CHAPTER ONE

INTRODUCTION
1.1. GENERAL

Concrete is the most versatile material used in the field of civil engineering. Concrete is used
in large quantities for mankind to build infrastructure. Usually, normal strength concrete is
employed for common purposes and mostly high strength concrete is used for tall and
important structures. Since concrete is a composite material, cracks and pores exist inside
concrete. So, it becomes necessary to investigate whether these cracks are stable or not.
Fracture mechanics is important in studying concrete behaviour under static loading. Cracks
are common but become important problem in structures. The severity of problem varies with
the type and importance of the structure.

Concrete has excellent shielding capability, fire rating, and long service life under normal
and accidental conditions. Further, ease in construction with relatively lower cost results in
its excessive usage as a building material in civil engineering structures. In-spite of such
salient features, the concrete structures generally consist of numerous micro-cracks that
might result in fracture of the concrete structures under service loads, accidental load and/or
exposure to regular environmental conditions. Thus, a micro-crack in concrete may become a
potential source of crack propagation leading to a growth of major crack. In order to prevent
the growth of major cracks, it is necessary to predict the failure mechanisms of structures, so
that the safety of concrete structures throughout the service life can be assured. The failure
mechanism can be studied by quantifying the energy consumed in crack propagation and
formation of new crack surfaces. In a concrete structure, the crack growth requires a certain
amount of energy that can only be studied through an energy-based propagation criterion,
which provides a fundamental basis for understanding the phenomenon of concrete fracture
mechanism.

Fracture is defined as the separation of a component into, at least, two parts. Fractures of
a material occur when sufficient stress and work are applied on the atomic level to break the
bonds that hold atoms together. The fracture behaviour in concrete is more complicated due
to its heterogeneous nature and the presence of large size fracture process zone (FPZ) at the
crack tip. Failure occurs for many reasons, including uncertainties in the loading or

1
environment, defects in materials, inadequacies in design and deficiencies in construction
and maintenance. Failure of a structure usually occurs due to catastrophic growth of cracks
resulting in localization of stresses and there by affects the serviceability of the structure.

1.2. FRACTALS AND FRACTAL DIMENSION

Fractals were invented by mathematician Benoit Mandelbrot in 1975. The term refers to any
boundary or surface which remains self-similar as the scale of the examination is magnified.
While most man-made objects have linear or smoothly curvilinear shapes, natural ones (such
as rivers, mountains, clouds, and fractures) are often rough, fragmented or discontinuous.
The word ‘fractal’ describes objects that are too irregular to fit into a traditional geometrical
setting. Fractal geometry provides a general framework for the study of such irregular sets.
The main tool of fractal geometry is the fractal dimension of the object in question. Fractal
analysis is a simple and powerful tool for quantifying the roughness and irregularities of
fractured surfaces.

As shown in Figure 1, the Von Koch flake is a typical fractal object which cannot be
described with Euclidean geometry. Firstly, when the number of iterations tends to be
indefinite, the total length increases indefinitely, whereas the surface is definite. Secondly,
the Von Koch flake is self-similar: each part of this object is the same as the whole for all
ranges of scale investigated. In this case, we can determine simply the fractal dimension, D,
which characterizes in fact the irregularity of the object. Fractal dimension, D is a decimal
number which exceeds the Euclidean dimension E. For the Von Koch flake shown in Figure
1, in which the initiator generates four patterns (N=4) with a ratio r for the new fractal
generator (r =1/3) the fractal dimension is found to be:

ln 𝑁 ln 4
𝐷= 1 = = 1.2619 (1.1)
ln(𝑟 ) ln 3

where N is the number of subparts at each step, and S is the scaling factor and is defined as
l/r where r is the length of each subpart relative to an initiator of unit length. An alternative
form for this equation is:

2
 𝑁 = 𝑆𝐷 (1.2)

Figure 1: Straight Line Initiator, Fractal Generator and Triadic Koch Curve

Ideal fractals such as the Koch curve are exactly self-similar while natural fractals, such
as a coastline are said to be statistically self similar; that is, upon magnification segments of the
coastline look alike, but never exactly similar to, segments at different scales. Therefore,
natural fractals exhibit fractal behaviour only within a specified range of resolutions (scales).
While self-similar shapes repeat exactly (ideal fractals) or statistically (natural fractals) under
magnification and scale equally in all directions, many objects repeat statistically only when
scaled by different amounts in different directions. This non-uniform scaling is known as self-
affinity.

A great deal of literature has been published on the use of fractal geometry to describe
and determine the scaling of surface roughness of various materials. However, very few
research studies on the application of fractal geometry to cementitious materials have been
published. Fractal geometry provides useful tools for the characterization of irregular and rough
surfaces. Irregularities produced during the fracturing process in cementitious materials are
manifested on the fracture surface; therefore, the applicability of fractal geometry in
characterizing these rough and tortuous surfaces may improve our understanding of the fracture
properties of such materials.

3
1.3. APPLICATION OF FRACTALS TO CONCRETE
The quantitative description of rough fracture surfaces of concrete has been an important
challenge for many years. Looking at the fracture surface of a concrete specimen, one
realizes that the self-affine geometry of crack faces results from the stochastic nature of the
crack growth. This is due to the heterogeneous nature of concrete that makes the crack
tortuous leading its way through weak bonds, voids, mortar and getting arrested on
encountering a hard aggregate forming crack face bridges. These mechanisms contribute to
the tendency of the crack to follow a tortuous path. The self-similarity contained in the
tortuous fracture surface of concrete makes it an ideal candidate to be considered as a fractal.
Further, the softening response itself has been treated as a singular fractal function by earlier
investigators. The very process of cracking and micro-cracking, could be considered very
close to the stick and slip process and therefore as a fractal. Therefore modelling a crack as a
fractal and characterizing it by a fractal dimension have become the focus of research in
recent years. Due to randomly distributed discontinuous flaws and high heterogeneity of the
internal structure of concrete, mechanical properties also randomly vary. Under the effect of
the same external force, the stress intensity factors to which different points in the concrete
are subjected are different. Hence the micro-cracks induced by the external force are
distributed discontinuously and randomly.
Most of current work on safety of civil engineering structures deals with evaluation of
structural reliability, and there is not much study on the assessment and detection of structural
damage. Nowadays, as more civil engineering structures reach to or close to their lifetime and
a great number of large-scaled civil engineering structures are being built worldwide, damage
assessment and detection technology of civil engineering structure has been of great
importance and needs to be urgently developed. Extraction of damage characteristic factors is
the core of the technology, and it effectively reflects degree of structural damage. For a large
amount of concrete structures in civil engineering, the most direct manifests of damage
occurrence are the appearance and broadening of cracks. Development or evolution state of
cracks affects the mechanical properties of structures directly. To some extent, it determines
the serviceability, reliability, safety and durability of the structures, and even becomes a major
measurement of structure lifespan. Therefore, it is of great practical value to assess the damage
magnitude according to crack development state of the damaged concrete structures. Because
high complexity involves in analyzing crack development of concrete structures, there is a
4
need to develop a rigorous analysis approach for crack development. Fractal geometry is an
effective tool to describe irregular, complex and disordered natural phenomenon and has been
applied in many engineering fields including concrete materials. It has become a prevalent
trend to describe the micro-pore structure and fracture surface of concrete materials using
fractal.

1.4. METHODS OF DETERMINING FRACTAL DIMENSION

Although the fractal dimension D of synthetic curves can be readily determined analytically,
the dimension of arbitrary curves requires the use of numerical methods. Three of the most
commonly used methods for measuring the fractal dimension of surface roughness are
reviewed next.

1.4.1. RULER METHOD

In the ruler method (Aviles and Scholz 1987; Mandelbrot 1983), a divider (that is ruler) of
given length S is used to determine the total length of the curve. Note that no fractional
lengths can be taken. Thus, for a given ruler length, the curve length L is given by SN, where
N is an integer. This operation is repeated for a ruler of smaller and smaller length S, and the
fractal dimension is given by:

N = sD

SN = S*(1/S)

But we know that, SN=L

Therefore, L = S1-D

InL = (1 - D) InS (1.3)

1.4.2. BOX COUNTING METHOD

In this method, a sequence of grids, each having a different cell size, is placed over maps of
the fracture surface traces and the number of cells intersected by fracture traces is counted.
The number of grid cells intersected by the profile N is then plotted on a log-log scale with
respect to the inverse of the grid cell size (1/r), where r is the grid cell size. If a linear
relationship is found, then the object is said to be fractal within the specified range of cell
sizes and the fractal dimension is the slope of the line.

5
1.4.3. SLIT ISLAND METHOD (SIM)
This method was proposed by Mandelbrot et al. In this method, the fractured surfaces are
coated with a layer of nickel or chromium to preserve the fractured surface profile. After
coating with nickel or chromium, the fracture surface is polished parallel to a reference plane.
As material is removed, ‘‘islands’’ of the base materials appear in a ‘‘sea’’ of the coating
material. Usually, the perimeter and area of the islands (generally 100 islands or more) are
measured by using an electronic planimeter, by digitizing the island perimeter on a computer,
or by quantitative image-analysis algorithms. The length of the coastlines of the slit islands is
generally measured with a fixed scale for each successive section. The fractal dimension D is
calculated by plotting the variation of the logarithm of the area (in ordinate) versus the
logarithm of the perimeter, for all the islands of each layer examined. D is obtained from the
slope 2/D’ of the straight line obtained by linear regression analysis. The fractal dimension,
D, was then obtained from the relation:

D = D’ + 1 (1.4)
The relationship between the area A and perimeter P of the islands formed is given in the
equation below:
A = P (2 /D’) (1.5)

1.4.3.1. MODIFIED SLIT ISLAND METHOD

In order to avoid indirect or destructive experimental procedures that are laborious, expensive
and generally unsuitable for concrete, the slit-island method was modified. In this method the
fracture surface is placed in non-staining dyed water, therefore when viewing from the top,
higher regions on the fracture surface appeared lighter than deeper regions. Proper care should
be taken while adjusting the concentration of the dye in order to make the highest and lowest
points of the fractured specimens clearly visible. A photo is taken from the top and analyzed by
an image analyzer. At each threshold level (intensity of light) increment, new islands appear
and their perimeter and area are measured. The results of the area–perimeter measurements are
plotted on a log – log scale and best fitted with linear regression. Then Equations (1.4) and
(1.5) are used to compute the fractal dimension. Figure 2 shows the fractured surface of a
wedge splitting specimen in the dyed water tank.

6
 Figure 2: Typical concrete specimen in the dyed water tank (Source: Issa et al., 2003)

1.4.4. 2D FAST FOURIER TRANSFORM METHOD

2D fast Fourier transform function calculates the surface fractal dimension of an image. Within
the Fourier domain, the spectral density has a power law dependence on the wave number. The
fractal dimension of the surface is related to the slope of the log–log plot of the 2D spectral
density versus wave number by:
8−𝑠𝑙𝑜𝑝𝑒
𝐷= (1.6)
2

In this method wave numbers and their corresponding spectral energy density of the image are
computed. This data is then plotted in log–log format and its slope is determined using a linear
least squares regression technique. From this slope the 2D statistical fractal dimension is
determined.

7
 CHAPTER TWO

LITERATURE REVIEW
2.1. GENERAL
Concrete structures were successfully designed and built long before the emergence of fracture
mechanics. Recent research has clearly demonstrated that both safety and economy of design
will benefit from the use of fracture mechanics. Fractal geometry is being considered as an
excellent tool to study the fracture surfaces of concrete.

2.2. LITERATURES REVIEWED

Saouma et al. (1990) used fractal geometry to characterize the roughness of cracked concrete
surfaces through a specially built profilometer. The fractal dimension is subsequently correlated
to the fracture toughness and direction of crack propagation. The results indicate that the
fracture surface is indeed fractal over two orders of magnitudes with a dimension of
approximately 1.20. Cubes of 36 in. were fractured by inflating a cylindrical packer in a four
inch diameter axial borehole, while the specimen was subjected to uniaxial compression
perpendicular to the borehole axis. The profilometer scans a sample by means of a motorized,
computer-controlled placement of a Linear Variable Differential Transformer (LVDT). LVDT
can travel along any desired path in the x-y plane. Measurement of the sample roughness in the
vertical z-direction is made by the LVDT. Following the experimental test, the fractal
dimension of the fracture surface roughness is determined using a computerized version of the
box method. They concluded that fracture surfaces in concrete are fractal over the range of
scales that have measured. The fractal dimension for the specimens tested is approximately
1.20 at the microscope scale and close to 1.10 at the macroscopic scale. The roughness of the
fracture does not vary significantly with fracture propagation direction.
Saouma et al. (1991) obtained fracture properties of concrete used in dams from laboratory
experiments of large wedge-splitting specimens with heights of 31 cm, 91 cm and 152 cm.
Various concrete mixes of maximum aggregate size of 19 mm, 38 mm and 76 mm were
investigated. The fracture toughness, KIc, and specific fracture energy, GF, are found to be
aggregate and specimen-size independent, only if certain minimum specimen dimensions are
exceeded. However, they appear to be dependent on the aggregate type and shape. For the two
types of aggregates used in this experimental program, more energy was required to fracture the
sub angular aggregates than to debond the rounded ones.
8
Issa and Hammad (1993) characterized the fracture surface in mortar in terms of fractal
geometry. They proposed a new technique for the evaluation of fractal dimension (D) of
fracture surface and a quantitative correlation between D and fracture toughness of
cementitious materials. The experimental program has been performed on compact tension
(CT) specimens (600 × 525 x 125mm) with three different aggregate sizes (d max = 4.7 mm, 18.8
mm and 37.5 mm). The fractal geometry concept is utilized in the evaluation of fracture surface
roughness. The specimens were sectioned perpendicular to crack plane and resulting profiles
were digitized by using an Image Analyzer. Fractal dimension D for each profile is evaluated
by the box counting method. Their results indicate that the concept of fractal geometry provides
useful tools for fracture surface characterization. Fractal dimension of the mortar fracture
appears to be ~2.1. The results also suggest that the fracture toughness can be correlated with
the fractal dimension of fracture surface.
Carpinteri et al. (1994) carried out the direct tensile testing of concrete in a stable manner and
the deformation eccentricity was adjusted continuously during the loading process. In this way,
the secondary flexural stresses were minimized so that it was possible to obtain nominal tensile
strength and fracture energy of the material. Both these properties appear to vary monotonically
with the diameter of the cylindrical specimens, i.e., the tensile strength decreases, whereas the
fracture energy increases. For disordered materials it is impossible to measure constant material
properties, unless we depart from integer dimensions of the material ligament at peak stress and
of the fracture surface at final rupture, in the framework of fractal geometry. In this way we can
define new tensile properties with physical dimensions depending on the fractal dimension of
the damaged microstructure, which turn out to be scale-invariant material constants.

Saouma and Barton (1994) had reported a correlation between fractal dimensions and fracture
properties in cementitious material. Wedge splitting tests of specimens were first performed to
determine fracture toughness K1c and fracture energy Gf. Subsequently, one of the split parts
was mapped using a profilometer to provide detailed one-dimensional profiles. Finally, the
fractal dimension of the profiles was determined by a specially developed computer program.
They concluded that the fracture surfaces are fractal over the measured range of scales, and the
fractal dimension is independent of crack trajectory. They reported a weak correlation between
the fracture properties and the fractal dimensions. The implications of the fractal nature of the
cracked surfaces on the fracture energy G f were discussed, as there was relationship between
fractal analysis and the size effect law.
9
Issa and Hammad (1994) carried out study on fractal characterization of concrete and mortar
by fracture trajectories. Compact tension fractured concrete specimens with three different
maximum aggregate sizes and a projected fracture area were analysed. Two methods, i.e., step
divider and variogram, were utilized to evaluate the fractal dimension of the fracture
trajectories. These trajectories were sectioned from a replica of the fracture surface, profiled
from a digitized image of the fracture surface, or generated via fractional Brownian motion
simulation. The fractal dimensions of concrete fracture trajectories calculated using these
methods were in the range of 1.07 to 1.25. It was found that the variogram method yielded
higher values than the step divider method. Synthetic profiles generated via fractional integrals
of Brownian motion showed a good visual similarity to the actual fracture profiles.

Charkaluk et al. (1998) used several experimental methods to determine the fractal dimension
according to the roughness of fracture surface. Three main methods are generally applied to
analyse the fractal dimension of the fractured surface of a material, i.e, Vertical section method,
Profilometer analysis and Slit Island Method (SIM). Many problems rose while correlating
fractal dimension of the fractured surfaces with the mechanical properties. Firstly, discrepancy
between the results was seen due to the methods used to calculate the fractal dimension.
Secondly, due to complexity of the mechanisms and the lack of precision of the methods, i.e., a
number of profiles had to be taken, resulting into variation of fractal dimension values. Thirdly,
experimental errors also lead to this variation. So it was concluded that further research was
necessary to improve the theoretical method and the derived algorithms before drawing
correlations between the mechanical properties and the fractal dimension.

Amparano et al. (2000) conducted experimental study the effect of aggregate content on
fracture behavior of concrete by testing 48 geometrically similar three-point bend concrete
beams. The test results are analyzed by using a size effect method, in which the fracture
behavior of concrete is characterized by two parameters: fracture energy G f and effective
fracture process zone cf. Test results showed that by increasing volume fraction of aggregate in
the range 45-75%: (1) the compressive strength of concrete decreases slightly (15%), and can
be practically considered to be a constant; (2) fracture energy Gf varies within 25%, and there
was not critical volume fraction which gives the maximum G f ; and (3) the size of the fracture
process zone decreases, which is explained by the change in coarseness of grain structures
defined in terms of mosaic patterns.

10
Issa et al. (2000) conducted an experimental investigation to study the influence of micro-
structural parameters (aggregate size) and macroscopic parameters (specimen dimensions) on
brittle fracture. Maximum aggregate size was used as a representative parameter of aggregate
distribution. Six groups of geometrically similar concrete specimens with various dimensions
and aggregate sizes were prepared. Three forms of comparisons of the test results with respect
to the specimen and aggregate sizes were adopted. The first corresponded to the various
specimen sizes cast with the same maximum aggregate size. The second comparison was based
on the geometrically identical specimens cast with various maximum aggregate sizes. The third
form of comparison dealt with complete geometrical similarity, i.e., specimen thickness to
maximum aggregate size ratio was identical.

The test results were also presented in the form of load-crack mouth opening
displacement curves, compliance data, surface measured crack length and crack trajectories as
well as calculated crack length, critical energy release rate, and fracture toughness, G1. From
the results, the following conclusions were obtained. (i) For a fixed specimen size, G1 increases
with an increase in the aggregate size (fracture surface roughness). (ii) For the same maximum
aggregate size specimens, the apparent toughness increases with specimen size. (iii) It was clear
that the rate of increase in G1, with respect to an increase of the dimensionless crack length (the
crack length normalized by the specimen width), increases with both specimen size and
maximum aggregate size increase.

Issa et al. (2003) addressed the potential application of fractal geometry to characterize the
fracture surface and to determine whether there is any correlation between fracture properties
and the roughness of the fracture surface. Fractured surfaces of three different size wedge-
splitting specimens, dimensions varying from (width x total depth x thickness) 410x410x50
mm to 1680x1680x200 mm with four different maximum aggregate sizes of 9.5, 19, 38, and 76
mm, were analysed using a modified slit-island technique. The fracture surface of the specimen
generated as a result of wedge-splitting test was sprayed with a very thin coating of white paint
in order to have uniform colour. The specimen was placed in a separate water tank and was left
to soak for a day or two so that it did not produce bubbles during the measurement procedure.
The fractured specimens were then placed in a tank filled with non-staining dyed water. The
entire setup, i.e., the tank and the camera, was placed in a black box in order to avoid any
artificial effects due to disturbance of light. Pictures were taken using macro-lenses. Black and

11
white pictures were then scanned and processed using an image analyser. For comparison, two
different techniques (area–perimeter and wave number–spectral energy density relationships)
were used for fractal analysis. They found that fractal dimension, i.e., roughness, increases with
an increase in both specimen and maximum aggregate size. A clear correlation exists between
roughness (fractal dimension) and fracture toughness. The tougher the material, the higher the
fractal dimension.

Cao and Ren (2006) experimentally investigated the fractal behavior of crack development in
damaged concrete materials and its application to structural damage assessment. The
experimental results demonstrated that the crack evolving is closely related to the
characteristics of fractal behavior, and the surface-crack distribution of the damaged concrete
structures possesses fractal characteristics. It thus indicates that the fractal may be capable of
performing quantitative characterization of crack development, which cannot be comprehended
by Euclidian geometry.

Wei et al. (2010) studied a series of reinforced concrete cracks by image graying, calibrating,
image denoising, image enhancement and image segmentation, and finally the binary image of
cracks is obtained. Using the Box Method, the fractal dimension of cracks is determined. The
relationships between the fractal dimension of surface cracks in the cracking beam under three-
point loading is studied along with the load and the deflection of the beam.

Konkol and Prokopski (2011) developed a new method of assessment of concrete porosity
diversification using the fractal dimension ‘D’, fractal dimension as an additional parameter for
characterizing the macro-pore structure with stereological methods. The analysis for porosity of
concrete was performed on the flat cross-sections of concrete specimens using image analysis.
By using basic morphological and punctual transformation along with proper preparation of
specimens, a binary image of macro-pores in the hardened concrete was obtained from the real
image of the specimen. They found that the greater the values of relative number of pore cross-
sections NA (or the number of pores N), the greater is the fractal dimension, D. They also
found that greater water-cement ratio, W/C, the smaller is the fractal dimension, D.

Hong-quan et al. (2011) experimentally studied the crack evolutions of reinforced concrete
beams with three different aggregate sizes under the concentrated loads. Using fractal theory,
the generation, development and distribution of the cracks on the reinforced concrete beams
were analyzed. The results showed that the cracks on the reinforced concrete beams have
12
fractal characteristics. On studying the fractal dimension of the cracks, the aggregate sizes have
significant effect to the cracks on the reinforced concrete beams. The relationships of the fractal
dimensions of the crack evolution in loading were described by the power function laws. The
results of their paper provide a new theoretical basis for selecting the appropriate aggregate size
to improve the strength of reinforced concrete beams.

Hong-quan and Jun (2011) used fractal geometry to study the crack evolving process of
reinforced concrete beams. The fractal dimensions on cracked surface of the reinforced
concrete beam and the mechanical properties of the beam have the linear relationships. In order
to compare the accuracy of the fractal dimensions, box counting method and the digital image
box method are used to calculate the fractal dimension. The cracks of the reinforced concrete
beam have a good fractal features. Hence, the fractal dimension can be used to characterize the
crack distribution on the surface of the reinforced concrete beam quantitatively. From the
results, it was observed that box counting method is more accurate in measuring fractal
dimension, but the processing is rather complex. Digital image method has a convenient
processing but the fractal dimension measured by the digital image method is smaller.

Beygi et al. (2013) conducted an experimental research to study the effect of water to cement
ratio on fracture parameters and brittleness of self-compacting concrete. They conducted three
point bending tests on 154 notched beams with different water to cement (w/c) ratios. The
specimens were made from mixes with various w/c ratios from 0.7 to 0.35. For all mixes,
common fracture parameters were determined using two different methods, the work-of-
fracture method (WFM) and the size effect method (SEM). Test results showed that a decrease
of w/c ratio from 0.7 to 0.35 in SCC resulted in the following: (a) the fracture toughness
increases linearly (b) the brittleness number is approximately doubled (c) the effective size of
the process zone cf in SEM and the characteristic length (lch) in WFM decrease which is
explained by the change in structural porosity of the aggregate–paste transition zone; and (d)
the fracture surface of concrete is roughly smoother, which can be attributed to the improved
bond strength between the aggregates and the paste.

Jin et al. (2013) presented a numerical analysis on the fractal characteristics of cracks and pore
structure of concrete with the help of digital image technology. The results show that concrete
cracks and the micro pore distribution of concrete are of fractal characteristics and the fractal
dimension ranges from 1 to 2. The fractal characteristics of pores in cracked concrete and un-
13
cracked concrete is similar. The fractal dimension of the micro pore structure in cracked
concrete is bigger than that in uncracked concrete. The cracks and micro-pore structure of
concrete show very clear fractal properties. The irregularity of concrete cracks and
homogeneous degree of micro pore distribution c a n be r e p r e s e nt e d with fractal principle.
Further, fractal dimensions can be taken as a characteristic parameter to describe the tortuosity
of cracks.

Karamloo et al. (2016) studied the influences of water to cement ratio on brittleness and
fracture parameters of self-compacting lightweight concrete. For this purpose, four mix
compositions with different water/cement ratios from 0.35 to 0.5 were prepared such that the
nominal maximum aggregate size and weight of coarse and fine aggregates were kept constant.
To determine the fracture parameters, twelve notched beam specimens were casted for each
mix and the results were analyzed by means of the size effect method. The obtained results
indicate that, there exists a relationship between the water/cement ratio, fracture behavior, and
mechanical properties of this material.

Wang et al. (2016) used the maximum loads of notched three-point-bend specimens with
constant height to determine the tensile strength and fracture toughness of concrete. Normally
the tensile strength and fracture toughness cannot be obtained from small three- point-bend
specimens because of the highly heterogeneous composite structures of concrete and formation
of fictitious crack prior to the maximum load. A simple fracture model is established, linking
fictitious crack to the maximum aggregate size so that the concrete structure and fictitious crack
formation can be considered together. Maximum loads of three- point-bend specimens with
different initial notches containing different influence of the tensile strength and fracture
toughness are then analyzed. Since, concrete is heterogeneous, inclusion of maximum
aggregate size, dmax in a final relation of concrete fracture modelling is necessary, which has
not been emphasized in the existing fracture models.

Uday (2017) experimentally determined the fracture energy (G F) by testing three point bend
concrete beams of same size but varying notch to depth ratios using work energy method
developed by RILEM. He conducted displacement controlled three point bend tests on 18
beams with three different notch to depth ratios and obtained the load-displacement graphs and
obtained the fracture energy. He found that as notch to depth ratio increases maximum load
carrying capacity decreases. He also concluded that fracture energy decreases as notch to depth
14
ratio increases.

Fayyad and Less (2017) experimentally investigated the cracking process in lightly reinforced
concrete (RC) beams and observed the details of the development of localized fracture process
zone. They aimed to investigate the relationships between beam height (120 mm, 220 mm and
320 mm), steel reinforcement ratio (0.1–0.5%), ductility and the onset of crack branching. RC
beams were tested in three-point bending and experimental surface strains and crack openings
were inferred using digital image correlation (DIC). It was found that the presence of the
reinforcement prevented premature fracture and led to crack branching where a single crack
bifurcated in the region of the compression zone. In the larger beams the branching developed
at a lower relative height and a larger reinforcement ratio led to a shallower branching angle.
These observations were associated with ductility measures for lightly reinforced concrete
beams.

Sadrmomtazi et al. (2019) performed three-point bending tests on 224 notched beams in order
to study the effect of different water to cement ratios (w/c=0.4–0.7) on the fracture
characteristics of Heavy-Weight Concrete (HWC). The fracture parameters were then analyzed
by Work of Fracture Method (WFM) and Size Effect Method (SEM). Heavyweight concretes
(HWC) are the type of concretes which are very effective for shielding against nuclear
radiation. Their properties and components of HWC consist of heavyweight aggregates and its
density is higher than normal concrete (2600 kg/m3). Fracture energy Gf considerably depends
on w/c ratio so that Gf decreases about 27% with increasing w/c ratio from 0.4 to 0.7. This can
be attributed to the fact that the content and size of pores in interfacial transition zone (ITZ)
increase with increasing w/c ratio which causes reduction of strength in both ITZ and cement
paste. Hence in concrete, fracture path goes around the aggregate which has lower strength.
Likewise, with decrease in w/c ratio, crack propagation occurs through the aggregates and also
smoother fracture surface with small fractal dimensions forms. In fact, the porosity in ITZ and
cement paste increases with increasing w/c ratio which causes the fracture mode from fracture
through aggregates to fracture around aggregates. Thus, he concluded that with increasing w/c
ratio, fractal dimensions increase.

Yin et al. (2019) evaluated the fracture properties of concrete using four-point bending test
where in the beam segment near the mid-span crack section is under pure bending. Five
different initial crack depth ratios varying from 0.2 to 0.6 for concrete specimens were
15
explored. The acquired values of load, crack mouth opening displacement and midspan
deflection were simultaneously recorded. The important fracture parameters such as critical
effective crack extension length, critical effective fracture toughness and fracture energy were
analyzed. Two approaches namely pure bending approach and four point bending approach
were implemented to determine the critical effective fracture toughness. The results indicate
that four-point bending test can be utilized to study the fracture properties of concrete.

2.3. SUMMARY
Fractals exhibit similar patterns at increasingly small scales called self similarity also known as
expanding symmetry or unfolding symmetry; if this replication is exactly the same at every
scale. The feature of "self-similarity“ is easily understood by analogy to zooming in with a lens
or other device that zooms in on digital images to uncover finer, previously invisible, new
structure. If this is done on fractals, however, no new detail appears; nothing changes and the
same pattern repeats over and over.

Looking at the fracture surface of a concrete specimen, one realizes that the self-affine
geometry of crack faces results from the stochastic nature of the crack growth. This is due to
the heterogeneous nature of concrete that makes the crack tortuous leading its way through
weak bonds, voids, mortar and getting arrested on encountering a hard aggregate forming crack
face bridges. These mechanisms contribute to the tendency of the crack to follow a tortuous
path. The self-similarity contained in the tortuous fracture surface of concrete makes it an ideal
candidate to be considered as a fractal.

From the above mentioned literatures, it is proven that the fractured surfaces of concrete are
essentially fractals and many of them have used the fractal dimension to characterize the
fracture properties of concrete. Fractal dimension is found to be a measure of the roughness of
the cracked concrete surface. Box counting method and slit island method are the most
common methods used to determine the fractal dimension of the fractured surface. The fractal
dimension thus obtained was correlated to fracture properties like fracture toughness and
stress intensity factor. It was found from the literature review that a good correlation exists
between fractal dimension and fracture properties of concrete.

16
2.4. OBJECTIVES OF THE WORK

 To study the application of fractals in concrete specimens.

 To relate the fracture energy to compressive strength of concrete.
 To relate the fracture energy to flexural strength of concrete.
 To study the effect of maximum aggregate size and specimen size on fracture properties
and fractal nature of concrete by referring the data from the work conducted by Issa et al
(2000).
 To determine the correlations between fracture properties and fractal characteristics, if
any, from the experimental work conducted by Issa et al (2003).

17
 CHAPTER THREE

METHODOLOGY AND EXPERIMENTAL RESULTS

3.1. EXPERIMENTAL PROGRAM

3.1.1. MATERIALS
Ordinary Portland Cement (OPC 53) conforming to Indian standard specification (IS:269-1983)
was used. The specific gravity of cement is 3.15. The maximum aggregate size of the coarse
aggregates used was 20mm and had a specific gravity of 2.6. The fine aggregates used
correspond to zone II of IS 383:1970 and had a specific gravity of 2.65.

3.1.2. MIX DESIGN

The concrete mix designs were prepared for grades M20, M30, M40 and M50 as per IS 10262:
2009. Detailed mix proportions are presented in Table 1 with quantities of concrete ingredients
per m3 of concrete for different grades. Superplasticizer (SP 430) is added to get desired
workability for M40 and M50 concrete mixes.

Table 1: Quantities of materials for 1 m3 of concrete

Mix Proportions Cement Coarse Fine Water Superplasticizer

Mix Aggregate Aggregate
(Cement: FA :CA:Water) (kg) (kg) (kg) (L) (kg)

M 20 1 : 1.62 : 3.4 : 0.49 403 1344 672 200 -

M 30 1: 1.4 : 2.68 : 0.43 432 1159 603 186 -

M 40 1: 1.8 : 3.25 : 0.4 383 1249 690 153 2.68

M 50 1 : 1.56 : 2.95 : 0.38 425 1252 662 162 2.98

18
3.1.3. MIXING, CASTING AND CURING OF SPECIMENS
Three cubes of size 150 mm for the purpose of determining the compressive strength of
concrete, two cylinders of diameter 150 mm and height 300 mm to determine the split- tensile
strength and three prism specimens of size 100 mm x 100 mm x 500 mm to conduct the three
point bending test were cast for four type of the concrete mixes. Midspan vertical notches were
done in the beam specimens by inserting a steel plate of required dimensions inside the tensile
face of the beams. The thickness and depth of the central notch are 2 mm and 15 mm
respectively. Beam geometry is given in Figure 3.

Figure 3: Geometry of the notched beam specimen

Metallic casting forms were greased with a release agent and fixed to a vibration
table before the start of casting. Concrete was poured in three layers and vibrated. Slump test
was done for each mix before the fresh concrete was placed carefully in the moulds with proper
compaction using table vibrator. Specimens were demoulded after 24 hours and immersed in
potable water for curing to avoid shrinkage and cracking. Compressive strength, split tensile
strength and flexure strength were determined after 28 days. The demoulded concrete
specimens are shown in Figure 4.

19
 Figure 4: Cube, cylinder and beam specimens after demoulding

3.1.4. TEST PROCEDURE

3.1.4.1. COMPRESSIVE STRENGTH TEST

The compressive strength is one of the most useful parameters of concrete which gives an
insight toward the mechanical and durability properties of concrete. Compressive strength is the
ability of material or structure to carry the load till failure. Compression testing of cube
specimens of size (150mm X 150mm X 150mm) was carried out in a compression testing
machine of capacity 2000 kN, as per IS 516: 1959. The compression testing machine is shown
in Figure 5. The load was applied without shock at the rate of 140 kg/sq.cm/min. Set of three
cubes were tested for each of the mix to determine the compressive strength at 28 days.
Ultimate load at which the specimens failed was noted down to find the compressive strength.

20
Compressive strength is calculated as the load applied till the point of failure to the resisting
cross-section area. Formula for calculation is:

𝑃
𝑓𝑐 = (3.1)
𝐴𝑐

where, fc = Compressive strength, P = Failure load, Ac = Cross-sectional area of the cube.

Figure 5: Compressive strength testing of concrete cube specimen

21
3.1.4.2. SPLIT TENSILE TEST
Splitting tensile strength test is a well-known indirect test used for determining the tensile
strength of concrete. The test consists of applying a compressive line load along the opposite
generators of a concrete cylinder placed with its axis horizontal between the compressive
platens. Splitting tensile test is conducted as per IS: 5816-1999. Load is applied at a rate of 1.2-
2.4 MPa/min. Laboratory setup of the experiment and failure of specimen is shown in Figure
4.2. Splitting tensile strength is calculated using the formula:

2𝑃
𝑓𝑡 = (3.2)
𝜋𝐷𝐿

where, P = Load at failure, d = Diameter of the cylinder, L = Length of the cylinder. The test
setup for split tensile specimen is shown in Figure 6.

Figure 6: Test setup for split tensile specimen

3.1.4.3. FLEXURE TEST

In order to evaluate the fracture parameters, for each mix, three notched beams with the same
dimensions of 100×100×500mm (depth × width × length) with a span equal to 400mm were
used. All tests were performed using Universal Testing Machine (UTM) with a maximum load
capacity of 300kN, controlled by a constant midspan displacement of 0.01 mm/sec for all
specimens. The loading was introduced at midspan of the beam. The deflection at centre of the

22
beam was measured at every load increment using the mechanical dial gauge with 0.01 mm
least count. Load increment on UTM was controlled so that the maximum load is reached in
15-20 minutes. The experimental set up is shown in Figure 7.

Figure 7: Experimental setup for flexure test

23
The modulus of rupture or flexure strength, σr can be calculated using the following formula:

3𝑃𝐿
𝜎𝑟 = (3.3)
2𝑏𝑑2

where P is the load at the fracture point, L is the span of the beam, b is the breadth and d is the
depth of the beam.

3.2. DETERMINATION OF FRACTURE ENERGY BY WORK OF FRACTURE

METHOD (WFM)
Various methods have been suggested by researchers in order to obtain the fracture parameters
of concrete. Work of fracture method (WFM) introduced by RILEM 50-FMC is one of the
simplest methods to determine the fracture energy (the most important fracture parameter of
concrete). The cohesive crack model is the basis of this method. This method, which was
originally developed for ceramics is the first method of testing for fracture properties of
concrete to be proposed as a standard (RILEM, 1985). The basis for applying this method to
concrete was developed by Hillerborg and his co-workers. Their method uses the "fictitious
crack" concept. The method can be applied to a variety of test specimen geometries but the
proposed standard uses a beam specimen loaded in three-point bending with a central edge
notch. RILEM technical committee recommended the guidelines for determination of fracture
energy of cementitious materials by conducting TPB test on notched beam. In order to obtain a
complete load - displacement curve, a closed loop servo hydraulic testing machine is needed.
The fracture behaviour of concrete characterized by the RILEM measures the average fracture
energy over the entire projected ligament area.

According to RILEM recommendation, the fracture energy is evaluated by dividing the

total applied energy with the projected ligament area. In WFM, performing three-point bending
test on notched beams, the total fracture energy was calculated as a sum of the area of the
complete load-deflection curve and work done by the deadweight of the specimen, according to
formula:

𝑊
𝐺𝐹 = (𝑑−𝑎𝐹 )𝑏 (3.4)
0

where WF is the total fracture energy consumed corresponding to the area under load-
displacement curve, b is the width of the beam, d is the beam height and a 0 is the notch depth.

24
3.3. COMPRESSIVE STRENGTH OF CONCRETE
The compressive strength of the concrete mixes was determined using three numbers of
standard cube specimens for each mix after curing for 28 days. The compressive strength value
of concrete mixes is presented in Table 2.

Table 2: Compressive strength values of different concrete mixes

Concrete mix Specimen Compressive strength Average compressive

(MPa) strength (MPa)
I 31.11
M20 II 30.67 31.11
III 31.55
I 41.78
M30 II 42.67 41.93
III 41.33
I 48.89
M40 II 48.44 49.33
III 50.67
I 59.55
M50 II 60.44 60.14
III 60.44

3.4. SPLIT TENSILE STRENGTH OF CONCRETE

The split tensile strength of the concrete mixes was determined using two numbers standard
cylindrical specimens for each mix after curing for 28 days. The split tensile strength of
concrete mixes is presented in Table 3.

Table 3: Split tensile strength values of concrete mixes

Concrete mix Specimen Split tensile strength Average split tensile

(MPa) strength (MPa)
M20 I 2.41 2.48
II 2.55
M30 I 3.4 3.47
II 3.54
M40 I 3.68 3.68
II 3.68
M50 I 4.11 4.18
II 4.25

25
3.5. FRACTURE PARAMETERS FROM THREE POINT BENDING TEST
All the prism specimens were tested for three-point bending. The deflection at the midpoint was
measured at each load increment and load-deflection graphs are plotted for each specimen. The
load- displacement plots for different concrete mixes are shown in the Figure 8-11. The load –
displacement readings of all the specimens are listed in Appendix A.

4
Load (kN)

0
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Displacement (mm)

Figure 8: Load – displacement plot for M20 grade concrete – specimen No. 3

26
 8

5
Load (kN)

0
0.1 0.3 0.5 0.7 0.9 1.1 1.3
Displacement (mm)

Figure 9: Load – displacement plot for M30 grade concrete- specimen No. 3

5
Load (kN)

0
0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5
Displacement (mm)

Figure 10: Load – displacement plot for M40 grade concrete- specimen No. 2

27
 9
8
7
6
Load (kN)

5
4
3
2
1
0
0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7
Displacement (mm)

Figure 11: Load – displacement plot for M50 grade concrete- specimen No. 1

Typical load-deflection plot comprises of three stages of behaviour. In the first phase,
the deflection increases linearly with the load while the notch is opened but does not extend. A
fracture process develops during the second phase where micro-cracks form and slow crack
growth is noticeable. In the third phase, known as strain softening, rapid crack growth is
apparent. Upon loading, the micro-cracks propagate and new macro-cracks are formed. The
load in the concrete decreases after the peak load as the deformation keeps on increasing. This
phenomenon is called softening. Strain softening of concrete occurs when the micro-cracks
coalesce to form a zone of damage, weakening the concrete.

3.6. ANALYSIS OF FRACTURE PARAMETERS FROM WFM

Total fracture energy (G F) in WFM, is the energy required to create a crack with unit surface
area. As shown in Eq. (3.4), the fracture energy (GF) was calculated as the total energy
consumption divided by the cross-sectional area of the ligament over the notch. The fracture
energy thus determined from Displacement plots of the beam specimens are presented in Table
4.

28
 Table 4: Fracture energy values of the concrete specimens

COEFFICIENT
OF
CONCRETE COMPRESSIVE FRACTURE ENERGY, GF AVERAGE
VARIATION
MIX STRENGTH (N/m) GF (N/m) (%)
(MPa)

Beam 1 Beam 2 Beam 3

M20 31.11 225.76 210.94 226.52 221.07 3.97

M30 41.93 399.06 435.09 444.07 426.07 5.59

M40 49.33 570.59 555.5 535.47 553.85 3.18

M50 60.14 675.61 710.28 686.26 690.72 2.57

It is clear from the above table that the fracture energy is a function of compressive
strength of concrete and it increases with increase in compressive strength of concrete. Based
on CEB-FIP, fracture energy is a function of compressive strength. Therefore, many
researchers have proposed relations between fracture energy and compressive strength. Some
of them have presented in Figure 12.

Figure 12: Relationship between GF and fc by different researchers

29
The plot between the compressive strength of concrete and the fracture energy is shown in
Figure 13. According to the obtained results in this study, the following relation between
fracture energy and compressive strength is presented:

𝐺𝐹 = 716.33 ln 𝑓𝑐 − 2243.5 (3.5)

where fc is compressive strength and GF is total fracture energy of concrete.

800

700
Fracture Energy, GF (N/m)

600

500

400

300

200

100

0
0 10 20 30 40 50 60 70

Compressive Strength, fc (MPa)

Figure 13: Correlation between fracture energy and compressive strength

3.7. CORRELATION BETWEEN FLEXURE STRENGTH AND FRACTURE

ENERGY
The modulus of rupture or flexure strength of each concrete specimen has been calculated using
Eq. (3.3). The values of modulus of rupture for each specimen along with its corresponding
fracture energy are listed out in Table 5.

30
 Table 5: Modulus of rupture for different concrete specimens

Concrete Peak load (kN) Modulus of rupture, σr (MPa) Average modulus Average GF
mix of rupture, σr (N/m)
(MPa)
Beam 1 Beam 2 Beam 3 Beam 1 Beam 2 Beam 3
M20 5.8 5.5 5.3 4.82 4.57 4.40 4.60 221.07
M30 6.6 6.8 7.0 5.48 5.65 5.81 5.65 426.07
M40 7.6 7.4 7.3 6.31 6.15 6.06 6.17 553.85
M50 7.9 8.0 8.2 6.56 6.64 6.81 6.67 690.72

From Table 5, it is clear that a correlation exists between flexure strength and fracture energy of
concrete. Figure 14 shows the graphical representation of fracture energy against flexure
strength. The data is best fitted with a quadratic relation that is given by:

GF = 30.722σr 2 - 119.22σr + 119.35 (3.6)

800

700

600
Fracture Energy, GF (N/m)

500

400

300

200

100

0
4 4.5 5 5.5 6 6.5 7
Modulus of rupture,σr (MPa)

Figure 14: Correlation between fracture energy and modulus of rupture of concrete

31
 CHAPTER FOUR

ANALYTICAL WORK FROM EXPERIMENTAL DATA BY ISSA ET AL

(2000, 2003)

4.1. SUMMARY OF THE WORK DONE BY ISSA ET AL. (2000, 2003)

Issa et al. (2000) experimentally investigated the influence of micro-structural parameters, such
as aggregate size, and macroscopic parameters, such as specimen dimensions, on brittle
fracture. Maximum aggregate size was used as a representative parameter of aggregate
distribution in their work. They had prepared six groups of geometrically similar concrete
specimens with various dimensions and aggregate sizes. Similarity of the specimens was
strictly maintained by scaling the specimen dimensions from one group to another by a factor
of two starting from a specimen size of (width x total depth x thickness) 105x105x12.5mm to
1680x1680x200 mm. Two separate sets of removable pre-cast notches were designed to
determine the effect of initial notch size. They have designed the loading fixture to have a
reproducible crack initiation and controlled crack growth. Several loading fixtures were
evaluated prior to selection of the one used in the experimental program. Quasi-static splitting
cyclic loading in edge cleavage configuration was applied. A servo-hydraulic Instron machine
was used for testing. The fracture process was monitored by optical and acoustic imaging
techniques. Then the fractured surfaces of three different size wedge-splitting specimens with
four different maximum aggregate sizes were analyzed using a modified slit-island technique.
They have conducted this fractal analysis to the potential application of fractal geometry to
characterize the fracture surface of concrete specimens.

4.2. EXPERIMENTAL PROGRAM

In the experiment conducted by Issa et al. (2000), they cast six sets of geometrically similar
concrete specimens of various sizes with various maximum aggregates. The geometrical
similarity was maintained by scaling up all the macroscopical dimensions of the specimens
from one set to the next by a factor of 2 beginning from specimens with sizes of 105x105x12.5
mm (width x depth x thickness) up to 1680x1680x200mm. Thus, the maximum scaling factor
was 24=16. They have also used an additional intermediate size specimen (S2) in the
experimental program. Various concrete mixes prepared by their team were based on Type I
Portland cement, regular river sand with a maximum size of 4.75mm, and coarse aggregates

32
with various maximum sizes, i.e., 4.75, 9.5, 19, 38, and 76mm. Details of the specimen
geometries are presented in Table 6 and Figure 15.

Table 6: Details of the specimen geometry

Specimen Total depth Depth d, Width W, Thickness t, Initial crack width a0,
Type H (mm) (mm) (mm) (mm) (mm)

0.15d 0.3d
Very large 1680 1400 1680 200 210 420
specimen
Large 840 700 840 100 105 210
specimen
Medium 420 350 420 50 53 105
specimen
Small 210 175 210 25 26 53
specimen

Figure 15: Specimen configuration of the wedge splitting test (Source: Issa et al. (2000))

Six different sets of moulds were prepared. Two separate sets of removable pre-cast notches
were designed that allow two relative notch depths of 0.15 and 0.30. The six sets of similar
specimens prepared by Issa et al. (2000) for testing are shown in Figure 16.

33
 Figure 16: Set of finished specimens (Source: Issa et al. (2000))

Bearings and rollers were used in the fixture to minimize friction. A downward vertical
load was applied to the rollers through a wedge with an angle of 8.75. The angle of the wedge
was optimized to reduce friction between the rollers and the wedge. The fixture and loading
setup were being designed by Issa et al. (2000) so that the applied vertical load by the actuator
would be translated as the splitting load to the specimen. They have used a servo-hydraulic 8500
Instron machine for testing the concrete specimens. The machine was equipped with
displacement, load, and strain channels to control the rate of displacement, loading, and strain,
respectively. They were able to monitor the fracture process by optical and acoustic imaging
systems. The downward force applied through a wedge produced splitting force acting through
the rollers. Two dynamic CMOD extensometers had been attached to the centreline of the
fixture and notch-tip to monitor the crack opening displacement during the test.

They conducted the testing in a displacement control mode. The rate of loading and
unloading was kept constant at 0.125mm/min (0.005 in/min). The specimens were loaded until
a crack advance occurred, resulting in a characteristic plateau on the load-displacement
diagram.

4.3. EXPERIMENTAL SETUP FOR FRACTAL ANALYSIS

In the study conducted by Issa et al. (2003) the non-destructive modified slit-island technique
was used to characterize the fracture surfaces of concrete specimens. From the specimens
mentioned earlier, three different sizes were used for fractal analysis, i.e., S 1, S3 and S4 with
four different maximum aggregate sizes of 9.5, 19, 38, and 76 mm. The fracture surface of the

34
specimen generated as a result of wedge-splitting test had been sprayed with a very thin coating
of white paint in order to have uniform colour. The specimen had been placed in a separate
water tank and was left to soak for a day or two so that it did not produce bubbles during the
measurement procedure. The fractured specimens were then placed in a tank filled with non-
staining dyed water. The concentration of the dye was adjusted so that the highest and lowest
points on the fracture surface were clearly visible. They placed the entire setup, i.e., the tank
and the camera, in a black box in order to avoid any artificial effects due to disturbance of light
were taken using macro-lenses. Black and white pictures were then scanned and processed
using an image analyzer. For comparison, two different techniques (area–perimeter and wave
number–spectral energy density relationships) were used by them for fractal analysis. The
image of the fractured surface was filtered to remove high spatial frequency noise.

Figure 17: Photographic view of the original fracture surface (Source: Issa et al. (2003))

Figure 18: Computer simulated fracture surface reconstructed on the basis of the image analysis
(Source: Issa et al. (2003))

35
4.4. COMPUTATIONAL TECHNIQUE
An image is represented by a set of pixels, each having a corresponding intensity value ranging
from 0 to 255, i.e., intensity variations from low (black) to high (white). For each applied
intensity range, new islands appeared and subsequently these islands grew and emerged upon
increasing the threshold values. For each intensity range, the area and perimeter of new islands
were automatically counted and recorded. The logarithmic areas of islands and their perimeters
appearing at different levels were best fitted by linear regression and the fractal dimension was
computed using Eq. (1.5). The technique employed by Issa et al. (2003) in wave number–
spectral energy density involves 2D fast Fourier transform procedures for processing the
images. Wave numbers and their corresponding spectral energy density values were recorded.
Within the Fourier domain, a power law relationship exists between the spectral energy density
and the wave number. The slope of the resulting log–log plot was used to compute the 2D
statistical fractal dimension of the image as presented in Eq. (1.6).

4.5. ANALYSIS OF THE DATA FROM THE WORK DONE BY ISSA ET AL. (2000)
To obtain the relation between fracture properties and fractal characteristics of concrete, the
data from the experimental work conducted by Issa et al. (2000, 2003) was referred. Different
correlations were obtained among the specimen characteristics, fracture properties and fractal
dimension of concrete by analysing the data. Various parameters under consideration were
specimen size, maximum aggregate size, fracture energy, critical energy release rate, critical
stress intensity factor and fractal dimension.

4.5.1. RELATION BETWEEN MAXIMUM AGGREGATE SIZE WITH FRACTURE

ENERGY AND STRESS INTENSITY FACTOR BY WORK OF FRACTURE
METHOD
The fracture energy was calculated as the area under load –displacement graph divided by the
ligament area (d x t). Fracture energy is the fundamental fracture parameter representing
cracking resistance and fracture toughness of concrete. Therefore it is generally considered as a
material property in concrete fracture mechanics. But from the literatures reviewed in the
previous chapter, it was observed that the experimentally determined fracture energy increases
with specimen size. The scale effects, i.e., dependence of fracture energy GF on specimen
dimension and maximum aggregate size is well pronounced in numbers reported in Table 7.

36
 Table 7: Fracture energy corresponding to different specimens

Specimen Specimen Size Maximum Average fracture

Type aggregate size, dmax energy, GF (J/m2)
(mm)
1680 x 1680 x 200 9.50 189.14
1680 x 1680 x 200 19.00 224.17
Very large
specimen 1680 x 1680 x 200 38.00 248.68
1680 x 1680 x 200 76.00 280.21
840 x 840 x 100 4.75 140.10
840 x 840 x 100 9.50 180.38
Large
specimen 840 x 840 x 100 19.00 199.65
840 x 840 x 100 38.00 231.17
840 x 840 x 100 76.00 252.19
420 x 420 x 50 4.75 134.85
420 x 420 x 50 9.50 143.61
Medium
specimen 420 x 420 x 50 19.00 182.14
420 x 420 x 50 38.00 199.65
210 x 210 x 25 4.75 108.83
Small
210 x 210 x 25 9.50 113.83
specimen
210 x 210 x 25 19.00 129.60

The graphical representation of fracture energy with respect to maximum aggregate size is
given in Figure 19. It can be observed that the fracture energy GF measured on larger
specimens is higher than that obtained from smaller specimens for all maximum aggregate size.
It can also be observed that for a fixed specimen size, the fracture energy GF increases with the
maximum aggregate size.

The pattern of fracture in concrete can be characterized by one or two cases or a

combination of both (a) The crack deflects around the aggregates and (b) The crack penetrates
through the aggregates. When a growing crack reaches an aggregate, it either goes around or
penetrates through it. The possibility of either or both patterns occurring depends on the relative
strength of cement paste and the aggregates. In case of a normal strength concrete where in the
cement paste is weaker than the aggregate, the crack pattern is a more tortuous path as the crack
deflects around the aggregates, instead of passing through them. Thus it can be concluded that,

37
the increase in fracture energy with increase in maximum aggregate size and specimen size is
due to the fact that the crack pattern is meandering path as the crack deflects around the
aggregates which tends to raise the energy consumption and toughness.

300
Average fracture energy, GF (J/m2)

250

200
Very large specimen
150
Large specimen

100 Medium specimen

Small specimen
50

0
0 20 40 60 80
Maximum aggregate size, dmax (mm)

Figure 19: Plot between maximum aggregate size and fracture energy

There exists a good correlation between these two parameters. The relation between fracture
energy and maximum aggregate size is shown in Table 8.

Table 8: Relation between fracture energy and maximum aggregate size

SPECIMEN TYPE EQUATIONS FIT (%)

Very large specimen GF = 42.952ln(dmax)+ 94.195 99.59

Large specimen GF = 39.67ln(dmax) + 83.893 98.63

Large specimen GF = -0.0672dmax2 + 4.9472dmax + 109.16 97.68

Small specimen GF = 1.4865dmax + 100.95 98.99

38
Stress intensity factor, KI is a fracture parameter that characterizes the stress distribution at the
crack tip. Stress intensity factor for a compact specimen as shown in Figure 20 is calculated
using the formulae given as below:

𝑎 𝐾𝐼 𝐵√𝑊
𝑓( ) = (4.1)
𝑊 𝑃

𝑎
𝑎 2+
𝑊 𝑎 𝑎 2 𝑎 3 𝑎 4
𝑓 (𝑊) = 𝑎 3/2
[0.886 + 4.64 (𝑊) − 13.32 (𝑊) + 14.72 (𝑊) − 5.60 (𝑊) ] (4.2)
(1− )
𝑊

Figure 20: Compact specimen

where B is the thickness of the specimen and a, W and P are as shown in Figure 20. The
computed values of stress intesity factor, KI are shown in Table 9. The graphical representation
of stress intensity factor with respect to maximum aggregate size is given in Figure 21. It can
be observed that the stress intensity factor, KI measured on larger specimens is higher than that
obtained from smaller specimens for all maximum aggregate size. It can also be observed that
for a fixed specimen size, the stress intensity factor, KI increases with the maximum aggregate
size.

39
 Table 9: Stress intensity factor for each of the specimens

Specimen Specimen Size Maximum a W B Maximum Stress

Type aggregate size, Load, P intensity
dmax (mm) (kN) factor, KI
(MNm-3/2)
1680 x 1680 x 200 9.50 210 1400 200 77.52 1.195
1680 x 1680 x 200 19.00 210 1400 200 84.75 1.306
Very
large 1680 x 1680 x 200 38.00 210 1400 200 89.86 1.385
specimen 1680 x 1680 x 200 76.00 210 1400 200 95.25 1.468
840 x 840 x 100 4.75 105 700 100 25.43 1.108
840 x 840 x 100 9.50 105 700 100 27.84 1.213
Large
specimen 840 x 840 x 100 19.00 105 700 100 29.37 1.280
840 x 840 x 100 38.00 105 700 100 32.10 1.399
840 x 840 x 100 76.00 105 700 100 34.85 1.519
420 x 420 x 50 4.75 53 350 50 6.43 0.797
420 x 420 x 50 9.50 53 350 50 8.21 1.017
Medium
specimen 420 x 420 x 50 19.00 53 350 50 8.92 1.105
420 x 420 x 50 38.00 53 350 50 10.12 1.254
210 x 210 x 25 4.75 26 175 25 2.13 0.739
Small
210 x 210 x 25 9.50 26 175 25 2.33 0.808
specimen
210 x 210 x 25 19.00 26 175 25 2.42 0.840

1.6
Stress intensity factor, KI (MNm-3/2)

1.5
1.4
1.3
1.2 Very large specimen
1.1 Large specimen

1 Medium specimen

0.9 Small specimen

0.8
0.7
0 20 40 60 80
Maximum aggregate size, dmax(mm)

Figure 21: Plot between stress intensity factor and maximum aggregate size
40
There exists a good correlation between stress intensity factor and maximum aggregate size.
The relation between these two parameters is shown in Table 10.

Table 10: Relation between stress intensity factor and maximum aggregate size

Specimen type Correlation Fit (%)

Very large specimen KI = 0.1296ln(dmax) + 0.9121 99.36

Large specimen KI = 0.1454ln(dmax) + 0.8756 99.11

Medium specimen KI = 0.2105ln(dmax) + 0.4964 97.15

Small specimen KI = 0.0729ln(dmax) + 0.6316 95.72

4.5.2. CRITICAL VALUES OF ENERGY RELEASE RATE, GIC AND STRESS

INTENSITY FACTOR, KIC DETERMINED BY SIZE EFFECT METHOD
Issa et al. (2000) performed two-dimensional finite element modelling of the above described
test configuration using the ‘FRANC2D’ (Cornell Fracture Group, 1995). They performed an
incremental solution with the geometry of the specimen and crack trajectory, modified at each
step to reflect the evolving configuration of the crack. Fracture mechanics parameters,
especially the stress intensity factor, KI and the energy release rate, GI had been computed at
each step of the analysis. Stress intensity factor, KI completely characterizes the stress
distribution at the crack tip. Fracture occurs at critical combination of stress and strain, i.e.,
fracture occurs when KI = K1C. Energy release rate, GI is a measure of the energy available for
an increment of crack extension. At the moment of fracture, GI = G1C, the critical energy
release rate. Both K1C and G1C are a measure of fracture toughness. The specimen configuration
employed in the computation replicated that used in the experimental program. The
experimental maximum load and crack-tip location at the end of each loading cycle resulting
from the quasi-static splitting tensile test (wedge splitting test) were used as the input data to
run the step by step analysis for each crack. The values of K1C and G1C observed at crack
initiation are summarized in Table 11.

41
 Table 11: values of K1C and G1C observed at crack initiation for different specimens

Maximum aggregate size,

Specimen Type dmax (mm) Average KIC (MNm-3/2) Average GIC (J/m2)
9.5 1.628 76.007
Very large specimen
19 1.700 96.573
(1680 x 1680 x 200)
38 1.781 99.475
76 1.779 95.972
4.75 1.263 47.285
9.5 1.519 66.2
Large specimen
19 1.546 71.629
(840 x 840 x 100)
38 1.583 78.634
76 1.619 83.888
4.75 1.124 38.004
9.5 1.443 59.545
Medium specimen
(420 x 420 x 50) 19 1.465 64.799
38 1.521 70.189
4.75 1.103 36.252
Small specimen
(210 x 210 x 25) 9.5 1.239 43.958
19 1.356 55.342

1.9

1.8

1.7
Average KIC (MNm-3/2)

1.6

1.5 Very large specimen

1.4 Large specimen

Medium specimen
1.3
Small specimen
1.2

1.1

1
0 20 40 60 80
Maximum aggregate size, dmax (mm)

Figure 22: Plot between Average KIC and Maximum aggregate size

42
The graphical representation of critical stress intensity factor, KIC with respect to maximum
aggregate size for different size specimens is shown in Figure 22. It can be observed that for the
same specimen size, KIC increases as the maximum aggregate size increases. It can also be
observed from the graph that, for the same maximum aggregate size, KIC increases with the
specimen size. The increase in the value of KIC with increasing maximum aggregate size is
associated with the increasing resistance encountered by the propagating crack. Similarly, for
different size specimens made with the same maximum aggregate size, KIC increases as the
specimen size increases. This is because of the fact that the propagating crack encounters more
resistance due to the presence of more coarse aggregate particles. The relation between critical
stress intensity factor and maximum aggregate size, dmax is shown in Table 12.

Table 12: Relation between critical stress intensity factor and maximum aggregate size

SPECIMEN TYPE EQUATION FIT (%)

Very large specimen KIC= -8*10-5dmax2 + 0.0093dmax + 1.5497 99.84

Large specimen KIC= 0.112ln(dmax) + 1.1764 75.73

Medium specimen KIC= -0.0007dmax2+ 0.0404dmax + 1.0123 79.47

Small specimen KIC= 0.1825ln(dmax) + 0.8218 99.81

The similar trend is shown in the case of critical energy release rate. The graphical
representation of critical energy release rate GIC with respect to maximum aggregate size for
different size specimens is shown in Figure 23.

43
 120

100

Average GIC (J/m2) 80

Very large specimen
60 Large specimen
Medium specimen
40
Small specimen
20

0
0 20 40 60 80
Maximum aggregate size, dmax (mm)

Figure 23: Plot between Average GIC and Maximum aggregate size

It can be observed that for the same specimen size, G IC increases as the maximum aggregate
size increases. It can also be observed from the graph that, for the same maximum aggregate
size, GIC increases with the specimen size. The reason for this size effect in the values of critical
energy release rate is the same as that for critical energy release rate since both are a measure of
fracture toughness. The relation between critical energy release rate and maximum aggregate
size, dmax is shown in Table 13.

Table 13: Relation between critical energy release rate and maximum aggregate size

SPECIMEN TYPE EQUATION FIT (%)

Very large specimen GIC= -0.015dmax2 + 1.5193dmax + 66.633 80.9

Large specimen GIC= 12.355ln(dmax) + 33.148 91.75

Medium specimen GIC= -0.053dmax2+ 3.1059dmax + 28.219 88.45

Small specimen GIC= 13.771ln(dmax) + 14.183 98.78

.

4.6. COMPUTATION OF FRACTAL DIMENSION

The fractured surfaces of three different size wedge-splitting specimens made with four
different maximum aggregate sizes were submitted to fractal analysis by Issa et al. (2003). The
fracture surface was reconstructed on the basis of the image analysis. Two techniques naming
slit-island method (Area – perimeter relationship) and 2D fast Fourier transform method (Wave
44
number- spectral energy density relationship) which are explained in the previous chapters
were used by them to compute the fractal dimension of the specimens. The fractal dimensions
thus obtained from the two techniques with specimen size and maximum aggregate size are
presented in the Table 14. The fractal dimensions for small specimens were calculated by
obtaining relationship between fractal dimension and t/d max (ratio of the thickness of specimen
to the maximum aggregate size) and extrapolating to get the corresponding fractal dimensions
for small sized specimens. The fractal dimensions computed from spectral energy density and
the wave number (Fourier transform procedure) are higher in magnitude compared to those
obtained by using the area–perimeter method.

Table 14: Fractal dimensions for different specimens (Source: Issa et al. (2003))

Maximum Fractal Dimension (D)

Specimen Type Specimen Size aggregate size,
dmax (mm)
Slit island 2D fast Fourier
method transform
method
1680 x 1680 x 200 9.5 2.19 2.28
1680 x 1680 x 200 19 2.22 2.31
Very large
specimen 1680 x 1680 x 200 38 2.25 2.35
1680 x 1680 x 200 76 2.29 2.39
840 x 840 x 100 9.5 2.15 2.26
840 x 840 x 100 19 2.19 2.29
Large specimen
840 x 840 x 100 38 2.21 2.32
840 x 840 x 100 76 2.24 2.34
420 x 420 x 50 9.5 2.15 2.25
Medium specimen 420 x 420 x 50 19 2.17 2.27
420 x 420 x 50 38 2.2 2.3
210 x 210 x 25 9.5 2.15 2.24
Small specimen 210 x 210 x 25 19 2.16 2.25

45
4.6.1. EFFECT OF SPECIMEN SIZE AND MAXIMUM AGGREGATE SIZE IN
FRACTAL DIMENSION
From Table 14, it is clear that a correlation exists for fractal dimension with specimen size and
aggregate size. A more simple understanding of the same can be concluded from the graphical
illustrations below.

2.3

2.28

2.26
Fractal dimension, D

2.24
Medium specimen
2.22
Very large specimen

2.2 Large specimen

Smallspecimen
2.18

2.16

2.14
0 20 40 60 80
Maximum aggregate size, dmax (mm)

Figure 24: Plot between Fractal dimension (from slit island method) and maximum aggregate size

Figure 24 presents the plot between fractal dimensions obtained from slit island method and
maximum aggregate size and it is evident that both are directly proportional. The correlations
between fractal dimension (from slit island method) and maximum aggregate size are listed in
Table 15 below.

46
 Table 15: Correlations between fractal dimension (from slit island method) and maximum
aggregate size

SPECIMEN TYPE EQUATIONS FIT (%)

Very large specimen D = 0.0476ln(dmax)+ 2.0808 99.45

Large specimen D = 0.0418ln(dmax)+ 2.0598 98.36

Medium specimen D = -0.00002dmax2 + 0.0026dmax + 2.1267 99.99

Small specimen D = 0.0144ln(dmax) + 2.1175 99.99

2.26

2.24
Fractal dimension, D

2.22

2.2 dmax=9.5
dmax=19
2.18
dmax=38
2.16

2.14
0 5 10 15 20 25
t/dmax

Figure 25: Plot between fractal dimension (from slit island method) and t/dmax

The plot between fractal dimension (from slit island method) and t/dmax is shown in Figure 25
where t is the specimen thickness and dmax is maximum aggregate size. The fractal dimensions
obtained from 2D fast Fourier transform method also show a similar trend with both specimen
size and maximum aggregate size. The plot between fractal dimensions obtained from 2D fast
Fourier transform method and maximum aggregate size is shown in Figure 26.

47
 2.4
2.38
2.36

Fractal dimension, D
2.34
2.32 Very large specimen
2.3 Large specimen
2.28 Medium specimen
2.26 Small specimen

2.24
2.22
0 20 40 60 80
Maximum aggregate size, dmax (mm)

Figure 26: Plot between fractal dimension (2D Fast Fourier Transform method) and maximum
aggregate size

The relationship between fractal dimension obtained from 2D fast Fourier transform method and
maximum aggregate size is given in Table 16.

Table 16: Relations between fractal dimension (2D Fast Fourier Transform method) and
maximum aggregate size

SPECIMEN TYPE EQUATIONS FIT (%)

Very large specimen D = 0.0534ln(dmax)+ 2.1568 99.56

Large specimen D = 0.039ln(dmax)+ 2.1743 99.18

Medium specimen D = 0.0361ln(dmax)+ 2.1671 98.68

Small specimen D = 0.0144ln(dmax) + 2.2075 99.99

Similarly, the plot between fractal dimension (from 2D fast Fourier transform method) and
t/dmax is shown in Figure 27 where t is the specimen thickness and d max is maximum aggregate
size.

48
 2.36

2.34

Fractal dimension, D 2.32

2.3 dmax=9.5
dmax=19
2.28 dmax=38

2.26

2.24
0 5 10 15 20 25
t/dmax

Figure 27: Plot between fractal dimension (from 2D Fast Fourier Transform method) and t/dmax

The plots obtained from both slit island method and 2D Fast Fourier Transform method follow
similar trend. From the comparison of fractal dimension obtained from both slit island method
and 2D fast Fourier transform method with the maximum aggregate size, dmax, it is observed
that, fractal dimension increases with increase in maximum aggregate size. The increase in
fractal dimension with increase in maximum size of aggregate can be attributed to the fact that
in specimens with larger size aggregates, bridging and other forms of crack face interaction
take place and cracks form a more tortuous path and the fractal dimension increases. It is
observed that, for the same maximum aggregate size, the fractal dimension increases as the
t/dmax ratio increases. This is attributed to that fact that the specimen becomes tougher (i.e., the
fracture crack path becomes more tortuous) with increase in size. An increase in fracture
toughness with an increase in specimen size is most probably associated with the ratio of
process zone and specimen size. The greater this ratio, the smaller the fracture resistance. Since
the process zone is relatively large in a small specimen, its fracture characteristic is lower. As
the specimen becomes bigger, the relative size of the process zone becomes smaller, and the
fracture characteristics rise.

49
4.6.2. CORRELATION BETWEEN FRACTAL DIMENSION AND FRACTURE
ENERGY AND CRITICAL ENERGY RELEASE RATE
For various specimens with different maximum aggregate sizes, the fracture energy was
computed as the area under splitting load–crack opening displacement envelopes. The average
values of fractal dimension and fracture energy for each type of the specimens are plotted in
Figure 28. It can be observed that the fractal dimension increases as the increase in fracture
energy. Fracture energy is a measure of fracture toughness and the above observation suggests
that higher fractal dimensions are associated with tougher fracture surfaces.

250
Average fracture energy, GF (J/m2)

200

150
Vey large specimen
Large specimen
100 Medium specimen
Small specimen
50

0 1

2.24 2.20 2.17 2.15

Fractal dimension, D

Figure 28: Plot between average fracture energy and fractal dimension

The fractal dimension is a measure of tortuosity of the crack pattern. More tortuous the
crack pattern, higher will be the fractal dimension. The coalescence of micro-cracks through a
higher tortuous path is always associated with higher energy dissipation. Hence, fracture energy
increases as the fractal dimension increases.

Similar trend was obtained between fractal dimension and critical energy release rate,
G1C. The average values of fractal dimension and critical energy release rate for each type of
the specimens are plotted in Figure 29. The critical energy release rate increases with increase

50
in fractal dimension. This is attributed to the fact that the fractal dimension increases with the
tortuosity of fracture path, so does the energy dissipation.

100
90
80
70
Average GIC (J/m2)

60 Vey large specimen

50 Large specimen
40 Medium specimen

30 Small specimen

20
10
0
2.24 2.20 1 2.17 2.15
Fractal dimension, D

Figure 29: Plot between critical energy release rate and fractal dimension

51
 CHAPTER SIX

CONCLUSIONS

Three point bending tests are done on concrete beam specimens to determine fracture properties
of concrete. Fracture energy is computed from the load-deflection graphs. The following
conclusions are made from the study:

 From the experimental results, the fracture energy, GF is observed to be increasing as the
compressive strength increases. There exists a clear correlation between the two.

 The trend of GF growth with compressive strength appears to be monotonous and

probably extends to the larger compressive strengths. However, extrapolation is a
dangerous proposition. It should be examined experimentally.

 The fracture energy increases as the flexure strength of concrete increases. The
correlation between the modulus of rupture and fracture energy is expressed in terms of a
polynomial function of second degree though the extrapolation of the same should be
experimentally validated.

From the experimental work conducted by Issa et al (2000,2003), appropriate data needed for
the study of fractals in concrete including fractal dimension, fracture energy, critical energy
release rate and critical stress intensity factor are analyzed and the following conclusions are
made:

 It is observed that the fracture energy, GF and stress intensity factor, KI measured on
larger specimens are higher than that obtained from smaller specimens for all maximum
aggregate size. It was also observed that for a fixed specimen size, the fracture energy G F
and stress intensity factor, KI increase with the maximum aggregate size. There exists
good correlation between GF and KI with maximum aggregate size.
 For the same specimen size, critical stress intensity factor, KIC and critical energy release
rate, GIC increase as the maximum aggregate size increases. Also, for the same maximum
aggregate size, KIC and GIC increase with the specimen size. The increase in fracture
toughness with increasing maximum aggregate size is associated with the increasing
resistance encountered by the propagating crack.

52
 The fracture surfaces obtained from the wedge splitting tests are indeed fractals in nature
with fractal dimensions ranging from 2.15 to 2.29 as obtained from the slit island method.
The values of fractal dimensions obtained from 2D Fast Fourier Transform method is
slightly higher than that obtained from slit island method and are ranging from 2.24 to
2.39.
 As the specimen size and maximum aggregate size increase, the fractal dimension also
increases. A good correlation is obtained between maximum aggregate size and fractal
dimension.
 Fracture energy and critical energy release rate increase as the fractal dimension
increases. This indicates that the fractal characteristics of fracture surface are related to
the fracture properties of concrete.

53
 FUTURE SCOPE OF THE STUDY

 The fractal dimensions of the beam specimens can be determined by different methods,
say, box counting method, slit island method etc.

 The crack mouth opening displacement (CMOD) can also be measured while
conducting the flexure test on the beam specimens. CMOD can be used to obtain
different fracture properties of concrete.

 Fractal dimensions can be correlated to different fracture properties of concrete.

 This study can be extended to beam specimens of different sizes.

54
 REFERENCES

1. Amparanoa F. E., Xi Y, Roh Y. (2000), “Experimental study on the effect of aggregate

content on fracture behavior of concrete.”Engineering Fracture Mechanics 67(1), 65–84.
2. Beygi M. H. A., Kazemi M. T., Nikbin I. M., Amiri J. V. (2013), “The effect of water to
cement ratio on fracture parameters and brittleness of self-compacting concrete.”
Materials and Design 50, 267–276.
3. Cao M., Ren Q. (2006), “Fractal Behavior of Concrete Crack and Its application to
Damage Assessment.” Key Engineering Materials 312, 325-332.
4. Carpinteri A., Ferrara G., Imperato L. (1994), “Scaling laws for strength and toughness of
disordered materials: a unified theory based on fractal geometry.” Engineering Fracture
Mechanics 48(5), 673-689.
5. Charkaluk E., Bigerelle M., Iost, A. (1998). “Fractals and fracture.” Engineering
Fracture Mechanics 61(1), 119-139.
6. Fayyad T. M., Less J. M. (2017), “Experimental investigation of crack propagation and
crack branching in lightly reinforced concrete beams using digital image correlation.”
Engineering Fracture Mechanics 182, 487–505.
7. Hammad A.M., Issa M.A. (1994), “Fractal dimension as a measure of roughness of
concrete fracture trajectories.” Advanced Cement Based Materials 1(4), 169-177.
8. Hong-quan S, Jun D, Jian G, Dong – liang F. (2011), “Fractal research on cracks of
reinforced concrete beams with different aggregate sizes.” Advanced Materials Research
250-253, 1818-1822.
9. Hong-quan S, Jun D. (2011), “The Comparison of Fractal Dimensions of Cracks on
Reinforced Concrete Beam”. Advanced Materials Research 291-294, 1126-1130.
10. Issa M.A., Hammad A.M. (1993), “Fractal characterization of fracture surfaces in
mortar.” Cement and Concrete research 23(1), 7-12.
11. Issa M.A., Hammad A.M. (1993), “Correlation between crack tortuosity and fracture
toughness in cementitious material.” International Journal of Fracture 60, 97-105.
12. Issa M.A., Hammad A.M. (1994), “Assessment and evaluation of fractal dimension of
concrete fracture surface digitized images.” Cement and Concrete Research 24(2), 325-
334.
13. Issa Mohsen A., Issa Mahmoud A., Islam Md. S., Chudnovsky, A. (2000), “Size effects

55
 in concrete fracture: Part I, experimental setup and observations.” International Journal
of Fracture 102(1), 1–24.
14. Issa Mohsen A., Issa Mahmoud A., Islam Md. S., Chudnovsky, A. (2000), “Size effects
in concrete fracture: Part II: Analysis of test results.” International Journal of Fracture
102(1), 25-42.
15. Issa Mohsen A., Issa Mahmoud A., Islam Md. S., Chudnovsky, A. (2003), “Fractal
dimension––a measure of fracture roughness and toughness of concrete.” Engineering
Fracture Mechanics 70(1), 125–137.
16. Jin X, Li B, Tian Y, Jin N, Duan A. (2013), “Study on Fractal Characteristics of Cracks
and Pore Structure of Concrete based on Digital Image Technology.” Research Journal of
Applied Sciences, Engineering and Technology 5(11), 3165-3171.
17. Karamloo M, Mazloom M, Payganeh G. (2016), “Influences of water to cement ratio on
brittleness and fracture parameters of self-compacting lightweight concrete.” Engineering
Fracture Mechanics 168(1), 227–241.
18. Konkol J, Prokopski G. (2011). “The use of fractal geometry for the assessment of the
diversification of macro-pores in concrete.” Image Analysis & Stereology 30(2), 89-100.
19. Naik Partha Uday. (2017). “Experimental Determination of Fracture Energy by RILEM
Method.” The International Journal of Engineering and Science (IJES) 6(3), 106-115.
20. Sadrmomtazia A, Lotfi-Omrana O, Nikbinb I. M. (2019), “On the fracture parameters of
heavy-weight magnetite concrete with different water-cement ratios through three
methods.” Engineering Fracture Mechanics 219, 106615.
21. Saouma V.E., Barton C.C., Gamaleldin N.A. (1990). “Fractal characterization of fracture
surfaces in concrete.” Engineering Fracture Mechanics 35(1-3), 47-53.
22. Saouma V.E., Broz J. J., Briihwiler E., Boggs H. L. (1991), “Effect of aggregate and
specimen size on fracture properties of dam concrete.” Journal of Materials in Civil
Engineering, 3(3), 204-218.
23. Saouma V.E., Barton C.C. (1994). “Fractals, fractures and size effects in concrete.”
Journal of Engineering Mechanics 120(4), 835-854.
24. Wang Y, Hu X, Liang L, Zhu W. (2016). “Determination of tensile strength and fracture
toughness of concrete using notched 3-p-b specimens.” Engineering Fracture Mechanics
160, 67–77.
25. Wei J, Cao L. F., Zeng Y. X, Dong R. X. (2010). “Study on surface cracks characteristics
56
 of reinforced concrete beam under three point loading.” International RILEM Conference
on Advances in Construction Materials, 594-601.
26. Yin Y, Qiao Y, Hu S. (2019), “Four-point bending tests for the fracture properties of
concrete.” Engineering Fracture Mechanics 211, 371–381.

57
 APPENDIX A

Table 17: Load – displacement values for beam M20- Specimen 1

M20- Specimen 1
Load (kN) Deflection (mm) Load (kN) Deflection (mm)
0 0 3.3 0.83
0.1 0.41 3.4 0.84
0.2 0.52 3.5 0.85
0.3 0.56 3.6 0.86
0.4 0.61 3.7 0.86
0.5 0.61 3.8 0.865
0.6 0.62 3.9 0.89
0.7 0.62 4 0.895
0.8 0.64 4.1 0.9
0.9 0.65 4.2 0.91
1 0.66 4.3 0.92
1.1 0.67 4.4 0.93
1.2 0.68 4.5 0.94
1.3 0.69 4.6 0.96
1.4 0.71 4.7 0.98
1.5 0.72 4.8 0.99
1.6 0.72 4.9 1
1.7 0.73 5 1.01
1.8 0.73 5.1 1.02
1.9 0.74 5.2 1.03
2 0.75 5.3 1.04
2.1 0.75 5.4 1.05
2.2 0.76 5.5 1.06
2.3 0.76 5.6 1.065
2.4 0.77 5.7 1.07
2.5 0.77 5.8 1.075
2.6 0.78 5.8 1.08
2.7 0.78 5.8 1.09
2.8 0.78 5.6 1.1
2.9 0.79 5.5 1.115
3 0.79 5.2 1.125
3.1 0.8 4.9 1.13
3.2 0.82 4.5 1.14

58
Table 18: Load – displacement values for beam M20- Specimen 2

M20- Specimen 2
Load (kN) Deflection (mm) Load (kN) Deflection (mm)
0 0 3.3 0.83
0.1 0.37 3.4 0.85
0.2 0.39 3.5 0.87
0.3 0.4 3.6 0.875
0.4 0.41 3.7 0.88
0.5 0.43 3.8 0.885
0.6 0.45 3.9 0.9
0.7 0.47 4 0.91
0.8 0.49 4.1 0.92
0.9 0.5 4.2 0.925
1 0.51 4.3 0.935
1.1 0.53 4.4 0.94
1.2 0.55 4.5 0.94
1.3 0.57 4.6 0.945
1.4 0.59 4.7 0.95
1.5 0.6 4.8 0.955
1.6 0.63 4.9 0.96
1.7 0.65 5 0.96
1.8 0.68 5.1 0.965
1.9 0.7 5.2 0.97
2 0.71 5.3 0.975
2.1 0.75 5.4 0.98
2.2 0.755 5.5 0.985
2.3 0.77 5.5 0.99
2.4 0.79 5.5 0.995
2.5 0.8 5.5 1
2.6 0.805 5.4 1.04
2.7 0.81 4.8 1.045
2.8 0.81 4.5 1.05
2.9 0.815 4.3 1.06
3 0.82 4.1 1.08
3.1 0.825 3.8 1.08
3.2 0.825

59
Table 19: Load – displacement values for beam M20- Specimen 3

M20- Specimen 3
Load (kN) Deflection (mm) Load (kN) Deflection (mm)
0 0 3.1 0.67
0.1 0.2 3.2 0.68
0.2 0.25 3.3 0.7
0.3 0.26 3.4 0.71
0.4 0.3 3.5 0.72
0.5 0.32 3.6 0.73
0.6 0.34 3.7 0.75
0.7 0.35 3.8 0.755
0.8 0.36 3.9 0.76
0.9 0.37 4 0.77
1 0.38 4.1 0.775
1.1 0.39 4.2 0.78
1.2 0.4 4.3 0.785
1.3 0.41 4.4 0.79
1.4 0.42 4.5 0.79
1.5 0.43 4.6 0.795
1.6 0.44 4.7 0.8
1.7 0.46 4.8 0.81
1.8 0.48 4.9 0.82
1.9 0.5 5 0.83
2 0.51 5.1 0.85
2.1 0.53 5.2 0.87
2.2 0.55 5.3 0.88
2.3 0.57 5.3 0.885
2.4 0.59 5.3 0.89
2.5 0.61 5.2 0.9
2.6 0.62 5 0.905
2.7 0.625 4.8 0.91
2.8 0.63 4.6 0.925
2.9 0.65 4.2 0.94
3 0.66 4 0.96

60
Table 20: Load – displacement values for beam M30- Specimen 1

M30- Specimen 1
Load (kN) Deflection (mm) Load (kN) Deflection (mm)
0 0 3.6 0.97
0.1 0.15 3.7 0.98
0.2 0.16 3.8 1
0.3 0.2 3.9 1.02
0.4 0.22 4 1.04
0.5 0.25 4.1 1.06
0.6 0.3 4.2 1.085
0.7 0.32 4.3 1.1
0.8 0.35 4.4 1.12
0.9 0.38 4.5 1.13
1 0.41 4.6 1.14
1.1 0.42 4.7 1.15
1.2 0.46 4.8 1.16
1.3 0.495 4.9 1.17
1.4 0.52 5 1.18
1.5 0.56 5.1 1.195
1.6 0.6 5.2 1.2
1.7 0.62 5.3 1.21
1.8 0.64 5.4 1.225
1.9 0.66 5.5 1.23
2 0.68 5.6 1.24
2.1 0.7 5.7 1.25
2.2 0.72 5.8 1.26
2.3 0.75 5.9 1.27
2.4 0.76 6 1.285
2.5 0.77 6.1 1.3
2.6 0.8 6.2 1.31
2.7 0.81 6.3 1.325
2.8 0.82 6.4 1.33
2.9 0.85 6.5 1.34
3 0.86 6.6 1.355
3.1 0.88 6.6 1.36
3.2 0.89 6.4 1.365
3.3 0.92 6.1 1.37
3.4 0.94 5.9 1.375
3.5 0.96

61
Table 21: Load – displacement values for beam M30- Specimen 2

M30- Specimen 2
Load (kN) Deflection (mm) Load (kN) Deflection (mm)
0 0 3.9 0.73
0.1 0.04 4 0.74
0.2 0.1 4.1 0.75
0.3 0.12 4.2 0.76
0.4 0.15 4.3 0.77
0.5 0.17 4.4 0.78
0.6 0.18 4.5 0.79
0.7 0.21 4.6 0.82
0.8 0.22 4.7 0.84
0.9 0.24 4.8 0.86
1 0.25 4.9 0.88
1.1 0.27 5 0.9
1.2 0.28 5.1 0.92
1.3 0.3 5.2 0.94
1.4 0.31 5.3 0.96
1.5 0.33 5.4 0.98
1.6 0.34 5.5 1
1.7 0.35 5.6 1.01
1.8 0.37 5.7 1.02
1.9 0.39 5.8 1.03
2 0.41 5.9 1.04
2.1 0.43 6 1.05
2.2 0.45 6.1 1.06
2.3 0.47 6.2 1.07
2.4 0.48 6.3 1.08
2.5 0.49 6.4 1.08
2.6 0.5 6.5 1.09
2.7 0.51 6.6 1.09
2.8 0.52 6.7 1.1
2.9 0.54 6.8 1.1
3 0.55 6.8 1.13
3.1 0.57 6.8 1.14
3.2 0.59 6.4 1.15
3.3 0.63 6.1 1.16
3.4 0.65 5.8 1.165
3.5 0.68 5.4 1.17
3.6 0.7 5.1 1.175
3.7 0.71 4.8 1.18
3.8 0.72
62
Table 22: Load – displacement values for beam M30- Specimen 3

M30- Specimen 3
Load (kN) Deflection (mm) Load (kN) Deflection (mm)
0 0 3.8 0.63
0.1 0.1 3.9 0.64
0.2 0.12 4 0.66
0.3 0.14 4.1 0.68
0.4 0.16 4.2 0.7
0.5 0.18 4.3 0.72
0.6 0.2 4.4 0.74
0.7 0.22 4.5 0.75
0.8 0.23 4.6 0.77
0.9 0.24 4.7 0.79
1 0.26 4.8 0.81
1.1 0.28 4.9 0.83
1.2 0.3 5 0.84
1.3 0.31 5.1 0.85
1.4 0.32 5.2 0.87
1.5 0.34 5.3 0.89
1.6 0.345 5.4 0.9
1.7 0.36 5.5 0.92
1.8 0.38 5.6 0.94
1.9 0.385 5.7 0.96
2 0.4 5.8 0.98
2.1 0.42 5.9 1
2.2 0.43 6 1.02
2.3 0.44 6.1 1.03
2.4 0.45 6.2 1.05
2.5 0.46 6.3 1.07
2.6 0.475 6.4 1.09
2.7 0.49 6.5 1.1
2.8 0.5 6.6 1.11
2.9 0.505 6.7 1.12
3 0.51 6.8 1.13
3.1 0.52 6.9 1.14
3.2 0.53 7 1.16
3.3 0.54 6.5 1.18
3.4 0.56 6.2 1.2
3.5 0.58 5.8 1.21
3.6 0.6 5.6 1.22
3.7 0.62

63
Table 23: Load – displacement values for beam M40- Specimen 1

M40- Specimen 1
Load (kN) Deflection (mm) Load (kN) Deflection (mm)
0 0 3.9 0.79
0.1 0.09 4 0.81
0.2 0.1 4.1 0.83
0.3 0.14 4.2 0.85
0.4 0.16 4.3 0.86
0.5 0.18 4.4 0.88
0.6 0.2 4.5 0.89
0.7 0.23 4.6 0.9
0.8 0.26 4.7 0.91
0.9 0.29 4.8 0.92
1 0.32 4.9 0.93
1.1 0.34 5 0.94
1.2 0.36 5.1 0.95
1.3 0.38 5.2 0.96
1.4 0.4 5.3 0.97
1.5 0.42 5.4 0.98
1.6 0.44 5.5 0.99
1.7 0.46 5.6 1
1.8 0.48 5.7 1.01
1.9 0.49 5.8 1.03
2 0.5 5.9 1.06
2.1 0.51 6 1.08
2.2 0.52 6.1 1.1
2.3 0.54 6.2 1.12
2.4 0.56 6.3 1.14
2.5 0.58 6.4 1.16
2.6 0.6 6.5 1.18
2.7 0.62 6.6 1.2
2.8 0.64 6.7 1.22
2.9 0.68 6.8 1.24
3 0.69 6.9 1.26
3.1 0.7 7 1.28
3.2 0.71 7.1 1.3
3.3 0.72 7.2 1.32
3.4 0.73 7.3 1.34
3.5 0.74 7.4 1.36
3.6 0.75 7.5 1.38
3.7 0.76 7.6 1.4
3.8 0.77
64
 Table 24: Load – displacement values for beam M40- Specimen 2

M40- Specimen 2
Load Deflection Load Deflection Load Deflection
(kN) (mm) (kN) (mm) (kN) (mm)
0 0 2.7 0.62 5.4 0.98
0.1 0.1 2.8 0.63 5.5 0.99
0.2 0.11 2.9 0.65 5.6 1
0.3 0.13 3 0.67 5.7 1.01
0.4 0.15 3.1 0.69 5.8 1.03
0.5 0.17 3.2 0.71 5.9 1.06
0.6 0.21 3.3 0.72 6 1.08
0.7 0.25 3.4 0.73 6.1 1.1
0.8 0.27 3.5 0.74 6.2 1.12
0.9 0.29 3.6 0.75 6.3 1.14
1 0.31 3.7 0.76 6.4 1.16
1.1 0.33 3.8 0.77 6.5 1.18
1.2 0.35 3.9 0.79 6.6 1.2
1.3 0.37 4 0.81 6.7 1.22
1.4 0.38 4.1 0.83 6.8 1.24
1.5 0.4 4.2 0.85 6.9 1.26
1.6 0.42 4.3 0.86 7 1.28
1.7 0.44 4.4 0.88 7.1 1.29
1.8 0.45 4.5 0.89 7.2 1.3
1.9 0.46 4.6 0.9 7.3 1.31
2 0.48 4.7 0.91 7.4 1.32
2.1 0.5 4.8 0.92 7.4 1.34
2.2 0.52 4.9 0.93 7.4 1.35
2.3 0.54 5 0.94 7.1 1.36
2.4 0.56 5.1 0.95 6.8 1.37
2.5 0.58 5.2 0.96 6.4 1.38
2.6 0.6 5.3 0.97

65
Table 25: Load – displacement values for beam M40- Specimen 3

M40- Specimen 3
Load (kN) Deflection (mm) Load (kN) Deflection (mm)
0 0 3.8 0.77
0.1 0.13 3.9 0.79
0.2 0.15 4 0.81
0.3 0.17 4.1 0.83
0.4 0.2 4.2 0.85
0.5 0.23 4.3 0.86
0.6 0.26 4.4 0.88
0.7 0.28 4.5 0.89
0.8 0.3 4.6 0.9
0.9 0.31 4.7 0.91
1 0.33 4.8 0.92
1.1 0.35 4.9 0.93
1.2 0.36 5 0.94
1.3 0.37 5.1 0.95
1.4 0.4 5.2 0.96
1.5 0.42 5.3 0.97
1.6 0.425 5.4 0.98
1.7 0.435 5.5 0.99
1.8 0.45 5.6 1
1.9 0.48 5.7 1.01
2 0.5 5.8 1.03
2.1 0.51 5.9 1.06
2.2 0.52 6 1.08
2.3 0.53 6.1 1.1
2.4 0.54 6.2 1.12
2.5 0.57 6.3 1.14
2.6 0.585 6.4 1.16
2.7 0.6 6.5 1.18
2.8 0.61 6.6 1.2
2.9 0.625 6.7 1.22
3 0.64 6.8 1.24
3.1 0.66 6.9 1.26
3.2 0.68 7 1.28
3.3 0.71 7.1 1.29
3.4 0.72 7.2 1.3
3.5 0.74 7.3 1.32
3.6 0.75 7.3 1.35
3.7 0.76 7.2 1.36

66
 Table 26: Load – displacement values for beam M50- Specimen 1

M50- Specimen 1
Load Deflection Load Deflection Load Deflection
(kN) (mm) (kN) (mm) (kN) (mm)
0 0 2.9 0.58 5.8 1.04
0.1 0.11 3 0.61 5.9 1.04
0.2 0.14 3.1 0.62 6 1.05
0.3 0.16 3.2 0.65 6.1 1.06
0.4 0.18 3.3 0.68 6.2 1.07
0.5 0.2 3.4 0.7 6.3 1.08
0.6 0.21 3.5 0.71 6.4 1.09
0.7 0.22 3.6 0.72 6.5 1.11
0.8 0.24 3.7 0.73 6.6 1.12
0.9 0.26 3.8 0.74 6.7 1.13
1 0.28 3.9 0.75 6.8 1.14
1.1 0.29 4 0.76 6.9 1.15
1.2 0.3 4.1 0.77 7 1.16
1.3 0.32 4.2 0.78 7.1 1.17
1.4 0.34 4.3 0.79 7.2 1.18
1.5 0.36 4.4 0.8 7.3 1.2
1.6 0.38 4.5 0.82 7.4 1.22
1.7 0.4 4.6 0.86 7.5 1.23
1.8 0.41 4.7 0.88 7.6 1.25
1.9 0.43 4.8 0.9 7.7 1.27
2 0.44 4.9 0.92 7.8 1.29
2.1 0.46 5 0.94 7.9 1.3
2.2 0.48 5.1 0.96 7.9 1.32
2.3 0.49 5.2 0.98 7.6 1.35
2.4 0.5 5.3 0.99 7.4 1.38
2.5 0.52 5.4 1 7 1.4
2.6 0.53 5.5 1.01 6.6 1.43
2.7 0.55 5.6 1.02 6.1 1.48
2.8 0.56 5.7 1.03

67
 Table 27: Load – displacement values for beam M50- Specimen 2

M50- Specimen 2
Load Deflection Load Deflection Load Deflection
(kN) (mm) (kN) (mm) (kN) (mm)
0 0 2.8 0.62 5.8 1.11
0.1 0.15 2.9 0.64 5.9 1.13
0.2 0.18 3 0.65 6 1.15
0.3 0.2 3.1 0.67 6.1 1.16
0.4 0.21 3.2 0.69 6.2 1.18
0.5 0.23 3.3 0.72 6.3 1.2
0.6 0.24 3.4 0.74 6.4 1.22
0.7 0.26 3.5 0.76 6.5 1.24
0.8 0.28 3.6 0.79 6.6 1.26
0.9 0.3 3.7 0.81 6.7 1.27
1 0.32 3.8 0.82 6.8 1.275
1.1 0.325 3.9 0.83 6.9 1.29
1.2 0.34 4 0.85 7 1.3
1.3 0.36 4.1 0.87 7.1 1.32
1.4 0.365 4.2 0.88 7.2 1.34
1.5 0.375 4.3 0.89 7.3 1.36
1.6 0.395 4.4 0.91 7.4 1.37
1.7 0.4 4.5 0.93 7.5 1.38
1.8 0.42 4.6 0.94 7.6 1.39
1.9 0.44 4.7 0.96 7.7 1.4
2 0.46 4.8 0.97 7.8 1.42
2.1 0.48 4.9 0.98 7.9 1.44
2.2 0.5 5 0.99 8 1.47
2.3 0.52 5.1 1 7.7 1.49
2.4 0.54 5.2 1.01 7.4 1.52
2.5 0.56 5.3 1.03 7 1.55
2.6 0.58 5.4 1.04 6.8 1.58
2.7 0.6 5.7 1.09

68
 Table 28: Load – displacement values for beam M50- Specimen 3

M50- Specimen 3
Load Deflection Load Deflection Load Deflection
(kN) (mm) (kN) (mm) (kN) (mm)
0 0 2.8 0.76 5.6 1.16
0.1 0.2 2.9 0.78 5.7 1.18
0.2 0.22 3 0.8 5.8 1.2
0.3 0.24 3.1 0.83 5.9 1.21
0.4 0.26 3.2 0.84 6 1.22
0.5 0.28 3.3 0.85 6.1 1.23
0.6 0.3 3.4 0.86 6.2 1.24
0.7 0.32 3.5 0.87 6.3 1.25
0.8 0.34 3.6 0.88 6.4 1.26
0.9 0.36 3.7 0.89 6.5 1.28
1 0.37 3.8 0.9 6.6 1.29
1.1 0.38 3.9 0.92 6.7 1.3
1.2 0.39 4 0.94 6.8 1.31
1.3 0.4 4.1 0.95 6.9 1.32
1.4 0.43 4.2 0.96 7 1.34
1.5 0.46 4.3 0.98 7.1 1.36
1.6 0.49 4.4 1 7.2 1.38
1.7 0.52 4.5 1.01 7.3 1.4
1.8 0.54 4.6 1.03 7.4 1.42
1.9 0.57 4.7 1.04 7.5 1.44
2 0.6 4.8 1.05 7.6 1.46
2.1 0.62 4.9 1.07 7.7 1.48
2.2 0.64 5 1.09 7.8 1.47
2.3 0.66 5.1 1.1 7.9 1.49
2.4 0.68 5.2 1.11 8 1.5
2.5 0.7 5.3 1.12 8.1 1.51
2.6 0.72 5.4 1.13 8.2 1.52
2.7 0.74 5.5 1.14

69