en ru

Projects

This page lists my projects (implemented and under developed). The aim, description and colleagues who helped me are given.

Table of Contents (TOC)

  1. Optical analyzer of granulometric composition
  2. Optical analyzer of quality of processing of mineral fertilizers with conditioning additives
  3. Analysis of mineral fertilizers by "fingerprints" technique (energy disperse X-ray fluorescent analysis of pressed granules)
  4. Modified salt index parameter for mineral fertilizers
  5. XRF analysis of mineral fertilizers with optical modification
  6. Flow analysis of phosphoric acid with ED XRF method
  7. Optical identification of the particle size of the pressed powder
  8. Using the concept of "corrected" analytical signal for metrological characteristics calculation of multidimensional methods of data analysis in analytical chemistry
  9. Optoelectronic microscope to study the properties of mineral fertilizer granules
  10. Techniques
  11. Research and Testing Center "Control Technology"
  12. Specializations (additional education and training)

1. Optical analyzer of granulometric composition

Back to TOC

automatic optical analyzer of granulometric composition automatic optical analyzer of granulometric composition
Fig. 1. Automatic optical analyzer of fertilizers granulometric composition (AGS) for conveyor belt and closed pipe. 1 - conveyor belt; 2 - rotary selection device; 3 - vibrating system; 4 - block of optical analysis.

The aim

Operational control of granulometric composition of industrially produced mineral fertilizers (optoelectronic method on the basis of machine vision with Python and OpenCv). It allows to reduce the amount of produced waste (fraction < 2 mm) and increase the controllability of the production process.

Brief project description

The system of the automated online control in industrial conditions is offered. The device consists of three independent parts: the system of sampling from the conveyor or pipe, the system of sampling to the field of analysis and the block of optical detection of granules.

Features of the design include online analysis of granulometric composition, the ability to analyze the shape and color of granule, work in industrial conditions and data transfer to the factory information system PISystem (SCADA).

Sampling is carried out by a robotic rotary system. Then, the sample is fed into the analysis area by the linear vibrations of the sample feed system. The entire sampling and feeding scheme is controlled by a computer, which is integrated in the optical block.

The algorithm for calculating the parameters of granule consists of obtaining an optical-electronic image (three-dimensional RGB matrix of pixel intensities), image preprocessing (smoothing, binaryization, morphology), calculation of closed contours, approximation of the found contours by ellipses and calculation of ellipses parameters (long and short axis, axis ratio and average RGB color intensity inside the ellipses).

All stages of obtaining and processing of the information are automated and implemented on Python 3.7 programming language (with PyQt5 UI). This devices successfully pass industrial tests and were deployed in JSC "Apatite" fertilizer facility (Cherepovets city, Russia).

Presentation of the project

Participants and they contribution

  • D.V. Yunovidov - idea development, software writing, project implementation, administration of devices (network, data transfer, operation system etc.).
  • V.A. Shabalov - creation of the device prototype (development and production of the vibration system of sample feeding and optical analysis unit).
  • K.A. Menshikov - implementation of electrical circuits and low level commutation.
  • M.N. Nadezhin, Y.A. Agafonov and V.V. Sokolov - assistance of project implementation in industrial practice.

2. Optical analyzer of quality of processing of mineral fertilizers with conditioning additives

Back to TOC

optical analyzer for the degree of processing of CA
Fig. 2. The optical analyzer of the conditioning additives quantity on granules. a) look of the device; b) photo of granules in ultraviolet (UV) light (the processed granules are fluorescent, the automatically recognized granules are circled by green contours).

The aim

Operative control of the degree of mineral fertilizers conditioning additives (CA) covering (optoelectronic method), which allows to reduce consumption of conditioning additives (expensive raw materials).

Brief project description

The method and device for control of the degree of mineral fertilizers CA covering was developed. The method consists of a specially designed analogue of the "darkroom" device. For illumination the light-emitting UV diodes is used (depends of CA type, in common case it is 360 nm).

The peculiarities include small time and cons of analysis, new "digitized" quality metrics for CA processing (uniformity of treatment), additional determination of granulometric composition, work in industrial conditions and data transfer to the SCADA factory information system (PI System).

The device may include built in computer or use laptop. Management and calculations are carried out through the special software "CA Calc" (Python 3.7). The offered device and technique allows to receive the digital data about CA processing quality.

The algorithm for calculating granule's parameters consists of selection of a representative sample of fertilizers (about 50 g., proof); measurement of granules in a special cuvette; granules recognition (similar to the project No 1) and calculation of the completeness of CA processing (the percentage of fluorescent granules VS others granules, as well as specific brightness of the image).

Presentation of the project

The contribution of the participants

  • D.V. Yunovidov - idea development, software writing and project development.
  • V.A. Shabalov - creation of a device prototype.
  • E.E. Sidorova - assistance in preparation of project presentations and writing scientific articles.
  • M.N. Nadezhin - assistance in implementation of the project in production practice.

3. Analysis of mineral fertilizers by "fingerprints" technique (energy disperse X-ray fluorescent analysis of pressed granules)

Back to TOC

analysis and identification of the fertilizer manufacturer
Fig. 3. "Fingerprint" analysis and identification of the fertilizers. a) software with clusters of fertilizers (manufacturer and brand); b) prepared fertilizer sample (pressed granules on boric acid substrates).

The aim

Express identification of mineral fertilizers (MF) manufacturer and "standartiness" with energy-disperse X-ray fluorescence analysis (ED XRF).

Brief project description

Next tasks were set and solved: * determine manufacturer based on XRF spectra; * determine if sample is related to good one ("standartiness"); * write software for specific XRF spectra processing; * develop technique for unknown sample identification; * describe XRF spectrometers for such project.

The described method allows to analyze different types of fertilizers by they "fingerprints" (using a unique ED XRF spectrum of each fertilizers type). This make possible to simplify the identification procedure (both for manufacturer and "standartiness").

The main source of information in this work is the ED XRF spectrometers ("REAN", "X-Spec" and "Panda" of JSC "Scientific Instruments"). Because of multidimensional clusterization, it is possible to simple press fertilizer granules on a boric acid substrate. Such samples are measured under the same conditions on a specific device. The entire spectrum was used as markers to classify the object (each energy channels, dispersions of channels in peaks and total dispersion). We can include in calculation obvious and not obvious properties of sample, with such approach was used.

It was found, that the studied fertilizer classes are perfectly divide into plane (2 main components). We accurate identified all encrypted objects (some of "PhosAgro" fertilizers) with using projection into two main components and \(chi^2\) statistics.

Next, special software was written, which allows to carry out all described approaches (include work with XRF spectrometers). For example, it automatically calculate distances to centers of clusters and chi-square statistics for unknown sample. Manufacturer and "standartiness" related to the smallest distance and \(chi^2\) value of pair "unknown sample - center of cluster".

The total time of analysis of one sample does not exceed 5 minutes. For analyses you need only boric acid and fertilizer granules.

Presentation of the project

Participants and they contribution

  • D.V. Yunovidov - idea development, software writing and project development.
  • A.S. Bakhvalov - ED XRD spectrometers supply for research.

4. Modified salt index parameter for mineral fertilizers.

Back to TOC

software registration certificate Fig. 4. Certificate of "Salt index calculator" software (modified salt index include).

The aim

Modification of the salt effects parameter of fertilizers (how strongly they absorb moisture from the soil and plants).

Brief project description

"Salt index" parameter already exist in scientific literature. However, this parameter, which was proposed by J. Mortvedev in 1943, is outdated and confusing a lot. Mineral fertilizers and industrial production has changed a lot since 1943 (e.g. types of production lines and fertilizer's brands). Moreover, you can find in literature, that there is a lot of misleading in calculation of this parameter (based on osmotic pressure or conductivity both soil solution and distilled solution).

It was decided to modify this parameter in view of modern requirements. Make its calculation more accurate and representative as well as make experimental technique of calculation (otherwise, we can't certify such method).

At the time it was created, the parameter reflects the osmotic pressure of 0.1 \% soil solution of fertilizers. But since that time, more and more scientist just measure conductivity of soil solutions (or even distilled solution) of fertilizers. Nowadays, salt index have a lot of table for major fertilizer's components to calculate it (in general, components with content more than 10 wt. % are used). From our point of view it is necessary to include all fertilizer's components with than 1 wt. \% of total mass (we propose correction for phase composition of mineral fertilizer, the "modified method"). In addition, the technique was developed, which is based on conductivity of soils solution.

The software "Salt Index Calculator" is written and registered for the offered "modified" theoretical method. The practical approach is implemented in the industrial technique in JSC "NIUIF".

Presentation of the project

  • Software "Salt Index Calculator", which was registered in the Software Federal Register of Russia.
  • The project of industrial technique of salt index measurements.

Participants and they contribution

  • D.V. Yunovidov - software writing and project development.
  • M.N. Nadezhin - help in writing software.
  • V.V. Sokolov and O.A. Abramova - assistance in the implementation of measurement technique.

5. XRF analysis of mineral fertilizers with adjustment of optical detected powder particles

Back to TOC

combined optical and ED XRF analysis Fig. 5. Principle of combined optical and X-ray fluorescent analysis of mineral fertilizers.

The aim

Comprehensive express analysis of mineral fertilizers for physical and chemical properties (explicit and implicit, such as nitrogen content, fractional composition and preliminary drying condition).

Brief project description

Statistical way for big data analysis of chemical and physical features of mineral fertilizers are considered (about 500 objects and 30 features). The range of such features is wide and combination of the ED XRF method and optical recognition of the sample surface were used to obtain these features. Multidimensional regression and classification algorithms quality with minimal "manual" data preparation is investigated. Optimal operation parameters for all considered algorithms are selected.

As a result, it was possible to determine the implicit properties of fertilizers, such as the nitrogen content (the signal of this element does not appear in the ED XRF spectrum), the preliminary drying condition of the sample (was dried or not) and the particle size of the pressed powder (Table 1).

Table 1. Predictions results of physical and chemical properties of mineral fertilizers.

N P K S Fraction Was dried or not
Values
Range
[0; 16] mass % [15; 52] mass % [0; 20] mass % [0; 20] mass %
granules;
pressed: granules or
powder < (500 mkm,
or < 100 mkm)
[0, 1]
Classification, F-metric (%)
Linear
99.31 99.78 99.59 99.56 92.40 72.94
Linear With L1
99.65 99.78 99.57 98.87 92.51 73.08
Linear With L2 99.65 99.78 100.0 98.99 91.33 68.46
Random
Forest
100.0 100.0 100.0 98.99
98.40 73.37
Most
Signif.
Feature
Cl (15.56)
Ca (17.04) Cl (14.31) Ca (20.47) counters
area (25.70)
P (11.49)
Regression, average absolute deviation (mass %)
Linear
0.39 1.12 0.29 0.70 - -
Linear With L1 0.60 1.32
0.42 1.04 - -
Linear With L2 0.36
1.11
0.29 0.70 - -
Most
Signif.
Feature
K (20.19) Sr (20.17) background
area (25.44)
K (17.46) - -

Presentation of the project

Participants and they contribution

  • D.V. Yunovidov - development of an idea, writing software, scientific articles, patent and project development.

6. Flow analysis of phosphoric acid with ED XRF method.

Back to TOC

flow cell for ED XRF Fig. 6. Energy dispersive (ED) X-ray fluorescence (XRF) spectrometer with a flow cell.

The aim

Rapid analysis of phosphoric acid in industrial conditions (P, S, Ca and Fe elements).

Brief project description

An algorithm for continuous X-ray fluorescence analysis of extraction phosphoric acid (EPA) was developed. The proposed approach was tested for extraction technological process of phosphoric acid from apatite (calcium, sulfur and phosphorus were determined). The proposed approach allows to control the quality of EPA before the neutralization stage (right before the mineral fertilizer is obtained).

The detection limits are 0.1 wt. \% for P, 0.4 wt.\% S, 0.1 wt.\% for Ca and 0.004 wt\% for Fe. Relative standard deviations for characteristic lines did not exceed 4.35 \%.

Moreover, the possibility of suspension detection in EPA was shown (by Ca and S signals). Software for calculations and spectrometer's control was created.

Presentation of the project

Participants and they contribution

  • D.V. Yunovidov - idea development, software writing and project implementation.
  • A.S. Bakhvalov - design and manufacture of cells.
  • V.V. Sokolov - assistance in implementation of the project in production practice.

7. Optical identification of the particle size of the pressed powder.

Back to TOC

optical particle size analysis
Fig. 7. Example of prepared samples and a computer processed sample surface to classify by particle size.

The aim

Powder's particles size automation detection method was developed. The source of information is the optical-electronic image in the RGB system (which is normalized to brightness, temperature and gradient of colors).

Brief project description

The way of powder's particles size classification is describe. The researched objects are industrially produced mineral fertilizers. Samples with different granules size (less than 100 microns, less than 500 microns and 2-5 mm granules) and different chemical composition (about five brands of mineral fertilizers) were considered. The sample's particles have an irregular shape close to spherical. Preliminary press in boric acid substrate was used to improve the accuracy of the analysis and eliminate the influence of the particle shape.

The essence of the proposed method is to obtain an optical-electronic image of the object with a resolution of at least 640x480 pixels (three-dimensional pixel intensity matrix in the Red-Green-Blue system). The area of the analysis is allocated from the image and converted into grayscale (two-dimensional matrix of pixel intensities with the resolution of at least 200x200 pixels). Next, proposed differentiation procedure allows to eliminate the influence of external illumination (gradient, temperature and brightness). The result is a "surface map" of the sample, which is reflected the defects of the pressed structure (patterns, which are responsible for the particle size of the pressed particles). Finally, samples are classified by the particle size, according to the patterns on the "surface map". Four classification algorithms (linear, linear with L1 and L2 regularization and nonlinear random forest algorithm) were studied as well.

All proposed approaches are automated and implemented in Python 3.7 programming language.

Presentation of the project

Participants and they contribution

  • D.V. Yunovidov - idea development, software writing and project implementation.

8. "Corrected" analytical signal concept for metrological characteristics calculation in analytical chemistry in case of multidimensional data analysis

Back to TOC

corrected analytical signal
Fig. 8. Example of the "corrected" analytical signal's concept in case of multiple regression for P determination in crystalline NP fertilizers.

The aim

Development of a concept, which allows to calculate the "corrected" analytical signal (using multivariate statistical methods) and use it in development and certification of chemical analysis techniques (with classical metrological approaches).

Brief project description

The concept of converting the multi-dimensional statistics results into the usual dependence "analytical signal - concentration" is proposed. The present approach will allow to develop techniques, which use multivariate statistical analysis with classical metrological approaches (e.g. RMG 61-2010 (in Russ.) etc.).

Presentation of the project

  • Industry-level techniques of analysis of \(AlF_3\) and crystalline MAPh (NP 12-52 fertilizer), which are used in JSC "NUIUF".

Participants and they contribution

  • D.V. Yunovidov - idea development.

9. Software for opto-electronic USB microscope to study the properties of mineral fertilizer granules.

Back to TOC

image from optical USB microscope Fig. 9. Example of caking granules (in software).

The aim

Develop of software, which allow to observe, record and analyze objects in time (color and surfaces).

Brief project description

Special software for objects analysis was developed. It allows to record video, analyze color, moisture (water drops) and contact area of granules (in total and in selected areas). Size of USB microscope allows to use it in climatic camera as well.

We was be able to show the effect of the "ring pore" with such approach (contact of two granules where moisture is accumulated and they stick to each other).

Presentation of the project -

Participants and they contribution

  • D.V. Yunovidov - idea development, software development and project implementation.

10. Techniques.

Back to TOC

techniques
Fig. 10. Example of using a corrected analytical signal after a multidimensional regression to determine fluorine in AlF3 (ED XRF).

The aim

Development and certification of mineral fertilizer techniques of analysis (at the industrial and state level).

Brief project description

Several of techniques were developed due to my regular laboratory work (Table 2). Furthermore, main part of them were certified at the industrial and state level. The methods were developed in accordance with Russian standard "RMG 61-2010".

Table 2. Developed techniques.

Title of method Basic method
Measuring range,
wt. %

Relative accuracy
value, ±δ, %
Analysis of fluorine in industrial aluminum fluoride (AlF3)
ED XRF 93.0 – 98.0 1.1
Analysis of Si impurities in industrial AlF3 ED XRF 0.08-0.40 24
P2O5 analysis in crystalline MAP ED XRF 48.0 – 62.0 1.0
Analysis of P, S, K in PKS fertilizers Wave XRF P2O5 12.5 – 25.0
K2O 13.5 – 33.0
S 4.0 – 12.0
1.5
3.0
3.6
P2O5 analysis in apatite concentrate ED XRF 37.0 – 40.0 1.2
Analysis of Al2O3 in nepheline and syenite concentrates ED XRF 26.0 – 30.0 1.0
Salt index analysis in fertilizers Potentiometry - -

Presentation of the project

  • JSC "NIUIF" everyday laboratory practice.

Participants and they contribution

  • D.V. Yunovidov - development of the techniques concept, collect of experimental data, statistic calculations.
  • E.V. Krylova - writing the text of techniques, statistic calculations.
  • O.A. Abramova - final verification the text of techniques, assistance in techniques development for industrial and state verification.

11. Scientific and research center "Control Technology"

Back to TOC

SRC
Fig. 11. Laboratory model for scientific and research center.

The aim

Creation of a scientific and research center (SRC) on the basis of the Department of Chemical Technology of Cherepovets State University (ChSU) to provide independent quality control of industrial products with an accent on international techniques and recommendations.

Brief project description

Scientific work, skills improvement (of outside and inside laboratory staff and students of CSU) are planned on the basis of this center. Carry out contracts with outside companies to provide flexible quality control of their products.

The payback of the project was calculated. The estimates of laboratory preparation and repairs were made. The project was approved by the technical and academic councils of ChSU. The repair of the laboratory should be completed by September 2020. Active work is planned for October 2020.

Presentation of the project

  • Councils of ChSU (technical and academic).
  • 1st place in the corporate competition "Young Leader" at JSC "Apatit" and 3rd place throughout "PhosAgro" Corporation (in Russ.), 2019, Vologda Region, Russia.

Participants and they contribution

  • D.V. Yunovidov - developing and implementation of the project. Organize of laboratory preparation and repair.
  • E.V. Tselikova - Vice-Rector for Science of ChSU, active assistance in promoting the project.
  • O.A. Kalko - Chief of Chemical Technology department of ChSU, active assistance in promoting the project.
  • K.V. Aksenchick - Assistant Professor of Chemical Technology department of ChSU, active assistance in promoting the project.

12. Specializations and certificates of additional education

Back to TOC

certificate certificate
certificate
Fig. 12. Final certificates.

The aim

Gaining Big Data Science skills (Python 3.7 and R).

Brief project description

  • "Machine learning and data analysis" (one year, online, Yandex and Moscow Institute of Physics and Technology, 6 courses). The diploma project consisted in realization of automatic definition of emotional of feedbacks on market.yandex.ru (sentimental analysis). The project is implemented as a web-service on flask (the service is launched locally).
  • Specialization "Data Science" (one year, online, Johns Hopkins University, 10 courses). The diploma project consisted in realization of user inputs auto-compleat (sentimental analysis). The project was implemented as a web-service on R (the service is launched locally).
  • Specialization "Data science" from Microsoft (one year, online, 11 courses). The diploma project was to study of rent loans in the USA.

Presentation of the project -

Participants and they contribution

  • D.V. Yunovidov - passing all courses and make all diploma projects.