R. A., 1992, ‘Softening of concrete loaded in compression’, PhD Thesis,
Eindhoven University of Technology, The Netherlands.
Vutukuri, V. S. & Hossaini, S. M. F., 1992, ‘Assessment of applicability of strength
criteria for rock and rock mass to coal pillars’, In: Proceedings of the 11th
International Conference on Ground Control in Mining, New South Wales,
Australia: University of Wollongong: pp. 1 – 8.
Watanabe, K., Niwa, J., Yokota, H., & Iwanami, M., 2004, ‘Experimental study on
stress-strain curve of concrete considering localized failure in compression’, Journal
of Advanced Concrete Technology, Vol. 2, No. 3, pp. 395 – 407.
Yamaguchi, E. & Chen, W. F., 1990, ‘Post-failure behavior of concrete materials in
compression’, Engineering Fracture Mechanics, Vol. 37, No. 5, pp. 1011 – 1023.
Yoshida, N., Morgenstern, N. R. & Chau, D. H., 1990, ‘A failure criterion for stiff soils
and rocks exhibiting softening’, Canadian Geotechnical Journal, 27 (2): pp. 195 –
202.
Yukutake, H., 1989, ‘Fracturing process of granite inferred from measurements of
spatial and temporal variations in velocity during triaxial deformations’, J.
Geophys. Res., 94, 15, pp. 639 – 651.
126

Bineshian Hossein 2014

Caricato da

Informazioni sul documento

Descrizione originale:

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Bineshian Hossein 2014

Caricato da

Copyright:

Formati disponibili

Failure and post-failure aspects of mechanical response of

concrete structures to compression and tension

This thesis is presented for the degree

School of Civil, Environmental and Mining Engineering

In total 102 uniaxial/triaxial compression/tension tests are conducted on 102 concrete

𝛽𝛽 A factor in Bineshian criterion that defines RUCS in terms of AUCS

Table 2.1. Failure criteria.

I would like to express my gratitude to my supervisors Prof. Arcady V. Dyskin from

Study of both failure and post-failure responses of concrete by experiments can be

Mechanical response of concrete to compression and tension in triaxial loading

• Developing a new comprehensive strength criterion for different types of

• Developing a new empirical concrete quality classification in terms of uniaxial

• A comprehensive literature review of mechanical properties and behaviour of

• Concrete batching, composition and mix design, and preparation in accordance

• Concrete specimens manufacturing and ends preparation in accordance with

• Uniaxial compression tests on prepared concrete specimens in accordance with

• Splitting tensile strength tests on prepared concrete specimens in accordance

• Brazilian tests on prepared concrete and rock specimens in accordance with

• Triaxial compression tests on prepared concrete specimens at desired confining

• Mathematical study, practice and modelling of the proposed strength criterion’s

• Calculations for suggested constant values for proposed strength criterion’s

• Preparing the suggested empirical classification for different concrete types in

• Pure study sector including literature review,

• Analytical and statistical investigations,

• Experimental sector including laboratory studies and field experience, and

• Numerical modelling and simulations using FEA software.

The thesis is organized as follows:

• Chapter 1 (the current chapter) is an introduction to the whole thesis containing

• Chapter 2 contains literature review conducted in this research including two

• Main parts of experimental and empirical investigation conducted in this

• New strength criterion for concrete is introduced in Chapter 4. Its performance

• Conclusions from all sectors of this research including experimental, empirical,

Pre-peak region of complete stress–strain curve of concrete is related to the mechanical

Intact 35% 50% 75% 100%

%75 Ultimate Stress

%50 Ultimate Stress

%35 Ultimate Stress

𝝈𝝈𝒄𝒄+𝝀𝝀𝝈𝝈𝟑𝟑 Non-linear Envelope

Mohr-Coulomb R^2 = 0.893

Mohr-Coulomb (1900) 𝜎𝜎1 = 𝐶𝐶 + 𝐵𝐵𝜎𝜎3    

Griffith (1921) (𝜎𝜎1 − 𝜎𝜎3 )2 = 8𝜎𝜎𝑡𝑡 (𝜎𝜎1 + 𝜎𝜎3 )    

Fairhurst (1964) (𝜎𝜎1 − 𝜎𝜎3 )2 = 𝑎𝑎 + 𝑏𝑏(𝜎𝜎1 + 𝜎𝜎3 )    

Hobbs (1964) 𝜎𝜎1 = 𝜎𝜎𝑐𝑐 + 𝜎𝜎3 + 𝐵𝐵𝜎𝜎3 𝛼𝛼    

Murrell (1965) 𝜎𝜎1 = 𝜎𝜎𝑐𝑐 + 𝐵𝐵𝜎𝜎3 𝛼𝛼    

Mogi (1966) 𝜎𝜎1 − 𝜎𝜎3 = 𝜎𝜎𝑐𝑐 + 𝐵𝐵𝜎𝜎3 𝛼𝛼    

Hobbs (1970) 𝜏𝜏 = 𝐶𝐶 + 𝑘𝑘𝜎𝜎𝑛𝑛 𝛼𝛼    

Hobbs (1971) 𝜎𝜎1 = 𝜎𝜎𝑐𝑐 + 𝑐𝑐𝜎𝜎3 𝑑𝑑    

Franklin (1971) 𝜎𝜎1 − 𝜎𝜎3 = 𝜎𝜎𝑐𝑐 1−𝐵𝐵 (𝜎𝜎1 − 𝜎𝜎3 )2    

𝜎𝜎1 − 𝜎𝜎3 𝜎𝜎3

Brook (1979) 𝜏𝜏⁄𝜎𝜎𝑐𝑐 = 𝐵𝐵(𝜎𝜎𝑛𝑛 ⁄𝜎𝜎𝑐𝑐 )𝛼𝛼    

Kwasniewski (1987) 𝜎𝜎1 = 𝜎𝜎𝑐𝑐 + 𝑑𝑑𝜎𝜎3    

Post-peak region of complete 𝜎𝜎 − 𝜀𝜀 curve represents the softening behaviour (Crowder

• Stiffness of the testing machine,

Pre-peak portion of 𝜎𝜎 − 𝜀𝜀 curves is independent of specimen slenderness ratio when

Softening branch is a structural property, depending on both the specimen configuration

In pre-peak regime, 𝜎𝜎 − 𝜀𝜀 diagrams should be used, whereas in post-peak regime where

Stress / Peak Stress

2.3.3. Strain Localization

Because of localization, the post-peak portion of stress–strain curve is also dependent

L = 50 mm L = 50 mm L = 100 mm L = 100 mm L = 200 mm L = 200 mm

2.3.4. Damage Plasticity

The CDP consists of a combination of non-associated multi-hardening plasticity and