=Paper=
{{Paper
|id=Vol-3156/paper20
|storemode=property
|title=Determining rock varieties on the basis of fuzzy clustering of ultrasonic measurement results
|pdfUrl=https://ceur-ws.org/Vol-3156/paper20.pdf
|volume=Vol-3156
|authors=Volodymyr Morkun,Gerhard Fischerauer,Natalia Morkun,Vitalii Tron,Alona Haponenko
|dblpUrl=https://dblp.org/rec/conf/intelitsis/MorkunFMTH22
}}
==Determining rock varieties on the basis of fuzzy clustering of ultrasonic measurement results==
https://ceur-ws.org/Vol-3156/paper20.pdf
Determining Rock Varieties on The Basis of Fuzzy Clustering of
Ultrasonic Measurement Results
Volodymyr Morkuna, Gerhard Fischerauerb, Natalia Morkuna, Vitaliy Trona and Alona
Haponenkoa
a
Kryvyi Rih National University, Vitalii Matusevich street,11, Kryvyi Rih, 50027, Ukraine
b
Bayreuth University, Universitätsstraße 30, Bayreuth, 95447, Germany
Abstract
Geophysical methods of investigating rocks are used to define the geological structure of
mineral resources by borehole sections, identify and evaluate iron ore raw materials.
Geophysical data can be also applied to designing, controlling and analyzing deposit mining
and engineering states of mine workings.
The research is aimed at improving accuracy of ultrasonic logging to determine physical-
mechanical and chemical-mineralogical characteristics of rocks based on the fuzzy clustering
of the results of ultrasonic measurements: velocity of longitudinal and transverse wave
propagation, the ultrasonic attenuation coefficient at fundamental frequency and higher
harmonics and relationship of these parameters.
Keywords 1
Iron ore, identification, ultrasound, fuzzy, clustering
1. Introduction
In Ukraine, balance reserves of iron ore amount to over 30 billion tons, which is sufficient to
supply mining and metallurgical enterprises for about 95-100 years ahead. Of ten large-scale iron ore
mining enterprises of the country, seven are located in Kryvyi Rih region: the CJSC Inhuletskyi GZK,
the PJSC Pivdennyi GZK, the Mining Department of the PJSC ArcelorMittal Kryvyi Rih, the PJSC
Tsentralnyi GZK, the PJSC Pivnichnyi GZK, the PJSC Kryvbasrudkom, and the PJSC Sukha Balka.
These enterprises of Kryvyi Rih iron ore basin provide more than 90% of Ukrainian metallurgical
enterprises’ needs in raw materials [1].
In general, iron ores are represented by three main types – rich martite, ferruginous quartzite and
brown ironstone. Ferruginous quartzites, which are the main reserve for developing the raw material
base of Kryvyi Rih basin, belong to the hard-rock geological and industrial type. Depending on the
availability of certain layers, magnetite ferruginous quartzites are divided into a number of varieties
characterized by a different technological value: magnetite quartzite containing almost no silicates
and carbonates; silicate-magnetite quartzite with subordinate quantity of silicate layers; hematite-iron-
mica-magnetite quartzite with subordinate quantity of hematite layers; magnetite-silicate quartzite
with subordinate quantity of magnetite layers, poor quartzite on the verge of industrial significance;
carbonate-magnetite quartzite; silicate-carbonate, or carbonate-silicate magnetite quartzite.
The efficiency of mining and processing enterprises depends on how accurately and timely they
receive information on geological and mineralogical types of iron ore raw materials that are extracted
or processed. To obtain information on geological and mineralogical types of iron ore raw materials,
there are applied methods of geophysical research based on various measurements of rock properties
IntelITSIS’2021: 3rd International Workshop on Intelligent Information Technologies and Systems of Information Security, March 23–25,
2021, Khmelnytskyi, Ukraine
EMAIL: morkunv@gmail.com (V. Morkun); MRT@uni-bayreuth.de (G. Fischerauer); nmorkun@gmail.com (N. Morkun);
vtron@knu.edu.ua (V.Tron); a.haponenko@protonmail.com (A. Haponenko)
ORCID: 0000-0003-1506-9759 (V. Morkun); 0000-0003-2000-4730 (G. Fischerauer); 0000-0002-1261-1170 (N. Morkun); 0000-0002-
6149-5794 (V.Tron); 0000-0003-1128-5163 (A. Haponenko)
©️ 2022 Copyright for this paper by its authors.
Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR Workshop Proceedings (CEUR-WS.org)
as well as related data obtained when performing various technological operations, such as drilling
boreholes.
Xiaowei Pan [2], Victor Mwango Bowa [3] suggest an approach to static optimization of drilling
processes in the context of ore mining and processing, yet little attention is paid to dynamic processes
occurring directly in borehole drilling. Z.Q. Yue, C.F. Lee, K.T. Law, and L.G. Lam [4] indicate that
the mechanical drilling rate is a sufficient dynamic indicator characterizing rock properties. The
disadvantage of this approach is that in the complex geological structure of iron ore deposits, this
indicator is insufficient and requires analysis of a large number of variables to confirm research
results [5,6].
H. Schunnesson [6] suggests determining a rock type in the course of drilling by two indicators – a
mechanical drilling (penetration) rate and torque. The disadvantage of this approach is that the
required accuracy of rock type recognition can be achieved in a binary geological structure, because
these parameters indicate homogeneity and strength of the rock quite accurately. With several
mineralogical and technological varieties available, this approach is difficult to apply.
Natalie Beattie [7] applies the following parameters to monitoring the drilling process: horizontal
and vertical vibrations, axial pressure, torque, drilling rate, rotation rate, etc. The disadvantage of this
approach is that among the controlled parameters there are only those of a drilling rig and no directly
or indirectly measured characteristics of the drilled rock. Jorge Martin [8] uses a similar approach to
identify a ternary geological structure using three types of drilled rocks. However, as in the above
work, rock varieties are mainly identified by a single indicator (strength) and drilling rig parameters
without measuring characteristics of the rock itself.
Ultrasonic measurements are promising for rapid determination of geological and mineralogical
types of iron ore raw materials.
Ultrasonic logging is based on elastic wave propagation in the studied medium to determine its
physical-mechanical and chemical-mineralogical characteristics [11,10]. Acoustic waves in solids can
be divided into volume, surface, waveguide and guided. Volume acoustic waves propagate throughout
the solid. The shape of the wave front is one of their distinguishing features [11,12]. They can be flat,
spherical, cylindrical, etc. Another feature is the direction of the shear vector of particles of the
medium, on this basis there are distinguished volume-longitudinal and volume-transverse waves.
Surface acoustic waves propagate near the free surface of a solid body or near the interface between
two different media [13]. Their phase velocity is parallel to this surface and their intensity decreases
rapidly with the depth of penetration into the volume of the solid. Waveguide acoustic waves can
exist in rods and thin layers as in waveguides, and guided waves – in protrusions or grooves of
different profiles on the surface of a solid body as in channels.
2. Proposed methodology
Let us consider the method of determining geological and mineralogical varieties of rocks based
on assessment of changes in the velocity of longitudinal and transverse volume ultrasonic waves, a
relationship of these values in the controlled medium, as well as a parameter characterizing the degree
of nonlinearity of this process – the value of ultrasound attenuation on fundamental and higher
harmonics.
In rocks, propagation velocities of elastic waves vary widely and depend on physical properties,
structure, texture, condition and other internal and external factors. Propagation velocities of elastic
waves in the unlimited elastic medium can be determined by formulas derived from wave equations
[14]. The velocity of the longitudinal wave in the bulk is
𝐸 1−𝜎
𝐶𝐿 = √ ∙ (1)
𝜌 (1 + 𝜎) ∙ (1 − 2𝜎)
where 𝜌 is the density of the medium, 𝐸 is the Young modulus, and 𝜎 is the Poisson ratio.
The transverse-wave propagation velocity is
𝐸 1 µ
𝐶𝑇 = √ ∙ =√ (2)
𝜌 2 ∙ (1 + 𝜎) 𝜌
where µ is the shear modulus.
The ratio of the velocity of longitudinal waves to that of the transverse ones is a function only of
the Poisson ratio of the rock:
𝐶𝐿 1−𝜎
= √2 ∙ 1−2𝜎 . (3)
𝐶𝑇
Propagation velocities of elastic waves in rocks vary significantly depending on mineral
composition, density, porosity, grain size, and other parameters. Their values increase from acidic
intrusive rocks to basic and ultrabasic ones. In igneous rocks, a decrease in values of propagation
velocities of elastic waves with increasing SiO2 content is noticeable. In rocks of basic composition,
the velocity of longitudinal waves in intrusive samples is on average 20% higher than in effusive
ones. As acidity increases, this difference decreases. There is no such relationship for transverse-wave
velocities [15]. The propagation velocities of longitudinal and transverse elastic waves for igneous
and metamorphic rocks are linearly related to density. No such regularities are established for
sedimentary rocks. The values of elastic wave velocities change due to variable structural and textural
characteristics of rocks in different places of sampling for testing. Thus, the coefficient of velocity
variation of longitudinal waves in rocks taken from one deposit amounts to up to 40% in clays and up
to 25% in limestones and dolomites. The distribution of values of elastic wave velocities in most
igneous and metamorphic rocks follows a Gaussian normal distribution [16,17].
Table 1 shows statistical characteristics of elastic properties of the most common rocks and
ultrasound velocities in them.
Table 1
Statistical characteristics of elastic properties of the most common rocks and ultrasound velocities in
them
Standard
Parameters Min Max Mean Median Variance
Deviation
CL, m/s 4450.0000 6320.0000 5562.2222 5930.0000 556094.4444 745.7174
CT, m/s 2780.0000 3370.0000 3123.3333 3140.0000 40175.0000 200.4370
ρ, g/cm3 2.5200 2.9600 2.7211 2.7000 0.0187 0.1366
E, hPa 47.7451 85,9804 66,9935 70,9804 159,3946 12,5008
μ, hPa 20.1961 32.7451 26.5359 27.0588 175.9804 4.1432
σ 0.1800 0.3200 0.2582 0.3050 0.0040 0.0636
As follows from Fig. 1 and Fig. 2, the velocities of longitudinal and transverse ultrasonic waves in
rocks are determined by their elastic characteristics:
• the velocity of longitudinal waves increases with the Young modulus and the Poisson
ratio;
• changes in the Poisson ratio from 0.1 to 0.4 increase the longitudinal wave velocity by
about 45%.
• the velocity of transverse waves increases with the Young modulus, but decreases with the
Poisson ratio.
The dependence on the material parameters is non-linear as shown by eqns. (1), (2), but it can be
linearized in sufficiently narrow intervals. Such regression lines have been drawn in Figs. 1 and 2.
Thus, the velocity of elastic wave propagation in rocks is determined by their elastic properties and
density. This velocity is almost independent of the wavelength, which enables using waves with any
frequency for research.
In contrast to the velocity of elastic wave propagation, the physical dispersion of which is almost
absent in most rocks, the attenuation coefficient is determined by the frequency of elastic oscillations.
In a wide range of frequencies from 1 Hz to 10 MHz, the attenuation coefficient а in different rocks
varies from 1•10−8 to 2•102 m-1. The attenuation decrement in the same frequency range varies from
1•10−2 to 1.0 on the average.
Figure 1: Dependence of the velocity of longitudinal ultrasonic waves, CL, on rock characteristics. The
shaded bands are the Working-Hotelling uncertainty intervals of the regression lines at the 95%
confidence level.
Figure 2: Dependence of the velocity of transverse ultrasonic waves, CT, on rock characteristics. The
shaded bands are the Working-Hotelling uncertainty intervals of the regression lines at the 95%
confidence level.
The attenuation coefficient increases with frequency. However, a clear unambiguous functional
dependence of attenuation on frequency for rocks is not established. For example, for granites in the
frequency range from 10 kHz to 1000 kHz, the best approximation is observed when describing the
frequency dependence by the quadratic function α=mf2, where m is the coefficient of proportionality.
In gabbro-diabases, quartzites, granite-gneisses, sandstones, shales and other rocks, the frequency
dependence in the range from 500 kHz to 5000 kHz obeys the law α=A1f+A2f2. This dependence is
observed for both longitudinal and transverse waves [15,16].
Table 2 contains statistical characteristics of elastic properties of the most common rocks and the
attenuation of ultrasonic waves in them. Table 3 summarizes the calculated correlation coefficients
between basic parameters of ultrasonic waves propagating in rocks considering their attenuation.
Table 2
Statistical characteristics of elastic properties of the most common rocks, velocities of propagation,
and attenuation of ultrasonic waves in them.
Standard
Parameters Min Max Mean Median Variance
Deviation
a, dB/m 4.0000 71.0000 35.6250 35.5000 563.1250 23.7303
CL, m/s 4144.0000 5714.0000 4768.8750 4620.0000 401339.5536 633.5137
CT, m/s 2343.0000 2857.0000 2652.6250 2731.0000 40868.5536 202.1597
3
ρ, g/cm 2.6000 3.0550 2.75350 2.7050 25.6934 0.1603
E, hPa 34.4400 59.8700 47.7438 47.9350 97.2451 9.8613
σ 0.1500 0.3300 0.2588 0.2650 0.0036 0.0601
Table 3
Correlation coefficients between basic parameters of ultrasonic wave propagation in rocks
considering their attenuation.
Parameters a, dB/m CL/CT CL, m/s CT, m/s ρ, g/cm3 E, kg/mm2 σ
a, dB/m 1.0000 0.0200 0.1905 0.3328 0.5132 0.3582 -0.1356
CL/CT 0.0200 1.0000 0.8379 0.2815 0.6962 0.6273 0.9732
CL, m/s 0.1905 0.8379 1.0000 0.7591 0.6393 0.9130 0.7743
CT, m/s 0.3328 0.2815 0.7591 1.0000 0.3105 0.8623 0.2014
ρ, g/cm3 0.5132 0.6962 0.6393 0.3105 1.0000 0.6514 0.6670
E, hPa 0.3582 0.6273 0.9130 0.8623 0.6514 1.0000 0.5697
σ -0.1356 0.9732 0.7743 0.2014 0.6670 0.5697 1.0000
In crystalline rocks, the attenuation coefficient of transverse waves is usually equal to that of
longitudinal waves or about 1.5–2 times higher. In wet clays and water-saturated sands, there is a
significant difference in attenuation coefficients of transverse and longitudinal waves (up to 5 or
more) [15,16]. It should be noted that when velocities of elastic wave propagation in many hard
monolithic rocks change by 40%–60%, their attenuation coefficients change by a factor of 2–4. Thus,
this indicates that the attenuation coefficient of elastic waves is a more sensitive parameter for
qualitative characteristics of rocks than the velocity of ultrasound. However, this factor causes a
dependence of attenuation of ultrasonic oscillations that propagate in the studied medium on a variety
of disturbing factors, such as changes or violations of its structure.
Ultrasonic methods of measuring characteristics of rocks with phase inhomogeneity such as cracks
and pores face a number of difficulties due to the fact that in this case ultrasonic propagation becomes
significantly nonlinear. In this case, the fundamental assumption that constitutive behaviour of a
material under study is linear-elastic which is justified for linear measurements is not true. Most rocks
are not linearly elastic even before formation of cracks. More importantly, nonlinear elasticity of
rocks increases significantly with available cracks and pores [18].
Porosity as presence of pores, cracks (voids) in rocks is one of their most important structural
features. Total (absolute, physical, complete) porosity is characterized by the ratio of the pore volume
to that of the whole rock. The coefficient of total porosity kt is a ratio of the volume of all pores Vр to
the total volume of the rock sample Vr in fractions or percentages
𝑉
𝑘𝑡 = 𝑉𝑝 ∙ 100. (4)
𝑟
According to their origin, pores and cracks can be primary or secondary. The first are pores and
cracks formed during sedimentation and formation of the rock massif. Secondary pores and cracks are
formed during post-diagenetic changes.
From research results [19] it can be concluded that velocities of longitudinal and transverse
ultrasonic waves have a significant dependence on porosity.
According to mechanics of elastic waves, ultrasound causes a slight deformation of tension and
compression inside the rock, with friction occurring during compression strains. The contact area
between crack surfaces and pore surfaces is a set of irregularities that are in close proximity to each
other, which, for example, corresponds to the grain contact pattern characteristic of sedimentary rocks
[20]. The energy dissipated at each irregularity can be written as follows [21]
𝑊𝑓 = 𝛼 𝜏𝑓 𝜎, (5)
where τf is shear stress on the crack surface caused by ultrasound, σ is compression stress, a is the
length of the contact between irregularities of the crack.
It follows from the above that available cracks and pores in the structure of the material cause
strongly nonlinear dynamic phenomena accompanying propagation of elastic waves, and,
consequently, changes in characteristics (propagation rate, attenuation) of this process.
Thus, there arises a need to study nonlinearity of the process of ultrasound propagation in rocks
and take into account the influence of cracks and pores on the results of measuring characteristics of
this process.
In the process of movement, the shape of the wave train changes: there is a decrease in the height
of the envelope of the ultrasonic signal while increasing its width (spread of the wave train in space
caused by dispersion properties of the medium).
Generation of higher harmonics is a traditional phenomenon, when the shape of the incident wave
is distorted by a nonlinear elastic response of the medium to this wave. When sending a purely
sinusoidal ultrasound signal with amplitude A0 at frequency f0, the detected signal passing through the
defect will be distorted, and the detected wave will have the amplitude component A1 at fundamental
frequency f0, the amplitude component A2 at the second harmonic frequency 2f0, A3 at the third
harmonic frequency 3f0, etc.
In [21] the solution of the second-order nonlinear wave equation is approximated to obtain the
following expression of the quadratic nonlinearity parameter:
8𝐴2
𝛽𝑛 = 2 2 , (6)
𝐴1 𝑘 𝑥
where A1 and A2 are the frequency amplitudes of the first and second harmonics of the recorded
signals, respectively; k is the wave number, x is the propagation distance.
Similarly, the third-order nonlinear coefficient of elasticity is expressed as
48𝐴3
𝛾= 3 3 , (7)
𝐴1 𝑘 𝑥
where A3 is the third harmonic amplitude.
As a rule, as the damage to the test material increases, the degree of nonlinear interaction also
increases, and energy is transferred from the fundamental frequency to higher harmonics. Thus,
damages and their quantitative characteristics can be assessed by measuring the nonlinearity of the
ultrasonic wave propagating through the material.
The longitudinal ultrasonic wave propagates through the contact between two hard surfaces of
medium particles. The availability of cracks and pores in the test samples causes modulation of the
propagating ultrasonic train. The formation of the harmonic frequency that is twice the frequency of
the main input signal is caused by an incorrect waveform in the time domain after a certain
propagation distance covered. The ultrasonic wave is distorted due to nonlinearity of the material, this
resulting in higher harmonics. Thus, the received signal consists of not only a fundamental frequency
wave, but also either the second wave or higher harmonics. The nonlinear parameter associated with
amplitudes of the fundamental wave and the second harmonic [21,22] is determined as follows:
𝐴2
𝛽= , (8)
𝐴1
where A1 and A2 are amplitudes of the fundamental wave and the second harmonic wave respectively.
Thus, the nonlinearity of propagation of ultrasonic waves in rock, which is determined by the
presence of cracks and pores, can be evaluated by specifying the amplitudes of the fundamental,
second, and higher harmonics. To obtain a qualitative assessment of nonlinearity of this process, all
the values should be normalized to represent only a relative change of the acoustic nonlinear response.
The obtained results support the conclusion that the second- and third-order nonlinearities should
be evaluated to adjust the velocities of ultrasonic waves in identifying geological and mineralogical
characteristics of rocks to eliminate errors associated with cracks and pores.
3. Results
Considering the above, the results of measuring the propagation of ultrasonic waves in iron ore
raw materials will be assumed to be the main parameters for identifying their mineralogical and
technological varieties by methods of crisp and fuzzy clustering.
Let us analyze characteristics of seven ore types that are mined and processed from one of Kryvyi
Rih iron ore basin deposits. Table 4 shows the results of this analysis. The following symbols of ore
types are adopted [1,12]: 1 – magnetite hornstones; 2 – silicate-carbonate-magnetite hornstones; 3 –
red-striped magnetite and hematite-magnetite hornstones; 4 – semi-oxidized and oxidized hornstones;
5 – silicate shales, non-metallic hornstones and quartz; 6 – magnetite-silicate-carbonate (poor)
hornstones; 7 – hematite-magnetite hornstones.
Table 4
Аnalysis results of different ore types
Quartz, Magnetite, Martite, Hematite, Siderite, Density,
Ore type
% % % % % kg/m3
1 63.7 30.9 0 1.4 3.8 3431
2 68.4 21.7 0 0.4 9.1 3248
3 64.5 30.2 0 1.5 3.8 3414
4 65.4 24.4 3.3 3.7 3.2 3412
5 74.6 4.5 0 0.7 20.2 2989
6 75.2 6.8 0 0.8 17.2 3009
7 60.8 31.4 0 5.4 2.5 3530
The data used to identify mineralogical and technological varieties of iron ore by crisp and fuzzy
clustering methods are observations of the physical process of ultrasonic wave propagation in the
studied medium: Each observation consists of n measured variables grouped into an n-dimensional
vector,
𝑥𝑘 = [𝑥𝑘1 , 𝑥𝑘2 , … , 𝑥𝑘𝑛 ]𝑇 , 𝑥𝑘 ∈ 𝑅 𝑛 . (9)
The set of N observations is denoted by:
𝑋 = {𝑥𝑘 |𝑘 = 1,2, … , 𝑁} , (10)
and presented as a matrix N×n:
𝑥11 … 𝑥1𝑛
𝑋=[ ⋮ ⋱ ⋮ ]. (11)
𝑥𝑁1 … 𝑥𝑁𝑛
The following rigid clustering methods have been studied: k-means and k-medoid; the fuzzy c-
means (FCM) algorithm; the Gustafson-Kessel method as an advanced standard fuzzy c-means
algorithm with the adaptive distance norm; the Gath-Geva method as a clustering algorithm of fuzzy
maximum likelihood estimates (FMLE) which uses the distance norm based on fuzzy maximum
likelihood estimates; the Fuzzy c-Shape method which replaces the norm of the internal product in the
FCM model with a shape-based distance function. Figure 3 shows the results of clustering by this
algorithm.
The research results indicate that the most effective method for solving this problem is the Fuzzy
c-Shape method, which identifies geological and mineralogical varieties of iron ore with a 0.91%
probability.
Figure 3: Сlustering results using the Fuzzy c-Shape algorithm
4. Conclusion
It is established that evaluated changes in the propagation velocity of longitudinal and transverse
ultrasonic bulk waves, the relationship of these values in the controlled medium and the parameter
characterizing the nonlinearity degree of this process – the value of ultrasonic attenuation at
fundamental and higher harmonics – can be used to identify geological and mineralogical varieties
(types) of iron ore raw materials.
The proposed method based on the results of ultrasonic measurements of rock characteristics and
the fuzzy inference allows identifying mineralogical and technological ore varieties in the rock massif
at the initial stage of the technological process of ore extraction and processing. The specified
varieties can be compared with relevant technological regulations and predetermined optimal
characteristics of technological units, thus ensuring achievement of set indicators of mining and
processing considering requirements of environmental protection and energy efficiency.
By comparative analysis of different crisp and fuzzy clustering algorithms, it is found that the best
results in terms of accuracy and efficiency are provided by the Fuzzy c-Shape method, which replaces
the norm of the internal product in the FCM model with a shape-based distance function.
5. References
1. S. Pysmenniy, N. Shvager, O. Shepel, K. Kovbyk, O. Dolgikh, Development of resource-
saving technology when mining ore bodies by blocks under rock pressure, E3S Web of
Conferences 166 (2020) 02006. doi: 10.1051/e3sconf/202016602006.
2. X. Pan, Optimization of mineral processing plant through ROM ore size, AGH Journal of
Mining and Geoengineering 36.4 (2012) 123-132.
3. V. M. Bowa, Optimization of blasting design parameters on open pit bench a case study of
nchanga open pits, International journal of scientific & technology research 4.9 (2015) 45-51.
4. Z.Q. Yue, C.F. Lee, K.T. Law, L.G. Lam, Automatic monitoring of rotary percussive drilling
for ground characterization, International Journal of Rock Mechanics & Mining Sciences 41
(2015) 573–612.
5. H. Schunnesson, Rock characterisation using percussive drilling, International Journal of Rock
Mechanics & Mining Sciences 35.6 (2014) 711–725.
6. H. Schunnesson, K. Holme, Drill monitoring for geological mine planning in the Viscaria
copper mine Sweden, CIM Bulletin 90.1030 (2015) 82-89.
7. N. Beattie, Monitoring-while-drilling for open-pit mining in a hard rock environment:
Master`s thesis. Kingston. Queen`s University. 2012. 127 p.
8. Jorge Martin. Application of Pattern Recognition Techniques to Monitoring-While-Drilling on
a Rotary Electric Blast Hole Drill at an Open-Pit Coal Mine, Master’s thesis, Queen`s
University, Kingston, 2013.
9. J. B. Segui, M. Higgins, Blast design using measurement while drilling parameters, Fragblast:
International Journal for Blasting and Fragmentation 6.3–4 (2012) 287–299.
10. M. J. Scoble, J. Peck, C. Hendricks, Correlation between rotary drill performance parameters
and borehole geophysical logging, Mining Science and Technology 8 (2012) 301-312.
11. V. Morkun, N. Morkun, V. Tron, Identification of control systems for ore-processing industry
aggregates based on nonparametric kernel estimators, Metallurgical and Mining Industry 7.1
(2015) 14-17.
12. V. Morkun, V. Tron, Automation of iron ore raw materials beneficiation with the operational
recognition of its varieties in process streams, Metallurgical and Mining Industry 6.6 (2014)
4-7.
13. V. Morkun, N. Morkun, A. Pikilnyak, The adaptive control for intensity of ultrasonic influence
on iron ore pulp, Metallurgical and Mining Industry 6.6 (2014) 8-11.
14. L. Gao, W. Zhang, W. Lu, X. Hu, H. Wu, J. Wang, B. Kong, Study on the effects of
temperature and immersion on the acoustic emission and electromagnetic radiation signals of
coal rock damage under load, Engineering Geology 297 (2022) 106503. doi:
https://doi.org/10.1016/j.enggeo.2021.106503.
15. Acoustic properties of rocks (2017). URL: http://ctcmetar.ru/volnovye-processy/9297-
akusticheskie-svoystva-gornyh-porod.html.
16. Acoustic velocity logging (2019). URL: http://fccland.ru/dobycha-nefti/6826-akusticheskiy-
karotazh-po-skorosti.html.
17. X. Kou, C. Pei, Z. Chen, Fully noncontact inspection of closed surface crack with nonlinear
laser ultrasonic testing method, Ultrasonics 114 (2021) 106426. doi:
10.1016/j.ultras.2021.106426.
18. H. Salehi, R. Burgueno, S. Chakrabartty, N. Lajnef, A. H. Alavi, A comprehensive review of
self-powered sensors in civil infrastructure: State-of-the-art and future research trends,
Engineering Structures 234 (2021) 111963. doi:
https://doi.org/10.1016/j.engstruct.2021.111963.
19. F. Saati, K.-A. Hoppe, S. Marburg, K. V. Horoshenkov, The accuracy of some models to
predict the acoustical properties of granular media, Applied Acoustics 185 (2022) 108358. doi:
10.1016/j.apacoust.2021.108358.
20. G. Liu, Z. Liu, J. Feng, Z. Song, Z. Liu, Experimental research on the ultrasonic attenuation
mechanism of coal, Journal of Geophysics and Engineering, 14 (2017) 502–512. doi:
10.1088/1742-2140/aa5f23.
21. E.V. Trifonov, Fully integrable one-dimensional nonlinear wave equation: Solution of a
general initial value problem, Partial Differential Equations in Applied Mathematics 5 (2022)
100268. doi: 10.1016/j.padiff.2022.100268.
22. M. L. Wang, J. P. Lynch, H. Sohn, Sensor Technologies for Civil Infrastructures, Woodhead
Publishing, 2014.