Particulate Science and Technology

An International Journal

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/upst20

Fractal dimension of crushing products: effects of

feed size distribution and feed rate

Iman Moradi & Mehdi Irannajad

To cite this article: Iman Moradi & Mehdi Irannajad (2020): Fractal dimension of crushing
products: effects of feed size distribution and feed rate, Particulate Science and Technology, DOI:
10.1080/02726351.2020.1856248

To link to this article: https://doi.org/10.1080/02726351.2020.1856248

Published online: 21 Dec 2020.

Submit your article to this journal

Article views: 33

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=upst20
PARTICULATE SCIENCE AND TECHNOLOGY
https://doi.org/10.1080/02726351.2020.1856248

Fractal dimension of crushing products: effects of feed size distribution and

feed rate
Iman Moradi and Mehdi Irannajad
Department of Mining and Metallurgical Engineering, Amirkabir University of Technology, Tehran, Iran

ABSTRACT KEYWORDS
In mineral crushing operations, the feed rate and feed size distribution affect the crushing rate Fractal dimension; particle
and particle size distribution of the crushed products. In the present study, copper ore feed size distribution; feed size
charges of different size distributions were prepared. These feed charges were fed to the jaw, distribution; crushing;
feed rate
cone, and roll crushers at different feed rates in the laboratory scales. The effects of these parame-
ters on the crushed particle size distributions and crushing rates were investigated and described
using the fractal geometry by the calculation of the product fractal dimensions. The fractal dimen-
sion was calculated based on a proportional relation between the size and weight of the particles.
The obtained fractal dimensions of the particle size distributions for the particles crushed by the
jaw, cone, and roll crushers were in the range of (2.16–2.38), (2.13–2.31), and (2.23–2.46) respect-
ively. In all of three types of crushers, the fractal dimensions and crushing rate of the ore were
increased with an increase in feeding feed size distribution and by increasing the amount of par-
ticles with sizes bigger than the setting size. Increasing the feed rate at the permitted capacity
had an inverse relation with the fractal dimension and led to a reduction in both the fractal
dimension and the crushing rate. However, an increase in the feed rate at the unpermitted cap-
acity, having a direct relationship to the fractal dimension, was resulted an increase in crushing
rate and fractal dimension of the ore due to the choke feeding.

1. Introduction networks have been used to study the size distribution of

particles obtained from the crushing operation (Hamzelo,
The feed rate and feed size distribution are considered as
Masinaei, and Mehrshad 2013). Besides, it has been shown
the important parameters of crushing operations in mineral
that the fractal functions in double-logarithmic coordinates
processing, the control of which facilitates the control and
can be also used for describing the crushed particle size dis-
establishment of the crushing operation (Sadrai et al. 2006;
tribution (Ahmed and Drrzymala 2005). The fractal geom-
Airikka 2015). The comminution operations usually account
etry is used as a novel method for linearization of the
for 30–50% of the mineral processing costs (Radziszewski
2000). In recent decades, it has been attempted to reduce particle size distribution curve (Lu et al. 2003; Casini,
the operational costs for the extraction of valuable minerals. Viggiani, and Springman 2013). Combining this novel
The comminution is an expensive operation entails nearly method with previous methods can yield an effective
3% of the total energy consumed worldwide (Sadrai et al. approach for determining the morphology of crushed par-
2011). The objective of crushing operations is to achieve the ticles and the mechanical characteristics of the comminution
desired feed size distribution for the next stages. Therefore, operations (Wang et al. 2015). The fractal dimension is a
it is necessary to take into consideration any parameter that good indicator for the evaluation of the crushing operation
can satisfy the intended particle size distribution by consum- since it yields a better interpretation of the crushing mech-
ing the least possible energy (Fuerstenau, Lutch, and anism, which has provided a theoretical basis for controlling
De 1999). the crushed particle size distribution (Cai, Xiong, and Lin
The amount of the crushed material after the crushing 2015). The use of a theoretical model in a crushing oper-
process can be examined by the size distribution of the par- ation on a material would make it possible to convert the
ticles under-study. To describe the size distribution of crushing probability distributions into a real simulated
crushed particles, Allen has presented statistical methods model of the operation (Delagrammatikas, Delagrammatikas,
such as Rosin-Rammler, Gates-Gaudin-Schuhmann, normal and Tsimas 2007). Through investigating a large number of
distribution, and so on. The logarithmic distributions are the samples and describing their particle size distribution using
most commonly used statistical methods used for the the fractal dimension, it was found out that the fractal
crushed particle size distribution (Allen 1997). Also, meth- dimension parameters featured good spatial structure and
ods such as image processing, machine vision, and neural were capable to describe spatial changes in the texture of the

CONTACT Mehdi Irannajad [email protected] Department of Mining and Metallurgical Engineering, Amirkabir University of Technology, Tehran, Iran.
ß 2020 Taylor & Francis Group, LLC
2 I. MORADI AND M. IRANNAJAD

studied material in the best possible way (Mohammadi et al. Filippov et al. 1999). In a study on samples with different
2019). Since the Euclidean geometry is unable to provide lithography, Zhang et al. showed that the size distribution of
precise description and measurement of the size and dimen- the crushed particles of these samples exhibited a fractal
sions of the objects and the irregular and complex natural behavior. The experiments were made under varied axial
phenomena, Mandelbrot (1982) proposed a fractal geometry stress conditions by a loading equipment of electro-
for describing and measuring the size and dimensions of hydraulic servo-motor test system (Zhang et al. 2017).
these phenomena and objects. The fractal objects and phe- Accordingly, as the pressure exerted on the samples was
nomena are made of small constituents so that each of these increased, the dispersion of the particle size distribution
small constituents resembles the whole set. Besides, these and, consequently, the fractal dimension increased propor-
fractal phenomena and objects have a constant (fixed) pat- tionally (Zhang et al. 2017). In another study (Tasdemir
tern, which is iterated in all the phases with a constant scale. 2009), five types of chromite ores underwent comminution
Such iterative pattern variations from parts to the whole with jaw, cone, and impact crushers as well as ball mill. This
(known as self-similarity) can be assessed by measuring the was followed by the investigation of the size distribution of
fractal dimension of an object or a phenomenon using vari- the particles resulted from the comminution process. Then,
ous methods including topological dimensions, Hausdorff, the fractal dimension of the crushed particles was obtained
self-similarity, capacity, information, correlation, and based on Turcotte’s method. It was found that different
Lyapunov. The topological dimension means that transfor- chromite types exhibited natural fractal behavior both in the
mations of a geometric object would cause no change in its case of the single-fractal (uni-fractal) and the multi-fractal
correlation. Concerning the topological dimension, the methods (Tasdemir 2009). An investigation of two types of
dimensions of an object or phenomenon are always ores, namely quartz and marble, in different crushing mech-
expressed by integers. The Hausdorff dimension is always anisms, using the linear regression method showed that each
defined with a scale, which divides small constituents in of these crushing mechanisms had its own specific and
such a way that their aggregation equals the whole phenom- unique domain, which was a function of the crushing
enon or object. Using this scale, a numerical value can be energy. Furthermore, the particle size of the input load and
assigned to the whole object or phenomenon. A phenom- the ore type significantly affected the particle size distribu-
enon or object is called fractal when its Hausdorff dimen- tion (Petrakis, Stamboliadis, and Komnitsas 2017).
sion is several times larger than its topological dimension. The particles are crushed into two forms of free and
The Hausdorff dimension has often a fractional concept, choke. This is termed arrested or free crushing as opposed
which is known as the fractal dimension (Mandelbrot 1982; to choked crushing, which occurs when the volume of a
Falconer 2003; Arhangel’skii and van Mill 2018). In material arriving at a particular cross section is greater than
Euclidean geometry, the point, line, surface, and volume that at leaving. In arrested crushing, the crushing is by the
have dimensions of 0, 1, 2, and 3, respectively; in fractal jaws only, whereas in choked crushing, the particles break
geometry, on the other hand, dimensions of the objects and another one. This inter-particle comminution can lead to
phenomena can take not only integers but also rational excessive production of fines, and if choking is severe, it can
quantities. Bartoli et al. (1991) showed that the dimension damage the crusher (Wills et al. 2016). In the roll crusher,
variations in the fractal geometry occur within the range of the choke feeding is not suitable as it tends to produce par-
0–3 so that a dimension greater than 3 is physically impos- ticles of irregular size (Gupta and Yan 2016). The concept
sible to occur for a phenomenon (Bartoli et al. 1991). of choke is very important in cone crushers because once
Calculating the fractal dimension requires establishing a the volume of a material in the chamber exceeds the min-
power-law relationship of a fractal distribution. The power imum sectional volume, it will build up above the choke
in this equation is indeed the fractal dimension. A fractal point. In this scenario, the choke point then controls the
dimension is a non-negative number that can take either flow of the material and ensures an orderly progression of
integer or rational values (Mandelbrot 1982). the material through the chamber, thus ensuring that the
A power equation can be obtained between the energy material passing through the chamber is exposed to an
consumption rate and the mineral crushing rate. appropriate number of breakage events. Alternatively, if a
Subsequently, the fractal dimension of crushed particles can cone crusher is being fed at rates lower than that dictated
be derived from this power equation. The crushing energy- by the choke point, then the rocks are free to rush through
relevant fractal dimension was shown to vary between 2 and the chamber and undergo a less-than-optimal number of
3 (Carpinteri and Pugno 2002). By investigating the con- breakage events. It is well documented that choke feeding
struction debris that remained from the destruction of maximizes throughput and size reduction (Evertsson 2000;
buildings by a vibratory hammer, Zhu calculated the fractal Elias, Chaffer, and Moya 2015; Dunne, Kawatra, and
dimension of the crushed particle size distribution and Young 2019).
obtained values ranging between 2.6 and 2.65. However, in It was found that the next stages of crushing and the
different tests at different frequencies of the vibratory mineral liberation degree of the ore should be predicted and
hammer, the fractal dimension changed from 2 to 2.9 controlled using the fractal dimension of the size distribu-
(Xu 2018). As the particles of material become finer, on the tion of the particles resulted from crushing (Cui, An, and
one hand, the energy consumption and the fractal dimen- Gong 2006). In order to calculate the fractal dimension of
sion of that mineral increase as well (Irannajad 1998; the particle size distribution for the ores with fractal
 PARTICULATE SCIENCE AND TECHNOLOGY 3

behavior, Turcott’s method is commonly used based on two point. According to this figure, the particle size of the
parameters of the weight and sizes of crushed particles studied ore sample was (50 þ 16) mm.
(Turcotte 1986). Variations in the fractal dimension of the
crushed particle size distribution were reported to depend
2.3. Mineralogical analysis
on the ore type, the amount of consumed energy, and
crushing mechanism (Blenkinsop 1991). An investigation on Regarding the fact that the present work aimed to investi-
the environmental materials revealed that these materials gate the fractal behavior of the particles, it was needed to
had fractal behavior so that their fractal dimension varied firstly know the minerals contained in the ore and, secondly,
within the range of 2.25–3.125. It was also shown that a the morphology of the particles (including particle sizes,
non-uniform particle size distribution affected significantly porosity, texture, particle size distribution, and so on). For
the obtained fractal dimensions (Zhang, Hu, and Liu 2004). this reason, the mineralogical analyses were accomplished by
In another study, multiple samples were collected from two methods of X-ray diffraction (XRD) and ore micro-
regions destructed by flood, and then the particle size distri- scopic examinations. The XRD results are presented in
bution of these samples was investigated via their fractal Figure 2 and Table 1. As indicated, the minerals constituting
dimension, which dimensions mostly ranged between 2.250 the ore sample can be sorted, based on their frequency, as
quartz, microcline, kaolinite, talc, and chlorite. Moreover,
and 2.798 (Zhong et al. 2019). The effects of ore type and
the XRD analysis ignores the quantity of the minerals that
feed size distribution on fractal dimensions of cone crusher
are, in terms of frequency, below a certain value (e.g., 3%).
products were studied by Moradi, Mohammadzadeh, and
The morphological studies were performed by preparing
Irannajad (2019). The fractal dimension for chromite ore
some polished and thin sections on the ore sample. The
was larger than that of the ilmenite ore and increasing the mineralogical results derived from the microscopic examina-
percentage of materials greater than the setting size of the tions were completely consistent with the XRD results.
crusher, produced more fine particles, resulting in increased Moreover, these examinations revealed that the ore sample
fractal dimensions (Moradi, Mohammadzadeh, and contains chalcopyrite, pyrite, and oligiste, in addition to the
Irannajad 2019). minerals presented in Table 1.
The results of some researches have been shown that the
comminution products does not always have fractal behav-
ior, which is called non-fractal size distribution. Results of 2.4. Chemical analysis
comminution for different materials revealed that some mill- The chemical composition of the ore sample was determined
ing materials usually had non-fractal behavior. Natural using the X-ray fluorescence (XRF) method and the results
examples of fragmentation always lack an infinite number of are provided in Table 2. The results indicate that among the
particles predicted by a true fractal array, but are commonly constituents of the sample, the highest concentrations belong
linear, with D 2
1.5 (corresponding to D 3
2.5) over an to SiO2, Al2O3, and Fe2O3, resulting from quartz, silicate
order of magnitude or more. Extensively fragmented rocks minerals, and pyrite, respectively. According to the mineral-
and minerals typically reach a comminution limit on the ogical analysis, chalcopyrite is a valuable mineral in the sam-
order of 1 lm, below which further brittle fracture could be ple. Besides, regarding the results of the chemical analysis
mechanically unfavorable (Stamboliadis, Petrakis, and indicating a CuO content of 0.64%, the stoichiometry calcu-
Pantelaki 2011; Glazner and Mills 2012). lations showed a content of 1.5% for the chalcopyrite grade
The present work aims to investigate the effect of the in the sample.
selected feed size distribution and different feed rates on the
performance of the jaw, cone, and roll crushers by the frac-
2.5. Particle size analysis
tal geometry. The investigated sample in the present work is
chalcopyrite ore obtained from Ghalezari mine, Khorasan The particle size analysis of the ore sample is presented in
province, Iran. Figure 3, in which a d80 of about 37,000 microns is obtained
for the ore sample.

2. Materials and methods

2.6. Research methodology
2.1. Materials
2.6.1. Arrangement of crushing operations
For the sample identification, the steps of sample prepar- The size distribution of the particles yielded by crushing
ation and mineralogical, chemical, and particle size analyses may be influenced by several parameters such as feed rate,
were examined on the ore sample. crushing mechanism, morphology of the ore pieces, feed
size distribution, and so on. Any change in each of these
2.2. Sample preparation parameters, and depending on the behavior of the ore in the
crushing stages, the particle size distribution would exhibit
The procedures of sample preparation for the present work, different distributions. Each crusher operates with a domin-
from Ghalezari copper ore mine, are shown in Figure 1. The ant mechanism (such as compression, impact, shear, and
ore sample for this study was obtained from the mixing abrasion). In the present study, the three crushers had the
4 I. MORADI AND M. IRANNAJAD

Figure 1. Partial flowsheet of mineral processing plant of Ghalezari copper ore.

Figure 2. Results of mineralogical analysis by XRD: (Quartz (Q), Microkline (M), Kaolinite (K), Talc (T) and Chlorite (Cl)).
 PARTICULATE SCIENCE AND TECHNOLOGY 5

Table 1. Mineralogical analysis of ore sample by XRD. Table 3. Weight and size distribution composition of jaw, cone and roll
Mineral Chemical formula crushers feeds.
Quartz SiO2 Size range (mm)
Kaolinite Al4[(OH)8Si4O10] Crusher Size fraction A Size fraction B Size fraction C
Microkline KalSi3O8
Talc Mg3(OH)2Si4O10 Jaw þ25.4 þ16  25.4 þ12.7  16
Chlorite (Mg, FeAl)6 (OH)8[Si4O10] Cone þ12.7  16 þ9.52  12.7 þ4.76  9.52
Roll þ4.76  9.52 þ4  4.76 þ2.38  4

Table 2. Chemical analysis of ore sample by XRF.

Element K2O S SiO2 Al2O3 MgO Table 4. Weight % of each size fraction in different feeds.
Percentage 2.63 2.69 62.7 10.6 2.20 Weight % in the feed
Element Other L.I.O CaO CuO Fe2O3 Feed Size fraction A Size fraction B Size fraction C
Percentage 1.35 5.18 3.61 0.64 8.4
1 33.33 33.33 33.33
2 40 40 20
3 40 20 40
100 4 20 40 40
90
Cumulative Weight Passing (%)

80
investigate the effect of particle sizes close to and larger than the
70
setting size of the cone crusher, it was fed with different weight
60
and size distribution compositions (Table 4).
50
As shown in Table 3, the roll crusher was fed with three
40
30
fractions of (9.52 þ 4.76), (4.76 þ 4), and (4 þ 2.38)
20
mm. The size of the setting of this crusher was 3 mm. To
10 investigate the effect of the particle sizes closer and larger
0 than the size of the setting on the PSD of the crushed prod-
1 10 100 1000 10000 100000 ucts, the roll crusher was fed with different weight and size
Size (micron) distribution compositions (Table 4).
Figure 3. Particle size analysis of ore sample. The particle size analysis and the fractal dimensions were
calculated and described for each crushed product. The particle
same mechanism, and the compressive force was the domin- size was analyzed using screen series of the ASTM standard,
ant force for crushing the ore. Here, the selected feed size which include 16,000, 12,700, 9,520, 4,760, 2,380, 2,000, 1,000,
distribution and feed rates were investigated as the main 850, 600, 500, 425, 300, 200, 150, 106, 75, and 53 microns for
establishing parameters of the crushing operations. The the jaw crusher, from 4,760 to 53 microns for the cone crusher,
maximum feeding capacity was 3.5 kg/min for the three and from 2,380 to 53 microns for the roll crusher.
laboratory crushers. To investigate the effect of feed rates on
the performance of these three crushers, the input feed was 2.7. Fractal approach
carried out at permitted capacities (1, 1.5, 2, 2.5, 3, and
3.5 kg/min). In addition, the effect of feed rate at unpermit- The fractal geometry is one of the methods that can be used
ted capacities was investigate by testing the feed rates of 4 for linearizing and description of the PSD (Flook 1978). For
and 4.5 kg/min. All the jaw, cone, and roll crushers were particles having a uniform shape and density, a proportional
successively operated in the open-circuit modes and particle relation between the size and number of particles can be
size distribution (PSD) was determined for each crusher. obtained as described in Relation (1). In the fractal function,
The effect of the selected feed size distribution was investi- N is the number of particles greater than R, R is the size,
gated on the PSD of the products. The weight and size dis- and D is the fractal dimension of particles (Mandelbrot
tribution composition of the feeds to the jaw, cone, and roll 1982; Turcotte 1986):
crushers are presented in Tables 3–5, respectively. NaRD (1)
According to Table 3, the jaw crusher was fed with differ-
ent percentages of weight from three fractions, namely Indeed, the use of this relation, which is based on the
(þ25.4), (25.4 þ 19), and (19 þ 16) mm. The setting size number and size of particles, is usually associated with
of this crusher was regulated to19 mm, thus d100 of output errors (Hyslip and Vallejo 1997). In order to resolve this
particles was 19 mm. According to Figure 1, the particle size deficiency and improve the relation, a proportional relation
of the ore sample was (50 þ 16) mm. was established between the size and weight of particles, as
To investigate the effect of the particle sizes closer and indicated in Relation (2) (Turcotte 1986; Tyler and
larger than the setting size on the PSD of the crushed prod- Wheatcraft 1992):
ucts, the jaw crusher was fed with different weight and size  3Df
MðKL > kÞ k
distribution compositions (Table 4). a , (2)
MT KL
According to Table 3, the cone crusher was fed with three
fractions of (16 þ 12.7), (12.7 þ 9.52), and (9.52 þ 4.76) where M(KL > k) is the cumulative weight of particles pass-
mm. The size of the setting of this crusher was 9.52 mm. To ing through each dimensional fraction, MT is the weight of
 6 I. MORADI AND M. IRANNAJAD

the total sample, k is the size of each dimensional fraction, 2 mm can enter the milling stage, a 2 mm screen was consid-
KL is the size of the largest dimensional fraction, and Df is ered as the target one. The amount of the material passing
the fractal dimension. through this screen was measured for each product to deter-
By converting the proportional Relation (2) into equality, mine the feed that produces the highest amount of material
Equation (3) is obtained as: for the next stage. The results obtained for the crushing
 3Df operations of jaw, cone, and roll crushers are presented sep-
MðKL > kÞ k
¼c , (3) arately in the following:
MT KL
where c is the equality constant. According to Equation (3),
3.1. Jaw crusher
instead of the number of particles, the cumulative weight of
particles passing from each fraction is used to calculate the To illustrate the results, only for the largest Df (feed 2 with
fractal dimension. By taking the Neperian logarithm (ln) of a feed rate of 4.5 kg/min), the middle Df (feed1 with a feed
both sides of Equation (3), Equation (4) is obtained as: rate of 1 kg/min), and the smallest Df (feed 4 with a feed
  rate of 3 kg/min), the crushed PSD diagrams for the prod-
MðKL > kÞ k
ln ¼ ð3  Df Þ ln þ ln c: (4) ucts are shown in Figure 4. The results of the other feeds
MT KL are indicated in Table 5. The fractal dimension was calcu-
Converting Equation (4) into a linear Equation yields lated using Equation (6), considering the slope line fitted at
Equation (5): the points obtained from different fractions. The correlation
y ¼ ax þ b, (5) coefficient of points in all of the experiments was greater
than R2 ¼ 0.95. The fractal dimensions obtained for the
where y ¼ln (M(KL > k))/MT, x ¼ ln(k/KL) , a ¼ 3-Df, and products of these feed rates with different feeds (Table 4)
b ¼ ln c (a constant number ¼ y-intercept) were equal to (2.24–2.31), (2.23–2.30), (2.21–2.28),
By considering Equations (4) and (5), the fractal dimen- (2.19–2.26), (2.16–2.23), (2.18–2.27), (2.26–2.33), and
sion is obtained from Equation (6): (2.29–2.38), respectively, as summarized in Table 5. These
Df ¼ 3  a (6) fractal dimensions conform to the axiom of fractal geometry
and indicate that the values for the PSD range between 2
From Equation (4), the PSD of crushed products can be and 3. As shown by these results, increasing the feed rate at
driven on a logarithmic-logarithmic scale. the permitted capacity of the crusher would result in a
Different studies on the fractal dimension calculation of reduction in the amount of the crushed material and the
PSD of crushed products have shown that the fractal dimen- fractal dimension due to the formation of a layer of the
sion varies in the range of 2–3. This implies that the energy material that prolongs its remaining time inside the crusher,
in the crushing operations is spent on creating further surfa- yielding a more uniform product. According to Figure 5,
ces on the particles (Df ¼ 2) and translating them to par-
ticles of reduced volumes (Df ¼ 3) (Turcotte 1992; Smallest Middle Largest
Carpinteri and Pugno 2003). −6 −5 −4 −3 −2 −1 0
0

3. Results and discussion −1

−2
Operations of comminution are very energy consuming and
ln(M/MT)

−3
aim to produce the feed required for the next grinding stage
while achieving the intended feed size distribution in a max- −4
imum possible amount of the material by spending a min- (Smallest) y = 0.84x - 0.99 −5
imum possible amount of energy. For this purpose, (Middle) y = 0.73x - 0.94 −6
parameters, such as feed rate, feed size distribution, and (Largest) y = 0.62x - 0.95
−7
mode of crushing (jaw, cone, and roll crushers), were inves- ln(k/KL)
tigated by crushing the ore sample as described in Section Figure 4. Particle size distributions for jaw crusher products of largest, middle,
2.1. Regarding the fact that only materials smaller than and smallest Df.

Table 5. Overall results of fractal dimensions and crushed material percentage passing thorough target screen for jaw crusher products.
1 2 3 4
Feed Feed rate (kg/min) Wt. % passing Df Wt. % passing Df Wt. % passing Df Wt. % passing Df
1 6.74 2.27 7.28 2.31 7.04 2.28 6.33 2.24
1.5 6.40 2.25 7.14 2.30 6.92 2.27 6.22 2.23
2 6.38 2.24 6.92 2.28 6.69 2.25 6.02 2.21
2.5 6.09 2.21 6.76 2.26 6.45 2.22 5.87 2.19
3 5.89 2.19 6.46 2.23 6.23 2.20 5.52 2.16
3.5 6.26 2.23 6.83 2.27 6.72 2.25 5.97 2.18
4 6.97 2.28 7.44 2.33 7.24 2.32 6.51 2.26
4.5 7.32 2.33 7.96 2.38 7.66 2.36 6.99 2.29
 PARTICULATE SCIENCE AND TECHNOLOGY 7

increasing the feed rate and reaching the maximum capacity 3.2. Cone crusher
of the crusher (i.e., 3.5 kg/min) will lead to increases in the
To illustrate the results, only for the largest Df (feed 2 with
material crushing rate and the fractal dimension of the
a feed rate of 4.5 kg/min), the middle Df (feed1 with a feed
crushed products. Furthermore, increasing the feed rate
rate of 1.5 kg/min), and the smallest Df (feed 4 with a feed
exceeding the permitted capacity of the crusher will result in
rate of 3 kg/min), the crushed PSD diagrams for the prod-
considerable increases in the material crushing rate and the
ucts are shown in Figure 6. The results of the other feeds
fractal dimension of the products; this can be due to the
are indicated in Table 6. The fractal dimension was calcu-
choking feeding, which leads to the production of finer par-
lated with regard to the slop line fitted at the points using
ticles. Based on the obtained results, therefore, there are
Equation (6). The correlation coefficient of points in all the
inverse and direct relationships between the feed rate and
experiments was greater than R2 ¼ 0.95. The calculated frac-
the fractal dimension at the permitted and unpermitted
tal dimensions for the products of the feed rates of 1, 1.5, 2,
capacities, respectively. Table 5 presents the amount of
2.5, 3, 3.5, 4, and 4.5 kg/min with different feeds were
material passing through the target screen. According to the
(2.21–2.27), (2.19–2.25), (2.17–2.22), (2.16–2.20),
results, increasing the feed rate would result in a decrease
(2.13–2.17), (2.16–2.21), (2.23–2.29), and (2.24–2.31),
and an increase in the amount of the material passing
respectively. The obtained fractal dimensions are in con-
through the target screen in the cases of permitted and
formity with the axiom of the fractal geometry, according to
unpermitted capacities, respectively. According to Table 5,
which the fractal dimension for the PSD is within the range
variations in the feed size distribution at different feed rates
of 2–3. All the results obtained for the cone crusher are pre-
revealed that an increase in the amount of the particles
sented in Table 6. Concerning the cone crusher, increasing
larger than the setting size leads to increased material crush-
the feed rate at the permitted capacity led to a reduction in
ing rate and the fractal dimension. These particles are
the material crushing rate and the fractal dimension similar
entangled between the jaws and the crushing forces are con-
to the jaw crusher. According to Figure 5, increasing the
verted from compression to abrasion and shear, leading to
feed rate and reaching the unpermitted capacity were associ-
the production of finer particles, thereby increasing the frac-
ated with increases in the material crushing rate and the
tal dimension. Thus, there is a direct relationship between
fractal dimension. In the case of changing the feed size dis-
the particle size and the fractal dimension.
tribution, an increase in the amount of the material with a
The decreased material crushing rate with an increase in
size larger than the setting size was associated with elevated
the feed rate at the permitted capacity may be due to the
material crushing rate and the fractal dimension at different
half-cycle, placement mode of the jaws, and prolonged
feed rates. The obtained results indicated that the fractal
remaining time of the material. Also, the reason for the
dimensions obtained for the products of the cone crusher
increased crushing rate of the material at the unpermitted
were smaller than those obtained for the jaw crusher, pos-
capacity is that the coarse-grained material blocks the crusher
sibly due to the particle size distribution uniformity of the
setting and, thereby, causes further crushing of the material.

Jaw Cone Roll Smallest Middle Largest

2.45 −5 −4 −3 −2 −1 0
0
2.4
−0.5
Fractal Dimension

2.35 −1
2.3 −1.5
−2
ln(M/MT)

2.25 −2.5
2.2 −3
(Smallest) y = 0.87x - 0.75 −3.5
2.15
(Middle) y = 0.78x - 0.76 −4
2.1 −4.5
(Largest) y = 0.69x - 0.66
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 −5
Feed Rate (kg/min) ln(k/KL)

Figure 5. Variation of fractal dimensions for feed 1 of crushers products by dif- Figure 6. Particle size distributions for cone crusher products of largest, middle,
ferent feed rates. and smallest Df.

Table 6. Overall results of fractal dimensions and crushed material percentage passing thorough target screen for cone crusher products.
1 2 3 4
Feed Feed rate (kg/min) Wt. % passing Df Wt. % passing Df Wt. % passing Df Wt. % passing Df
1 19.19 2.23 20.02 2.27 19.87 2.26 18.77 2.21
1.5 18.94 2.22 19.69 2.25 19.22 2.24 18.41 2.19
2 18.63 2.20 19.15 2.22 18.84 2.21 18.08 2.17
2.5 18.13 2.17 18.89 2.20 18.41 2.18 17.88 2.16
3 17.76 2.15 18.32 2.17 18.05 2.16 17.31 2.13
3.5 18.21 2.18 19.09 2.21 18.73 2.20 17.94 2.16
4 19.23 2.25 20.33 2.29 20.20 2.28 19 2.23
4.5 19.68 2.27 20.92 2.31 20.78 2.30 19.25 2.24
8 I. MORADI AND M. IRANNAJAD

products of the cone crusher. Table 6 presents the amount products of different feeds conform to the axiom of the frac-
of material passing through the target screen. According to tal geometry, suggesting that the fractal dimension for the
the results, an increase in the feed rate resulted in a decrease PSD ranges from 2 to 3. All the results obtained for the roll
and an increase in the amount of the material passing crusher are presented in Table 7. Similar to the results
through the target screen in the cases of permitted and obtained for the jaw and cone crushers, increasing the feed
unpermitted capacities, respectively. According to Table 6, rate at the permitted capacity was associated with decreases
changing the feed size distributions at different feed rates in the material crushing rate and the fractal dimension of
revealed that the material crushing rate and the fractal the products. This can be due to the prolonged crushing
dimension would be increased with an increase in the time of the material, which leads to further uniformity of
amount of the material with a size larger than that of the PSD of the products. Moreover, an increase in the feed
the crusher setting. These particles are entangled between rate augments the amount of the material among the rollers
the cones and the crushing forces are converted from com- and the rollers are engaged initially in crushing the particles
pression to abrasion and shear, leading to the production of with larger sizes. As such, smaller particles find the oppor-
finer particles, thereby raising the fractal dimension. tunity to quickly pass through the rollers, resulting in
decreases in material crushing rate and the fractal dimen-
sions of the products. According to Figure 5, increasing the
3.3. Roll crusher feed rate exceeding the permitted capacity of the crusher
To illustrate the results, only for the largest Df (feed 2 with raises the material crushing rate and the fractal dimension
a feed rate of 4.5 kg/min), the middle Df (feed1 with a feed due to the occurrence of the choke feeding. Furthermore,
rate of 1.5 kg/min), and the smallest Df (feed 4 with a feed the results indicate that an increase in the amount of the
rate of 3 kg/min), the crushed PSD diagrams for the prod- materials with sizes larger than the setting size leads to ele-
vations in material crushing rate and the fractal dimension
ucts are show in Figure 7. The results of other feeds are
of the products for different feeds in the case of changing
indicated in Table 7. The fractal dimension was calculated
the size distribution of the feeds. Table 7 demonstrates the
with regard to the slope line fitted at the points using
amount of material passing through the target screen. As
Equation (6). The correlation coefficient of points in all the
indicated by the results, increasing the feed rate would cause
experiments was greater than R2 ¼ 0.97. The calculated frac-
a decrease and an increase in the amount of the material
tal dimensions for the products of the feed rates of 1, 1.5, 2,
passing through the target screen in the cases of permitted
2.5, 3, 3.5, 4, and 4.5 kg/min with different feeds were
and unpermitted capacities, respectively. According to Table
(2.32–2.41), (2.29–2.38), (2.28–2.36), (2.26–2.34),
7, in the case of changing the feed size distributions at dif-
(2.23–2.32), (2.28–2.37), (2.34–2.43), and (2.38–2.46),
ferent feed rates, increasing the amount of the material with
respectively. The fractal dimensions obtained for the
sizes larger than that of the crusher setting would be associ-
ated with increases in the crushing rate of the material and
Smallest middle Largest the fractal dimensions. These particles are entangled between
−4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 the rollers and the crushing forces are converted from com-
0 pression to abrasion and shear, leading to the production of
−0.5 finer and irregular particles, thereby increasing the frac-
tal dimension.
−1
ln(M/MT)

−1.5

−2 4. Conclusion
(Smallest) y = 0.77x - 0.14
−2.5 In the present work, the effects of the feed rate and the
(Middle) y = 0.66x - 0.03
−3 selected feed size distribution on the crushed PSD were
(Largest) y = 0.54x + 0.15
−3.5
investigated using the fractal geometry. For this purpose, a
ln(k/KL) qualitative parameter (crushed PSD) was described and
Figure 7. Particle size distributions for roll crusher products of largest, middle, investigated by a quantitative parameter (the fractal dimen-
and smallest Df. sion). The results can be summarized as follows:

Table 7. Overall results of fractal dimensions and crushed material percentage passing thorough target screen for roll crusher products.
1 2 3 4
Feed Feed rate (kg/min) Wt. % passing Df Wt. % passing Df Wt. % passing Df Wt. % passing Df
1 90.86 2.36 97.66 2.41 95.44 2.39 87.61 2.32
1.5 88.73 2.34 94.29 2.38 93.20 2.37 84.45 2.29
2 85.44 2.31 92.17 2.36 90.07 2.34 83.28 2.28
2.5 83.21 2.29 90.06 2.34 89.18 2.33 81.31 2.26
3 80.14 2.26 87.82 2.32 86.34 2.30 78 2.23
3.5 87.91 2.31 93.19 2.37 90.40 2.34 84.11 2.28
4 90.61 2.37 98.18 2.43 95.81 2.40 89.42 2.34
4.5 94.80 2.41 99.37 2.46 98.77 2.45 94.62 2.38
 PARTICULATE SCIENCE AND TECHNOLOGY 9

 By considering an appropriate pattern for the lines fitted, Numerical and Analytical Methods in Geomechanics 26 (5):499–513.
with a high correlation coefficient, on the points [Database] doi: 10.1002/nag.209.
Carpinteri, A., and N. Pugno. 2003. A multifractal comminution
obtained from the particle size distributions and a simi-
approach for drilling scaling laws. Powder Technology 131 (1):93–8.
larity in the particle size analysis diagrams for the three doi: 10.1016/S0032-5910(02)00335-2.
crushers’ products of, the ore sample has a self-similar Casini, F., G. M. B. Viggiani, and S. M. Springman. 2013. Breakage of
and fractal behavior. an artificial crushablematerial under loading. GranularMatter 15 (5):
 The fractal dimensions for the PSDs were obtained 661–73.
within the range of 2.13–2.46, which conforms to the Cui, L., L. An, and W. Gong. 2006. Effects of process parameters on
the comminution capability of high pressure water jet mill.
axiom of the fractal geometry (2–3). International Journal of Mineral Processing. 81 (2):113–21. doi: 10.
 Increasing the feed rate at the permitted capacity of the 1016/j.minpro.2006.07.005.
crushers leads to decreased fractal dimension of the Delagrammatikas, G., M. Delagrammatikas, and S. Tsimas. 2007.
products due to decreased dispersion of the particle size Particle size distributions a new approach. Powder Technology 176
distribution PSD and diminished amount of fine par- (2–3):57–65. doi: 10.1016/j.powtec.2007.01.026.
ticles. Consequently, the fractal dimension and the feed Dunne, C. R., K. S. Kawatra, and A. C. Young. 2019. SME mineral
processing extractive metallurgy handbook. Englewood, CO: Society
rate have an inverse relationship at the permit- for Mining, Metallurgy & Exploration.
ted capacity. Elias, I., S. Chaffer, and E. Moya. 2015. Improvements in service life
 By increasing the feed rate exceeding the permitted and cost reduction for gyratory primary crushers through mantle
capacities of the crushers, a considerable amount of fine and concaves optimization: A case study. Proceedings of Iron Ore
particles is produced as a result of the choke feeding, 2015, Australasian Institute of Mining and Metallurgy, Melbourne,
Victoria, 115–27.
leading to an increase in the fractal dimension with a
Evertsson, C. M. 2000. Cone crusher performance. Ph.D. thesis,
direct relationship between the fractal dimension and the Chalmers University of Technology, Sweden.
choke feeding. Falconer, K. 2003. Fractal geometry mathematical foundations and
 Variations in the feed size distribution revealed that ele- applications. Hoboken: John Wiley and Sons.
vated amount of large particles in the feed was associated Filippov, L. O., R. Joussemet, M. Irannajad, R. Houot, and A. Thomas.
with increased rate of material crushing and the frac- 1999. An approach of the whiteness quantification of crushed and
floated talc concentrate. Powder Technology 105 (1–3):106–12. doi:
tal dimension. 10.1016/S0032-5910(99)00124-2.
 The fractal dimensions of cone crushed particles had the Flook, A. G. 1978. The use of dilation logic on the quantimet to
lowest value, possibly due to the functional, remaining achieve fractal dimension characterization of textured surfaces.
time, and particle size distribution uniformity of the Powder Technology 21 (2):295–8. doi: 10.1016/0032-5910(78)80099-0.
products of this type of crusher. Fuerstenau, D. W., J. J. Lutch, and A. De. 1999. The effect of ball size
on the energy efficiency of hybrid high-pressure roll mill/ball mill
 The fractal dimension was shown to have the highest
grinding. Powder Technology 105 (1–3):199–204. doi: 10.1016/S0032-
value in the roll crusher, which can be due to the cre- 5910(99)00138-2.
ation of fine-grained materials and more limited particle Glazner, A. F., and R. D. Mills. 2012. Interpreting two-dimensional
size distribution of the products of this crusher. cuts through broken geologic objects: Fractal and non-fractal size
distributions. Geosphere 8 (4):902–14. doi: 10.1130/GES00731.1.
Gupta, A., and D. Yan. 2016. Mineral processing design and operations.
ORCID 2nd ed., Chapter 6 – Roll crushers, 169–88. Amsterdam: Elsevier.
Hamzelo, E., M. Masinaei, and N. Mehrshad. 2013. Discrimination and
Mehdi Irannajad http://orcid.org/0000-0002-5469-084X Identification of Hydrothermal Alteration Types Related to Porphyry
Copper Deposits, Using Hyperion Hyperspectral Data. Iranian
Journal of Mining Engineering (IRJME) 8 (19):25–33. (In Persian)
References Hyslip, J., and L. E. Vallejo. 1997. Fractal analysis of the roughness and
size distribution of granular materials. Engineering Geology. 48
Ahmed, H. A. M., and J. Drrzymala. 2005. Two-dimensional fractal lin-
earization of distribution curves. Physicochemical Problems of (3–4):231–44. doi: 10.1016/S0013-7952(97)00046-X.
Mineral Processing. 39:129–39. Irannajad, M. 1998. Chapter vii: Study of Particles Morphology by
Airikka, P. 2015. Automatic feed rate control with feed-forward for Fractal Approach. PhD thesis, entitled: Valorization of talcose min-
crushing and screening processes. IFAC-PapersOnline 48–17:149–54. ing rejects by physico-chemical (flotation) processing: Influence of
Allen, T. 1997. Particle size measurement. Powder sampling and particle treatments on the whiteness of concentrates. Department of
size measurement. Powder technology series, Vol. 1. London: Environment and Mineralurgy (Laboratoire Environnment et
Chapman and Hall. Mineralurgie (L.E.M.), National Superior School of Applied Geology
Arhangel’skii, A. V., and J. van Mill. 2018. Some aspects of dimension (Ecole Nationale Superieure de Geologie Applique (ENSG),
theory for topological groups. Indagationes Mathematica 29:202–25. University of Lorraine (ex. Lorraine National Polytechnics Institute
Bartoli, F., R. Philippy, M. Doirisse, S. Niquet, and M. Dubuit. 1991. (Institut National Polytechnique de Lorraine (INPL), Nancy, France.
Structure and selfsimilarity in silty and sandy soils: The fractal Lu, P., I. F. Jefferson, M. S. Rosenbaum, and I. J. Smalley. 2003. Fractal
approach. Journal of Soil Science 42 (2):167–85. doi: 10.1111/j.1365- characteristics of loess formation: Evidence from laboratory experi-
2389.1991.tb00399.x. ments. Engineering Geology 69 (3–4):287–93. doi: 10.1016/S0013-
Blenkinsop, T. G. 1991. Cataclasis and processes of particle size reduc- 7952(02)00287-9.
tion. Pure and Applied Geophysics. 136:1–33. Mandelbrot, B. B. 1982. The fractal geometry of nature. New York:
Cai, G., Y. Xiong, and L. Lin. 2015. Experimental study on the fractal Freeman.
characteristics of rocks crushing. MATEC Web of Conferences, 25, Mohammadi, M., M. Shabanpour, M. H. Mohammadi, and N.
02008. Davatgar. 2019. Characterizing spatial variability of soil textural frac-
Carpinteri, A., and N. Pugno. 2002. A fractal comminution approach tions and fractal parameters derived from particle size distributions.
to evaluate the drilling energy dissipation. International Journal for Pedosphere 29 (2):224–34. doi: 10.1016/S1002-0160(17)60425-9.
10 I. MORADI AND M. IRANNAJAD

Moradi, I., M. H. Mohammadzadeh, and M. Irannajad. 2019. Turcotte, D. L. 1992. Fractals and chaos in geology and geophysics.
Calculation of fractal dimension of crushing product by variation of Cambridge: Cambridge University Press.
ore type and feed size distribution. The 2nd Iranian Conference of Tyler, S. W., and S. W. Wheatcraft. 1992. Fractal scaling of soil par-
Green Mining & Mineral Industry. Zanjan, Iran. 18 December. ticle-size distribution: Analyses and limitations. Soil Science Society
Petrakis, E., E. Stamboliadis, and K. Komnitsas. 2017. Evaluation of the of America Journal 56 (2):362–9. doi: 10.2136/sssaj1992.
relationship between energy input and particle size distribution in 03615995005600020005x.
comminution with the use of piecewise regression analysis. Wang, Y., W. Dan, Y. Xu, and Y. Xi. 2015. Fractal and morphological
Particulate Science and Technology 35 (4):479–89. doi: 10.1080/ characteristics of single marble particle crushing in uniaxial com-
02726351.2016.1168894. pression tests. Advances in Materials Science and Engineering 2015:
Radziszewski, P. 2000. Developing an experimental procedure for 1–10.
charge media wear prediction. Minerals Engineering 13 (8–9): Wills, B. A., J. A. Finch, F. C. I. M. Frsc, and P. Eng. 2016. Chapter 6
949–61. doi: 10.1016/S0892-6875(00)00080-7. – Crushers. In Wills’ mineral processing technology: An introduction
Sadrai, S., J. A. Meech, M. Ghomshei, F. Sassani, and D. Tromans.
to the practical aspects of ore treatment and mineral recovery, 8th
2006. Influence of impact velocity on fragmentation and the energy
ed., ed. B. A. Wills and J. A. Finch, 123–46. Amsterdam; Boston,
efficiency of comminution. International Journal of Impact
MA: Elsevier/BH.
Engineering. 33 (1–12):723–34. doi: 10.1016/j.ijimpeng.2006.09.063.
Xu, Y. 2018. Fractal dimension of demolition waste fragmentation and
Sadrai, S.,. J. A. Meech, D. Tromans, and F. Sassani. 2011. Energy effi-
cient comminution under high velocity impact fragmentation. its implication of compactness. Powder Technology 339:922–9. doi:
Minerals Engineering 24 (10):1053–61. doi: 10.1016/j.mineng.2011. 10.1016/j.powtec.2018.08.071.
05.006. Zhang, J. R., Z. L. Hu, and Z. D. Liu. 2004. Fractal features character-
Stamboliadis, E., E. Petrakis, and O. Pantelaki. 2011. Variation of ized by particle size distribution of eco-material for erosion control
dimensional properties of particulate materials during grinding and of cutting slope. http://docplayer.net.
their non-fractal nature. Canadian Metallurgical Quarterly 50 (2): Zhang, J., M. Li, Z. Liu, and N. Zhou. 2017. Fractal characteristics of
119–26. doi: 10.1179/000844311X12949291727970. crushed particles of coal gangue under compaction. Powder
Tasdemir, A. 2009. Fractal evaluation of particle size distributions of Technology 305:12–8. doi: 10.1016/j.powtec.2016.09.049.
chromites in different comminution environments. Minerals Zhong, W., N. He, T. Cosgrove, Y. J. Zhu, and L. Fu. 2019. Analysis
Engineering 22:156–67. of the correlation between fractal dimension of gravelly soil
Turcotte, D. L. 1986. Fractals and fragmentation. Journal of and debris-flow initiation through in-situ experiments. http://www.
Geophysical Research 91 (B2):1921–6. doi: 10.1029/JB091iB02p01921. aloki.hu.