Content uploaded by Yurii Kuchirka
Author content
All content in this area was uploaded by Yurii Kuchirka on Feb 10, 2022
Content may be subject to copyright.
Yu.M. Kuchirka, E.T. Volodarsky
Methods and objects of chemical analysis, 2019, Vol. 14, No. 1, 15–20
An essential part of modern quality management system in cement production is state-of-the-art radiation
measurement technologies based on methods of neutron activation, X-ray uorescence and X-ray diffraction
chemical analysis of substance. The high speed and accuracy of measuring the characteristics of raw materials
and nished products can be achieved by their complex application, thereby ensuring an improving in the level
of automation of cement production and the quality of cement in general. The main stages of the portlandcement
production process by dry method are considered, including the keypoints of quality control of their implementation
by applying mentioned radiation methods of chemical analysis of raw materials and completed product. The
metrological problems of their practical implementation in continuous cement production are analyzed, in particular
problems of uncertainty assessment, static and dynamic calibration and increase of accuracy of measuring
systems that implement neutron activation analysis methods. Shown the directions of their improvement by the
use of alternative neutron sources, methods of Monte Carlo N-Particle Transport Code for neutron activation
analysis physical processes simulation and machine learning for the efcient processing of spectral characteristics
of investigated substances.
Keywords: cement, neutron activation analysis, PGNAA, PFTNAA, XRF, uncertainty
15
Cement Quality Control by using Modern Radiation Methods of
Chemical Analysis in the Process of its Production
Yu.M. Kuchirka*†, E.T. Volodarsky‡
† Ivano-Frankivsk National Technical University of Oil and Gas,15, Karpatska st., Ivano-Frankivsk, Ukraine, 76019;
‡ National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, 37, Prosp. Peremohy, Kyiv, Ukraine,
03056; *e-mail: kuchirka@nung.edu.ua
Recieved: February 01, 2019; Accepted: March 01, 2019
DOI: 10.17721/moca.2019.15-20
Невід’ємною складовою сучасної системи управління якістю продукції у цементному виробництві є
передові вимірювальні технології, які базуються на радіаційних методах нейтронно-активаційного,
рентгенофлуоресцентного та рентгенодифракційного хімічного аналізу cкладу речовини. Висока
швидкодія та точність вимірювання характеристик сировини та готової продукції досягається
саме комплексним їх застосуванням, тим самим забезпечуючи зростання рівня автоматизації
цементного виробництва та якості цементної продукції в цілому. Було розглянуто основні етапи
процесу виробництва портландцементу сухим способом, в т.ч. ключові моменти контролю якості
їх проведення шляхом застосування згаданих радіаційних методів хімічного аналізу сировини
та готового продукту. Проаналізовані метрологічні проблеми їх практичного впровадження
у неперервному цементному виробництві, зокрема оцінки невизначеності, статичного та
динамічного калібрування, а також підвищення точності вимірювальних систем, що реалізують
методи нейтронно-активаційного аналізу. Показані напрями їх подальшого удосконалення шляхом
застосування альтернативних нейтронних джерел, використання методів Монте-Карло для
моделювання фізичних процесів нейтронно-активаційного аналізу та машинного навчання для
ефективної обробки спектральних характеристик досліджуваних речовин.
Ключові слова: цемент, нейтронно-активаційний аналіз, PGNAA, PFTNAA, XRF, невизначеність
Контроль якості цементу за допомогою сучасних радіаційних
методів хімічного аналізу у процесі його виробництва
Кучірка Ю. М.*†, Володарський Є.Т.‡
† Івано-Франківський національний технічний університет нафти і газу, 15, вул. Карпатська, м. Івано-Франківськ,
Україна, 76019;
‡ Національний технічний університет України «Київський політехнічний інститут імені Ігоря Сікорського», 37,
просп. Перемоги, м. Київ, Україна, 03056; *e-mail: kuchirka@nung.edu.ua
Надійшла: 01 лютого 2019 р; Прийнята: 01 березня 2019 р
DOI: 10.17721/moca.2019.15-20
Cement quality control by using modern radiation methods of chemical analysis in the process of its production
© Methods and objects of chemical analysis, 2019, Vol. 14, No. 1, 15–2016
Ensuring the high competitiveness of enterprises
in a highly competitive and complex world market
of cement products requires a constant increase in
productivity and quality control of cement production.
To this end, considerable attention is paid to increasing
the speed and accuracy of the quality management
systems for raw materials and nished products at
every stage of cement production. The main stages
of the technological process of production of Portland
cement by the dry method and its quality control can
be described as follows (Fig.1). At the initial stage,
with the help of specialized drilling rigs (1), the
extraction of raw materials in the quarry (2) is carried
out, depending on the predened conditions of its
occurrence and properties. So, before the start of
extraction, the opening, alternation and thickness of
various layers of raw materials, the angle of inclination
of these layers, the length, shape and internal structure
of the deposits, the level of groundwater, and the
like are estimated. Moreover, important parameters
are both the chemical composition and the physical
properties of the raw materials. The latter include:
rock hardness, angle of repose (the angle that can
form a slope of free-owing bulk material in a state of
equilibrium with a horizontal plane), moisture capacity,
water permeability, expansion (degree of volume
increase after extraction compared to the rock mass),
weight, rock strength at its compression and impact.
Drilling machines (screw, roller cone, pneumatic,
hydraulic, and combined), excavators, self-propelled
combines based on rotary excavators (for soft rocks
(chalk, clay)) and others are used as installations for
the extraction of raw materials. To control the quality
of raw materials coming from career, it is periodically
sampled (ex situ method) and there is made a XRF/
XRD chemical analysis [1, 2].
From the quarry (2), the roughly crushed raw
materials are transported to the main crusher (3),
where they are ground into small fractions. Three main
groups of crushers are used for this purpose: impact
hammer crushers; jaw, cone crushers; roll and gear
roll crushers. The principle of their work is based on
the physical methods of crushing and cutting the rock.
The choice of a specic type of crusher is based on the
optimal combination of requirements for its cost, wear
resistance, energy efciency, the maximum allowable
size of the input and output particles, hardness,
tensile strength, and the moisture capacity of the
raw material. The crusher system can include one
or several crushers to gradually achieve the required
particle size of the raw materials. The quality control of
the crushing process is ensured by the design of the
crushers, which are equipped with special output grids
that let through only raw material particles whose size
does not exceed the allowable value.
Fig.1. Quality control of raw materials and nished products in cement production.
© Methods and objects of chemical analysis, 2019, Vol. 14, No. 1, 15–20
Yu.M. Kuchirka, E.T. Volodarsky
17
The raw materials of a career eld can be quite
heterogeneous (non-homogenized), especially if it
comes from different elds. Therefore, to homogenize
it, the crushed raw material is transported by conveyor
to the pre-blending (4). It includes the stacker reclaimer
system and homogenization composition. The stacker
reclaimer mixes the raw materials by placing them
in piles and by gradual selection according to the
Chevron or Windrow methods [3]. The effect of mixing
H is usually determined by the ratio of the standard
deviation of one of the signicant chemical parameters
of the incoming raw materials Sin to the standard
deviation of this parameter of the raw material Sout,
that is: H =Sin/Sout. After mixing and achieving the
desired effect of H, the raw material is accumulated in
the homogenizing storage.
It should be noted that during transportation from
the crusher (3) to the pre-blending (4), all raw material
undergoes chemical analysis (in situ method) using
a PGNAA or PFTNAA string analyzer [4, 5], which
determines its elemental composition with high
accuracy and in real time (every 1-2 min). The obtained
information on the chemical composition of the raw
material allows not only determining the standard
deviation of the concentration of any of its chemical
elements Sin and, thus, setting requirements for the
process of its further homogenization, but also taking
into account the real chemical composition when
mixing its various components. As a result of mixing,
the raw material becomes more homogenized, and the
parameters of its chemical composition are averaged.
To control the quality of the raw material mixing process,
its periodic sampling and chemical XRF/XRD analysis
can be carried out, which allows determining the
standard deviation Sout and, accordingly, the H value
and thereby evaluating the quality of the preliminary
homogenization process. The raw materials from the
composition of homogenization, depending on their
type (marl, limestone, etc.), are transported to the
appropriate raw mix component bins (5). The dosage
of each type of raw material passes automatically
by regulating the speed of conveyor strings from the
output of the corresponding raw mix component bin
to the mixer. Determination of the concentration of all
components of the raw material in the mixture and its
quality control is carried out every 1-2 minutes using
a PGNAA/PFTNAA analyzer. To ensure the stability
of dosing, the obtained values are averaged over a
certain period of time (usually at least 5 values within
5-10 minutes) and serve as the basis of calculations
for adjusting the raw mix component bin in real time.
The raw mix obtained after dosing is ground in a
raw mill (6). Three types of mills [3] can be used: ball,
vertical roller (VRM) and hydraulic roll presses (often in
combination with ball mills). Together with the raw mill
(6), there is a separator (7), which serves to separate
the non-solid product and return it to the grinding mills.
Quality control of the nished raw mix is carried out
every 1-2 hours using the technology of XRF/XRD
chemical analysis. It should be noted that the results of
XRF analysis also serve to automatically correct (every
1–2 hours) shows of the aforementioned PGNAA/
PFTNAA analyzer to ensure maximum accuracy of
dosing of raw materials components. Periodically, in
order to improve the accuracy of dosing, an additional
dynamic calibration of the PGNAA/PFTNAA analyzer
is carried out, based on the average value of its
shows and XRF measurements. For temporary buffer
storage and additional mixing (in order to achieve the
highest degree of homogenization), the raw meal is
transported to the Homogenizing Silos (8). To control
the quality of the homogenization process, a sampling
and XRF/XRD chemical analysis of raw meal is carried
out every 1-2 hours at the silos outlet. After that, it is
fed for heat treatment to a pre-heating tower and pre-
calcining based on a cyclone heat exchanger system
(9). To control the quality of raw meal after preliminary
heat treatment and calcination (decarbonization),
its regular sampling and XRF/XRD analysis are
carried out. Next, the raw meal is transported to the
rotary kiln for clinker burning (10). To create a high
temperature (1700 - 1900 ° C) in a rotary kiln there can
be used traditional types of fuel (coal or natural gas)
as well as alternative sources [6], primarily industrial
and utility wastes (waste tires, solvents, plastic,
impregnated sawdust, wood, paper, cardboard, etc.).
Coal comes from a separate composition (11) for
grinding in a coal mill (12) and enters the rotary kiln
(10) already in the pulverized state. The quality control
of coal is also provided by the of XRF/XRD analysis.
After heat treatment, the clinker enters the clinker
cooler (13) for its quick cooling to a temperature of
50-60° C. At its outlet (every 1 – 2 hours), XRF/XRD
analysis and quality control of the clinker burning
process is carried out, and then it goes to the clinker
silos (14) for temporary storage. It should be noted
that the air heated in the kiln (10) is used in the
process of preparing the raw meal in the system of
cyclone heat exchangers of the pre-heating tower (9),
as well as for its drying in the mill (6). Gypsum (from
a gypsum bin (15)) and other additives (zeolite, slag,
etc.) are added to the clinker, which comes from silos
(14), in predetermined proportions, depending on the
type of cement. Therefore, the mixture obtained is
ground in a cement mill (16) using an air separator
(17), which serves to return non-bulk cement to the
mill for grinding. Every 2–4 hours, there is carried
out the quality control of the nished cement at the
outlet of the cement mill, and the cement is supplied
for temporary storage in cement silos (18) for further
packaging or loading to consumers in bulk form (19).
Before shipment or packaging, nal quality control
of cement is carried out at the outlet of cement silos,
which includes its full XRF/XRD chemical analysis
and evaluation of all parameters in accordance with
national and international criteria for the quality of
cement products.
By analyzing this process of cement production,
Cement quality control by using modern radiation methods of chemical analysis in the process of its production
© Methods and objects of chemical analysis, 2019, Vol. 14, No. 1, 15–2018
we can conclude that its integral component is a
quality management system based on the integrated
application of advanced XRF and PGNAA/PFTNAA
methods for chemical analysis of cement at the main
stages of its production. Thus, the quality of cement
products largely depends on the accuracy and speed
of measuring systems that implement these methods.
At the same time, the practical introduction of XRF/
XRD and PGNAA/PFTNAA methods in the cement
industry is accompanied by a number of complex
metrological issues. These include, rst of all, the
problem of assessing the accuracy of the results of
PGNAA/PFTNAA and XRF/XRD chemical analysis of
raw materials under continuous production conditions
[2, 7, 8]. The main components of the uncertainty of
measurement results by these methods include:
- the uncertainty component associated with the
in situ measuring methods (PGNAA/PFTNAA) and
ex situ (XRF) in the continuous motion of the raw
material, in particular the difference in the averaged
value of PGNAA/PFTNAA analysis measurements for
a certain time interval from the mean value of the XRF
sampling analysis for the same period of time;
- the uncertainty component due to the natural
inhomogeneity of the material being studied in the
stream, as well as the inuence of the processing
procedures of the sample for conducting laboratory
chemical analysis (XRF/XRD);
- analytical uncertainty determined by the
metrological characteristics of the measuring system
itself (including the uncertainty of its static and dynamic
calibration), which implements an appropriate method
of chemical analysis (PGNAA, PFTNAA or XRF),
as well as additional equipment that provides it
with the necessary input parameters for the correct
processing and interpretation of the measured data,
in particular the speed and productivity of the string
conveyor, the temperature of the detecting unit for
the implementation of PGNAA/PFTNAA analysis or
equipment parameters for sampling preparation of
material for its XRF analysis.
- component of the uncertainty that occurs when
comparing the results of PGNAA/PFTNAA and XRF
analysis during the dynamic calibration of PGNAA/
PFTNAA systems based on the results of XRF
analysis, due to the fundamental difference between
their physical principles.
In addition to estimating the uncertainty and
comparing the results of PGNAA/PFTNAA and XRF
analysis under the production conditions, there is
equally important scientic problem, i.e. to increase
its accuracy and speed, which inuences the quality
of raw materials and nished cement products. To this
end, today the most advanced methods of machine
learning, including articial neural networks (ANNs),
[9-10] are being actively implemented. The result of
their application depends to a large extent on the
quality and quantity of the data set for the study of
ANN. Given the complexity of obtaining a large
number of experimental data for this purpose, there
are applied mathematical modeling algorithms, in
particular, the Monte Carlo-Library Least-Squares,
MCLLS [9] and the family of programs for modeling
the transfer of ionizing radiation (neutrons, photons,
electrons) in material systems (Monte Carlo N-Particle
Transport Code, MCNP) [10]. To improve the quality
of the measured signal, advanced digital processing
methods are applied based on Kalman lters (with
XRF analysis [11]), Fourier methods [12] and
wavelet transforms [13], and also empirical modular
decomposition (EMD) for nonstationary processes
(PGNAA/PFTNAA) [14].
It should also be noted that their precision
calibration in production conditions is an integral
part of ensuring the high accuracy of the systems
implementing the PGNAA/PFTNAA chemical analysis
methods. Such a process traditionally includes the
preliminary calibration of the analyzer using standard
samples at the factory, as well as calibration with
the same samples in the production conditions [15].
Moreover, calibration in the production conditions
requires a long stop of the conveyor line of the relevant
process (in some cases up to 48 hours) and is carried
out periodically when replacing radionuclide sources
of neutron radiation (every 2.5 years when using
Californium-252), or unscheduled at a substantial
change in the geometry of measurement or detection
unit. In any case, the component composition, the
form of standard samples, the stationary mode of their
calibration can not fully meet the real conditions of ow
chemical analysis of raw materials in the production,
which causes an additional component of the
uncertainty of such analysis results. As noted above,
for the partial solution of this problem, the averaged
measurements of the PGNAA/PFTNAA systems
(typically 30-60 values) are regularly (every 1-2 hours)
corrected using the XRF analysis. However, the XRF
method requires procedures for preliminary sampling,
milling and pressing of the resulting powder of the
test material on the appropriate equipment. Moreover,
under real production conditions for the XRF analysis,
samples of raw our are taken after grinding in the mill
(6) and this method only determines the composition
of its surface layer, while PGNAA/PFTNAA chemical
analysis is carried out for the entire volume of the
raw material mixtures to the stage of its grinding.
For this reason, dynamic calibration based on XRF
measurements will have additional uncertainty in
PGNAA/PFTNAA analysis results. In addition, such an
approach naturally requires the joint use of PGNAA/
PFTNAA and XRF analyzers, despite their high cost. In
assessing the latter, it is necessary to take into account
not only the initial cost of the measuring instruments
themselves and their metrological characteristics, but
also the high operating costs of them. The latter may
include costs for the regular replacement of expensive
fuel neutron sources (PGNAA/PFTNAA), periodic
calibration and maintenance of analyzers, radiation
© Methods and objects of chemical analysis, 2019, Vol. 14, No. 1, 15–20
Yu.M. Kuchirka, E.T. Volodarsky
19
safety and physical protection during their operation in
accordance with national and international standards.
In addition to improving the accuracy of chemical
analysis of raw materials by applying the most
advanced achievements of digital signal processing
and high-speed electronic systems to implement
them, scientists are actively working to improve
PGNAA/PFTNAA analyzers using MCNP methods for
modeling physical processes based on them [16-19]. In
particular, the direction of using alternative sources of
neutron radiation instead of classical Californium-252
[4, 16, 17], the newest materials and the design of the
main blocks of PGNAA/PFTNAA analyzers [4, 18-20]
is extremely promising. These works are aimed both at
improving the accuracy and precision of measurement,
as well as reducing capital and operating costs when
introducing PGNAA/PFTNAA methods in production.
In particular, in [16], modeling was carried out and
there were shown practical advantages of replacing
a short-lived (half-life 2.65 years) neutron source Сf-
252 with a long-lived Am-Be (half-life 433 years) with a
concomitant modernization of the neutron (moderator)
of high-density polyethylene and with a lead sheath
for radiation protection, which could ensure almost
constant neutron ux throughout the lifetime of the
PGNAA analyzer (20 years) and thereby increase its
accuracy; while in [18, 19], MCNP simulations were
performed on the effective choice of material and
thickness of the reector, absorber and moderator
of the neutron ux in order to increase the reduced
gamma radiation, and hence the sensitivity of the
measurement using the PGNAA method with Сf-252;
there is shown in [20] the use of the MCNP method
for modeling a reector and a moderator when using
the D-T generator [4] as a neutron source, which is
the basis of the PFTNAA technology to increase the
neutron ux density on the sample under study and,
as a consequence, increase the sensitivity of this
method.
Thus, based on the above, we can conclude that
the development of new and improvement of existing
methods for estimating uncertainty, improving the
accuracy of measurement, calibration and design of
PGNAA/PFTNAA analyzers using modern methods
of machine learning and mathematical modeling of
the processes of ionizing radiation with substance is
the actual scientic problem. Moreover, these studies
should be aimed not only at improving the accuracy
and precision of the above-mentioned methods and
their integrated metrological support as part of quality
management systems for cement products, but also
at reducing capital and operating costs in order to
increase the economic attractiveness of using such
technologies in the cement industry.
References
1. Streli, C.; Wobrauschek, P.; Kregsamer, P.
X-Ray Fluorescence Spectroscopy, Applications. In
Encyclopedia of Spectroscopy and Spectrometry;
2016; 707–715.
2. Jacobs, R.; Regis, S. Automated Cement
Plant Quality Control: In-Situ versus Extractive Sam-
pling and Instrumentation. In 2005 IEEE Cement In-
dustry Technical Conference Record; 2005; Vol. 2005,
11–22.
3. Chatterjee, A. K. Cement production tech-
nology: principles and practice. CRC Press, Taylor &
Francis Group: Boca Raton, FL, 2018, 439.
4. Lim, C. S. Recent Developments in Neu-
tron-Induced Gamma Activation for on-Line Multiele-
mental Analysis in Industry. J. Radioanal. Nucl. Chem.
2004, 262 (2), 525–532.
5. Hall, J. C. Some Modern Applications. Rous-
seau. De Gruyter : Berlin, Boston 2015, 138–148.
6. Georgiopoulou, M.; Lyberatos, G. Life Cycle
Assessment of the Use of Alternative Fuels in Cement
Kilns: A Case Study. J. Environ. Manage. 2018, 216,
224–234.
7. Lyman, G.; Asbury, J. Determination of the
Precision of Sampling Systems and On-Line Analy-
sers. TOS forum 2015, 2015 (5), 111-118.
8. Lombard, F. A New Approach to Determining
the Precision and Bias of On-Line Gauges. Chemom.
Intell. Lab. Syst. 2003, 69 (1–2), 77–87.
9. Du, H.; Zhang, R.-X.; Yang, J.-B.; et al. Re-
search on PGNAA Adaptive Analysis Method with BP
Neural Network. Mod. Phys. Lett. B 2016, 30 (32n33),
1650386.
10. Eftekhari Zadeh, E.; Feghhi, S. A. H.; Roshani,
G. H.; et al. Application of Articial Neural Network in
Precise Prediction of Cement Elements Percentages
Based on the Neutron Activation Analysis. Eur. Phys.
J. Plus 2016, 131 (5), 167–174.
11. Nakhaei, F.; Sam, A.; Mosavi, M. R.; et al.
Prediction of XRF Analyzers Error for Elements On-
Line Assaying Using Kalman Filter. Int. J. Min. Sci.
Technol. 2012, 22 (4), 595–601.
12. Pinault, J. L.; Solis, J. The Optimization of
Gamma Spectra Processing in Prompt Gamma Neu-
tron Activation Analysis (PGNAA). Nucl. Instruments
Methods Phys. Res. Sect. B Beam Interact. with
Mater. Atoms 2009, 267 (7), 1139–1148.
13. Barzilov, A.; Kessler, B.; Womble, P. C. Anal-
ysis of 14-MeV Neutron Induced Gamma-Ray Spectra
Using Multiwavelets. Radiat. Meas. 2015, 79, 43–49.
14. Bayat, E.; Afarideh, H.; Davani, F. A.; et al.
Noise Reduction of PGNAA Spectrum Using Empirical
Mode Decomposition Technique. Radiat. Phys. Chem.
2018, 149, 49–53.
15. Oliveira, C.; Salgado, J. Calibration Curves of
a PGNAA System for Cement Raw Material Analysis
Using the MCNP Code. Appl. Radiat. Isot. 1998, 49
(12), 1685–1689.
16. Hadad, K.; Sadeghpour, H.; Nematollahi, M.
R. Source Substitution and Optimization of CBXTM
Cement quality control by using modern radiation methods of chemical analysis in the process of its production
© Methods and objects of chemical analysis, 2019, Vol. 14, No. 1, 15–2020
PGNAA Analyzer in Kangan Cement Plant. Prog.
Nucl. Energy 2016, 90, 204–211.
17. Ghal-Eh, N.; Ahmadi, P.; Doost-Mohammadi,
V. A Quantitative PGNAA Study for Use in Aqueous
Solution Measurements Using Am–Be Neutron
Source and BGO Scintillation Detector. Nucl. Instru-
ments Methods Phys. Res. Sect. A Accel. Spectrome-
ters, Detect. Assoc. Equip. 2016, 808, 123–127.
18. Yang, J. B.; Tuo, X. G.; Li, Z.; et al. Optimized
Design of a PGNAA System for Cement Analysis
Using Monte Carlo Simulations. Nucl. Technol. 2013,
184 (2), 233–238.
19. Hadad, K.; Nematollahi, M.; Sadeghpour,
H.; et al. Moderation and Shielding Optimization for
a 252Cf Based Prompt Gamma Neutron Activation
Analyzer System. Int. J. Hydrogen Energy 2016, 41
(17), 7221–7226.
20. Shan, Q.; Chu, S.; Jia, W. Monte Carlo Sim-
ulation of Moderator and Reector in Coal Analyzer
Based on a D-T Neutron Generator. Appl. Radiat. Isot.
2015, 105, 204–208.
