Young scientists’ researches and achievements in science

DONETSK NATIONAL TECHNICAL UNIVERSITY

Faculty of Computer Science and Technology English Language Department

(YOUNG SCIENTISTS’ SCIENTIFIC AND TECHNICAL CONFERENCE)

April 16, 2020 DONETSK

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

ГОСУДАРСТВЕННОЕ ОБРАЗОВАТЕЛЬНОЕ УЧРЕЖДЕНИЕ ВЫСШЕГО ПРОФЕССИОНАЛЬНОГО ОБРАЗОВАНИЯ ДОНЕЦКИЙ НАЦИОНАЛЬНЫЙ ТЕХНИЧЕСКИЙ УНИВЕРСИТЕТ ФАКУЛЬТЕТ КОМПЬЮТЕРНЫХ НАУК И ТЕХНОЛОГИЙ КАФЕДРА АНГЛИЙСКОГО ЯЗЫКА

СБОРНИК ДОКЛАДОВ НАУЧНО-ТЕХНИЧЕСКОЙ КОНФЕРЕНЦИИ ДЛЯ МОЛОДЫХ УЧЕНЫХ «YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE» (16 АПРЕЛЯ 2020)

Донецк ДонНТУ, 2020 2

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

УДК 001.8 ББК 72 Н34 Young scientists’ researches and achievements in science: сборник докладов научно-технической конференции для молодых учёных (Донецк, 16 апреля 2020 г.) / ответств. за вып. Е.Н. Кушниренко. – Донецк: ДонНТУ, 2020 – 160 с. Сборник подготовлен по результатам научно-технической конференции «Young scientists’ researches and achievements in science». Материалы, вошедшие в сборник, представлены студентами, магистрантами и аспирантами ГОУ ВПО «Донецкий национальный технический университет» и другими зарубежными авторами. Статьи печатаются в авторской редакции. Рекомендовано к изданию на заседании совета факультета компьютерных наук и технологий. Протокол №4 от «22» мая 2020г. Председатель оргкомитета конференции: Соснина Л.В. – д.филол.н., доцент, профессор кафедры английского языка ДонНТУ Члены оргкомитета: Каверина О.Г. – д.пед.н., профессор, зав.кафедрой английского языка ДонНТУ Горбылева Е.В. – доцент кафедры английского языка ДонНТУ Гильманова Р.Р.– ст. преподаватель кафедры английского языка ДонНТУ Ответственный секретарь оргкомитета: Кушниренко Е.Н. – ст. преподаватель кафедры английского языка ДонНТУ Адрес оргкомитетa: г. Донецк, ул. Артема, 131 Донецкий национальный технический университет, 11-й учебный корпус, факультет компьютерных наук и технологий, кафедра английского языка, ком.

248. Справки по телефонам: (062)301-03-74. e-mail: kaf_engl-2017@mail.ru © Авторы статей, 2020 © ГОУ ВПО «Донецкий национальный технический университет», 2020 3

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

CONTENTS

Allaberdina A. Z. ......................................................................................................... 6 FEATURES OF CONSUMPTION OF YOUTH SLENG IN MODERN RUSSIAN AND ENGLISH LANGUAGES Babak B. N., Vygovskaya D. D., Vygovsky D. D., Kaverina O.G. .............................. 9 DETERMINATION OF THE COMPLEX OF SOCIO-ECONOMIC INFORMATION IN THE DESIGN OF TECHNOLOGICAL SCHEME OF A COAL MINE Bibik A.S., Tartakovsky Ye.A., Sarmakeeva A.S., Plesovskikh K.Yu., Revina N.V. 12 AUTOCORRELATION OF RETURNS IN MAJOR CRYPTOCURRENCY MARKETS Bydash A.R., Vinnichenko N.G., Kushnirenko Ye.N. .............................................. 25 OVERVIEW OF METHODS AND TECHNOLOGIES OF PRODUCING OXYGEN-SENSITIVE ELEMENTS FOR MEASURING DISSOLVED OXYGEN IN A RESERVOIR Chernyshov B.S., Fedyaev O. I., Girovskaya I.V. ..................................................... 29 MODERN SOLUTION TO FACE RECOGNITION PROBLEMS BASED ON NEURAL NETWORKS Frolov I.V., Sokolova O.V. ......................................................................................... 35 THE DEVELOPMENT AND USE OF ARTIFICIAL INTELLIGENCE TECHNOLOGIES Gallyamova E. D. ........................................................................................................ 38 INFLUENCE OF INTERNET RESOURCES ON MODERN COMMUNICATIONS Kharadzha T.A. ........................................................................................................... 41 DIRECTIONS FOR IMPROVING LOGISTIC SYSTEMS IN THE FOOD INDUSTRY Kolbasov S.Y., Kushnirenko Ye.N. ............................................................................ 47 SOLVING THE PROBLEM OF PATTERN RECOGNITION USING CONVOLUTIONAL NEURAL NETWORKS Kolomoets A.S., Skazhenik V.B., Kaverina O.G. ...................................................... 52 COMPUTER MODELING OF MINING DEVELOPMENT STRATEGIES IN COAL MINES Komarichev R.E., Girovskaya I.V. ............................................................................. 56 GENERAL-PURPOSE COMPUTING ON GPU WITH CUDA Kosogor D.V., Borshсh I.V. ....................................................................................... 62 THEORETICAL ASPECTS OF MODERN STRATEGIC MANAGEMENT OF THE ENTERPRISE Kukhta S.S., Sokolova O.V. ...................................................................................... 68 THE FIVE-LEVEL DATA PROTECTION MODEL Kupich V.A., Butuzova L.F., Boyko V.N. .................................................................. 72 THE PECULIARITIES OF THE LOW-TEMPERATURE SEMI-COKE PROCESS AND METHODS INFLUENCED THE YIELD AND PROPERTIES OF ITS PRODUCTS 4

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Kurchenko E.N., Kalinihin O.N., Boyko V.N. .......................................................... 76 CO-DISPOSAL OF WASTE FROM COKE PLANTS AND MUNICIPAL SOLID WASTE COMPONENTS Lyutova E.I., Kolomoytseva I.A., Gilmanova R.R..................................................... 81 OVERVIEW OF SPAM FILTERING ALGORITHMS Mokrushin D.A, Bykov V.V. ...................................................................................... 86 RESULTS OF EXPERIMENTAL STUDIES OF BRAKING QUALITIES OF CARS ON THE LINE OF INSTRUMENTAL CONTROL Nemov G.Y., Neyezhmakov S.V., Paniotova L.N. ..................................................... 94 AUTOMATED CONTROL SYSTEM OF AIR DISTRIBUTION IN THE MINE VENTILATION SYSTEM Pilipenko A. S., Girovskaya I. V. ............................................................................... 99 BEST DISCOVERIES OF YOUNG SCIENTISTS IN RUSSIA Rudak L.V., Fedyaev O.I., Girovskaya I.V. ............................................................. 105 STUDENT ATTENDANCE CONTROL SYSTEM BASED ON NEURAL NETWORK RECOGNITION Rusakov V.O., Petrenko U.A.. Boyko V.N. .............................................................. 112 SUBSTANTIATION OF THE MINING WORKING’S CONSTRUCTION ON THE BASIS OF ANCHORS, ATTACHED TO THE MINING ARRAY WITHOUT BINDING COMPOSITIONS Savchenko L.A., Borshсh I.V. .................................................................................. 118 ESSENCE AND TYPES OF STRATEGIES FOR THE DEVELOPMENT OF CARBON INDUSTRY ENTERPRISES Smeshnaya A.V., Kushnirenko Ye.N. ..................................................................... 122 MAGNETIC FIELDS USED IN ELECTROMAGNETIC FLOW METERS Strizhko M. A., Sukov S. F., Paniotova L. N. .......................................................... 127 SMART TRAFFIC LIGHT CONTROL SYSTEM FOR URBAN TRAFFIC FLOWS Sukhanov A.A., Sokolova O.V. ............................................................................... 130 THE ROLE OF AUTOMATION IN THE MODERN WORLD Tomilov M. K., Prilipko Y. S., Boyko V. N. ............................................................ 133 NANOSTRUCTURED CERAMIC FUNCTIONAL MATERIALS Yakhina I.E. .............................................................................................................. 138 NEOLOGISMS AS A MEANS OF LANGUAGE DEVELOPMENT Yasinskaya Ya.O., Sokolova O.V. ............................................................................ 141 THE DEVELOPMENT OF COMPUTER TECHNOLOGIES IN RUSSIA Yasnitsky M.V., Vasyaeva T.A., Revina N.V. ......................................................... 144 WEB-ORIENTED SYSTEM FOR OPTIMUM SELECTION OF TEMPORARY ACCOMMODATION Zaglada E.A. ............................................................................................................ 149 ANALYSIS OF INDUSTRIAL ENTERPRISES SOCIAL RESPONSIBILITY SYSTEM IN THE DONETSK PEOPLE'S REPUBLIC Zemlianskiy D.A., Kolomoytseva I.A., Gilmanova R.R. ......................................... 154 CLASSIFICATION OF TEXTS 5

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

UDC 81'27 FEATURES OF CONSUMPTION OF YOUTH SLENG IN MODERN RUSSIAN AND ENGLISH LANGUAGES Allaberdina A. Z. alllaberdina@mail.ru

Abstract. Slang is a kind of vocabulary that is characteristic for a certain circle of people, interconnected by age, profession, gender, etc. The most popular is youth slang – these are words that have a neutral meaning, or have an emotional (positive or negative) color. Studies have shown that youth slang in the English language has influenced the Russian language, since most of the words in Russian youth vocabulary are borrowed from the English language.

Keywords: slang, jargon, vocabulary, youth slang, word.

Slang is one of the most interesting language systems in modern linguistics. The study of the concept of «slang» was carried out by many scientists: I.V. Arnold, V.A. Khomyakov, T.V. Belyaeva, M.M. Makovsky, E. Partridge, O. Jesperson, H. Mencken, S.I. Hayakawa, R. Spears and others. According to most linguists, slang is a common word in a state of constant change, which is characterized by emotionality and expressiveness.

It comes from the English slang «jargon», from an indefinite form. This word is found in 1756 in its meaning. «Dictionary of thieves and vagrants»; since 1801, applies to the jargon of any professional or social group.

The origins of Russian slang lie in the school jargon of the mid-19th century, when the creation of isolation in closed educational institutions led to the formation of a kind of vocabulary. N.V. Gogol gave the first name to youth jargon, calling it a «technical word» Slang is popular among young people up to 30-35 years old. The meaning of youth slang words is often incomprehensible to older people. Youth slang is lively, popular, large in composition and motley. There are words that are specific to a certain age, social group or cultural area.

It is necessary to distinguish between the concepts of slang and jargon. Jargon means a more established expression, many of the words in it are very old. Slang appears in a certain group and is available to a certain circle of people. Slang is widely used among young people, as young people often come up with new words, their own unique and understandable language.

The reason for the popularity of jargon among young people is that in this way they try to attract attention, to assert themselves.

6

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

According to M. Goldenkov, young people from poor neighborhoods and slums, getting into high society, use slang, and the people around them remember this. Youth slang is found among fans of music artists, especially rock music performers [1].

Unlike jargon, slang selects popular words from the jargon of various groups of society. Slang differs from jargon in its massive use. Slang can be used by people of different groups, different educational status and professions. These can be people with a criminal past, as well as cultural and intelligent. Slang usually reflects recent trends in the vocabulary of new supermodels and is often associated with youth speech.

It should be noted that young people quickly remember and actively use slang in real life. Usually, young people get information about slang while watching movies, when communicating on social networks.

Among the words of the Russian youth slang, words of a neutral meaning are common, such as: «nerd» – «a little thing», «go ahead» – «say», «voice», «Technician» – «a person who is inclined to technical sciences», «Humanitarian» – «a person who likes humanitarian objects», «crap» – «lies», «stir up» – «make friends», «scoreboard» – «face».

Slang can express an emotional assessment: positive or negative. So, being used as an emotionally positive assessment, it gives brightness and expressiveness to speech: «crash» – «the person you like», «hype» – «fashionable», «stir up» – «organize», «anneal» – «have fun time», «Otpad» – «good, best, beautiful» – «well done», etc.

Slang, which has an emotionally negative connotation, refers in most cases to crude slang: «roll cotton» – «idle», «Zashkvar» – «something stupid, lost its relevance», «kipish» – «disassembly», etc.

Youth gross slang expresses the desire to reduce the quality of the face, its social significance. So, in youth speech, such words as «teacher» – «teacher», «cop» – «policeman» as a degradation of a person’s status, reduction of his social status.

Currently, English is one of the most widely spoken languages, which is why borrowed youth slang from the English language most often comes to modern Russian.

The lack of a standardized translation from English has led to the emergence of so many youth slang.

For example, the word «loser»№» – (from the English «luser») – «loser», Nice – (from the English «nice») – «beautiful, good, beautiful, pretty», «all shook up» – «shucher, vanity», «easy» – (from the English «easy») – «easy», «Stonik» – clothes of the Stone Island brand, «Flip» – (from the English «flip») – «coup, somersault», «check in» (From the English «Check in») – Register. Designate your location.

The English language is characterized by the use of youth slang. Many youth slangs in the English language are derogatory, offensive. Usually slang expresses a negative value in describing a person’s quality: «chicken» – «coward», «bigmouth» – «talker», «egghead» – «egghead», «deadhead» – «bore» or actions, actions of a person: «goof off» – «fool around», «wasting time», «show off» – «show off, show off», «beat one's brains» – «puzzle; brains roll to one side», «bite the bullet» – «swallow the pill; to reconcile», «be nuts about» – «to be crazy about smth.; go crazy on smb.», «blow the whistle on» – «knock on smth; to lay smb., to inform the interested parties», «pin 7

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

smb’s ears» – give on the brain; set the heat / bashing», «take a powder» – «pull up stakes; to rinse off», «sell a pup» – «to foist junk», «piece of cake» – «just to spit, a couple of trifles».

Slang is not only words that can humiliate a person, but also praise him. So, the positive qualities of a person can be expressed in slang words in Russian: «lighter» – «a cheerful sociable girl who knows how to cheer up others», «high» – «good, beautiful, excellent», in English: «awesome, bomb, biggity / diggity, bommy, fantabulous» – «amazing, amazing».

English slang is characterized by brevity, therefore it is often used in the form of abbreviations, especially: gonna – going to, wanna – want to, ye – yes, U – you, dunno – don't know. These abbreviations are often found in text messages, and this slang is called mobile slang: AFAIR – as far as I remember, as far as I can remember, AGF – assume good faith – let's hope for the best, BBIAB – soon to be back – I'll be back soon, BFF – best friends forever – Best friends forever, FYEO – just for your eyes – just for you, OMG – Oh, my God! So, slang is an integral part of the life of young people, gives speech emotionality and expressiveness. The study allows us to conclude that English youth slang had an impact on Russian slang, since most of the Russian words are borrowed from English.

It should be noted that English youth slang is well known and widely used. However, the slang of Russian youth has the most diverse semantic connotations and many words are incomprehensible to the majority of the population. Of course, slang plays a very important role in the life of young people, as it helps young people to communicate with each other and facilitates the process of learning borrowed English vocabulary, thereby expanding the vocabulary.

References 1. Голденков М. А. «Осторожно HOT DOG! Современный активный english». 2-е издание, испр. и доп. – М.: ЧеРо, 1999. – 272с.

2. Маковский М.М. Современный английский сленг: Онтология, структура, этимология. – Изд. 6, стереотип. – URSS. – 2020. – 168 с.

3. Хомяков В.А. Введение в изучение слэнга – основного компонента английского просторечия. – Изд. 2. – URSS. –2009. – 104 с.

4. Jesperson O. Growth and Structure of the English Language / O. Jesperson. – London, 1998. – 98p. Аннотация. Сленг – своеобразная лексика, характерная для определенного круга людей, связанных между собой возрастом, профессией, гендером и т.д. Наиболее популярным считается молодежный сленг – это слова, которые имеют нейтральное значение, или имеют эмоциональную (положительную или отрицательную) окраску. Исследования показали, что молодежный сленг в английском языке оказал влияние на русский язык, поскольку большинство слов в русской молодежной лексике заимствованы из английского языка. Ключевые слова. Сленг, жаргон, лексика, молодежный сленг, слово.

8

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Сведения об авторе: Аллабердина Альфида Зиннатовна – студент 2 курса факультета философии и социологии БашГУ, г. Уфа

UDC 622.272 DETERMINATION OF THE COMPLEX OF SOCIO-ECONOMIC INFORMATION IN THE DESIGN OF TECHNOLOGICAL SCHEME OF A COAL MINE Babak B. N., Vygovskaya D. D., Vygovsky D. D., Kaverina O.G. Bogdan1314@yandex.ua

Abstract. The article provides recommendations and approaches for selecting socio-economic information used in the design of the technological scheme of a coal mine, taking into account the reliability, accuracy and dynamism of the accepted parameters and factors.

Keywords: source information, operating mode, cost parameters, reliability, dynamic indicators.

The complex of socio-economic information includes information describing the amount of labor, material, and energy resources at the mine, the efficiency of their use, and spending standards. This complex also includes data on the operating mode of the mine, average wages, coefficients of comparative economic efficiency, the wholesale price of coal, and norms for spending time, energy, and materials.

In accordance with the resolutions of the management of the DPR coal industry, the following mode of operation of mines, treatment and preparatory faces is currently maintained.

The number of working days in a year is considered 300; number of shifts for the extraction of coal per day to three; the duration of the working shifts in the mines or for mines with especially harmful and heavy labor conditions – 6 hours for other mines – 7h; duration of the work shift on the surface – 8h.

In mines with especially harmful and heavy labor conditions in Stopes is as follows: when the reservoir is not dangerous on sudden emissions, as well as when working in nevybocovala the areas defined by the forecast of coal seams – three mining and one of the repair and preparatory; in the development of the layers, dangerous on sudden emissions of coal and gas as well as coal seams steep and steeply inclined drop, requiring measures for dust control – two production shifts and one for special events, as well as one of the repair and preparatory shift.

In mines with normal working conditions, the daily operation of the treatment faces is set with two mining and one repair and preparatory shifts.

9

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

In the preparation faces of mines with particularly harmful and difficult working conditions, three shifts (6 hours each) are set directly for the workings and one for special events and repair and preparatory work. For preparatory faces combined with cleaning faces, the number of shifts for carrying out workings should be determined by calculation.

When modeling and designing coal mines, the most significant part of the socioeconomic data consists of cost parameters (indicators) used in the implementation of mining and construction and installation works, the purchase of equipment, the implementation of production processes, etc.

The most important elements of information support for optimal design of coal mines are cost indicators and they are the basis for any technical and economic calculations.

Cost indicators are aggregated parameters of the cost of any type of work in a coal mine, refer to a certain unit of work volume. Such parameters of the unit of work volume are the cross-section area or output volume (m2, m3), the mass of coal produced (t), time (day, year), etc. Moreover, if you know the values of the volume of work and the cost indicator, it is easy to calculate (model) the total cost of production of work over a period.

Cost parameters (indicators) are subject to increased requirements for reliability, simplicity of calculations and dynamism of their use. They should provide the necessary accuracy in determining costs in various project variants and allow detailed analysis of the results of calculations.

The reliability of the indicator is provided by the fact that its values should be determined under the condition of using advanced equipment, technology and organization of work, the application of scientifically justified standards for the consumption of materials, energy and other technological parameters.

The simplicity of calculations should be provided by the nature of the dependencies included in the target function with factors affecting the level of the target function.

The number of accepted factors should not be large, so that the cost indicators are easily used in calculations.

Taking into account the dynamism of cost parameters (indicators) is related to their structure and the purpose of the task when accepting numeric values included in the parameter calculation formula. It is necessary to find such analytical forms of the cost indicator that reflect the dependence of costs on the main parameters of the designed object by applying coefficients, the value of which can and should be combined and summarized in tables. This approach makes it easy to change and update their values.

This ensures the required accuracy and reliability over time, and takes into account technical progress when performing design and construction work.

Cost parameters (indicators) that have a statistical basis, as well as a structure of correlation dependencies, provide less dynamism, reliability and accuracy of the indicator.

10

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

The accuracy of cost parameters (indicators) can be achieved by fully accounting for the accepted cost elements, taking into account factors that affect costs, as well as taking into account the design conditions for real mining conditions under which specific work is performed. If the value parameter (indicator) included in the formula is uncertain, its accuracy argument may decrease due to an increase in the number of such arguments (factors).

When designing, cost parameters (indicators) are expressed as functional dependencies of the unit costs for the production of the accepted type of work, initial information, and optimized variables. This requires a large amount of research and calculation work.

The methods of mathematical statistics used, direct calculations must be made by dividing expenses into groups. As input data for statistical analysis, it is necessary to use individual prices for types of mining operations, the results of calculations, the actual data of the coal mine on the costs of repairing and maintaining the workings, etc.

Cost parameters (indicators) should be constantly adjusted and refined due to the state and changes in socio-economic conditions over time affecting the initial information of the accepted factors.

References 1. Malkin A. S., Puchkov A. S., Salamatin A. G., Eremeev V. G. Design of mines. Textbook for universities. Under the editorship of L. A. Puchkov. M. – AGN Publishing house, 2000. – 375p. Аннотация. В статье приводятся рекомендации и подходы по отбору социально-экономической информации используемой при проектировании технологической схемы угольной шахты с учетом надежности, точности и динамичности принимаемых параметров и факторов. Ключевые слова: Исходная информация, режим работы, стоимостные параметры, надежность, динамичность показателей. Сведения об авторах: Бабак Богдан Николаевич – аспирант кафедры РМПИ, ДонНТУ Выговский Даниил Данилович – доцент, к.т.н., кафедра РМПИ, ДонНТУ Выговская Даниэла Данииловна – доцент, к.т.н., кафедра РМПИ, ДонНТУ Каверина Ольга Геннадиевна – профессор кафедры, доктор педагогических наук, кафедра английского языка, ДонНТУ

11

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

UDC 662.749.33 AUTOCORRELATION OF RETURNS IN MAJOR CRYPTOCURRENCY MARKETS Bibik A.S., Tartakovsky Ye.A., Sarmakeeva A.S., Plesovskikh K.Yu., Revina N.V. saninstein@gmail.com

Abstract. This paper is the first of a series of short articles that explore the efficiency of major cryptocurrency markets. A number of statistical tests and properties of statistical distributions will be used to assess if cryptocurrency markets are efficient, and how their efficiency changes over time. In this paper, we analyze autocorrelation of returns in major cryptocurrency markets using the following methods: Pearson’s autocorrelation coefficient of different orders, Ljung-Box test, and first-order Pearson’s autocorrelation coefficient in a rolling window.

Keywords: bitcoin, cryptocurrency, autocorrelation, efficient market hypothesis In this paper, we use benchmarks provided by the efficient market hypothesis (EMH) [7] to estimate the efficiency of cryptocurrency markets.

The cryptocurrency market is a new and fast-growing sector that adopts features from the financial markets established earlier. It has both advantages and disadvantages in comparison to the traditional markets. The main advantages are worldwide 24/7/365 trading and fast (often immediate) execution and settlement.

The prevalent disadvantages are associated with the immaturity of the cryptocurrency markets. There are not enough established tools, more services and businesses are needed to engender competition and boost reliability, and the cryptocurrency market does not yet interface comfortably with the established traditional markets. For example, there were no prime brokers, and until recently, no custodians that provided services for cryptocurrency markets.

The market also has completely novel elements because the currencies being traded are issued on public blockchain technology, as described by Nakamoto in the original Bitcoin technical paper, [13] and are therefore not controlled by any government or institution. All transactions are immutable, visible to everyone, and remain stored as long as the underlying blockchain exists.

The Efficient Market Hypothesis has been the central proposition of finance since the early 1970s, and it is one of the most controversial and well-studied propositions in all the social and economic sciences. It states that asset prices reflect all available information as Malkiel’s and Fama’s research showed in [9], and that profiting from predicting price movements is very difficult and unlikely as Clarke et al. state in [4].

Over time, the attempts to prove or refute the EMH have been inconsistent, and we still cannot reliably claim that the EMH describes financial markets well enough to 12

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

make investment decisions. However, it provides us with a convenient unified test suite to detect when a market is definitely not efficient. That information can be used to direct further research on how to use the found inefficiency profitably . By now, the majority of established financial markets have been analyzed by many researchers.

Among others, Sewell in [17] explores the market efficiency of the Dow Jones Industrial Average (DJIA), and Zeman et al. [20] do the same for EUR/USD foreign exchange market.

State-of-the-art The Efficient Market Hypothesis is defined as follows. A market is said to be efficient with respect to an information set if the price would be unaffected by revealing the information set to all market participants, as the Malkiel research showed [10], i.e., if the price fully reflects that information set [9]. The classic taxonomy of information sets, according to Roberts [15] and Malkiel and Fama [9], consists of the following: • Weak form of efficiency. The information set only includes the history of prices.

• Semi-strong form of efficiency. The information set includes all the information known to all market participants (publicly available information).

• Strong form of efficiency. The information set includes all the information known to any market participant (private information).

There exists some controversy over the efficiency of the cryptocurrency market based on the proponents and opponents of the EMH. For example, Nadarajah and Chu [12], Bariviera [2], and Tiwari et al. [18] conclude that the Bitcoin market is almost efficient. In contrast, Jiang et al. [5], Chean et al. [3], and Al-Yahyaee et al. [1] present skeptical empirical results that do not support the EMH for this market [14].

Dr. Andrew Urquhart in his published work in 2016 [19] examined the Bitcoin Market during that period and concluded that the market was not efficient, but was moving towards efficiency. In his work, he investigated the behavior of returns using the Ljung-Box test, runs tests, the dispersion ratio test, and the BDS test. The analysis showed that the Bitcoin market did not exhibit weak efficiency during the entire sampling period, but several market efficiency tests show that the market could become more efficient over time, suggesting that Bitcoin’s yield was random later in the period.

However, Bitcoin’s inefficiency was quite strong. Three years have passed since that article was written and the situation in the Bitcoin market has changed significantly.

Hence, we revisit the analysis to bring the information up to date.

One of the most recent publications on this topic by Ladislav Kristoufek and Miloslav Vovrda [6] is devoted to testing whether the examined Bitcoin, DASH, Litecoin, Monero, Ripple and Stellar were efficient, and the overall level of efficiency in the cryptocurrency market. The authors utilized the Efficiency Index comprising long-range dependence, fractal dimension, entropy components, and descriptive statistics. They came to the following conclusions: • Historically, every cryptocurrency was inefficient over the analyzed period.

• Efficiency and ranking were dependent on the quote currency (US dollar or Bitcoin).

13

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

• Most of the coins and tokens were efficient between July 2017 and June 2018.

• The least efficient coins were Ethereum and Litecoin, while DASH was found to be the most efficient cryptocurrency.

Another recent paper by Akihiko Noda [14] studies whether the market efficiency of major cryptocurrencies changes over time based on the adaptive market hypothesis (AMH). Empirical results in this work have shown that the degree of market efficiency changes over time, that the level of efficiency of the Bitcoin market is higher than other markets during most periods, and the efficiency of the cryptocurrency market has improved. Adaptive Market Hypothesis (AMH) embraces EMH as an idealization that is economically unrealizable but which serves as a useful benchmark for measuring relative efficiency.

Presently, even with the cryptocurrency market taking root and occupying a prominent position in global finance, there is little research on it as a full-fledged player, hence the relevance of present study. First, we elaborate on how the statistics were calculated, then provide an analysis for the last 5 years, from 2014-07-01 to 201907-01. In conclusion, we evaluate and discuss the obtained results.

Research hypothesis Consider the BTC/USD, ETH/USD, ETH/BTC markets on Bitfinex, and the XBT/USD market on Bitmex. Let us assume the markets to be efficient as a null hypothesis. For a market to be efficient according to EMH: • There should be no statistically significant autocorrelation of any order on any time frame.

• The p-value of Ljung-Box test should not drop below 0.05 on any lag and any time frame.

• Rolling autocorrelation should display the same stable, close to zero results on any time frame.

Pearson’s autocorrelation coefficient A necessary but not sufficient condition for EMH to hold is that time series has no autocorrelation of any order. Let 𝑋 be a stochastic process and 𝑡 a point in time, then 𝑋𝑡 is the value produced by a given run of the process at time 𝑡. Suppose that 𝑋𝑡 has mean 𝜇 and variance 𝜎 2 at time 𝑡 for each 𝑡. Then the Pearson autocorrelation coefficient between times 𝑡1 and 𝑡2 is defined by: 𝑅(𝑡1 , 𝑡2 ) =

𝐸[(𝑋𝑡1 −𝜇𝑡1 )(𝑋𝑡2 −𝜇𝑡2 )] 𝜎𝑡1 𝜎𝑡2

,

(1)

where 𝐸 – the expected value operator.

Ljung-Box test The null hypothesis of the Ljung-Box test [8], where 𝐻0 signifies that the data are independently distributed (i.e., the correlations in the sample are taken as 0, so that any observed correlations in the data result from the randomness of the sampling process). The alternate hypothesis, 𝐻𝑎 – the data are not independently distributed; they exhibit serial correlation.

14

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

A significant p-value in this test rejects the null hypothesis that the time series is not autocorrelated. To run the Ljung-Box test, we calculate the statistic 𝑄. For a time series of length 𝑛: ̂2 𝜌

𝑄(ℎ) = 𝑛(𝑛 + 2) ∑ℎ𝑗=1 𝑘 .

(2) 𝑛−𝑘 Where 𝜌̂𝑘2 – is the sample autocorrelation at lag 𝑘, and ℎ is the number of lags being 2 tested. Under 𝐻0 , the statistic 𝑄 asymptotically follows a 𝜒(ℎ) . For significance level 𝛼, the critical region for rejection of the hypothesis of randomness is: 2 𝑄 > 𝜒1−𝛼,ℎ , (3) 2 where 𝜒1−𝛼,ℎ is the 1 − 𝛼 – quantile of the chi-squared distribution with ℎ degrees of freedom.

Pearson’s autocorrelation coefficient in a rolling window Additionally, we compute first-order Pearson’s autocorrelation coefficient in a rolling 1-year window. That will allow us to see how the coefficient changes over time and whether there is a stable deviation from zero. If there is, that would contradict the null hypothesis as well.

Results All calculations were made using Python programming language, statsmodels [11], and pandas [16] libraries. The four currency pairs chosen to carry out this research were selected for the following reasons: • BTC/USD on Bitfinex exchange: one of the oldest and most liquid spot cryptocurrency markets.

• XBT/USD perpetual swap on Bitmex futures exchange: the most liquid cryptocurrency market in the world (as of the date this paper was written) with 2.20 billion USD traded daily, according to their reports.

• ETH/USD on Bitfinex exchange: the second-biggest cryptocurrency and one of the most liquid spot markets.

• ETH/BTC on Bitfinex exchange: a cross-pair between the first and secondmost liquid cryptocurrencies. This pair exhibits significantly different behavior compared to USD-denominated pairs.

For each market, we conducted all tests using four different time frames: 5minute, 1-hour, 1-day, and 1-week bars.

Pearson autocorrelation coefficient was measured for the BTC/USD, ETH/USD, ETH/BTC markets on Bitfinex exchange and the XBT/USD market on Bitmex exchange, each on 5m, 1H, 1D and 1W time frames. Results are presented in figures 1 to 4.

15

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 1 – Autocorrelation of BTC/USD, from 2014-07-01 to 2019-07-01

Figure 2 – Autocorrelation of ETH/USD, from 2016-03-09 to 2019-07-01

16

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 3 – Autocorrelation of ETH/BTC, from 2016-03-09 to 2019-07-01

Figure 4 – Autocorrelation of XBT/USD, from 2017-10-12 to 2019-07-01

17

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

As shown in figures 1 and 4, statistically significant lower-order negative autocorrelation was found in the BTC/USD and XBT/USD markets for the 5m and 1H time frames. In contrast, on both ETH pairs in figures 2 and 3, first-order autocorrelation was found to be positive for the 5m and 1H time frames. Second-order autocorrelation was found to be negative. For all four markets, we observed statistically significant positive autocorrelation on the 16th lag of the 1H time frame. For the 1D and 1W time frames, there is no autocorrelation level of any order on any market that significantly exceeds the confidence interval.

According to our observations, none of the four explored markets on lower time frames satisfy the no-autocorrelation condition required for a market to be considered efficient.

In addition to Pearson’s autocorrelation, Ljung-Box test’s p-value was calculated for a number of lags 𝑛 = 30 on the BTC/USD, ETH/USD, ETH/BTC markets on Bitfinex exchange and the XBT/USD market on Bitmex exchange, each on 5m, 1H, 1D and 1W time frames. For the 5m and 1H time frames for each market, pvalues remained zero, 𝑝 = 0, signaling that the null hypothesis of no autocorrelation was rejected on all lags starting from the first at the threshold equal to 0.05. Results for the 1D and 1W time frames are presented in figures 5 to 8.

Figure 5 – Ljung-Box test on BTC/USD, from 2014-07-01 to 2019-07-01

18

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 6 – Ljung-Box test on ETH/USD, from 2016-03-09 to 2019-07-01

Figure 7 – Ljung-Box test on ETH/BTC, from 2016-03-09 to 2019-07-01

Figure 8 – Ljung-Box test on XBT/USD, from 2017-10-12 to 2019-07-01 19

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

To reiterate, on the 5m and 1H time frames, the null hypothesis of no autocorrelation was rejected for every market on all lags, signifying that the Ljung-Box test detected the presence of autocorrelation on all four markets on lower time frames.

These results are consistent with the results observed by measuring Pearson’s autocorrelation coefficient.

The picture is not as clear for the 1D and 1W time frames. For the XBT/USD pair, the p-value never dropped below the threshold on any lag. For the BTC/USD pair, the p-value was above the threshold for lower lags, dropping below the threshold for several lags and rising again later. This observation, while statistically significant, is probably too unstable to be usable. For the ETH/USD pair on the 1W time frame, the majority of lags-starting with the first–had p-value below the threshold, signifying the presence of autocorrelation. For the ETH/BTC pair on the 1D time frame, all lags except one showed p-value below the threshold.

For the last experiment, we explored how first-order autocorrelation coefficient value changes over time. To do that, we calculated first-order autocorrelation coefficient in a rolling 1-year long window on the BTC/USD, ETH/USD, and ETH/BTC markets on Bitfinex exchange and the XBT/USD market on Bitmex exchange, each on 5m, 1H, 1D and 1W time frames.

Figure 9 – Rolling autocorrelation on BTC/USD, from 2014-07-01 to 2019-07-01, 1year window 20

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 10 – Rolling autocorrelation on ETH/USD, from 2016-03-09 to 2019-07-01, 1-year window

Figure 11 – Rolling autocorrelation on ETH/BTC, from 2016-03-09 to 2019-07-01, 1-year window 21

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 12 – Rolling autocorrelation on XBT/USD, from 2017-10-12 to 2019-07-01, 1-year window From the results in figures 9 to 12, it can be seen that first-order autocorrelation coefficient for the 1H time frame stays negative over time for every market. In contrast, for the 1W time frame, it stays positive most of the time in all markets. For the 5m time frame, the pictures differ between ETH and BTC pairs. Rolling autocorrelation of ETH pairs tends to stay positive over time for the 5m time frame, while for the BTC (XBT) pairs, it tends to stay negative. For the 1D time frame, rolling autocorrelation on every market behaves differently.

In this paper, we explored the behavior of autocorrelation tests on the BTC/USD, ETH/USD, ETH/BTC markets on Bitfinex exchange and the XBT/USD market on Bitmex exchange, each on 5m, 1H, 1D and 1W time frames. To do so, we conducted three different experiments, measuring Pearson’s autocorrelation coefficient of different orders, p-values of Ljung-Box test up to 30 lags, and first-order Pearson’s autocorrelation coefficient value in a rolling 1-year window. The initial hypotheses were as follows: • There should be no statistically significant autocorrelation of any order on any time frame if the market is efficient, according to EMH. • The p-value of Ljung-Box test should not drop below 0.05 on any lag and any time frame if the market is efficient, according to EMH as well.

• Rolling autocorrelation should display the same stable, close to zero results on any time frame, which follows from the first point.

According to observed results, all three hypotheses can be rejected for every market on 5m and 1H time frames. We found statistically significant autocorrelation 22

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

of lower orders for all considered markets on 5m and 1H time frames by calculating Pearson’s autocorrelation coefficient. The same observation was confirmed by LjungBox test. Rolling autocorrelation confirmed these observations to be present over time.

Results for 1D and 1W time frames are less clear and not stable and uniform enough to reject EMH or draw any practical conclusions, though rolling autocorrelation on the 1W time frame shows interesting tendency to be strongly positive on all markets.

Even though all three hypotheses were rejected and the markets were found to be inefficient according to EMH, these findings alone are sufficient to extract profits from crypto markets. However, rolling autocorrelation might constitute a useful feature to improve the predictive power of machine learning models.

In our future work, we plan to explore whether long memory is present in the same markets as well as to measure statistical characteristics of the distribution of returns. These articles are going to be used as the baseline for further research into how we can work with financial data in ways that will present markets as less efficient and more predictable.

References 1. Al-Yahyaee K. H., Mensi W., and Yoon S.-M. Efficiency, multifractality, and the long-memory property of the bitcoin market: A comparative analysis with stock, currency, and gold markets. // Finance Research Letters, 27:228-234, 2018.

2. Bariviera A. F.. The inefficiency of bitcoin revisited: A dynamic approach. // Economics Letters, 161:1-4, 2017.

3. Cheah E. T., Mishra T., Parhi M., and Zhang Z.. Long memory interdependency and inefficiency in bitcoin markets // Economics Letters, 167:18-25, 2018.

4. Clarke J., Jandik T., and Mandelker G. The efficient markets hypothesis // Expert financial planning: Advice from industry leaders, p. 126-141, 2001.

5. Jiang Y., Nie H., and Ruan W. Time-varying long-term memory in bitcoin market. // Finance Research Letters, 25: 280-284, 2018.

6. Kristoufek L. and Vovrda M. Herding, minority game, market clearing and efficient markets in a simple spin model framework // Communications in Nonlinear Science and Numerical Simulation, 54: 148-155, 2018.

7. Laffont J. J. and Maskin E. S. The efficient market hypothesis and insider trading on the stock market // Journal of Political Economy, 98(1): 70-93, 1990.

8. Ljung G. M. and Box G. E. On a measure of lack of fit in time series models // Biometrika, 65(2): 297-303, 1978.

9. Malkiel B. G. and Fama E. F. Efficient capital markets: A review of theory and empirical work // The journal of Finance, 25(2): 383-417, 1970.

10. Malkiel B. Efficient market hypothesis // P., M. Milgate, and J. Eatwell (eds.), New Palgrave Dictionary of Money and Finance, 1992.

11. McKinney W. et al. Data structures for statistical computing in python. In // Proceedings of the 9th Python in Science Conference, volume 445, pages 51–56.

Austin, TX, 2010.

23

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

12. Nadarajah S. and Chu J. On the inefficiency of bitcoin. // Economics Letters, 150: 6–9, 2017.

13. Nakamoto S. Bitcoin: A peer-to-peer electronic cash system. Technical report, Manubot, 2019.

14. Noda A. On the time-varying efficiency of cryptocurrency markets. // arXiv preprint arXiv:1904.09403, 2019.

15. Roberts H. Statistical versus clinical prediction of the stock market. // Unpublished manuscript, 1967.

16. Seabold S. and Perktold J. Statsmodels: Econometric and statistical modeling with python. In // 9th Python in Science Conference, 2010.

17. Sewell M. The efficient market hypothesis: Empirical evidence. // International Journal of Statistics and Probability, 1(2): 164, 2012.

18. Tiwari A. K., Jana R., Das D., and Roubaud D. Informational efficiency of bitcoin. // Economics Letters, 163: 106–109, 2018.

19. Urquhart A. The inefficiency of bitcoin. // Economics Letters, 148:80–82, 2016.

20. Zeman P., M.Marš13̆053′fk, et al. High-frequency data and the effectiveness of the spot exchange rate eur/usd. // Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis, 61(7): 2965-2971, 2013. Аннотация. Эта статья является первой из серии коротких статей, в которых рассматривается эффективность основных рынков криптовалюты. Ряд статистических тестов и свойства статистических распределений будут использоваться для оценки эффективности рынков криптовалюты и изменения их эффективности во времени. В этой статье мы анализируем автокорреляцию доходностей на основных рынках криптовалюты, используя следующие методы: коэффициент автокорреляции Пирсона разных ордеров, тест ЛьюнгаБокса и коэффициент автокорреляции Пирсона первого порядка в скользящем окне. Ключевые слова: биткойн, криптовалюта, автокорреляция, гипотеза эффективного рынка. Сведения об авторах: Бибик Александр Сергеевич – студент группы ИСм-19, ДонНТУ Ревина Наталья Владимировна – ст.преподаватель, ДонНТУ Сармакеева Анастасия Семеновна – студент, Институт математики и механики им. Н.Н.Красовского УрО РАН Тартаковский Евгений Александрович – основатель, 3Jane Плесовских Ксения Юрьевна– исследователь, Research Center for Digital Assets and Blockchain

24

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

UDK 556.552, 574.633 OVERVIEW OF METHODS AND TECHNOLOGIES OF PRODUCING OXYGEN-SENSITIVE ELEMENTS FOR MEASURING DISSOLVED OXYGEN IN A RESERVOIR Bydash A.R., Vinnichenko N.G., Kushnirenko Ye.N. an_bydash@mail.ru

Abstract. The article observes the methods and technologies for producing oxygen-sensitive elements for measuring dissolved oxygen in a reservoir. The article discusses the issues of obtaining oxygen-sensitive elements for measuring the amount of dissolved oxygen in water. It was made a comparison of existing methods for measuring dissolved oxygen, the method for the best achievement of the result for this process was selected, and the technology for producing a fluorescence oxygen-sensing element was described.

Key words: dissolved oxygen, fluorescence method, sol-gel process.

In view of fish shortage that has arisen, the way to improve the existing fishing industry is relevant for Donbass. One of the very first criteria that workers encounter is the determination of dissolved oxygen in artificial ponds for the normal functioning of fish. By-turn, the accuracy of the determination of dissolved oxygen directly depends on the device used.

Based on the foregoing, the task is to analyze the existing methods for measuring the amount of dissolved oxygen in a reservoir and to describe the best option for obtaining an oxygen-sensing element for a method in which the best measurement result could be achieved.

When studying the concentration of dissolved oxygen in water, various methods are used. One of them is the titration method [1]. This method was one of the first, and although it is considered one of the oldest methods, it is still relevant. The essence of this method is the reaction of dissolved oxygen in a sample with freshly precipitated Mn (II) hydroxide, which is formed by the addition of sodium or potassium hydroxide to manganese sulfate. Acidification and oxidation of iodide by a higher-valence manganese compound result in the release of iodine in oxygen-equivalent amounts.

When choosing a method, you should take into account an important factor – the environment where the study will be conducted. The Winkler Method, when used in natural waters, showed the presence of numerous interferences.

The next proposed method for determining the oxygen concentration was the pyrophosphate measuring method [2]. It uses the same reaction of Mn (II) oxidation with dissolved oxygen up to Mn (III) in the alkaline medium, which was the basis for the Winkler method. However, due to the presence of sodium pyrophosphate in the 25

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

solution, the precipitate is dissolved, since the pyrophosphate complexes Mn(II) and Mn (III) are soluble in water.

The main advantage of this method is that it can be used in the presence of many substances that react with iodine or iodide ions, in particular, in the presence of nitrites, and thus interfere with the determination of oxygen by Winkler method.

A photometric method based on the monovalent copper ion oxidation into a divalent copper ion with oxygen contained in the sample is recommended for determining the concentration of molecular oxygen in the certification of calibration solutions [3]. Photometric methods for determining the concentration of molecular oxygen in mixtures have the following advantages: high sensitivity and selectivity; the ability to create universal designs of analyzers with several indicator solutions to determine a number of micro-mixtures simultaneously. Disadvantages of the photometric method for determining the concentration of molecular oxygen are bulky hardware design and low reliability.

Now, one of the best research methods is the method of fluorescence. This method is based on the interaction of oxygen molecules with luminescent indicators. It is characterized by a particularly high speed and sensitivity, which by-turn leads to the most widespread use of it in recent researches. One of the main advantages of this method is its high sensitivity that allows to detect several billionths of a gram of a luminescent substance in the desired sample and several times as high as the sensitivity of other methods.

The fluorescence method for determining dissolved oxygen in water is the most practical in both quality of measuring and its maintenance. It makes it possible to reduce both the control period and the sample collection time that can become a determining factor for a quick problem solution.

Sensors based on the fluorescent method work in difficult conditions as any contamination affects the accuracy of measurements and has serious consequences [4].

The sensor for measuring the concentration of oxygen dissolved in water consists of two main components: 1) Sensor cover with a layer of phosphor applied to a transparent substrate; 2) Sensor housing with red and blue light-emitting diode, photodiode and signal Converter.

When measured, a blue light-emitting diode emits a pulse of light that passes through a transparent substrate and is partially absorbed by the phosphor. As mentioned above, this transparent substrate contains a fluorescent film in which oxygen atoms interact with the phosphor.

The process of obtaining such film that is the process of obtaining a luminescent oxygen sensitive element has two stages: 1) Obtaining a porous matrix.

2) Immobilization of the indicator into the matrix.

There are cases when these processes are combined. Each step of the process affects the properties of the sensor element overall. The sensor element matrices into 26

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

which the luminescent indicator is adsorbed can be either thin films (for ash and solgel processes) or volume-porous elements.

Now, the most promising direction in the development of luminescent sensitive elements of oxygen sensors is the use of ash technologies [5]. In comparison with existing technologies, the matrix derived with the help of ash technology has a number of advantages. They are large surface area, the predetermined surface roughness, high chemical, photochemical and temperature stability.

The possibility to use the sol-gel synthesis process at normal (low) temperatures allows creating favourable conditions for the immobilization of organic molecules in an inorganic glass structure. In addition, the sol-gel process simplifies obtaining the specified coating properties that determine the critical parameters of the sensor such as sensitivity and response time [6]. One more feature of the sol-gel technology is the possibility to miniaturize the sensor's sensitive element up to micro-dimensions. This process can be used to create up to 0.2 microns thick films that provide the sensor short response time.

An important characteristic of the oxygen-sensitive element is its ability to transfer the process of operation and storage invariably [7]. However, it is the ability of fluorescent sensors to preserve the metrological characteristics of the sensor element during long storage and operation helped them become popular and compete with electrochemical detectors widely used before.

The peculiarity of sensitive element position in relation to the photo detector and excitation source is no less important than the choice of technology and method for analysis. To some extent, this determines the choice of the method and scheme for registering fluorescence quenching. The choice of circuit depends on the sensor and its operating conditions. Most often, the signal at which occur the fluorescence measuring is registered in transmitted and reflected light, at an angle of 90 degrees to the exciting radiation.

At present, basic directions of fluorescent sensors development can be distinguished: 1) synthesis of new stable indicators with high quantum outputs; 2) creating Sensitive Elements with weakly temperature – dependent (ideallyindependent) parameters; 3) obtaining Sensitive Elements with hydrophobic properties; 4) creating materials with stable luminescence characteristics that are independent on environmental influences and can be used to produce support elements; 5) development of new methods for sustainable implementation of indicators in the matrix in order to improve the temporary stability of the sensor; 6) introduction of components that increase the sensitivity and sensor response time in the operating range of oxygen concentrations; 7) creation of highly selective Sensitive Elements.

In addition, the development of Sensitive Elements lines is promising. Such sensors can be created by using: 27

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

1) luminescent indicators with various properties for different analytes immobilized in a single matrix; 2) coatings with various permeability for different analytes; 3) a set of selective sensors (sensor line) based on modern Microtechnologies.

To sum up, the creating luminescent oxygen sensors is undoubtedly promising, as with all the accompanying problems it reflects the trend of development of modern instrumentation aimed at creating microsensors and sensor multisystems based on advanced science-intensive technologies.

Conclusion Various methods for determining the concentration of dissolved oxygen in reservoirs have been investigated. Comparing of obtained results when used these methods showed that the most promising in the development of an electronic system will be the use of fluorescence method and luminescent oxygen sensor. This electronic system will determine the amount of dissolved oxygen in the fishery reservoir, taking into account the season, the impact of external and internal factors.

References 1. Агасян П.К. Основы электрохимических методов анализа. Химия, 1984. С. 168.

2. Замышляева М.Г. Михеева А.А. Очистка производственных сточных вод. Москва, 1969. С. 201.

3. Люминесцентные сенсоры кислорода [Электронный ресурс]/ Ин-т аналит. прибор-ия (РАН). 2009.URL: http://www.microsystems.ru/files/publ/15.htm (дата обращения 25.03.2020) 4. Люминесцентный метод измерения растворенного кислорода в воде [Электронный ресурс]//Персональный сайт ЭкоИнструмент. 2007. URL: http://www.ecoinstrument.com.ua/ldo-lyuminescentnyj-metod-izmereniyarastvorennogo-kisloroda-v-vode (дата обращения: 29.03.2020) 5. Прокофьев. М.А. Метод определения химического потребления кислорода. Москва: Стандартинформ. - 2014. – С. 10.

6. Скуг Д. Уэст Д. Основы аналитической химии. Казань: Изд-во МИР.

1979. С. 408-410.

7. Шуваева О.В. Современное состояние и проблемы элементарного анализа вод различной природы. Новосибирск: ГПНТБ СО РАН. – 1996. – С. 48. Аннотация. В статье рассматриваются вопросы получения кислородочувствительных элементов, для измерения количества растворенного кислорода в воде. Проведено сравнение существующих методов для измерения растворенного кислорода, выбран метод для наилучшего достижения результата для данного процесса, а также описана технология получения люминесцентного кислородочувтсвительного элемента. Ключевые слова: растворенный кислород, флуоресцентный метод, зольгельный процесс.

28

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Сведения об авторах: Быдаш Анастасия Романовна – студент группы ПСм-19, ДонНТУ Винниченко Николай Григорьевич – к.т.н., доцент, доцент кафедры «Электронная техника», ДонНТУ Кушниренко Елена Николаевна – ст.преподаватель кафедры английского языка, ДонНТУ UDC 004.93’12 MODERN SOLUTION TO FACE RECOGNITION PROBLEMS BASED ON NEURAL NETWORKS Chernyshov B.S., Fedyaev O. I., Girovskaya I.V. adobe.a.dope@gmail.com

Abstract. The article describes the modern ways of solving the following face recognition problem: selecting the face and aligning it, generating a vector identifier of the face from the resulting image, and identifying a person by the identifier.

Keywords: face recognition, neural network, training, classifier, MTCNN, dataset.

In the modern world, face recognition systems are very popular and are in demand everywhere where a person is involved. They can be implemented, for example, in security systems in order to identify unwanted persons who penetrate the protected area. The requirements for such a system are quite serious: everything needs to work quickly and at the same time show high recognition accuracy [1]. It is these indicators that are key in the development of such a system. In addition, there are a number of problems that developers face, such as the illumination of an object, the resolution of a recognizable image, camera angle, and others.

Solution architecture 1. Cascading neural network – to search for faces in the photo.

2. Convolutional NS – to generate a vector face identifier.

3. Classifier based on the support vector method – determining a person by the identifier of a person.

Cascade neural network The first problem is the allocation of the person’s face and its subsequent alignment. To date, the complexity of its solution lies in the possibility of the presence in the image of various angles, light sources and overlapping mutual arrangement of faces.

Having examined the existing software implementations that allow us to solve the problems of face extraction and alignment, we identified 3 approaches: Viola-Jones 29

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

in the implementation of OpenCV, HOG in the implementation of dlib and R-FCN (convolutional cascade neural network). On the FDDB dataset, the selected algorithms demonstrated the following results (Fig. 1).

Figure 1 – FDDB Algorithm Results The best result in the tests performed was demonstrated by R-FCN. However, the runtime was 120 ms on the GPU, which is low for a real-time system. Therefore, it was decided to find a faster alternative with similar accuracy. As a result, a solution was found using MTCNN, showing 51ms on FDDB, but slightly inferior in accuracy – 90%.

This structure uses a cascading architecture with three stages of carefully designed deep convolutional neural networks to predict the location of the face and its key points. Five points stand out on the face: two on the eyes and the corners of the lips, and one on the tip of the nose (see figure 2). Dots help align the face to the fullface position, which in the future will facilitate the task of face recognition.

Figure2 – Face alignment process 30

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Convolutional Neural Network The next problem is the conversion of the selected person to the form of a vector identifier. The most optimal means of solving the problem are convolutional neural networks.

Convolutional neural network – a special architecture of artificial neural networks, proposed by Jan Lekun in 1988 and aimed at efficient image recognition, is part of deep learning technologies. It uses some features of the human visual system, in which the so-called simple cells, which react to straight lines at different angles, and complex cells, the reaction of which is associated with the activation of a certain set of simple cells, were discovered. Thus, the idea of convolutional neural networks is to alternate convolution layers and subsampling layers or pooling layers. The network structure is unidirectional (without feedbacks), basically multilayer. For training, standard methods are used, most often the method of back propagation of error. The activation function of neurons (transfer function) – any, according to the choice of the researcher.

The architecture of the network got its name because of the convolution operation, the essence of which is that each image fragment is multiplied by the matrix (core) of the convolution element by element, and the result is summed and written to the same position in the output image.

In an ordinary perceptron, which is a fully connected neural network, each neuron is connected to all neurons of the previous layer, and each connection has its own personal weight coefficient. In the convolutional neural network, the convolution operation uses only a limited matrix of small-sized weights, which are «moved» throughout the processed layer (at the very beginning – directly in the input image), forming an activation signal for each next layer neuron with a similar position after each shift. That is, for different neurons of the output layer, the same weight matrix is used, which is also called the convolution core. It is interpreted as graphic coding of a feature, for example, the presence of an inclined line at a certain angle. Then the next layer, resulting from the convolution operation by such a weight matrix, shows the presence of this feature in the layer being processed and its coordinates, forming the so-called feature map. Naturally, in a convolutional neural network, the set of weights is not one, but a whole gamma encoding image elements (for example, lines and arcs at different angles). Moreover, such convolution kernels are not laid in advance by the researcher, but are formed independently by training the network using the classical method of back propagation of error. The passage with each set of weights forms its own instance of the feature map, making the neural network multichannel (many independent feature maps on one layer). It should also be noted that when sorting a layer with a weight matrix, it is usually moved not by a full step (the size of this matrix), but by a small distance. So, for example, with a 5 × 5 weight matrix dimension, it is shifted by one or two neurons (pixels) instead of five so as not to «step over» the desired sign.

The subsampling operation (English subsampling, English pooling, also translated as «subsampling operation» or merging operation), reduces the dimension 31

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

of the generated feature maps. In this network architecture, it is believed that information about the fact of the presence of the desired attribute is more important than the exact knowledge of its coordinates, therefore, the maximum is selected from several neighboring neurons of the feature map and taken as one neuron of a compacted feature map of smaller dimension. Due to this operation, in addition to accelerating further calculations, the network becomes more invariant to the scale of the input image.

Consider the typical structure of the convolutional neural network in more detail.

The network consists of a large number of layers. After the initial layer (input image), the signal passes through a series of convolutional layers, in which the convolution itself and alternating sampling (pooling) alternate. Alternating layers allows you to make «feature maps» from feature maps; on each subsequent layer, the map decreases in size, but the number of channels increases. In practice, this means the ability to recognize complex feature hierarchies. Usually, after passing through several layers, a feature map degenerates into a vector or even a scalar, but there are hundreds of feature maps. At the output of the convolutional layers of the network, several layers of a fullyconnected neural network (perceptron) are additionally installed, to the input of which terminal cards of attributes are fed.

In the final implementation, the architecture of ResNet-152 with modifications developed by Google was used. It contains 152 layers, it is calculated for 11.3 billion floating point operations, which is much less than that of competitors, with a superior number of layers (see Fig. 3).

Figure 3 – Comparison of VGG-19 and ResNet-152 architectures 32

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Classification of vectors The next problem in solving the face recognition problem is the classification of vector identifiers of faces obtained because of the convolutional neural network. To solve this problem, you need to use the classifier.

Classification is one of the sections of machine learning dedicated to solving the following problem. There are many objects (situations) divided in some way into classes. A finite set of objects is given for which it is known which classes they belong to. This set is called the training set. The class affiliation of the remaining objects is not known. It is required to construct an algorithm capable of classifying an arbitrary object from the original set.

Classifier – an algorithm in machine learning that allows you to solve the problem of classifying an object.

Among the available implementations of classifiers, the most suitable for solving our problem are: a classifier based on the support vector method (SVC) and logistic regression.

The support vector machine method (English SVM, support vector machine) is a set of similar teaching algorithms with a teacher used for classification and regression analysis problems. Belongs to the family of linear classifiers and can also be considered as a special case of Tikhonov regularization. A special property of the support vector method is a continuous decrease in the empirical classification error and an increase in the gap; therefore, the method is also known as the classifier method with a maximum gap.

The main idea of the method is the translation of the original vectors into a space of higher dimension and the search for a separating hyperplane with a maximum gap in this space. Two parallel hyperplanes are constructed on both sides of the hyperplane separating the classes. A separating hyperplane is a hyperplane that maximizes the distance to two parallel hyperplanes. The algorithm works under the assumption that the greater the difference or distance between these parallel hyperplanes, the smaller the average error of the classifier.

Logistic regression or logit regression (English logit model) is a statistical model used to predict the probability of an event occurring by fitting data to a logistic curve (see Fig. 4).

Figure4 – Logistic curve 33

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Among these two classifiers, SVC was chosen for accuracy and runtime indicators.

Conclusion This article examined the most modern approaches to implementing the solution to the problem of face recognition. The most optimal version of the program using the technologies described in the article was also selected and implemented.

References 1. Кулябичев Ю.П. Обзор методов идентификации людей на основе изображений лиц /Пивторацкая С.В // Московский инженерно-физический институт 2. Уоссермен Ф. Нейрокомпьютерная техника: теория и практика / Ф. Уоссермен. М. //Перевод на русский язык, Ю. А. Зуев, В. А. Точенов, 1992.

3. Форсайт Д.А. Компьютерное зрение: современный подход / Понс Ж.// пер. с англ. – М.: Издательский дом «Вильямс», 2004.

4. Hurieh Khalajzadeh FaceRecognition using Convolutional Neural Network and Simple Logistic Classifier. // Intelligent Systems Laboratory (ISLAB) Faculty of Electrical & Computer Engineering K.N. Toosi University of Technology, Tehran, Iran 5. Moghaddam B.IEEE Transactions on Pattern Analysis and Machine Intelligence. / Pentland A // 1997. V. 19. P. 696.

6. Samaria F. Face Recognition Using Hidden Markov Models // PhD thesis, Engineering Department, Cambridge University, 1994.

7. WenYi ZhaoImagebased Face Recognition Issues and Methods /Rama Chellappa. // Sarno Corporation Center for Automation Research, Washington Road University of Maryl Аннотация. В статье описаны современные способы решения следующих проблем задачи распознавания лиц: выделение и выравнивание лица, генерация из получившегося изображения векторного идентификатора лица и определение человека по идентификатору. Ключевые слова: распознавание лиц, нейросеть, обучение, классификатор, MTCNN, датасет. Сведения об авторах: Чернышов Богдан Сергеевич – студент группы ПИм-19, ДонНТУ Федяев Олег Иванович – к.т.н., доцент кафедры «Программная инженерия», ДонНТУ Гировская И.В. – старший преподаватель кафедры английского языка ДонНТУ

34

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

UDC 004.852 THE DEVELOPMENT AND USE OF ARTIFICIAL INTELLIGENCE TECHNOLOGIES Frolov I.V., Sokolova O.V. f-r-o-l-o-v2000@mail.ru

Abstract. The history of emergence and application of Artificial Intelligence technologies is considered in the article. Advantages and disadvantages of this concept are revealed. The author describes an important role of Artificial Intelligence in definite spheres of our life and emphasizes that this technology will not be able to replace a human being in the nearest future.

Keywords: intelligence technologies, Artificial Intelligence (AI), computer, information, application, data, neural network, tasks, capabilities.

Nowadays the expression «Artificial Intelligence» is heard much more often than before. The news appears on the network that «artificial intelligence» plays chess with world champions in this discipline, calculates and processes a huge capacity of data, makes predictions, etc.

Artificial Intelligence (AI) is not a format or a function, but primarily a process and an ability to think and analyze the data. People think that the expression «Artificial Intelligence» represents smart humanoid robots conquering the world. However, AI is not intended to replace human beings. On the contrary, its goal is to expand human skills and capabilities making it a valuable resource [2].

The article focuses on the phenomenon of AI due to the importance of its role in our lives. The aim of the work is to study the history of the emergence and the development of AI. We shall analyze the structure and the areas of application, we shall identify advantages and disadvantages of AI. As it was mentioned above, AI is an ability of a computer to analyze the data.

Artificial intelligence allows computers to learn based on their own experience, to adapt to the given parameters and to perform the tasks that were possible for human beings only. The possibility of deep learning and natural language processing is extremely important in many cases of AI implementation (from computer chess programs to unmanned vehicles). Due to these technologies, computers can be «taught» to perform certain tasks by processing a large amount of data and identifying their patterns.

The term «artificial intelligence» appeared in 1956, but AI technology has just become popular today. It is caused with increasing data amounts, improving algorithms, optimizing computing power and data storage facilities.

The first researches in the field of AI started in the 50s of the last century and were aimed to solve the problems and to develop symbolic computing systems. In the 35

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

60s, this field attracted the interest of the US Defense Department: American military began training computers to simulate human mental activity. For example, the Department of Advanced Research Projects (DARPA) completed a series of virtual street map projects in the 1970s. DARPA specialists managed to create intelligent personal assistants in 2003, long before the modern Siri, Alexa and Cortana appeared.

These works became the basis for the principles of automation and the formal logic of reasoning used in modern computers, in particular, in decision support systems and smart search systems, designed to complement and amplify human capabilities.

Although AI is often represented in the form of humanoid robots exciting power over the world in science fiction films and novels, at this stage the development of AI technology, AI are not so scary and far from smart. On the contrary, the development of Artificial Intelligence allows these technologies to bring real benefits in all sectors of the economy [1].

Why is Artificial Intelligence so important? It is known that AI allows to automate repetitive learning and to search processes through the use of data. However, AI differs from robotics, which is based on the use of hardware. The goal of AI is not automation of manual work, but the reliable and continuous execution of numerous large-scale computerized tasks. Such automation requires human participation for the initial setup of the system and the correct formulation of questions.

AI makes existing products intelligent. As a rule, AI technology is not implemented as a separate application. AI functionality is integrated into existing products, providing means for their improvement like Siri technology added to Apple's devices of new generation. Automation, communication platforms, bots and smart computers combined with large amounts of data can improve various technologies used at home and in offices: from security data analysis systems to investment analysis devices.

AI is adapted thanks to progressive learning algorithms so that further programming is based on data. AI detects structure and patterns in the data that allow the algorithm to study a specific skill: the algorithm becomes a classifier or predictor.

Thus, according to the same principle by which an algorithm study a game of chess, he can learn to offer suitable products online.

AI provides much deeper analysis of large amounts of data using neural networks with many hidden levels. A neural network is one of the methods of machine learning. This is a mathematical model designed on the principle of organization and functioning of biological neural networks – nerve cells networks of a living organism and its software or hardware implementation too. A few years ago, the development of a fraud detection system with five hidden levels was almost impossible. Everything has changed due to the increase of computing power and the advent rise of «big data». For models of deep learning, a huge amount of data is necessary because they are trained on their basis. Therefore, the more data are, the more accurate model is.

Deep neural networks allow AI to achieve an unprecedented level of accuracy.

For example, working with Alexa Google Search and Google Photos service is based on deep training. The more often we use these tools, the more effective they become.

36

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

In medicine, the diagnosis of cancerous tumors with the help of MRI images using AI technologies (deep learning, image classification, objects recognition) is not inferior in accuracy to the conclusions of highly qualified radiologists.

AI maximizes the benefits of data. With the appearance of self-learning algorithms the data itself becomes an object of intellectual property. The data contains necessary answers. It means that you just need to find them using AI technology. As the data play much more important role than ever before it can be considered as a competitive advantage. When the same technologies are used in competitive conditions the one with the most accurate data will win.

AI functionality is widely demanded in different spheres, especially for question-answer systems. In industry, AI can analyze data in manufacturing with the connected equipment and to predict load and demand with the help of recurrent networks (a special kind of deep learning network used for work with serial data). In trade AI helps to purchase online with individually selected recommendations and allows the sellers to discuss the shopping with the customers. Besides, AI technologies can optimize processes of product management and placement. In the sports coaches receive reports with camera shots and sensor indicators which help to decide how better organize the game and how to optimize the players’ layout and strategy of the game.

There are a lot of such examples of using this technology [3].

AI has its drawbacks like any technology. Artificial intelligence technologies can change any industry, but their capabilities are not unlimited.

The main disadvantage of AI is that training is possible only based on data. It is impossible in any other ways. It means that any inaccuracies in data will affect the results. New levels of prognostication or analysis need to be added separately. Modern AI systems are intended for definite tasks. A system configured to detect fraud will not be able to drive a car or to provide law assistance. Moreover, AI system designed to detect fraud in the health field will not be able to detect frauds with taxes or claims for guarantees with the same degree of accuracy [4].

In other words, these systems have very narrow specialization. They are designed to perform only one specific task and they are not able to perform multiplechoice tasks as people do. In addition, self-learning systems are not autonomous. The images of AI technology that we see on television and cinemas are still elements of fiction. Nevertheless, computers capable of analyzing complex data to learn and improve specific skills are not unusual [4].

We can conclude that the goal of AI is to provide the operation of software products capable of analyzing input data and interpreting the results. Artificial intelligence is a tool providing more intuitive process of human interaction with programs. It is assistance for us in making decisions within certain tasks. AI cannot replace a man and it will not become such in the nearest future.

References 1. Мюллер Дж., Массарон Л. Искусственный интеллект для чайников. – М.: Диалектика, 2019. – 384 с.

37

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

2. Электронный ресурс// Свободная энциклопедия «Википедия». 2001.

URL: https://ru.wikipedia.org/wiki/Искусственный_интеллект 3. Электронный ресурс// «SAS»- аналитические решения и ПО. 2001. URL: https://www.sas.com/ru_ru/insights/articles/analytics/what-is-artificialintelligence.html 4. Электронный ресурс// Сайт «Oracle».

2000.

URL: https://www.oracle.com/ru/artificial-intelligence/what-is-artificial-intelligence.html Аннотация: В докладе рассматривается история возникновения и применения технологий искусственного интеллекта (ИИ). Раскрыты плюсы и минусы этой концепции. Автор описывает важную роль искусственного интеллекта в определённых сферах нашей жизни и подчёркивает, что данная технология не сможет заменить человека в ближайшем будущем. Ключевые слова: технологии искусственного интеллекта, искусственный интеллект (ИИ), компьютер, информация, приложение, данные, нейронная сеть, задачи, возможности. Сведения об авторах: Фролов Илья Вячеславович – студент группы ПИ-18в, ДонНТУ Соколова Ольга Викторовна – ассистент кафедры английского языка ДонНТУ

UDC 316.77 INFLUENCE OF INTERNET RESOURCES ON MODERN COMMUNICATIONS Gallyamova E. D. gallyamovaelizaveta@gmail.com

Abstract: Internet communication is one of the conditions and results of the formation of a globalizing world. Thanks to it, the coherence of information flows is ensured.

The Internet in the modern world provides many new ways of communication.

But also, it can be noted that the role of the Internet does not end solely in maintaining interpersonal communication between people, but also includes building communication between entire companies and cultures.

Keywords: Internet, virtual reality, communication, Internet communication, Internet community, web technologies.

Communication between people is the foundation for the unification of society.

In today's world, many communications flow on the Internet. The Internet has become 38

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

a virtual reality in which there is an interaction between people with different statuses and roles. The Internet makes communication easier, faster and more affordable. Now people on different continents can exchange information without leaving home, and spending only a few minutes of their time on it.

Therefore, to assess the relevance of the Internet in society, we must recall the specific characteristics of the Internet as a technology. Then we must place this in the context of the transformation of the overall social structure, as well as in connection with the culture characteristic of this social structure. Indeed, we live in a new social structure, a global network society, which is characterized by the growth of a new culture, a culture of autonomy.

Our society is a network society; that is, a society built around personal and organizational networks based on digital networks transmitted over the Internet.

Moreover, since networks are global and know no boundaries, a network society is a global network society. Academic studies have found that the Internet does not isolate people and does not reduce their sociability; it actually enhances sociability. In addition, a major study by Michael Willmott for the British Computer Society (Trajectory Partnership 2010) showed a positive correlation for both individuals and countries. We studied the frequency and intensity of the use of the Internet, as well as psychological indicators of personal happiness.

Controlling other research factors has shown that using the Internet empowers people by reinforcing their sense of security and personal freedom. This effect is especially positive for lower income people and for women. Age does not affect a positive relationship, it is important for all ages. Why women? Because they are at the center of their family’s network, the Internet helps them organize their lives. In addition, it helps them overcome their isolation, especially in patriarchal societies.

With the help of the Internet, people can stay in touch with friends and family regardless of time and place. It helps develop proximity and unity in a growing mobile world. On the other hand, it also creates more opportunities for strangers to meet because of common interests, as well as for the development of new friendships and other relationships between random acquaintances.

Geography has always limited the social boundaries of people. It’s easier and faster to keep in touch with someone who lives nearby physically. The class also helped identify social circles of people with economic, sexual, racial, religious, and national factors that act as barriers to global communication.

The Internet did not destroy these restrictions, but reduced their relevance, allowing people to participate together in one virtual arena. Thanks to the Internet, people can communicate with each other from almost anywhere in the world where there is access to the Internet. The result was a mixture of traditional cultures, when like-minded people are looking for each other on the Internet. Subsequently, interestbased cultures became much stronger. This helps isolated people find communities and broaden their horizons.

The Internet provides people with new means of communication to express themselves. Web technologies help media lovers create and distribute their videos.

39

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

These technologies help artists exhibit, perfect, and even create their art. Authors can find their voice using unlimited web resources, activists can organize their communities, and chronically busy people get information and express themselves in 140 characters thanks to Twitter. In addition, the development of Internet memes has created a new social currency that people can use to communicate with each other, creating completely new social structures that help shape the future of society.

The Internet has accelerated the pace of business and expanded the possibilities of what is possible in the workplace. The most obvious benefit is that people have the best opportunities for direct communication, including email, text messaging and video conferencing. Long-distance calls, which usually took place over the phone within a few days, can end in a few minutes. Email filtering helps the most important messages get the priority they need. However, the benefits go even deeper than that. Thanks to the Internet, modern companies can exist in several places, and each site can easily connect to others via the Internet. Good IT practices help automate many business processes, such as online sales. This allows computers to do most of the work, reducing the chance of human errors.

The greatest development in the field of services provided via the Internet was received by financial services, which include the following activities: banking services, the services sector in the foreign exchange and stock markets, and Internet insurance.

Summing up, we can say that Internet communication has led to the emergence of a fundamentally new form of social interaction – the Internet community. Online communities can arise based on various network resources: forums, social networks, online games, chats, blog platforms, video conferencing, etc. Sufficient conditions are a unifying topic, the opportunity to speak out, and regular interactions with other users.

At the same time, an individual joins in communication only when a topic that is relevant to himself appears and interacts only with those members of his social network that arouse his interest. Communication in communities allows Internet users to satisfy their need for communication, to identify themselves with groups that are close in their interests.

References 1. Болгов Р. В. Сообщества пользователей интернет-проектов // Вестник МГИМО-университета. 2013. Сергеев Е. Ю. Средства массовой коммуникации в условиях глобализации // Общество. Среда. Развитие. – 2009.

2. Горошко Е. И. Информационно-коммуникативное общество в гендерном измерении. Харьков: ФЛП Либуркина Л. М. – 2009.

3. Еляков А. Д. Дефицит и избыток информации в современном социуме // Социологические исследования. – 2010.

5. Лавренчук Е. А. Аутопойезис социальных сетей интернеткоммуникаций // Вестник РГГУ. Серия «Философия: Социология». – №12: – 2009.

40

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

6. Митичева Т. И., Маслова В. С. Проблема активности молодежи в социальных сообществах интернет-сети // Альманах современной науки и образования. Тамбов: Грамота, №5(95) – 2015.

7. Соловьев Д. Как меняются люди под воздействием медиа? URL: http://cossa.ru/articles/234/13669/ Аннотация: Интернет-коммуникация является одним из условий и результатов становления глобализирующегося мира. Благодаря ей обеспечивается связанность информационными потоками. Интернет в современном мире дает множество новых способов коммуникации. Но так же, можно отметить, что роль интернета не заканчивается исключительно на поддержании межличностного общения между людьми, а еще включает в себя построение коммуникаций между целыми компаниями и культурами. Ключевые слова: Интернет, виртуальная реальность, коммуникация, интернет-коммуникация, интернет-сообщество, веб-технологии. Сведения об авторах: Галлямова Елизавета Дамировна, студент 2 курса факультета философии и социологии БашГУ, г. Уфа

UDC 338.242.2 DIRECTIONS FOR IMPROVING LOGISTIC SYSTEMS IN THE FOOD INDUSTRY Kharadzha T.A. tanyakharadzha@rambler.ru

Abstract. This article focuses on the analysis of the features of foreign experience in the formation of logistics systems at the present stage of development of the world economy, the analysis of the current economic condition of the industry of the Donetsk People's Republic, which characterizes the conditions for the formation of supply chains. Based on the analysis, promising directions for the development of the food industry`s logistic systems in the region are formulated.

Keywords: supply chain, logistic system, material flows, production evaluation, management, analysis.

In the enterprise of the industrial sector in modern conditions, it is difficult to carry out successful, competitive, production activities, based only on improving the technology of the production process and minimizing costs. Given the unstable 41

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

economic situation in the region, there is a need to build non-standard systems for organizing the production cycle, improving the competitive advantages of not only the final product, but also the functioning system of the entire enterprise as a whole through the creation of a new field of activity, diversification and integration of production, the formation and rational management of material flows reorganization of supply chains.

At the same time, the formation of material flows does not require solid investments, it helps to increase competitiveness, production efficiency by improving the reliability of supplies, optimizing labor resources, and increasing production capacities.

Considering that the efficiency of the use of production resources depends on the quality and rhythm of the supply of raw materials for processing, the main obstacle to the development of the enterprise today is the problem of providing it with raw materials. Recently, the capacity utilization at the enterprises of the Donetsk region has sharply decreased, which entails the deterioration of production relations and the prevalence of tolling raw materials in the supply chain. The replacement of commodity-money relations with natural ones is caused by the deterioration of the economic and economic activity of the manufacturing enterprise, a decrease in its share in the formation of profit due to its displacement from the commodity market. All of the above determines the relevance of research and the search for solutions to these problems.

Features of the formation of material flows in supply chains are investigated in the works of many domestic and foreign scientists: Sergeeva V.I. [3], Gordon M.P., Somova V.E., Kolesnikova S.N., Karnaukhova S. B., Semenenko A.I., Kozlova V.K., Uvarova S.A., Firona H.E., Van Rosta S., Waitina T., Rodnikova A.N., Kostoglodova D.D., Ashikina B.A. and many others [1, 6-8]. In most cases, the policy for the formation of all types of logistics flows provides for their regulation in the relationship at the enterprise, the features of the formation of their individual types are investigated. However, the issues of the formation, analysis and management of material flows in the supply chains of a manufacturing enterprise in a tense socioeconomic situation are not fully disclosed.

The formation of production relations in the Donetsk People's Republic contributed to the formation of consumer priority, with the usual set of needs for society, but did not fully influence the formation of the right logistic infrastructure that allows manufacturers to satisfy the market demand at the level of world standards.

Despite the fact that many manufacturers and sellers invest maximum effort and money, forming their own logistics networks and complexes, the lack of their own centralized republican logistics system reinforces the gap in creating an efficient and modern market for goods and services. These circumstances actualize the need to create a republican logistics model in order to increase competitiveness. The formation of an effective logistics system is a long process that requires competent and clear preliminary research, analysis and justification of not only theoretical aspects, but also taking into account the practical features of the business environment.

In terms of the development of the logistics system, our region is currently significantly behind developed countries. Recently, however, there has been a positive 42

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

trend of formation and development in terms of growth in government investment in the construction and restoration of the industrial sector of the republic, the restoration of roads and rail links, the development of postal logistics infrastructure. Such events are systemic elements of building a nationwide logistic model that is being formed, including taking into account advanced foreign ideas and features of the Donetsk region, which occupies one of the most geographically attractive territories. Based on this, the formation of the republican logistics system, which currently far does not meet modern standards, can be accelerated by using not only specific logistics models, but also the experience of their implementation and functioning in advanced foreign countries. A comprehensive analysis of the main advantages and disadvantages of existing systems implemented and operated in specific areas, taking into account their specifics, will help to formulate their own approach to the formation of a nationwide logistics system aimed at the formation and development of external relations.

Historically, Donbass has always been considered one of the largest regions in Europe. Industry has always been the main sector of the economy of the Donetsk region. The key components of Donbass industry were metallurgical, fuel and energy, chemical, engineering, construction complexes, as well as transport, light and food industries, supplemented by significant scientific and technical potential. The industry of Donbass included a huge number of enterprises in both the mining and processing industries. The first group consisted of coalmines, as well as enterprises engaged in the extraction of iron ore, mercury, rock salt, building materials, etc. The group of processing industries included machine-building, chemical, and metallurgical plants, as well as enterprises of light and food industries. In total, in the beginning of 2014, over 28 thousand enterprises functioned in the territory of Donetsk region, of which about two thousand belonged to the rank of large, strategically significant. The enterprises carried out labor activities of more than 900 thousand specialists, providing for that period about 20% of the country's GDP. Among the largest enterprises of the industrial sector of the Donetsk region, there were enterprises that have no analogues in the world, and most enterprises of the CIS countries cannot carry out their production activities without raw materials extracted or produced at these enterprises.

During 2014, industrial production decreased by 30%, in 2015, dozens of strategic enterprises mothballed their activities, and some others completely closed production. All this was due to a lack of material and raw material base, a sharp migration of the population, a lack of demand and a market for finished products, the destruction of transport links, etc.

The basis of the industrial potential of the DPR in 2019 was mechanical engineering, the chemical and food industries. The list of machine-building enterprises of the Republic includes 38 enterprises, of which such plants of the DPR as Donetskormash, Yasinovatsky, Gorlovsky and Makeevsky machine-building plants, Donetsk plant «Prodmash», NPP «Energy» and several others. The leading position for 2019, in terms of the number of enterprises, is occupied by the food industry – 44 business entities, including Trading House Gornyak LLC, Your Producer LLC, SLAVOLIA GROUP LLC, KOLBIKO FIRM, etc. [6].

43

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

At the beginning of 2019, in the DPR, the leaders in the production of products were metallurgy – 37.6%, energy – 25.9% and food industry – 10.6%. In March 2018, in terms of sales of products, food industry enterprises rose from 5th to 3rd place in the Republic and in 2019 continue to hold it, second only to metallurgical production and the electric power industry. According to the number of registered enterprises in the republic, the food industry is leading. Of the 538 enterprises registered in 2019 in the field of activity of the Ministry of Industry and Trade of the DPR, 224 (42%) enterprises account for the food industry [2].

Currently, the DNI food industry is represented by 140 enterprises of the milling, pasta, baking, confectionery, dairy, oil and fat, meat processing, fish processing, food taste, brewing, wine-making industries, as well as the production of alcoholic beverages and soft drinks. According to the Main Statistics Office of the DPR, the food industry is the only industry in the DPR, where current production volumes have increased significantly compared to last year [2].

In 2019, the volumes of production of the following products increased: sausages – by 25.2%; dairy products – by 27.9%; confectionery products – by 30.2%; meat and offal - by 40.8%; fresh, chilled and frozen poultry meat – by 41.4%; pasta – 2 times; unrefined sunflower oil – 8.3 times. Import deliveries of finished food products decreased in January-August 2019 compared to the same period in 2018 by 16.5%, amounting to 215.9 million rubles. Over half of the republic’s exports of these products are confectionery and meat [5].

It should be noted that in 2017-2018, the production of socially significant food products was a priority for the Republic, and in 2019, manufacturers began to pay more attention to expanding the range of products.

A number of enterprises resumed production activities. In March 2017, it launched production of Donetsk Brewery LLC and domestic beer appeared on the consumer market of the Republic. It is also worth noting that the Republic has formed a complete production chain for the production of alcoholic beverages, which allows us to produce quality products at affordable prices.

Stable work of food industry enterprises depends on the availability of agricultural raw materials. For this reason, certain difficulties have recently been felt, in particular, by enterprises of the dairy industry, milk yield is declining annually. Over the past year, they decreased in the republic by 3%, since the beginning of this year – by another 6%. If in 2018 milk sales prices decreased from March to August, then in 2019 already in June they began to rise. In August, agricultural enterprises sold it more expensive than in May, by 23%. In January-August, average selling prices for milk increased 1.6 times in comparison with January-August last year.

The financial problems of many food enterprises are associated with the rise in price of agricultural raw materials and energy. Manufacturers of food products are often limited in the corresponding increase in prices for finished products, because the food industry is a socially significant industry. If in 2018, the profit of food enterprises of the republic amounted to about 214.7 million rubles, then in 2019 – 34 million rubles. The difficult financial condition of enterprises negatively affected their 44

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

investment activity. Investments in fixed assets of the food industry of the republic, amounting to 100.1 million rubles in 2018, decreased 2.9 times compared to 2017.

The state of the DPR food industry today continues to be characterized by the presence of a number of difficult to resolve problems, bottlenecks that impede the development of its competitive potential. There remains a high proportion of old food enterprises with a low technological level of production that do not require highly qualified personnel. Rising prices for raw materials, tightening customs legislation, incomplete business complexes, open production chains, insufficient agricultural production, ensuring the sale of products, professional employment, and the low solvency of the DPR population put many food production companies on the brink of survival.

High quality products, being one of the main factors in the development of food industry enterprises, ensures their competitiveness and economic efficiency. Studies conducted at the main enterprises of the food industry of the Republic showed that the main factors affecting the quality of food sold are technical, organizational, logistic, socio-economic, legal factors that have different effects on the quality of finished products.

These factors in combination determine the level of organization of production and labor of workers and are the basis for improving product quality. In turn, the high quality of the products is the most important factor in increasing its competitiveness.

In addition to improving product quality, a large reserve of competitiveness growth lies in improving the marketing policy of the enterprise, the effectiveness of which allows you to generate high demand for products and increase customer satisfaction. Other factors affect the competitiveness of products, such as the price of the product, its packaging, its own logistics, consumption structure, reputation of the enterprise, and the behavior of competitors. A study of the activities of the food industry enterprises in the DPR showed that the main and priority areas for increasing the competitiveness of manufactured products are improving the state of factors that determine the quality of products, as well as improving marketing activities at enterprises.

Recommended measures for each specified area of improving food quality are as follows: - increase the level of material and technical base of production; - implementation of a quality control system; - development of programs to improve the quality of raw materials and their implementation in production activities; - improvement of the organization of labor and production; - advanced training of personnel and quality of work of employees; - modernization of the marketing policy of the enterprise.

The study showed that the level of development of the logistics system in the Donetsk region needs to actively stimulate the formation and development of supply chains. A comprehensive analysis of the main advantages and disadvantages of existing systems implemented and operated in specific areas, taking into account their specifics, made it possible to formulate an approach to the development of the logistics system 45

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

with the involvement of external relations. Studies in this direction at the main enterprises of the food industry of the Republic showed that the main factors affecting the quality of food sold are technical, organizational, logistic, socio-economic, legal factors that have different effects on the quality of finished products.

References 1. Водолазская Н.В. Методика выбора транспортно - логистических систем энергоемких производств с использованием сетевых методов. //Проблемы развития транспортной логистики: тез. докл. 3-ой медунар. научн.-практ. конф. «Интер – ТРАНСЛОГ’2011» 25-30 сентября 2011 г. / Одесса, Украина – Несебр, Болгария. – Одесса : ОНМУ, 2011. – С. 165 – 168.

2. Главное управление статистики Донецкой Народной Республики официальный сайт // URL: http://glavstat.govdnr.ru.

3. Дыбская В.В., Сергеев В.И. Мировые тренды развития управления цепями поставок // Логистика и управление цепями поставок. – 2018. – № 2 (85).

–С. 3-14.

4. Каталог предприятий Донецкий Народной Республики // Министерство промышленности и торговли Донецкой Народной Республики официальный сайт // URL: http://mptdnr.ru.

5. Министерство экономического развития Донецкой Народной Республики официальный сайт // URL: http://mer.govdnr.ru. // URL: http://mptdnr.ru.

6. Промышленный потенциал Донецкой Народной Республики // Министерство промышленности и торговли Донецкой Народной Республики официальный сайт 7. Vodolazskaya N. Models of network planning and management of powerconsuming industries / N. Vodolazskaya // Application of newtechnologies in management. ANTiM 2009. Proceedings. Vol.2. – Vrnjačka Banja. Serbia, 2009. P. 811-818.

8. Vodolazskaya N. To a question of providing a sustainable development of regional production systems of various level / N. Vodolazskaya // Wspόlpraca Europejka (European cooperation) (ISSN 2449-7320). Vol. 8 (15). Warszawa, Polska.

2016. Р.64-70 Аннотация. В настоящей статье основное внимание уделяется анализу особенностей зарубежного опыта формирования логистических систем на современном этапе развития мировой экономики, анализу современного экономического состояния промышленности Донецкой Народной Республики, характеризующего условия формирования цепей поставок. На основе выполненного анализа сформулированы перспективные направления развития логистических систем пищевой промышленности региона. Ключевые слова: цепь поставок, логистическая система, материальные потоки, оценка производства, управление, анализ.

46

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Сведения об авторе: Хараджа Татьяна Андреевна – аспирант, кафедра «Экономика предприятия и инноватика», ДонНТУ.

UDC 004.8, 004.93 SOLVING THE PROBLEM OF PATTERN RECOGNITION USING CONVOLUTIONAL NEURAL NETWORKS Kolbasov S.Y., Kushnirenko Ye.N. kolbasovsergey02@gmail.com

Abstract. The work is dedicated to an overview of the solution to the problem of pattern recognition using convolutional neural networks. The article also describes the problems of the pattern recognition and the main steps of convolutional neural networks.

Keywords: convolutional neural networks, pattern recognition, computer vision, classification.

In the digital age, significant discoveries and achievements are made every year, which undoubtedly improve, facilitate and extend human life. In an attempt to achieve the set goals, it sometimes takes several generations of scientific activity, this is due to many reasons, such as the lack of various technologies, methods of observation, and maintenance, which will allow us to collect huge statistical data during empirical observations. One such problem is pattern recognition. The need for such recognition arises in a variety of areas – from military affairs and security systems to the digitization of analog signals. There are a huge number of tasks, the solution of which is difficult, and sometimes impossible, without the participation of specialized programs for pattern recognition:  optical character recognition;  barcode recognition;  license plate recognition;  face recognition;  speech recognition;  image recognition;  classification of documents; Studying the features of solving the problem of pattern recognition using convolutional neural networks.

Pattern recognition The creation of devices that perform the function of recognizing objects in a video stream, in most cases, will allow replacing a person with a specialized machine.

47

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Currently, many scientists are working in the field of research to solve the problem of automatic recognition of objects in the video stream. This is stimulated by the practical requirements associated with the creation of military and commercial systems.

There are several serious problems in the object recognition problem:  strong dependence on the initial parameters (you need to know quite a lot about those objects that need to be searched in the video stream);  object identification (various similar objects may be similar, for example, a motorcycle and a bicycle);  intraclass variability.

The main task of recognition is the need to answer the question «Where in the image (and is there) the desired object?». The structural approach involves an answer in the form of a bounding box around an object (Fig.1). The task is complicated by complex backgrounds in the images and a sufficiently large intraclass variability of objects. For example, there are hundreds of types of ordinary chairs.

Figure 1 – Bounding box around recognized object (face) Formally, the pattern recognition problem can be described by two devices: a sensor and a classifier. The sensor converts the physical characteristics of the object to be recognized into a set of features x = (x1 ... xn) that characterize the given object [1].

Let X be the set of descriptions of objects. W = (w1 ... wm) is a finite set of class numbers. We assume that the recognition system makes a mistake if it assigns to the class wi an object that actually belongs to a class other than wi. The recognition system R1 is better than the system R2, in the event that the probability of an error for the system R1 is less than for the system R2. The sensor provides information in the form of a vector x = (x1, x2, ... xn), where n is the number of measured characteristics of each 48

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

physical object. It is assumed that the vector of dimensions (features) x belongs to one of the M classes w1 ... wm.

The decisive function is a function y(x) that relates x exactly to one of M given classes. The optimal function is the one that gives the least probability of error for all permissible values of x.

A function y assigns a collection x to the class wi if and only if the inequality p(x | wi) > p(x | wj) ∀i ≠ j holds.

Convolutional neural network The best results in the field of face recognition were shown by the convolutional neural network (CNN), which is a logical development of the ideas of such neural network architectures as cognitron and neocognitron. The success is due to the possibility of taking into account the two-dimensional image topology, in contrast to the multilayer perceptron.

CNNs provide partial resistance to changes in scale, displacement, rotation, change of angle and other distortions. CNNs combine three architectural ideas to ensure invariance to scaling, rotation, shifting, and spatial distortions:  local receptive fields (provide local two-dimensional connectivity of neurons);  general synaptic coefficients (provide detection of some features anywhere in the image and reduce the total number of weighting factors);  hierarchical organization with spatial subsamples.

Now, convolutional neural network and its modifications are considered the best algorithms for finding objects on the scene in terms of accuracy and speed. Since 2012, neural networks have been at the forefront of ImageNet's renowned international pattern recognition competition.

CNN is such a special kind of feedforward neural networks. Feedforward means that the variable neurons in this network are divided into groups called layers.

Moreover, when such a layered neural network is applied to the data, the activation of the layers (the value of these variables) is calculated sequentially: first, the activation value of the first layer, then the activation value of the second layer, and so on until the last layer. The activation of the last layer serves as the output of the neural network, and in this neural network there are many parameters, each layer has its own parameters that determine how the activation of the next layer depends on the activation of the previous layer. In addition, what’s even more important, activations inside one layer can be counted in parallel, at the same time, they are independent of each other, and this leads to the fact that such neural networks can be very conveniently and efficiently calculated on modern processors, including graphic coprocessors.

The architecture of the CNN can contain a large number of different layers, but the main layers, which are usually present, are the convolutional, pooling (subsampling) and fully connected layers (Fig.2).

49

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 2 – Architecture of the convolutional neural network The convolutional layer is named after the convolution operation, which is the main process of this layer. The work of the layer is to use a filter (usually a 3x3 matrix), the values of which are multiplied by the initial values of the image pixels, after which these values are summed. Each unique position of the source image produces a number, the totality of such numbers is called a feature map.

The pooling layer (subsampling) performs the operation of downsampling the spatial dimensions (width and height) of the original image, as a result of which the volume is reduced [3]. That is, non-linear compaction of the feature map is performed.

Usually a pooling layer uses a 2x2 window with a step of 2, that is, at each step only unique pixels enter the window. However, different step sizes and window sizes may be used in various architectures. Compaction is performed by selecting pixels that have fallen into the window at the current step. The following basic methods are used for selection: selecting the pixel with the maximum value or finding the average value among all the pixels that fell into the window. The main logic of the pooling layer is that if some features were detected on the previous convolutional layer, then such a detailed image is not required for further processing, therefore it is compressed to a less detailed one.

The fully connected layer outputs an N-dimensional vector (N is the number of classes) to determine the desired class, that is, it performs the classification. The work of this layer is organized by referring to the output of the previous layer (feature map) and determining the properties that are most characteristic of a particular class. The fully connected layer, in essence, is a multilayer perceptron. CNN usually contains several such layers. However, the use of fully connected layers also imposes a restriction on the size of input images – they all must be the same size, so some modern CNNs do not use this layer and perform classification in other ways, which allows using images of any size during both training and classification.

Training the neural network that is, setting its parameters, takes place on a large number of training data as follows: for each training example, it is known what needs to be obtained at the output. The current state of the neural network is taken and applied 50

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

to a specific training example, then it looks at what happened at the output and how it differs from what was expected. Further, by means of simple mathematical manipulations called the error back propagation method, we can understand how exactly it is necessary to modify the parameters of each layer so that the output starts to get something more similar to what is expected [2]. And the neural network, in turn, looks at individual examples, or more often at sets of examples at the same time, in order to increase the accuracy of calculations, and, looking at these separate examples or packages of examples, it gradually adjusts its parameters and ensures that it starts to predict what’s in the training examples need.

The advantages of this approach are a number of features.

1. Compared to a fully connected neural network (such as a perceptron), there are much fewer customizable weights, since one core of the weights is used entirely for the entire image, instead of having its own personal weighting factors for each pixel of the input image. This encourages the neural network during training to generalize the information displayed, rather than per-pixel memory of each shown image in a myriad of weighting factors, as the perceptron does.

2. Convenient parallelization of calculations, and therefore, the ability to implement algorithms and network training on GPUs. 3. Relative resistance to rotation and shifting of the recognized image.

4. Training with the classic error backward propagation method.

However, with such huge advantages, there are a number of significant disadvantages:  poor translational invariance and lack of orientation information (CNNs have problems turning objects or changing lighting conditions – to solve these problems somehow – data augmentation is applied);  strong loss of information on pooling.

Conclusions. In the course of this study, the problem of pattern recognition was presented as one of the methods for classifying and identifying objects, the importance of solving this problem for the development of modern science. As a solution to this problem, convolutional neural networks were presented, description of the algorithm, the advantages and disadvantages of this approach to pattern recognition was given.

References 1. Ту Дж. Принципы распознавания образов / Дж. Ту, Р. Гонсалес. – СПб: Мир, 1978. – 414 с.

2. Рашид Т. Создаем нейронную сеть / Т. Рашид. – СПб: Альфа-книга, 2017. – 274 с.

3. Lawrence S. Face recognition: a convolutional neural-network approach / S. Lawrence, C.L. Giles, A.C. Tsoi, A.D. Back // IEEE Transactions on Neural Networks – 1997. – Vol. 8(1) – P. 98-113. Аннотация. Работа посвящена обзору решения проблемы распознавания образов при помощи сверточных нейронных сетей. Приведены основные 51

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

проблемы задачи распознавания образов, описаны основные шаги работы сверточных нейронных сетей. Ключевые слова: сверточные нейронные сети, распознавание образов, компьютерное зрение, классификация. Сведения об авторах: Колбасов Сергей Юрьевич – студент группы ПОИСм-19, ДонНТУ Кушниренко Елена Николаевна – ст. преподаватель кафедры английского языка, ДонНТУ

UDC 622.012 COMPUTER MODELING OF MINING DEVELOPMENT STRATEGIES IN COAL MINES Kolomoets A.S., Skazhenik V.B., Kaverina O.G. nastya.kolomoets@yandex.ru

Abstract. In order to justify the need for dynamic adjustment of design decisions at the coal mine, the analysis of deviations in the actual terms of working off the faces from the planned ones was carried out. The proposed modelling approaches and justification of policies for the development of mining.

Keywords: mine, mining, strategies, computer modeling, systems analysis.

The purpose of the work is to substantiate methodological approaches to computer modeling of mining development strategies in a coalmine.

Objectives:  analysis of existing software tools for reservoir modeling;  construction of a model for the development of mining operations of a coal mine;  comparison of possible mining development strategies.

Common systems for modeling deposits in the world, as a rule, cover the entire complex of geological and mining tasks. The analysis identifies the following systems for coal deposits: Geovia Minex – is a specialized package for geological modeling, design and planning of reservoir mineral deposits. Unlike most traditional mining and geological packages, Minex does not work with wireframe models, but with grid surfaces, which greatly simplifies the process of constructing strata of various capacities.

Surpac – allows you to create skeleton geological models by geological cuts. The bedrocks will be rebuilt based on surveyors' profiles or geological sections, and qualitative characteristics will be assigned by interpolation to a block data model.

52

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Micromine – provides construction of models of deposits of minerals both ore and coal deposits. Depending on the parameters of bedding and the chosen modeling method, the following ways of interpreting coal seams are possible:  Classical delineation.  Interpretation of the roof / sole.  Interpretation on the central line.

In this regard, the software complex Shakhta 3D developed with the participation of DonNTU employees was used for modeling coalmines.

The geological model of coal seams is created on the basis of reconnaissance data. At the same time, both geological reports data and data presented on the mine work plans can be used. The model in the form of surfaces may reflect data on geological disturbances.

For simplified visualization of the coalmine, it is possible to model the surface and formations based on the isohypse of the seams. Through the triangulation of a network of points, a wireframe model of surfaces is formed. Provides visualization of reserves using a different color gamut in order to improve the perception of the model.

The existing mining network is being built in accordance with the mining plan.

Digitize the sides of the workings on the soil, and then in the program for a given section and shape of the workings are formed three-dimensional objects.

A set of graphical tasks for building a coalmine model and ways to solve them are presented in table 1.

The software package contains special tools for building project workings. For reservoir and field design workings, the construction procedure is different. First, this is due to the fact that the trajectory of reservoir workings must correspond to the surface of the formation and with complex hypsometry, the development route can have a complex shape. For such situations, the mode of linking the workings to the surface is implemented. To build a mine, you need to set the direction vector, the length of the mine, and its cross section. Field workings, as a rule, cannot be uniquely linked to existing surfaces, so they are linked to a pre-defined plane.

At the «Komsomolets Donbassa» mine coal mining is carried out in 7 mining slaughter on 4 layers. For the timely preparation of stocks in a simultaneous operation can be up to 20 preparatory faces. Under such conditions, it is difficult to make decisions on the possible development of mining operations.

53

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Table 1 – Set of graphical tasks for building a coal mine model

Using the model, three possible mining development strategies were analyzed in table 2.

54

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Table 2 – Possible option for the development of mining operations

The design indicators of possible options are shown in table 3.

Table 3 – Results of comparison of options

55

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Conclusion.  approaches to the construction of a model of a coal mine allowing to reflect the development of a mining facility in time and space  quickly calculated the indicators of the options  selected the most effective option.

References 1. Бесперстов А. С. Моделирование пластовых месторождений при помощи ГГИС Micromine // Горная промышленность. 2011. № 5(99). С. 86-88.

2. Васильев И. Д., Мельникова О. А. Решение прикладных горногеологических задач с использованием программного обеспечения Gemcom // Горная промышленность. 2011. № 5(99). С. 90-92.

3. Стагурова О. В. Опыт применения горно-геологической системы GEMS на предприятиях СНГ // Горная промышленность. 2005. № 5. С. 36-39. Аннотация. С целью обоснования необходимости динамической корректировки проектных решений на угольной шахте проведен анализ отклонений фактических сроков отработки забоев от плановых. Предложены подходы к моделированию и обоснованию стратегий развития горных работ. Ключевые слова: шахта, добыча, стратегии, компьютерное моделирование, анализ систем Сведения об авторах: Коломоец Анастасия Сергеевна – аспирант 1 курса, ДонНТУ Скаженик Владимир Борисович – к.т.н., доцент, доцент кафедры «Управление производством», ДонНТУ Каверина Ольга Геннадиевна – д.пед.н., проф., профессор кафедры английской филологии, ДонНТУ

DC 004.31:519.683.8 GENERAL-PURPOSE COMPUTING ON GPU WITH CUDA Komarichev R.E., Girovskaya I.V. komarichev_pi15b@mail.ru

Abstract. The common differences between CPU and GPU are described in the article, the work of CUDA is reviewed. Two common GPGPU tasks are analyzed and explained as an example.

Keywords: parallel computing, gpu, gpgpu, multiprocessor, cuda.

56

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Everyone who has ever tried programming, started with small scripts that resolved simple tasks. Such programs do not need high computing power and can be easily executed on CPU (central processing unit) with a single core. You can walk a long way never facing a problem of power lack. However, there are so many cases where sequential solution won't be quick enough to be used in real-life situations.

It is a well-known fact that CPU and GPU (graphics processing unit) are both processors, but they're optimized for different types of objectives.

The main task of CPU is to execute chains of instructions as quickly as possible.

It is designed to perform several chains at the same time or to split one chain of instructions into many and after that merge them together again. Each instruction depends on another one and that's why CPU has so few computing cores. All the emphasis is on speed of execution and reduction of downtime that is achieved with cache and pipelines.

The main task of GPU is graphics rendering and visual effects. In fact, it works on a huge number of independent tasks, so it has much more memory but not as fast as CPU has. Also modern GPU has thousands of computing cores whereas CPU contains 2 – 8.

There are many differences in multithreading support. CPU executes 1 – 2 computing threads per core. GPU can launch several thousands threads for each using its multiprocessor. Switching between threads on CPU costs hundreds of clock cycles, GPU switches several threads in a single clock cycle. In CPU most of chip area is occupied by instruction buffers, hardware branch prediction and huge cache sizes, while in GPU most of area is execution units (Fig. 1) [1].

Figure 1 – CPU and GPU chips One of the most common operation performsed on GPU much faster than on CPU is matrix multiplication. Let's say we have two square matrices A and B and their product is matrix C. According to the rules of matrix multiplication, each C's element is sum of products A's row and B's column.

For N=100 we have to perform (100+99)*100*100=1990000 math operations, not to mention index increments. Pretty much to compute sequentially, and the bigger N the longer it takes to compute. Each C's element is independent, so we don't have to 57

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

wait for the C1,1 result to start computation of C1,2 or any other. Thus, we can efficiently sequence these tasks.

The popular technology for such purposes today is CUDA – parallel computing architecture developed by NVIDIA Corporation that can significantly increase computing performance using GPUs. It is widely used by software developers, scientists and researchers in various fields such as video and image processing, computational biology and chemistry, modeling fluid dynamics, reconstructing images obtained by computed tomography, seismic analysis and more. To use this technology you need any NVIDIA GeForce 400 series video card or later and C/C++ programming skills. Specifications of latest cards are described below (Table 1) [2-4].

Table 1 – Specifications of latest NVIDIA video cards Model

Number of CUDA cores

Base Clock (MHz)

Memory amount

GTX 1050

640 768

1354 1392

2 GB GDDR5 3 GB GDDR5

GTX 1050 Ti

768

1290

4 GB GDDR5

GTX 1060

1152 1280

1506 1506

3 GB GDDR5 6 GB GDDR5X

GTX 1070

1920

1506

8 GB GDDR5

GTX 1070 Ti

2432

1607

8 GB GDDR5

GTX 1080

2560

1607

8 GB GDDR5X

GTX 1080 Ti

3584

1481

11 GB GDDR5X

GTX 1650

896 896

1485 1410

4 GB GDDR5 4 GB GDDR6

GTX 1650 SUPER

1280

1530

4 GB GDDR6

GTX 1660

1480

1530

6 GB GDDR5

GTX 1660 Ti

1536

1500

6 GB GDDR6

GTX 1660 SUPER

1408

1530

6 GB GDDR6

RTX 2060

1920

1365

6 GB GDDR6

RTX 2070

2304

1410

8 GB GDDR6

RTX 2070 SUPER

2560

1605

8 GB GDDR6

RTX 2080

2944

1515

8 GB GDDR6

RTX 2080 SUPER

3072

1650

8 GB GDDR6

58

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

CUDA C++ extends C++ and allows us to define C++ functions, called kernels, that are xecuted N times in parallel by N different CUDA threads. To define the count of parallel threads and distribute data CUDA used concepts of blocks and grids. Call to kernel launches grid of blocks, each block has many parallel threads (Fig. 2).

Figure 2 – Grid of thread blocks Let's do some experiments and see how much benefit we can get with CUDA. The hardware is Intel Core i7-8750H vs NVIDIA GeForce RTX 2060. In each experiment we will multiply two square matrices of size NxN on CPU (single thread), GPU, and measure the time elapsed (Table 2). Blocks will always have dimension 32x32 and grid size will be calculated according to this fact and to size of matrices so that each element of C is computed in its own thread.

Table 2 – Time costs of matrix multiplication on CPU and GPU N

CPU, ms

GPU, ms

Acceleration

256

101

2

x50,5

512

542

9

x60,2

768

1789

29

x61,7

1024

5013

70

x71,6

1280

10278

177

x77,3

1536

20925

220

x95,1

1792

38214

314

x121,7

2048

59879

420

x145,3

59

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

As we can see, using GPU operation takes much less time. Figure 3 illustrates the result acceleration.

Figure 3 – Matrix multiplication on CPU and GPU. Time vs N (above) and acceleration vs N (below) In fact, it should be noted that computing on GPU requires additional overhead on memory exchange. Every time we want to perform GPU computation, we have to send input data from host memory to device memory and, when it is done, send output data back, whereas when using CPU, all data is on host all the time. In this experiment we have few data and its transfers almost didn't take time and results are still relevant.

However, we should keep this in mind.

To see the influence of memory exchanging let's do another experiment. We have two enormous arrays of numbers A and B and want to find array C of the same 60

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

size where Ci = Ai + Bi. In this case the number of operations per element is 1 (just one sum), unlike the previous one, where it was 2N-1 (calculate N products and find their sum). Thus, one element is faster to compute and we can take much larger N without increasing the experiment duration (Table 3).

Table 3 – Time costs of array summing on CPU and GPU GPU w/o One transfer GPU with CPU, transferring, of N transferring, Acceleration ms ms elements, ms ms

N 230 = 1073741824

2374

42

165

537

x4,4

231 = 2147483648

4714

90

316

1038

x4,5

232 = 4294967296

9431

176

640

2096

x4,5

233 = 8589934592

18907

366

1290

4236

x4,5

234 = 17179869184

37678

706

2650

8656

x4,4

As we can see, execution time increases linearly (Fig. 4) because different N does not affect the amount of work for one element. Acceleration now is much worse because of the need to transfer huge amount of data between the host and the device.

Figure 4 – Time costs of arrays summing on CPU and GPU The conclusion is that the efficiency of GPU-accelerated computing is directly proportional to the number of operations performed independently on an individual set of data.

References 1. CUDA C++ Programming Guide [Электронный ресурс] // CUDA Toolkit Documentation.

URL: https://docs.nvidia.com/cuda/cuda-c-programming61

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

guide/index.html (дата обращения: 15.04.2020).

2. GeForce 10 Series Graphics Cards [Электронный ресурс] // NVIDIA. URL: https://www.nvidia.com/en-us/geforce/10-series/ (дата обращения 15.04.2020).

3. GeForce GTX 16 Series Graphics Card [Электронный ресурс] // NVIDIA. URL: https://www.nvidia.com/en-us/geforce/graphics-cards/16-series/ (дата обращения 15.04.2020).

4. GeForce RTX 20 Series and 20 SUPER Graphics Cards [Электронный ресурс] // NVIDIA. URL: https://www.nvidia.com/en-us/geforce/20-series/ (дата обращения 15.04.2020). Аннотация. В статье описаны общие различия CPU и GPU, выполнен обзор базовых принципов работы CUDA. На примере проанализированы и объяснены две распространённые задачи GPGPU. Ключевые слова: параллельные вычисления, gpu, gpgpu, мультипроцессор, cuda. Сведения об авторах: Комаричев Руслан Евгеньевич – студент группы ПИм-19, ДонНТУ Гировская Ирина Валерьевна – ст. преподаватель кафедры английского языка, ДонНТУ

UDC 658.51 THEORETICAL ASPECTS OF MODERN STRATEGIC MANAGEMENT OF THE ENTERPRISE Kosogor D.V., Borshсh I.V. diana.kosogor@gmail.com

Abstract. The article discusses some theoretical aspects related to the strategy of the enterprise and strategic management of a modern enterprise. Definitions of the concept of «strategic management» of various authors, characteristic features of the strategy and the evolutionary path of development of strategic enterprise management are presented.

Keywords: strategic management, strategy, enterprise.

Today the strategic management of the enterprise is one of the main factors in ensuring its competitiveness in the modern market, therefore the issues of strategic management have become relevant for Russian enterprises. Successful business management, effective management at all organizational levels and rational allocation 62

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

of resources are the result of strategic enterprise management. The importance of strategic management of the enterprise is also due to the instability of the external environment, which entails unforeseen threats and possible risks. s in order to respond in a timely manner to changes in the trends in the market of goods and services, to be able to anticipate, prevent or reduce the consequences of possible risks, and create conditions for organizational development.

The development of strategic management is determined by the historical framework, that is, the transition to a market economy, and, consequently, the rejection of the traditional five-year plans and the transition to the realization of the need for the functioning of the national economy in accordance with economic laws. The determining event in the establishment of strategic management was the realization of the inappropriateness of using long-term planning and the use of a constant reassessment of the initial concepts of development of economic systems, taking into account changes in the internal and external environment.

A significant contribution to the development of strategic management was made by I. Ansoff, M. Porter, H. Wissema, D. Cleland, W. King, J. B. Quinn, G. Alstrend, S. Goshal, G. Mintzberg, A. J. Strickland, V. Gerasimchuk, S. Oborskaya, S. Shershneva, S. Popov, M. Tulenkov, A. Vikhansky and others. Approaches to the definition of the theory of strategic management are presented in table 1 [1].

Table 1 – Approaches to the definition of the theory of strategic management Author 1 I. Ansoff

J.M. Higgins H. Wissema

A.A. Thompson A.J. Strickland

Definition 2 Strategic management as an activity related to the determination of the goals and objectives of the organization and ensuring the relationship between the organization and the external environment, consistent with its internal capabilities and allows you to remain susceptible to external requirements.

Strategic management as a management process with the goal of fulfilling the organization’s mission by managing the interaction of the organization with its environment.

Strategic management as a management style and methods of communication, information transfer, decision making and planning, with the help of which the management apparatus and line managers make timely and concretize decisions regarding the goals of entrepreneurial activity; as predictions of the strategic orientation of all employees and the organization of plans for units responsible for implementing the goals of the company.

Strategic management as a continuous process of company development, setting goals, formulating a strategy, implementing a strategic plan and evaluating activities, implementing and correcting strategies.

63

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Z.S. Shershneva S.V. Oborskaya A.S. Vihansky

Strategic management as a broad-spectrum, formal-behavioral management process that helps to formulate and implement effective strategies that help balance the relationship between the organization, including its individual parts, and the external environment, as well as achieve the goals.

Strategic management as the management of an organization that is based on human capital, directs production activities to consumer demands, flexibly responds and makes timely changes in the organization, makes it possible to gain competitive advantages and survive in the long term.

Thus, strategic management is a process of making and implementing strategic decisions, the central link of which is strategic choice, based on a comparison of the enterprise’s own resource potential with the capabilities and threats of the external environment in which it operates. Strategic management in an enterprise is a set of actions and decisions taken by management that lead to the development of specific strategies designed to help him achieve his goals.

The strategy, in turn, is a general program of actions that identifies priority areas of the enterprise, its main problems and alternatives for solving them, as well as resources to achieve the main goal. It formulates the main goals and the main ways to achieve them in such a way that the company receives a single direction of movement and a vector of development.

In order to determine the organization’s behavior strategy and put this strategy into practice, management must have an in-depth understanding of both the organization’s internal environment, its potential and development trends, and the external environment, its development trends, and the place the organization occupies in it.

The main classification of existing strategies of organizations can be presented in the form of a table 2 [3].

Table 2 – Classification of strategies Enterprise Strategies Types of strategies General strategies  Growth strategy  Stabilization or limited growth strategy  Survival or reduction strategy Functional strategies  Marketing Strategy  Financial strategy  Innovation Strategy  Production Strategy  HR strategy  Social strategy 64

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

 Environmental strategy Competitive strategies  Leader Strategy  Leader follow-up strategy  Challenging Strategy  Nicher Strategy  Cost Optimization Strategy  Differentiation Strategy  Strategy for concentrating efforts on the segment Commodity Market  Commodity-market strategies for each product Strategies  Commodity market strategies for each market Corporate strategy is the strategy of the entire company as a whole, it includes a business strategy for each area of the company’s activities, a functional strategy for each functional unit of the company and an operational strategy for the main structural units within the functional units of the company. As a rule, a large enterprise has three levels of strategic decisions: corporate, business, functional; therefore, the corporate strategy of the enterprise also covers these levels.

The structure of corporate strategy in general terms for the enterprise is presented in fig. 1.

Corporate strategy

Corporate Responsibility

Business Strategy

Responsibility of company managers (subsidiaries)

Functional strategies (marketing, manufacturing, finance, staff)

Responsibility of Heads of Departments

Operational strategies

Responsibility of company managers

Figure 1 – The structure of corporate strategy in General It should be noted that the strategy:  gives a definition of the main directions and ways to achieve the goals of strengthening, growth and ensuring the survival of the organization in the long term based on the concentration of efforts on certain priorities; 65

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

 is a way of establishing the interaction of the company with the external environment;  is formed on the basis of generalized and insufficiently accurate information in connection with the instability of the external environment;  is constantly updated in the process of activity, which should contribute to a well-established feedback;  the enterprise has a complex internal structure because of multipurpose nature;  is the basis for the development of strategic plans, projects and programs, which are a systematic characteristic of areas and tools for enterprise development;  is the basis for the formation and implementation of changes in the organizational structure of the enterprise, a generalized core of the activity of all levels and strategic management links;  is the main substantive element of the activities of senior management personnel;  acts as a factor in stabilizing relations in the organization in view of understanding the content and goals of organizational activity at all levels;  allows you to establish effective motivation, accounting, control, analysis, acting as a standard that defines development and allows you to evaluate the results achieved [2].

In the past, enterprise management changed significantly almost every decade.

The conditions for doing business changed, and enterprises faced the task of resolving issues of achieving goals in a new way, and of approaching the search for means of survival in the competition. And each time the concept of “strategic” enterprise management acquired a special meaning, often diametrically opposed to that which was previously invested in it.

The evolutionary development of the concept of strategic management can be briefly presented in the form of a table 3 [3].

Table 3 – The evolution of the concept of strategic management Period The first half of the ХХ century

Concept basis Using long-term planning. The beginning of practical application in management

70s

Practical application of strategic planning

80s

The formation of the concept of strategic management

The end of the 90s – the

Strategic entrepreneurship concept 66

Background of change The need for a quality response to change internal and external environment The problem of growth and development in the face of increased competition Globalization processes, the formation of transnational

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

beginning of the XXI century Present

corporations, fierce market competition

Innovation Management Strategy Concept

The requirements of the modern globalized economic space, the development of scientific and technical progress

Taking into account the analysis of the table. 1.3, we can conclude that the concept of strategic management is constantly improved and developed, adapting to the requirements of the modern globalized economic space, the development of scientific and technical progress, the emergence of new technologies, the introduction of innovations, the transformation of enterprise management as a process, etc. Thus, it is advisable for Russian enterprises to recognize the importance of strategic management and conduct activities based on the concept of an innovation management strategy. All world trends indicate that business development is no longer driven by spontaneity as a success factor and only the transition to strategic enterprise management is the first step to future success.

References 1. Касьян Л. Е. Методические подходы к определению сущности стратегического управления предприятием / Л. Е. Касьян, В. В. Бугас // Научный вестник Международного гуманитарного университета. – 2013. – №3. – С. 94-96.

2. Кайлюк Е. Н. Стратегический менеджмент: учеб. пособие. / Е. М. Кайлюк, В. М. Андреева, В. В. Гриненко; Харьк. нац. акад. гор. хоз-ва. – М .: ХНАГХ, 2010. – 279 с.

3. Попрозман Н. В. Концептуальные элементы стратегического управления / Н. В. Попрозман, О. И. Попрозман // Формирование рыночных отношений в Украине. – 2015. – №4 (167). – С. 18-22. Аннотация. В статье рассмотрены некоторые теоретические аспекты, касающиеся стратегии предприятия и стратегического управления современным предприятием. Представлены определения понятия «стратегическое управление» разных авторов, характерные признаки стратегии и эволюционный путь развития стратегического управления предприятием. Ключевые слова: стратегическое управление, стратегия, предприятие. Сведения об авторах: Косогор Диана Вячеславовна – студент группы МОм-19, ДонНТУ Борщ Ирина Владимировна – ст. преподаватель кафедры английского языка, ДонНТУ 67

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

UDC 004.056.55 THE FIVE-LEVEL DATA PROTECTION MODEL Kukhta S.S., Sokolova O.V. kukhtass@mail.ru

Abstract. The five-level data protection model is described in the article. The basic idea of this approach is a message breaking-up into small pieces and their distribution over few different communication channels. This technique can provide the impossibility of reading the message by the interceptor, when he has just one part of the data. The author touches upon the advantages and disadvantages of this model of information transmission.

Keywords: data protection, cryptography, steganography, message interception, information transmission.

Nowadays the protection of transmitting and storing information is based upon the principles created in cryptography and steganography. Using cryptographic methods a protected message is converted into unreadable one without key character set. Steganographic techniques allow us to create a private link, which is hard to be detected even with the use of special information processing methods. The placement of hidden information in the containers occurs with the help of the key set too.

Specialists have made a lot of effective cryptographic attacks on famous cyphers and steganographic methods of protection. The presence of successful attacks indicates the vulnerability of the existing principles of information protection.

The development of information security tools and tools for cyphers attacks, methods of hiding messages have competitive (iterations) character. Usually a widespread cypher is effectively attacked a few years later after its creation and its application gradually fades away. The method of brainstorm concerning the creation of new protection algorithms is stimulated by the international competitions.

A fundamentally new approach of information security may become the method of forming several levels of message protection (Fig. 1). Figure 1 shows five possible information protection barriers. In this case, a message is any transmitting (storing) information or data [1].

68

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 1 – Possible information protection barriers Spatial spraying of information may become one of additional information protection barriers (besides cryptographic and steganographic). The basic idea of spatial spraying of information is in a message breaking up into some much smaller components (sentences, words, symbols, blocks of characters, groups of bytes, bytes, groups of bits, bits) and their delivering in pieces and distributing over several communicational channels (K1…Kn). The interception of all message components by breaker C (Fig. 2) is complicated by the fact that correspondents A and B have the opportunity to use several telecommunication channels available for them (radio, satellite, wire, cable, radio relay).

Figure 2 – Interception of a message by breaker C 69

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Information transmission in global networks is possible due to many existing services (e-mail, messengers, chats, forums, blogs, distributed databases WWW, etc.).

The use of cellular communication allows us to spray a message over several MMS or SMS and transmit them using a large number of telephone channels.

Besides three levels of protection of transmitted information mentioned above another level can be created (the fourth one), which is technically and algorithmically harmoniously combined with the considered barriers. This is a temporary message breaking-up (data transmission according to a pre-agreed schedule).

The spatial and temporal message breaking-up fits together perfectly and complements one another. These two barriers can be represented as a single one and can be called as a spatio-temporal spraying of a message. The idea of spatio-temporal spraying of a message is illustrated in Figure 3.

Figure 3 – Spatio-temporal spraying Information blocks No.1…9 that contain sending message, are transmitted in pseudo-random order by communication channels (K1…Kn). Moments of transmission of message blocks are pseudo-random too. Information blocks transmission is mixed with sending masked (misinforming) blocks No.10…16. Blocks translation order, channel numbers and temporary windows are set with the help of a secret key.

It is worth to admit that if only one telecommunication channel is used for a link a spatio-temporal barrier changes into temporal one.

The fifth (algorithmic) barrier is based on special way of processing of transmitting blocks, when absence of even one of intercepted blocks causes cryptanalyst insurmountable difficulties. The closest ideology (in design and use) to this barrier are encryption modes, which are described in the domestic standard GOST 28147-89, the American standard DES and many publications [2]. The fifth barrier is like a continuation of the cryptographic one, but methodologically it is advisable to divide it into a separate level of protection.

70

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Thus, the principle of complex, multi-level information protection can be implemented by creating various barriers. In each case the participants reasonably choose a sufficient degree of message security (the number of barriers used, the type of cypher, the key length, the types of steganographic containers, the methods for steganographically embedding information in containers, the type and number of communication channels or storage media used, the number of time windows for transmitting information).

It is clear that the implementation of proposed measures of high cryptographic stability is accompanied by an increase of the time transmission, the number of errors in a message transmission. All these complicate the transmission of a message and reduce the convenience of information storage. It is possible to regulate the degree of message protection (that means the variety of time of message transmission and changing service properties) by a selective use of some of barriers (not all of them) but only a sufficient part of them. On-line transmission of information (when the value of information is calculated in hours and even minutes) can occur with a minimum numbers of used barriers.

References 1. Алексеев, А.П. Сокрытие информации в звуковых WAV-файлах [Электронный ресурс] : методические указания по проведению лаб. работ по дисциплине «Информатика» / А.А. Аленин, А.П. Алексеев . – Самара : ИУНЛ ПГУТИ, 2010 . – 11 с. – Режим доступа: https://rucont.ru/efd/278728 2. Алексеев, А.П. Стеганографические и криптографические методы защиты информации [Электронный ресурс] : учеб. пособие по дисциплине «Информатика» / В.В. Орлов, А.П. Алексеев . – Самара : ИУНЛ ПГУТИ, 2010 . – 289 с. : ил. – ISBN 978-5-904029-12-8. – Режим доступа: https://rucont.ru/efd/278727 Аннотация. В статье описана пятиуровневая модель защиты данных/информации. Основная идея данного подхода состоит в разделении сообщения на небольшие части и распределение их по нескольким каналам связи. Данная техника может обеспечить невозможность прочтения сообщения перехватчиком, когда у него есть не все данные, а только одна их часть. Автор рассматривает преимущества и недостатки данной модели передачи информации.

Keywords: защита данных, криптография, стеганография, перехват сообщений, передача информации. Сведения об авторах: Кухта Сергей Сергеевич – студент группы МИДм-19, ДонНТУ Соколова Ольга Викторовна – преподаватель кафедры английского языка, ДонНТУ 71

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

UDC 662.749.33 THE PECULIARITIES OF THE LOW-TEMPERATURE SEMI-COKE PROCESS AND METHODS INFLUENCED THE YIELD AND PROPERTIES OF ITS PRODUCTS Kupich V.A., Butuzova L.F., Boyko V.N. the-best-hidan@mail.ru

Abstract. The article considers the process of low-temperature semi-coking, analyzes the methods of production of semi-coking, and studies the properties of the products obtained, as well as their application.

Keywords: semi-coking, pyrolysis, tar, solid combustible fossil, volatile.

Semi-coking is the process of thermal processing of solid fuels (coal, brown coal, and shale) without air access to temperatures of 500-600°C. For semi-coking, mainly coals with a high yield of volatile substances are used, giving a large yield of the primary resin. The yield of the primary resin and semi-coke depends on the quality of the feedstock, the design and the mode of the furnaces. Most of the volatile substances released during low-temperature pyrolysis, with the exception of free moisture vapors, are formed in the hottest areas of the plastic layer. The products of semi-coking are semi-coke, primary resin, pyrogenetic water, and primary gas. Semi-coking products are called primary products, since they do not undergo far-reaching thermal decomposition processes.

The process is carried out in special semi-coking furnaces. Depending on the heating method, all existing semi-coking furnaces can be divided into two groups: with external heating and with internal heating.

Semi-coke is the main product of the low-temperature pyrolysis process, with a solid residue of up to 90% from coal weight. Semi-coke fineness depends on its size, strength, thermal stability (change in strength, the occurrence of deformation and disintegration into smaller pieces when heated), of the original bulk material coal and semi-coking technology (furnace device and mechanical loads into pieces of coal inside the furnace, speed of heat and temperature.) Usually coals with the size of 2080 mm are used in furnaces for semi-coking.

Semi coke is widely used as an energy-industrial fuel for direct combustion in the furnaces or industrial units such as power, cement, glass and ceramic plants.

Consumption of lumpy half-coke is the most common. Due to its special features, semicoke during combustion makes it possible to provide higher temperatures in furnaces with less fuel consumption.

Semi-coke is also used as household fuel. In the countries where the use of semicoke in household stove is common, a number of special requirements are made to 72

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

semi-coke, primarily ensuring smokeless combustion, smooth lumpiness, etc.

Therefore, only some varieties are used for household purposes.

Recently, semi-coke has been used as a semi processed unit in the production of molded metallurgical coke. Brown-coal coke can be used in the charge for coking in chamber furnaces, where it in some cases successfully replaces the thinning components of the charge. A significant effect is achieved when the crushed semi-coke is injected into the blast furnace, where it plays the role of fuel, as well as a chemical reagent, which allows you to save a significant amount of blast-furnace coke from expensive and scarce sintering coals.

The primary resin is also called «tar» or «semi-coking resin». Liquid products that condense from the vapor-gas phase formed during the semi-coking of combustible minerals are called primary resin. Primary resin differ in composition and properties depending on the nature of the fossil fuels. The density of primary resin is close to the density of water and varies between 0.95-1.05 g / cm3, so the resin are poorly protected from water. The yield of the primary resin is an important characteristic of the semicoking process. Modern high-speed processes are focused on obtaining the maximum possible number of primary resins. Their yield depends on both the genetic characteristics of the solid combustible fossil (SCF) and the technological parameters of the thermal processing. Resins are dark brown liquids; the density of the resins varies from 0.95 to 1.1 g / cm3 depending on the method of semi-coking. The composition of semi-coke resins can include up to 35% phenols, 3-5% olefins, up to 10% naphthenic, 15-25% aromatic hydrocarbons, 1-2% organic bases and 2-10% paraffin hydrocarbons.

Semi-coke resins can be used as raw materials for motor fuel, phenol, paraffin, and aromatic hydrocarbons production. Phenol is used in the production of plastic, lacquer, synthetic fibers, and pharmaceuticals. Paraffin serves as a raw material for the surfactants and detergents production.

Primary gas is a mixture of gaseous products formed during semi-coking. After gas gasoline extracting, resin products consists mainly of methane, its homologues, other hydrocarbons and hydrogen. The composition is also determined by the type of SCF that is subjected to semi-coking. The composition of the primary gas is characterized by a high content of methane and its homologues, which contributes to a high combustion temperature. The main amount of semi-coke gas is spent on heating the fuel and other needs at the enterprise where the semi-coke is carried out. The excess of semi-coke gas can be used as household fuel, as well as for organic synthesis.

Factors affecting the yield and quality of semi-coking products.

The yield and quality of semi-coking products depend on the properties of the fuel being processed, the heating conditions, in particular, the rate of heat input, the final heating temperature, and the pressure. The type of furnaces used, the heating method, the residence time of volatile substances in high-temperature zones, and other factors that determine the uniformity of the temperature field and influence the formation of final products are also of great importance. The yield of resins during semi-coking of brown coals of various types can vary widely – from 4.5 to 15-17%, the yield of resin formed during semi-coking of coal also depends on the characteristics 73

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

of their structure and varies from 1.5 to 20% and above. As metamorphism increases (from gas coals to lean coals), the yield of semi-coke resin decreases. The only exception is fatty coal, which often produces as many primary resins as gas coal when heated to 600°C. The final heating temperature of the fuel significantly affects the yield and properties of the semi-coking products, since high-temperature transformations occur as the heat is supplied. Usually the final temperature of semi-coking is 600°C. At this temperature the resin formation processes are almost completed. Higher-temperature more than 600°C transformations are typical for the transition stage of semi-coke to coke (dehydrogenation, dealkylation, the interaction of hydrogen with nitrogencontaining heterocycles, their subsequent splitting and the formation of ammonia and molecular nitrogen). An additional amount of gases (hydrogen, methane, ammonia, nitrogen, etc.) is formed, the output of the solid residue decreases, and its quality changes. As the temperature increases, the reactivity of the product decreases and the structural strength increases. An increase of the final heating temperature in the real process affects the yield and composition of the resins. To increase the temperature in the load, it is necessary to increase the temperature of the heat transfer gas in furnaces with internal heat supply or the temperature of the heating walls in furnaces with external heating. It causes additional pyrolysis of volatile substances.

Thus, an increase in temperature above 600°C leads to a decrease in the yield of solid product and resins, and an increase in the amount of gases.

An important factor is the heating rate. With more intensive heat input, the output of the semi-coke decreases and the output of the resin increases. When the volatile components pass through the fuel layers, they undergo greater pyrolysis both in the fuel mass, mixing with the more heated coolant, and at the heating walls. As a result of secondary pyrolysis, the resin yield can significantly decrease and the amount of gaseous products can increase.

A noticeable effect on the yield of semi-coking products is provided by the size of the pieces of processed fuel. Usually, the larger the size of the pieces, the less resin but more semi-coke is formed.

A similar effect on the semi-coking process is exerted by an increase in pressure: the yield of the resin decreases, but an additional amount of semi-coking and gaseous products is formed. When the pressure increases, not only the yield of semi-coke increases, but its strength increases as well. It is explained by the difficulty of separating volatiles, increasing their impact with solid and non-volatile liquid-phase products.

There are two main methods of semi-coking, in accordance with the method of heating furnaces for semi-coking, when the obtained products are quantitatively and qualitatively different. Stoves heating can be: internal and external. The simplest type of semi-coking is semi-coking in a vertical (shaft) furnace, in which fuel is supplied from above, and from below the preheated heat carrier gas is supplied evenly throughout the entire section of the furnace. After passing through the thickness of the fuel load, the gas gives its physical heat, due to which the process of semi-coking 74

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

occurs, and is removed at the top, taking from the furnace liquid and gaseous products of semi-coking. This relatively rapid removal of volatile parts of the distillation creates favorable conditions for obtaining primary products. The higher the gas velocity in the furnace is, the more favorable conditions for obtaining primary products are. It is not always possible to supply a large amount of heat transfer gas to the furnace, since the resistance of the fuel layer increases in proportion to the squared velocity as the gas speed increases, and this greatly complicates the process. In industrial conditions, where you have to deal with different in size pieces fuel, compliance with the same regime as for the large fuel or for the fine one is impossible for the same reasons. The most common coolant is gas, produced in the same semi-coke oven.

The internal heating of the furnaces makes it possible to semi-coke some varieties of sintering coals. Due to the fact that the gaseous heat transfer supplied into the furnace reduces the partial pressure of the coal liquid and gaseous products formed during the decomposition of coal, their removal in vapor form occurs before reaching the temperature of the plastic state of the coal and therefore the semi-coke is poorly sintered. If there is not enough coolant, i.e. at low speeds, the bitumen begins to evaporate at the temperature of the formation of a plastic layer, pieces of fuel stick together and coal freezes in the furnace.

In furnaces with external heating, the coal is poured into a chamber whose walls are heated by flue gases produced in a separate furnace. After giving heat through the furnace wall of the fuel load, the combustion products are removed to the atmosphere, and the liquid and gaseous products of semi-coking are removed to the condensation equipment through a special hole in the top of the furnace. The fuel is heated from the wall to the center of the furnace. Semi-coke is therefore unevenly coked throughout the entire thickness of the load and the least amount of volatiles is contained at the walls.

The greater the thickness of the heated fuel layer is the less uniform of the semi-coke will be. Therefore, the design of furnaces with external heating is tended to choose the smallest possible thickness of the fuel layer. As a result, the fuel decomposes more evenly and completely.

Thus, such issues as the process of low-temperature coking, its features, methods of semi-coking, product characteristics, as well as their application, factors affecting the output and properties of products were considered in this work.

References 1. Агроскин А.А. Химия и технология угля. – М.: Недра, 1969.

2. Глущенко И.М. Теоретические основы твердых горючих ископаемых. – М.: Металлургия, 1990.

3. Кафтанов С.В. Общая химическая технология топлива. – М.: Госхимиздат, 1971.

4. Луазон Р., Фош П., Буайе А. Кокс – М.: Металлургия, 1975.

5. Макаров Г.Н., Харлампович Г.Д. Химическая технология твердых горючих ископаемых. – М.: Химия, 1986.

75

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Аннотация. В статье рассмотрен процесс низкотемпературного полукоксования, проанализированы методы производства полукокса и изучены свойства полученных продуктов, а также их применение. Ключевые слова: полукоксование, пиролиз, смола, твердое горючее ископаемое, летучие. Сведения об авторах: Купич Валентин Андреевич – студент группы ХТм-19, ДонНТУ Бутузова Людмила Федоровна – д.хим.н., профессор, кафедра «Химическая технология топлива», ДонНТУ Бойко Виктория Николаевна – ст.преподаватель кафедры английского языка, ДонНТУ

UDC 628.54:628.47 CO-DISPOSAL OF WASTE FROM COKE PLANTS AND MUNICIPAL SOLID WASTE COMPONENTS Kurchenko E.N., Kalinihin O.N., Boyko V.N. elenakurchenko96@mail.ru

Abstract. The article analyzes difficult situation in the field of solid household and industrial waste disposal over the past three years. Features of the implementation of the production of RDF fuel based on household waste components and bottoms are described.

Keywords: solid household waste, RDF-fuel, cube residues, technology, burning.

The choice of research topics is associated with exacerbation of economic, environmental and social problems due to an increase in the volume of produced and utilized solid waste and industrial wastes, the need to search for and introduce more advanced, environmentally and economically sound waste disposal plants.

Currently, there are huge accumulations of municipal solid waste (MSW) in Donetsk region, because there are no fully-fledged effective methods of waste disposal and recycling.

In 2018, 750539.468 tons of hazard class IV wastes were generated, 99.1% of which is disposed in specially designated places and facilities. The rest of the waste is disposed in unorganized storage areas. Wastes of the IV hazard class make up 99.95% of the annual volume of waste of all hazard classes.

76

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Today, one of the main tasks relevant worldwide is the processing and disposal of municipal and industrial waste. Waste management statistics in Europe are shown in Figure 1 [1].

Figure 1 – Household waste treatment in the countries of the European Union Figures released by Eurostat have revealed the enormous gulf between countries in the western and eastern halves of the European Union when it comes to recycling. Germany is leading the way, recycling 62 percent of its waste, incinerating 37 percent and landfilling just one percent.

The Netherlands comes a close second with the same level of landfill and 61 and 38 percent levels of recycling and incineration respectively. Belgium rounds off the top three, recycling 56 percent of its waste.

Sweden, which has been making headlines recently regarding the efficiency of its waste to energy programme, comes fourth. Even though the figures show Sweden 77

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

as landfilling one percent of its waste, it has emerged that the country is actually importing 800,000 tons of rubbish to power its waste to energy plants – burned waste powers approximately 250,000 Swedish homes.

Unfortunately, Sweden’s excellent example shows no signs of being repeated in Eastern Europe where the level of landfill rises dramatically. Croatia, the EU’s newest member, is the fourth worst offender for recycling – 92 percent of Croatian trash ends up in landfill, while eight percent is recycled.

The accumulation and untimely removal of waste creates an environmental hazard to public health due to the content of a large amount of organic substances in them, which form harmful chemical compounds when decomposed.

Energy recovery from waste is the conversion of non-recyclable waste materials into usable heat, electricity or fuel through a variety of processes, including combustion, gasification, pyrolisis, anaerobic digestion and landfill gas recovery. This process is often called energy waste.

MSW can be considered as a strategic resource for generating heat and electricity.

European countries are moving away from the practice of burning household waste, replacing this method with modern resource and energy-saving technologies, alternative energy sources and the reuse of raw materials, which acts as a method of resources saving and the environment preservation.

Refuse Derived Fuel (RDF) is one of such technologies when the fuel is produced from various types of waste, such as municipal solid waste and industrial waste. The main consumers of this fuel are the cement industry, power stations and heating plants [2].

Considering the morphological composition of solid waste, about 25% of the volume of waste transported to landfill can be involved in the production of RDF fuel.

The advantages are that the amount of waste disposed and the consumption of exported energy sources are reduced.

Table 1 shows the indicators of the components of municipal solid waste in Donetsk region, which are of interest as components of RDF fuel [4].

Table 1 – Indicators of waste fractions in MSW № 1 2

Type of waste Paper, cardboard Wood

% mass 20-25 2-3

QP, MJ / kg 13,4 14,7

The caloric content of fuel depends on the amount of combustible fractions in MSW; therefore, it is reasonably to add components with high calorific value to the waste. As such a component, tarry wastes of coke-chemical wastes (CCP), such as coal fusas, polymers, bottoms, etc. can be used.

78

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Let us consider still bottoms of the crude benzene rectification workshop as an additive to fuel briquettes based on solid waste components. They are a suitable waste due to their low ash content with a small amount of sulfur, the value of which is quite high in fossil fuel.

Abroad still bottoms of rectification are pyrolyzed in the presence of water vapor and hydrogen, followed by catalytic dehydrogenation of the gaseous pyrolysis products. Binding residues are obtained from still bottoms. Binding residues are used to form binders for the manufacture of anti-corrosion dyes, cladding plates and other building materials.

The bottoms of the crude benzene rectification workshop are a mixture of products of different polymerization depths of unsaturated compounds with benzene hydrocarbons, thiophene and its homologues, as well as aromatic hydrocarbons with a boiling point above 200 °C, extracted from absorption oil while crude benzene production. In addition, they contain mineral impurities, which are a mixture of sodium sulfate (3.7%), free alkali (0.1%) and sodium sulfonates (14.5%) formed during alkaline neutralization of the fraction after acid washing [3].

Alternative fuel, based on the components of solid waste and bottoms of the crude benzene rectification workshop at coke plants, can be used as a more affordable fuel in terms of price policy, which is not inferior in terms of technical characteristics to brown and hard coal.

The technological process of obtaining RDF consists of two operations: waste crushing and separation of ferrous metals, but if we are limited to only these two operations, the RDF obtained in this case will contain many ballast fractions and be of low quality.

Therefore, additional machines, mechanisms and assemblies are used in the production of granular fuel, that allow enriching, granulating and briquetting fuel from waste, while increasing investment and operating costs, but the resulting fuel has significantly better quality. Schematic diagram of the production of granular fuel is shown in Figure 2.

The method of producing RDF fuel depends on the type of waste, its composition, as well as the subsequent method of use as the main or additional (together with the main – coal, peat, etc.) fuel [4]. The heat of combustion of granular fuel varies 5300 – 17700 kJ / kg.

Many boiler plants need only a little modernization to work on granular fuel, because they are equipped with devices for removing slag and fly ash.

79

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 2 – The main stages of the production of RDF fuel Modern thermal processes are environmentally friendly during the heat treatment of prepared solid waste, in accordance with technological standards and using modern gas cleaning methods (in turn, the efficiency of gas cleaning is largely determined by the implementation of the so-called primary measures in the thermal process). In this case, according to German practice, industrial emissions are well below the limits regulated by strict environmental legislation [5].

As can be seen from the above, it can be concluded that the use of a binder based on bottoms will significantly reduce the cost of briquettes while maintenance their operational qualities. Due to the co-disposal of solid waste and coke plants waste, they become the category of commercial products.

References 1. Recycling Remains a Rarity in Eastern Europe [Электронный ресурс] – Режим доступа: https://www.statista.com/chart/1312/recycling-remains-a-rarity-ineastern-europe/ –17.01.2019 – Загл. с экрана.

2. Зарубежная практика использования альтернативного топлива из отходов для цементной промышленности/ И.В. Ламзина, В.Ф. Желтобрюхов, И.Г. Шайхиев // Вестн. технолог.ун-та. Т.18, №17. – 2015. – С. 85-88.

3. Лазорин С.Н. Обезвреживание отходов коксохимических заводов/ С.Н. Лазорин, Т.И. Паннов, В.И.Литвиненко. –М.: Металлургия, 1977. – 239 с.

4. Сметанин В. И. Защита окружающей среды от отходов производства и потребления: Учебники и учеб. пособия для студентов высш. учеб. заведений. – М.: Колос, 2000. –232 с.

5. Термическая переработка ТБО [Электронный ресурс] – Режим доступа: https://ztbo.ru/o-tbo/lit/texnologii-otxodov/termicheskaya-pererabotka-tbo – 17.02.2019 – Загл. с экрана.

80

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Аннотация. В статье проанализирована ситуация сложившаяся в сфере утилизации твёрдых бытовых и промышленных отходов за последние 3 года. Описаны особенности реализации производства RDF-топлива на основе компонентов твердых бытовых отходов и кубовых остатков. Ключевые слова: твёрдые бытовые отходы, RDF-топливо, кубовые остатки, технологии, сжигание. Сведения об авторах: Курченко Елена Николаевна – студент группы ИЗОСм-19, ДонНТУ Калинихин Олег Николаевич – к.т.н., доцент, кафедра «Прикладная экология и охрана окружающей среды», ДонНТУ Бойко Виктория Николаевна – ст. преподаватель кафедры английского языка, ДонНТУ

UDC 004.912 OVERVIEW OF SPAM FILTERING ALGORITHMS Lyutova E.I., Kolomoytseva I.A., Gilmanova R.R. ktrnlutova@gmail.com

Abstract. This article presents an overview and analysis of spam detection algorithms. The mathematical model of the algorithms and their classification are presented. Promising areas of research on the exact identification of spam are highlighted.

Keywords: spam, classification, machine learning, spam identification algorithms, probabilistic classifiers, linear classifiers Spam is information that is widely distributed to people who do not want to receive it. In order to combat spam some measures are being taken to eliminate it, but the schedule given in Figure 1 shows their inefficiency.

Figure 1 – Schedule of spam in 2018 year 81

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Formal problem statement of information classification Let D = {𝑑 (1) ,….,𝑑 (𝑁) } - be a set of documents, C = {𝑐1 ,….,𝑐𝑀 } be a set of categories, Φ be an objective function that, by a pair ( 𝑑 (𝑖) , 𝑐𝑗 ), determines whether a document 𝑑 (𝑖) belongs to a category 𝑐𝑗 (1 or True) or not (0 or False). The classification ̃ , as close as possible to the Ф [2]. In this problem is to construct a function Ф formulation of the problem, it should be noted that there is no additional information about the categories and documents except ones that can be extracted from the document itself. If the classifier gives an exact answer: : D  C  {0, 1}, (1) then the classification is called exact. If the classifier determines the degree of similarity ( Categorization Status Value ) of the document: CSV: D  [0, 1], (2 ) then the classification is called threshold [5].

With these concepts in mind, let us consider spam filtering algorithms that are based on machine learning.

Naive Bayesian model The naive Bayes model is based on the Bayes theorem: 𝑃(𝑑 |𝑐 )𝑃(𝑐) P(c|d)= (3) 𝑃(𝑑)

where,  P (c | d) is the probability that document d belongs to class c, that is we need to calculate it;  P (d | c) is probability to meet document d among all documents of class c;  P (c) is unconditional probability of meeting a document of class c in the document corpus;  P (d) is unconditional probability of document d in the body of documents.

Bayes' theorem allows us to rearrange the cause and effect. Knowing the probability with which the cause leads to a certain event, the theorem allows us to calculate the probability that it is this reason that led to the event observed.

The purpose of text classification according to Bayes' theorem is to understand what class a document belongs to, so you do not need the probability of the document but the most probable class. Bayesian classifier uses maximum a posteriori estimation to determine the most probable class. Roughly speaking, this is a class with maximum probability [7]. 𝑃(𝑑 |𝑐 )𝑃(𝑐) С𝑚𝑎𝑝 = 𝑎𝑟𝑔 max (4) 𝑐∈𝐶

𝑃(𝑑)

It is necessary to calculate the probability for all category classes and choose the class with maximum probability. It is worth paying attention to the fact, that the 82

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

denominator (the document likelihood) in the formula (4) is the constant and can not effect the class ranking, so it can be omitted. = 𝑎𝑟𝑔 max[𝑃(𝑑|𝑐)𝑃(𝑐)] (5) 𝑐∈𝐶 С 𝑚𝑎𝑝 The «naivety» of the classifier model is that it uses the naive assumption that the words in the text of the document are independent of each other. The «naivety» of the classifier model is that it uses the naive assumption that the words in the text of the document are independent of each other. Based on this, the conditional probability of a document is approximated by the multiplication of the conditional probabilities of all the words included in the document. 𝑃(𝑑|𝑐) ≈ 𝑃(𝑤1 |𝑐)𝑃(𝑤2 |𝑐) … 𝑃(𝑤𝑛 |𝑐) = ∏𝑛𝑖=1 𝑃(𝑤𝑖 |𝑐) (6) Where 𝑤𝑖 is the number of times word i - th occurs in the documents of class c.

Substituting the resulting expression into formula (5) we obtain: С 𝑚𝑎𝑝

= 𝑎𝑟𝑔 max[𝑃(𝑐) ∏𝑛𝑖=1 𝑃(𝑤𝑖 |𝑐)] 𝑐∈𝐶

(7)

In the formula (7), the logarithm of these probabilities is most often used instead of the probabilities themselves. Since the logarithm is a monotonically increasing function, class c with the highest probability logarithm will remain the most probable. Then: С𝑚𝑎𝑝 = 𝑎𝑟𝑔 max[𝑙𝑜𝑔𝑃(𝑐) + ∑𝑛𝑖=1 𝑙𝑜𝑔𝑃(𝑤𝑖 |𝑐)] (8) 𝑐∈𝐶

Then, you need to assess the probability of the category class and the words in the class. They are calculated by the formulas (9) and (10): 𝐷 𝑃(𝑐) = 𝑐 (9) 𝐷

where 𝐷𝑐 is the number of the documents belonging to class c.

D is the total number of documents in the training set. 𝑊 𝑃(𝑤𝑖 |𝑐) = ∑ 𝑖𝑐 𝑖′∈𝑉 𝑊𝑖′𝑐

(10)

where V is the dictionary of the documents corpus (a list of all unique words).

The next step is to apply additive smoothing (Laplace smoothing). It is necessary for words that did not occur in the training set of the classifier to have no zero probability. Therefore, it is necessary to add 1 to the frequency of each word. 𝑊 +1 𝑊 +1 𝑃(𝑤𝑖 |𝑐) = ∑ 𝑖𝑐 = |𝑉|+∑ 𝑖𝑐 (11) 𝑖′∈𝑉(𝑊𝑖′𝑐 +1)

𝑖′∈𝑉 𝑊𝑖′𝑐

Substituting all the obtained probability estimates (i.e. formulas 9, 11) we obtain the final formula for the model of naive Bayes classifier. 𝐷 𝑊 +1 С𝑚𝑎𝑝 = 𝑎𝑟𝑔 max[log 𝑐 + ∑𝑛𝑖=1 log |𝑉|+∑ 𝑖𝑐 ] (12) 𝑐∈𝐶

𝐷

83

𝑖′∈𝑉 𝑊𝑖′𝑐

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

The Bayesian method is used to detect unauthorized advertising mailings by email (spam). At the same time, the training base is considered, i.e. two arrays of emails, one of which is composed of spam, and the other of ordinary emails. For each of the cases, the frequency of use of each word is calculated, after which a weight estimate (from 0 to 1) is calculated, which characterizes the conditional probability that a message with this word is spam. Weight values close to 0.5 are not taken into account under integrated counting, therefore words with such weights are ignored and removed from dictionaries [3].

The advantages of this approach to spam filtering are: - the method of naive Bayes classifier is easy to use; - It is highly effective in filtering spam; - high speed of calculation; This method has a large number of modern filters spam (Mozilla Thunderbird 0.8, BayesIt ! 0.6.0 , SpamAssassin 3.0.0 rc3 ) [8].

However, the method has a fundamental flaw. Is is based on the assumption that some words are more likely to be found in spam, and others in ordinary letters. It is ineffective if this assumption is incorrect. However, as practice shows, even a person is not able to determine such spam «by eye», i.e. only by reading the letter and understanding its meaning. There is a Bayesian poisoning method that allows you to add a lot of extra text, sometimes carefully chosen to «trick» the filter.

Another, non-fundamental drawback associated with the implementation is that the method works only with texts. Knowing this limitation, you can put advertising information into pictures, the text in the letter is either missing or does not make any sense. To combat this Bayesian poisoning method we have to use means to detect pictures in texts, or the old filtration techniques, i.e. a «black list» and regular expression (as such letters are often stereotyped in forms).

Method of support vectors The method of Support Vector Machine (SVM) belongs to the group of boundary classification methods. It determines the belonging of objects to category classes using the boundaries of regions. We will consider only the binary classification of the method. The classification is based on only two categories c and c, taking into account the fact that this approach can be extended to any finite number of categories. Suppose that each classification object is a vector in N-dimensional space. Each coordinate of the vector is a certain feature. The larger the feature is quantitatively, the more this feature is expressed in the object. The advantages of SVM method are:  it is one of the best quality methods;  it enables to work with a small set of data for training;  it gives reducibility to the convex optimization problem having a unique solution.

The disadvantages of the method is its difficult interpretability of the algorithm parameters and the instability with respect to outliers in the source data [1].

84

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Algorithm of k - nearest neighbors The method of nearest neighbors is a simple metric classifier based on assessing the similarity of objects. The object classified belongs to the class that includes the objects of the training set that are close to it. One of the simple systematization algorithms in real problems often turns out to be ineffective for this reason. In addition to the accuracy of systematization, the question of the classifier is considered to be the rate of systematization: if there are N objects in the training set, M objects in the test set, but the dimension of space is K, then the number of actions to systematize the test set can be estimated as O (K * M * N). When identifying spam, this method does not achieve high accuracy, which makes it quite inefficient when filtering unnecessary information [4].

Advantages of the method are: ability to update the training sample without retraining the classifier; algorithm stability to abnormal outliers in the source data;  relatively simple software implementation of the algorithm; easy interpretability of the results of the algorithm; good training in the case of linearly inseparable samples.

The disadvantages of the method are: representativeness of the data set used for the algorithm; high dependence of classification results on the chosen metric; long duration of work due to the need for a complete enumeration of the training sample; impossibility of solving problems of large dimension by the number of classes and documents [6].

Conclusions In this article, classification algorithms for identifying spam have been analyzed. The shortcomings and advantages of the approaches considered have been identified and studied in detail.

The results of the review can be presented in the following way of research: researching poorly studied algorithms; developing combined methods.

Bayesian classifier method is of particular interest, since it has a higher reliability and is quite simple to implement. The research prospects are the development of our own algorithm of spam classification based on the naive Bayesian classifier.

References 1. Алгоритм Роккио [электронный ресурс], – Режим доступа: https://ozlib.com/867902/informatika/klassifikatsiya_obucheniem_drugie_algoritmy 2. Ландэ Д.В., Снарский А.А., Безсуднов И.В. Интернетика:Навигация в сложных сетях: модели и алгоритмы. Книжный дом «Либроком», 2009. – 77 с.

85

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

3. Ландэ Д.В., Снарский А.А., Безсуднов И.В. Интернетика:Навигация в сложных сетях: модели и алгоритмы. Книжный дом «Либроком», 2009. – 88-89с..

4. Метод ближайших соседей [электронный ресурс], – Режим доступа: http://www.machinelearning.ru/wiki/index.php?title=Метод_ближайшего_соседа 5. Методы автоматической классификации текстов [электронный ресурс], – Режим доступа: http://swsys.ru/files/2017-1/_85-99.pdf – 86 с.

2. Методы автоматической классификации текстов [электронный ресурс], – Режим доступа: http://swsys.ru/files/2017-1/_85-99.pdf – 88 с.

3. Наивный байесовский классификатор [электронный ресурс], – Режим доступа: http://bazhenov.me/blog/2012/06/11/naive-bayes.html.

4. Применимость байесовского классификатора для задачи определения спама [электронный ресурс], – Режим доступа: https://securelist.ru/primenimostbajesovskogo-klassifik/475/. Аннотация. В данной статье приведен обзор и анализ алгоритмов обнаружения спама. Представлено описание математической модели алгоритмов, а также дана их классификация. Выделены перспективные направления исследования точной идентификации спама. Ключевые слова: спам, классификация, машинное обучение, алгоритмы идентификации спама, вероятностные классификаторы, линейные классификаторы. Сведения об авторах: Лютова Екатерина Игоревна – студент группы ПИм-19, ДонНТУ Коломойцева Ирина Александровна – ст. преподаватель, кафедра «Программная инженерия», ДонНТУ. Гильманова Роза Разимовна – ст. преподаватель кафедры английского языка, ДонНТУ.

UDC 621.2.082.18 RESULTS OF EXPERIMENTAL STUDIES OF BRAKING QUALITIES OF CARS ON THE LINE OF INSTRUMENTAL CONTROL Mokrushin D.A, Bykov V.V. dmitrijmokrushin94@gmail.com

Abstract. This article presents the results of experimental studies of M1 category cars on the instrumental control line of the Diagnostics laboratory of the Donetsk National Technical University Automobile and Road Institute GOUVPO, confirming 86

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

the need to improve road safety due to the diagnostic method of the M1 category car brake system on a roller stand.

Keywords: braking distance, brake disc, clutch coefficient, runout, braking forces.

In recent years, there has been a constant increase in the road speed limit, and traffic intensity. However, some negative trends in motorization are also evident – this is a significant number of road traffic accidents (RTAs). A lot of factors affect the accidents: the condition of the roads, traffic intensity, lighting, level of professional training and many others. One of the most important places among them is the technical condition of cars. Тest result of 900 cars

– Serviceable,

– Other faults,

– Runout,

– Different thickness

Figure 1 – Technical Review Results In this regard, it is necessary to increase the level of active and passive safety systems of wheeled vehicles (WV), to improve the methods and means of determining their technical condition based on high-quality and informative diagnostics.

In this paper, it is made an attempt to improve road safety due to the existing method for diagnosing the brake system of cars of category M1 on a power roller stand.

Experimental studies using the bench test method for assessing the braking properties of a car were carried out on a BSA 250 power roller bench of the German company BOSCH, which is part of the SDL 260 diagnostic line in the Diagnostics Laboratory of the Automobile Transport Department of Automobile and Road Institute. A Vauxhall Vectra C car, equipped with an anti-lock system, was used in the 87

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

experiments. The measurements were carried out in a state of full and curb mass. The force on the brake pedal was measured using a pedometer with a remote control unit for the diagnostic line.

Figure 2 – Brake pedal force sensor with remote control (SDL260).

A methodology for testing cars of category M1 was developed to qualitatively determine the technical condition of a car brake system with ABS:  install a car on the SDL 260;

88

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 3 – Installing a Vauxhall Vectra C in a tool control line SDL 260.  prepare the car VAUXHALL VECTRA C for diagnosis at the stand;  connect the diagnostic cable UBOX 2 of the system tester KTS 520 to the diagnostic OBD connector of the vehicle and perform identification, run the program for diagnosing the ABS TRW 430 control unit;

Figure 4 – Diagnostics of the control unit ABS TRW 430 a car VAUXHALL VECTRA C.    algorithm; 

check the fault memory of the ABS control unit; if there are malfunctions, eliminate them using SIS CAS; carry out functional tests on the brake stand according to the ESI [tronic] check the actual parameters on the BSA 250 (with the engine running);

89

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 5 – Developed braking force of the front, rear axle and parking brake.

Figure 6 – Разница тормозных усилий и зависимость тормозного усилия от степени нажатия на педаль.  

set the test diagnostic mode for CBS with ABS; check the criteria for the effectiveness of the brake system on the stand;

90

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 7 – The coefficient of uneven braking forces of the front axle and rear axle.

Figure 8 – The specific braking force and the coefficient of unevenness of the braking forces are maximum. 

enter the verification data into the PC memory and print;

91

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 9 – Test report.

The proposed technique allows us to evaluate the braking performance of vehicles equipped with anti-lock braking systems.

Summing up, we can draw the following conclusions: 1. State Standard Specifications R 33997-2016 prescribes the determination of only the total specific braking force and does not take into account the presence of regulating devices (brake force regulators and anti-lock braking systems) in the brake system drive, as well as the load state of vehicles. It is not consistent with the certification requirements of UNECE Regulation No. 13 regarding the distribution of specific braking forces of the front and rear axles for all vehicle loading conditions.

2. 88% of tested vehicles do not meet the requirements of State Standard Specifications R 33997-2016.

References 1. Бухарин Н.А. Тормозные системы автомобилей / Н.А. Бухарин. – М.– Л.: Машгиз, 1950. – 291 с. 8. Бухарин Н.А. Испытание автомобиля с использованием электрических методов измерений / Н.А. Бухарин, В. К. Голяк. – М.: Машгиз, 1962. – 228 с.

2. ГОСТ 33997-2016 «Автотранспортные средства. Требования безопасности к техническому состоянию и методы проверки» – [Дата введения 2018–02–01] 92

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

3. Говорущенко Н.Я. Диагностика технического состояния автомобилей / Н.Я. Говорущенко. – М: Транспорт. 1970. – 254 с.

4. Мокрушин, Д.А. Оценка тормозных свойств автомобилей категории М1 по результатам инструментального контроля / Д.А. Мокрушин, В.В. Быков, В.Э Волошин. – Горловка : ОБРАЗОВАНИЕ, НАУКА И МОЛОДЕЖЬ – 2017 Керчь, 26 октября 2017 г., Керчь, 26 октября 2017 г.

5. Мокрушин, Д.А. Оценка тормозного пути автомобиля категории М1 с АБС при дорожных испытаниях / Д.А. Мокрушин, В.В. Быков. – Горловка : Научно-технические аспекты развития автотранспортного комплекса Горловка, Горловка, 24 мая 2018 г.. Аннотация. В данной статье приведены результаты экспериментальных исследований автомобилей категории М1 на линии инструментального контроля лаборатории «Диагностики» Автомобильно-дорожного института ГОУВПО «Донецкий национальный технический университет», подтверждающие необходимость повысить безопасность дорожного движения благодаря методике диагностики тормозной системы автомобилей категории M1 на роликовом стенде. Ключевые слова: Тормозной путь, тормозной диск, коэффициент сцепления, биение, тормозные силы. Сведения об авторах: Мокрушин Дмитрий Александрович – аспирант, ДонНТУ. Быков Валерий Васильевич – к.т.н., доцент кафедры «Автомобильный транспорт», декан дорожно-транспортного факультета, АДИ ГОУВПО «ДонНТУ»

93

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

UDC 622.455.22-52 AUTOMATED CONTROL SYSTEM OF AIR DISTRIBUTION IN THE MINE VENTILATION SYSTEM Nemov G.Y., Neyezhmakov S.V., Paniotova L.N. nemovyura@gmail.com

Abstract. In the process of mining operations and lengthening the excavation passages the aerodynamic drag is continuously changing. In this connection the deviation of the actual airflow into each passage of the ventilation network takes place from a predetermined value in an upward or downward direction. To ensure fresh air jet of a required amount in the passages of a network the automated control system of air distribution in the mine ventilation network has been developed. The ventilation system model has been worked out to investigate the technological process of air distribution in the mine ventilation network and to train students to work with a PLC PLC150 -220 and a pressure transducer PD150-DIV200P. Keywords: ventilation network, automated control system, air distribution, air barriers, Programmable Logic Controller.

In mining operations and the excavation passages, the aerodynamic drag is continuously changing [2]. In this case, the deviation of the actual airflow into each passage of the ventilation network takes place from a predetermined value in an upward or downward direction. As a result of this process there may not be enough air in one of the passages of network, whereas in another passage there will be an excess of air, which is forbidden by safety rules [3].

To solve the task of ensuring fresh air jet in passages the automatic control system of air distribution in the mine ventilation network has been developed to allow airflow regulation in three passages.

Figure 1 shows a block diagram of the automated air distribution control system in the ventilation shaft.

There are airflow meter SDSV 01.01.01-M, a programmable controller TX9042 and air barriers in the mine workings. The signal from the SDSV 01.01.01 M enters the TX9042. The TX9042 is powered by the PM-6321/13.5-A.01. The information from the TX9042 is transmitted through RS-485 Modbus protocol to a programmable logic controller PLC150-220, which is located on a remote control of the mine manager. The airflow meter SDSV 01.01.01 M and PD150-DI10,0K pressure converter are disposed in the suction channel to measure airflow of an outgoing jet and the vacuum value connected to PLC150-220.

Airflow rate and vacuum adjustment in the mine workings are carried out by affecting PLC150-220 and TX9042 on the main ventilation fan guide device and ventilation barriers.

94

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 1 – Block diagram of the automated air distribution control system in the ventilation shaft 95

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 2 shows a flowchart of the airflow adjustment.

Figure 2 – Flowchart of the airflow adjustment 96

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

During operation air distribution automated control system in the mine ventilation network determines the state of the mine atmosphere in each passage of the ventilation network, pointing out the passage with the highest air shortage. After that, the position of the regulatory mechanism in this passage is determined. If the damper is closed or in an intermediate position, the control signal is sent to open it.

When the barrier is fully open, the amount of air entering the passage with a shortage is regulated by overlapping adjacent passages, in case of safe regulation of excess air, the damper being in the open or intermediate position.

If there is no regulation reserve in the adjacent passages of the network and there is insufficient air in the regulated passage the overall consumption is increasing by opening the main ventilation fan guide device.

If there no insufficient air passage and another passage having excess air, the overall air consumption is reduced by closing the main ventilation fan guide device. It will result in air redistribution between passages and detection of the network passage with air shortage [1].

To reduce the number of commutations of actuators, a dead zone (2-3%) is introduced together with time delay to redistribute air in the network passages after determining the actuator control signal (barrier or guide device).

To confirm the system efficiency the model of the mine ventilation network was elaborated Figure 3 is the technological scheme of the system.

Figure 3 – Flow diagram of the maket of the mine ventilation network The fan 1 makes a directed air flow taken from the atmosphere through the inlet 14. The flow is distributed into three parallel passages. Speed sensors 2,5,8,11 measure flow rate at the outlet of the ventilation network and in parallel passages respectively.

Dumpers 3,6,9,12 are designed to change the ventilation network resistance and at the same time they are the sources of disturbing influence. The programmable logic controller PLK150-220 sets the required network resistance by affecting the controlled barriers 4,7,10,13. The degree of closing dumpers is determined by the resistive 97

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

position sensors 15-22. Air pressure is measured by pressure converters PD150DIV200P installed at the entrance to the ventilation network 23, at the intersection of parallel passages 24, 25 and at the exit from the network 26.

Thus, the proposed system of automated control of air distribution in the mine ventilation network solves the problem of providing fresh air stream of three passages of the network in the required amount, and the model of the ventilation system allows you to explore the technological process of air distribution in the mine ventilation network as well as to train students to work with a PLC PLC150 -220 and a pressure transducer PD150-DIV200P. References 1. Абрамов, Ф.А. Автоматизация проветривания шахт / Ф.А. Абрамов, В.А. Бойко – Киев: Наук. думка, 1967 2. Аэрология горных предприятий / К.3. Ушаков [и др.]: Учебник для вузов.

– 3-е изд., перераб. и доп. – М.: Недра, 1987. – 421 с.

3. Правила безопасности в угольных шахтах [Электронный ресурс] : утв. приказом Гос. Комитета горного и тех. надзора ДНР и Мин-вом угля и энергетики ДНР 18.04.2016 г. № 36/208 : ввод в действие 17.05.2016. – Донецк, 2016. – Режим доступа: https://doc.minsvyazdnr.ru/docs/2476. – Загл. с экрана Аннотация. В процессе проведения горных работ и удлинения горных выработок аэродинамическое сопротивление непрерывно изменяется. В связи с этим происходит отклонение фактического потока воздуха в каждой ветви вентиляционной сети от заданного значения в сторону увеличения или снижения. Для обеспечения ветвей сети струёй свежего воздуха необходимого объёма, разработана автоматизированная система управления распределением воздуха в шахтной вентиляционной сети. Модель вентиляционной системы разработана для исследования технологического процесса распределения воздуха в шахтной вентиляционной сети и обучения студентов работе с ПЛК ПЛК150 -220 и датчиком давления ПД150-ДИВ200P. Ключевые слова: система автоматизированного управления, программируемый логический контроллер, вентиляционная сеть, воздухораспределение, воздушные заслоны. Сведения об авторах: Немов Георгий Юрьевич – студент группы АУПм-19, ДонНТУ Неежмаков Сергей Владимирович – к.т.н., доцент кафедры «Горная электротехника и автоматика им. Р.М.Лейбова», ДонНТУ Паниотова Любовь Николаевна – ст.преподаватель кафедры английского языка, ДонНТУ

98

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

UDC 001.6 BEST DISCOVERIES OF YOUNG SCIENTISTS IN RUSSIA Pilipenko A. S., Girovskaya I. V. ArtemPilipenko5@gmail.com

Abstract. Every day, many interesting things are invented in the world that will make life easier for people. There are many young scientists in Russia; some of them are laureates of the Presidential Prize in Science and Innovation for Young Scientists for 2018. Let's consider the best inventions: «The solution for everything in the world», «Remote detection of explosives», «Regeneration of nerve cells», «Biological protection of plants».

Keywords: inventions, Russian scientists, young scientists, award, laureates, 2018 Opening number 1. The solution for everything Mathematician Ivan Oseledets from the Skolkovo Institute of Science and Technology, created to create revolutionary computing technologies and solve various problems in physics, chemistry, biology and data analysis based on the so-called tensor expansions. He developed an algorithm to solve problems of a huge class.

«The work of Ivan Oseledets is included in classical standard textbooks around the world, in American textbooks he became a classification of this very important and new direction», presidential aide Andrei Fursenko said at a press conference.

In this article, I described my work as follows: «I am engaged in computational mathematics and developed algorithms. There are general mathematical principles. In my case, tensor sophisticated methods. In this article, I lead a scientific group, and we mainly apply these methods for data analysis, for tasks, for artificial intelligence and for deep neural dialing, where there are a huge number of parameters that should be able to be compactly represented».

«Imagine that you have some amount that you must predict. » For example, a person’s life span, explains Ivan Oseledets. – There are indirect signs by which you can determine what will happen: human income, environmental situation. If there are only two characters, then a hundred options will be obtained, if 3 – then a thousand. I have developed new tools for working with huge data arrays».

These methods are applicable in many areas. Only an approximate value of the parameters is given. Developed by scientists approaches to increase the speed of work on the Internet.

«In the future, we see two huge directions of development – the application of our aspect for the analysis of various data, for example, images, videos, texts, sound information». The second direction is the modeling of complex optional dynamic systems, for example, for the tasks of the oil industry, the creation of composite materials and so on, » said Oseledets [1].

99

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 2 – Ivan Oseledets Opening number 2. Remote Explosive Detection EvgenyGorlov and Victor Zharkov from V.E. Zuevsky Institute of Atmospheric Optics SB RAS, Presidential Prize Scientists have created ultra-sensitive laser systems that can detect not only explosives, but also vapors or traces of micron-sized explosives up to 50 m away.The sensitivity of the devices is comparable to that of a dog’s nose.

«The closest foreign analogues in sensitivity are three orders of magnitude lower. That is, a thousand times», – said winner EvgenyGorlov.

«All of the well-known methods that are used today are contact: you need to take a sample or a dog to get closer. «The main advantage of our device is its high sensitivity», – said EvgenyGorlov. «Our device can be used covertly, from afar».

A situation that has no analogues in the world is a new step in the fight against terrorism. Scientists hope that in the near future the safety of people at the airport, train station and transport will be ensured. The device works even in conditions of heavy passenger traffic. Siberian inventors are sure that the result will be intermediate, and in the future they will be able to increase the sensitivity of the installation by at least 10 times [2], [3].

100

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 2 – Yevgeny Gorlov and Viktor Zharkov

Figure 3 – Remote Explosive Detection

101

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Opening number 3. Nerve cell regeneration Scientist from Vladivostok VyacheslavDyachuk, Senior Researcher A.V. The Russian Academic University received a prize for the discovery of new processes of development of the nervous system of invertebrates and vertebrates.

Scientists have proven the ability of auxiliary cells of the nervous system (glial cells) to transform into various types of cells of vertebrate embryos. He discovered chameleon cells. «This is the medicine of the future. This type of cells, which we found, can greatly help in the regeneration of tissues», VyacheslavDyachuk explains his discovery.

The discovery may be useful in medicine, it allows you to regulate the processes of development of the nervous system and their treatment: both in the case of «failures» and in older people suffering from Parkinson's disease, Alzheimer's disease and other neurodegenerative diseases [4] [5].

Figure 3 – VyacheslavDyachuk Opening number 4. Biological plant protection It should be noted Ekaterina Grisanova (32 years old, studied at the Novosibirsk State Agrarian University). It has opened up new possibilities for protecting agricultural and forestry from the bacteria Bacillus thuringiensis. Catherine made the discovery necessary for many, experimenting with the larvae of various insects and butterflies. This can help insect pests create fertilizer resistance.

«There are chemical insecticides based on chemistry. And there is biological plant protection. These drugs are environmentally friendly, they do not get into food», explains Ekaterina Grizanova.

102

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

However, these drugs are gradually becoming ineffective. «Our colleagues have studied mutations well. We found that these people have a sustainable nature and they have their own effectiveness in using biological products», Catherine explained.

Grisanova helped in creating an unusual new type of fertilizer that will be safe for others, including nature and humans. Today, experts from the NSAU have already developed a comprehensive essential tool for combating harmful plants based on this invention, which has no world analogues. The insecticide contains bacteria that cause insect diseases and secondary plant metabolites that weaken the pest (its protective systems).

«The complex was created from entomopathogenic bacteria that cause diseases in insects and secondary metabolites of plants – low molecular weight substances (they do not study in growth, reproduction and development) ». The discovery has no direct analogues. According to Ekaterina Grisanova, this insecticide will work much more efficiently – plant metabolites weaken the insect protection system» [6].

Figure 4 – Ekaterina Grizanova References 1. Екатерина Гризанова - самый молодой лауреат президентской премии: scientificrussia.ru. – 2015 [Электронный ресурс]. Дата обновления: 07.02.2019. – URL: https://scientificrussia.ru/articles/ekaterina-grizanova-samyj-molodoj-laureatprezidentskoj-premii (дата обращения: 12.04.2020).

2. И все-таки они восстанавливаются: scientificrussia.ru. – 2015 103

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

[Электронный ресурс]. Дата обновления: 22.09.2019. – URL: https://scientificrussia.ru/articles/i-vse-taki-oni-vosstanavlivayutsya(дата обращения: 12.04.2020).

3. Оселедец Иван Валерьевич: Википедия. – 2001 [Электронныйресурс]. Дата обновления: 25.03.2020. – URL: https://ru.wikipedia.org/wiki/%D0%9E%D1%81%D0%B5%D0%BB%D0%B5%D0 %B4%D0%B5%D1%86,_%D0%98%D0%B2%D0%B0%D0%BD_%D0%92%D0% B0%D0%BB%D0%B5%D1%80%D1%8C%D0%B5%D0%B2%D0%B8%D1%87 (дата обращения: 11.04.2020).

4. Сибирские лидары чувствуют взрывчатку на расстоянии 50 метров: pikabu.ru. – 2008 [Электронный ресурс]. Дата обновления: 08.02.2019. – URL: https://pikabu.ru/story/sibirskie_lidaryi_chuvstvuyut_vzryivchatku_na_rasstoyanii_5 0_metrov_6519009(дата обращения: 11.04.2020).

5. Томские ученые удостоены награды президента за метод дистанционного обнаружения взрывчатки: vesti.ru. – 2005 [Электронный ресурс]. Дата обновления: 07.02.2019.

– URL: https://www.vesti.ru/doc.html?id=3113744(дата обращения: 11.04.2020).

6. Ученые открыли механизм, который поможет излечить болезнь Альцгеймера: ria.ru. –v 2004 [Электронный ресурс]. Дата обновления: 07.02.2019.

– URL: https://ria.ru/20190207/1550508138.html(дата обращения: 12.04.2020). Аннотация. Каждый день в мире изобретается много интересных вещей, что облегчат жизнь людям. В России много молодых ученых, некоторые из них являются лауреатами Президентской премии в области науки и инноваций для молодых ученых за 2018 год. Рассмотрим лучшие изобретения: «Решение для всего в мире», «Дистанционное обнаружение взрывчатых веществ», «Регенерация нервных клеток», «Биологическая защита растений». Ключевые слова: изобретения, российские ученые, молодые ученые, лауреаты премии, 2018 год. Сведения об авторах: Пилипенко Артём Сергеевич – студент группы ПИм-19, факультет компьютерных наук и технологий по специальности «Программная инженерия», ДонНТУ. Гировская Ирина Валерьевна – старший преподаватель кафедры английского языка, ДонНТУ.

104

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

UDC 004.93’12 STUDENT ATTENDANCE CONTROL SYSTEM BASED ON NEURAL NETWORK RECOGNITION Rudak L.V., Fedyaev O.I., Girovskaya I.V. leo2598@yandex.ua

Abstract. Student’s attendance control system based on neural network face recognition is described in the article. The students’ attendance automation problem was analyzed, the main solving methods were considered and the most appropriate was chosen in this work.

Keywords: face recognition, neural network, computer vision, fitting.

For university teachers, the need to independently monitor student attendance has become an integral part of conducting classes. However, this process takes too much time for various reasons.

Today, existing technologies make it possible to make student attendance control automated. There are several ways to solve this problem.

Known implementations Flexible graph comparison method (Elastic graph matching).The essence of the method is reduced to elastic comparison of graphs describing facial images. Faces are represented as graphs with weighted vertices and edges. At the recognition stage, one of the graphs – the reference graph – remains unchanged, while the other is deformed in order to best fit the first. In such recognition systems, graphs can be either a rectangular lattice or a structure formed by characteristic (anthropometric) points of the face [1][2].

Figure 1 – Graph based on anthropometric points of the face

105

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

At the vertices of the graph, the values of the signs are calculated, most often they use the complex values of the Gabor filters or their ordered sets – Gabor wavelets (Gabor builds), which are calculated locally in some local region of the graph vertex by convolving the pixel brightness values with Gabor filters.

Method cons: high computational complexity of the recognition procedure. Low manufacturability when remembering new standards. The linear dependence of the operating time on the size of the database of persons.

Neural networks. Neural networks are trained on a set of training examples. The essence of training is to adjust the weights of interneuron connections in the process of solving the optimization problem by the gradient descent method. In the process of learning the NN, the key features are automatically extracted, their importance is determined, and the relationships between them are built. It is assumed that a trained NN will be able to apply the experience gained in the learning process to unknown images due to generalizing abilities.

The best results in the field of face recognition (according to the analysis of publications) were shown by the Convolutional Neural Network or the convolutional neural network (hereinafter – CNN), which is a logical development of the ideas of such NN architectures as the cognitron and neocognitron. The success is due to the possibility of taking into account the two-dimensional image topology, in contrast to the multilayer perceptron.

Distinctive features of the CNN are local receptor fields (provide local twodimensional connectivity of neurons), total weights (provide detection of some features anywhere in the image) and hierarchical organization with spatial sampling (spatial subsampling). Thanks to these innovations, the CNN provides partial resistance to changes in scale, displacement, rotation, change of angle and other distortions.

Figure 2 – Architecture of a convolutional neural network

106

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Method cons: adding a new reference person to the database requires a complete retraining of the network on the entire available set (a rather lengthy procedure, depending on the sample size from 1 hour to several days). Mathematical problems associated with learning: getting into a local optimum, choosing the optimal optimization step, retraining, etc. It is difficult to formalize the stage of choosing a network architecture (the number of neurons, layers, the nature of the connections).

Summarizing all of the above, we can conclude that the National Assembly is a «black box» with difficultly interpreted work results.

Hidden Markov models (HMM).One of the statistical methods of face recognition are hidden Markov models (HMM) with discrete time [3]. HMM use the statistical properties of signals and take directly into account their spatial characteristics. The model elements are: a lot of hidden states, a lot of observable states, a matrix of transition probabilities, the initial probability of states. Each has its own Markov model. When recognizing an object, the Markov models generated for a given base of objects are checked and the maximum of the observed probabilities is found that the sequence of observations for this object is generated by the corresponding model.

To date, it has not been possible to find an example of the commercial use of HMM for face recognition.

Method cons: it is necessary to select model parameters for each database, the HMM does not have a discriminating ability, that is, the training algorithm only maximizes the response of each image to its model, but does not minimize the response to other models.

Principal component analysis (PCA).One of the most famous and welldeveloped is the principal component analysis (PCA) method, based on the KarunenLoev transform.

Initially, the method of principal components began to be used in statistics to reduce the space of signs without significant loss of information. In the face recognition problem, it is mainly used to represent a face image with a small dimension vector (main components), which is then compared with the reference vectors stored in the database.

The main goal of the principal component method is to significantly reduce the dimension of the feature space in such a way that it describes the «typical» images belonging to many faces as best as possible. Using this method, it is possible to identify various variability in the training sample of facial images and describe this variability in the basis of several orthogonal vectors, which are called eigenface.

Active Appearance Models (AAM). Active Appearance Models (AAM) are statistical models of images that can be adapted to a real image through various deformations. This type of model in a two-dimensional version was proposed by Tim Kuts and Chris Taylor in 1998. Initially, active appearance models were used to evaluate the parameters of facial images.

The active appearance model contains two types of parameters: parameters associated with the shape (shape parameters), and parameters associated with the 107

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

statistical model of image pixels or texture (appearance parameters). Before use, the model must be trained on a set of pre-marked images. Image marking is done manually.

Each label has its own number and defines a characteristic point that the model will have to find during adaptation to a new image.

Figure 3 – An example of marking a face image of 68 points forming the AAM shape Selected Implementation The most optimal option is to use the combined method, which includes computer vision, a deep convolutional neural network and a classifier.

108

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

The first step involves finding a face in the image. The implementation technology for solving this problem is OpenCV [4] [5]. However, face recognition consists of many interrelated subtasks:  find all faces in the picture;  learn to recognize every face, even if it is strangely turned, or if the lighting is poor - because it's still the same person;  identify unique facial features that distinguish one person from others, for example, eye size, face shape and so on;  compare the revealed unique features of this person with all the people that the system already knows in order to understand who is shown in the photo.

To find faces in the image, you need to make the image black and white, then look at each pixel in the image, one after the other [6].

The purpose of the algorithm is to find out how dark the current pixel is compared to neighboring ones. Then the direction vector is specified in which the image becomes darker: After completing this procedure for each individual pixel in the image, replace each pixel with a vector. These vectors are called gradients, and they show the direction from bright pixels to dark throughout the image [7]. However, preserving the gradient for each pixel is too much information. It would be better if we could see the main direction of the change in light at a higher level, thus obtaining a common image canvas.

This requires dividing the image into small squares of size N x N pixels each. In each square, it is necessary to calculate how many gradient points are rotated in each of the main directions (up, up, right, right, etc.). Then you need to replace this square in the image with vectors directed to the same place as the majority.

To find a face in a HOG image, all you need to do is find the part of the image that is most similar to the famous HOG pattern obtained from many other faces during training: The second step involves encoding the face. The solution is to create a convolutional neural network of deep learning. But instead of training the network to recognize objects in the image, you need to train it to create 128 dimensions for each face. During network training, three persons are analyzed simultaneously:  image of already known person;  other Image of the same person;  image of other person.

The algorithm then looks at the measurements it takes for each of these three images. Then he sets up the neural network a little to make sure that the measurements created for images # 1 and # 2 will be more similar, and the measurements for # 2 and # 3 less similar.

By repeating this step millions of times for millions of images of thousands of different people, the neural network learns to reliably create 128 dimensions for each person. Any ten different images of the same person should give approximately the same measurements. Received 128 measurements of each face is called a card.

109

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

This process of training a convolutional neural network to obtain a face map requires a lot of data and a powerful computer. Even with the expensive NVIDIA Tesla graphics card, it takes 24 hours of continuous training to get good accuracy. But once the network is trained, it will be able to generate measurements for any person, even one who sees for the first time. Therefore, this needs to be done only once.

Figure 4 – Network result example He third stage is the classification of a person by image encoding. All that needs to be done is to find a person in our database of famous people whose measurements are closest to those obtained by us. [7] This can be done using any basic machine learning classification algorithm. We will use a simple linear classifier – logistic regression.

It is necessary to train a classifier that will take measurements from the image being checked, and will tell which person we know most resembles him. The launch and training of the classifier takes milliseconds, and the output is the name of the person.

110

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Of the two above, the second method was chosen because of a number of advantages:  work time;  no need to re-train the neural network;  resource cost.

Conclusions In this work, we consider and analyze the advantages and disadvantages of existing methods of face recognition to solve the problem of monitoring student attendance. Based on the analyzed data and comparison, the optimal method for solving the problem was chosen.

References 1. Обзор методов идентификации людей на основе изображений лиц / Ю.П. Кулябичев, С.В. Пивторацкая 2. Уоссермен Ф. Нейрокомпьютерная техника: теория и практика / Ф. Уоссермен. М., 1992.

3. Форсайт Д.А., Понс Ж. Компьютерное зрение: современный подход / Д.А. Форсайт, Ж. Понс: пер. с англ. – М.: Издательский дом «Вильямс», 2004.

4. Face Recognition using Convolutional Neural Network and Simple Logistic Classifier /Hurieh Khalajzadeh.

5. Imagebased Face Recognition Issues and Methods // WenYi Zhao, Rama Chellappa 6. Moghaddam B. and Pentland A. // IEEE Transactions on Pattern Analysis and Machine Intelligence. 1997. V. 19. P. 696.

7. Samaria F. Face Recognition Using Hidden Markov Models // PhD thesis, Engineering Department, Cambridge University, 1994. Аннотация. В данной работе выполнен анализ проблемы автоматизации контроля посещаемости студентов, рассмотрены существующие методы её решения, а также был выбран наиболее подходящий из них. Ключевые слова: распознавание лиц, нейросеть, компьютерное зрение, обучение. Сведения об авторах: Рудак Леонид Викторович – ст.гр. ПИм-19, ДонНТУ Федяев Олег Иванович – доцент кафедры программной инженерии, ДонНТУ Гировская И.В. – старший преподаватель кафедры английского языка ДонНТУ

111

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

UDC 662.749.33 SUBSTANTIATION OF THE MINING WORKING’S CONSTRUCTION ON THE BASIS OF ANCHORS, ATTACHED TO THE MINING ARRAY WITHOUT BINDING COMPOSITIONS Rusakov V.O., Petrenko U.A.. Boyko V.N. rusakov-94@list.ru

Abstract: This report describes the methods and means of securing the mine workings with anchor support according to the literature sourses. Justification of new concepts for the use of anchor fastening, its constructive manufacturing and methods of the anchor’s securing in the rock mass.

Keywords: anchor, metal consumption, binder, the mechanism of anchor’s operation.

Nowadays more than 80% of all supported mine workings in the mine are secured by a metal arched pliable support. Moreover, more than 50% of the length of these workings are deformed.

As a mounting system, the arch support has several disadvantages. In fact, it does not support the mine working until the host rocks collapse and begin to shift into the mine working, loading the roof supports.

So the support works in a passive mode and does not prevent the destruction of the enclosing array. In addition, the main disadvantages of the arch support using are: 1.Large metal consumption 2. The support is not included in the work immediately after exposure of the rock contour of the mine working.

3. The inability to fully mechanize the fastening process (tightening of the frames and backing up of fixing space are done manually. The complexity of the process of securing the workings with arch support reaches 80% of the total laboriousness of the workings).

4. The traditional construction of the arch lining does not meet the conditions of its loading (there is no coaxially between the direction of compliance of the lining and the direction of the largest displacements of the output contour).

You can fundamentally improve the technical and economic performance of the mine by applying anchor support [2].

This locking system has a number of advantages that are proven by practice and include: 1.

Improving the safety of work, disposed by the elimination of industrial injuries caused by the collapse of rocks in the process of conducting and operating mine workings.

2.

An increase in 1.5-2.0 times the pace of construction workings.

112

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

3.

The reduction in 5-10 times of material and labor costs for fixing workings.

4.

The reduction in the transportation volume of fasteners and a decrease in the manual labor share of in the process of installing supports.

5.

More efficient use of the cross section of the workings by reducing the loss of working area.

6.

Reducing of repairing cost of mine workings during their operation. So, for example, the use of anchor fastening in UK mines allowed to reduce the share of the cost of mining in the cost of 1 ton of coal from 42% for metal fastening to 15% for anchor. At the same time, the pace of mine workings amounted to 650-680 m / month.

For the widespread introduction of international experience in Ukrainian mines, by order of the Minister of Coal Industry, in 1997, the Anker program was distributed, in which one of the priority areas for reducing coal production costs was the development, manufacture and implementation of new technologies for using anchor supports.

Currently, the calculation of anchor support’s parameters is made in accordance with the requirements of regulatory documents [10].

The calculation is carried out in the following sequence. At the beginning, the average weighted strength of the rocks containing the excavation is determined. Then, the expected displacements of the excavation contour are calculated, the magnitude of which determines the load on the anchor support. At the same time, the normative compliance of the anchor rod is taken into account, with a value of 1.0% of its length.

The number of anchors is determined by dividing of the expected load on the bearing capacity of the anchor. The resulting number of anchors is ramdomly distributed into the roof and the sides of the mine working.

Let us analyze this technique in detail. Let's start with the mechanism of the anchor support’s operation. The existing traditional ideas about the operation of anchor supports according to the «Hemming» and «Sewing» schemes do not correspond to reality, because a zone of inelastic deformations is formed around the workings at great depths, which is laid down in [10estion about the role of anchor support. After all, the anchor is installed in a tunnel, in an array that is not yet destroyed. What kind of load on the anchor and its bearing capacity can be discussed? Because the contact area of the anchor lining on the contour is not comparable with the frame lining, and the passive role of the lining, i.e. lining perceives the load on the circuit according to deformations development deeper in the array. According to the existing method, the greater the bearing capacity of the anchor, the lower their installation density. This is acceptable for frame structures, and for anchor support this can lead to the paradoxical conclusion that the working can be supported by 2-3 or even 1 anchor.

Consider the following aspect of the existing methodology. Calculating the load on the anchor support, possible deformations of the anchor are laid within 1% of its length, i.e. 20-25 mm. The operating experience of the workings fixed with anchor support shows that the displacements of their contour exceed 100 and more mm [4].

113

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Obviously, the role of anchors is not limited only to load perception. If we analyze the anchors’ using experience in the Donbass mines, you can notice such an important pattern that in the workings secured with anchor support, the deformation of rocks has decreased and is generally absent, although this phenomenon is observed under similar conditions with frame support.

In our opinion, the mechanism of anchor lining operation consists not only in the idea of anchors as a supporting structure such as a frame, but as elements that change the structure of the array and prevent its destruction, i.e. the formation of the zone of inelastic deformations around the working. From these positions, the absence of deformations in the workings fixed with arch support is easily explained.

In Donetsk National Technical University is proposed a new concept for the use and operation of anchor fastening in workings. The new concept is based on the principle: how many anchors need to be installed so that the array does not collapse or collapse within specified limits?

Based on this concept, the currently used radial arrangement of anchors is not the most rational since the area of influence of anchors on the array in this case is minimal. In this regard, the spatial patterns of rock anchoring developed at Donetsk National Technical University allow using the minimum number of anchors to maximize the load-bearing capacity of the rock mass [3].

Such an arrangement of the anchor support allows it to be used not only as a power element that prevents stratification of rocks and their displacement into the cavity of the mine, but also as an element that provides communication between individual fragments of destroyed rocks in all directions (radial, tangential and along the axis of the mine). This provides a significant increase in the carrying capacity of the anchored shell of the destroyed rocks by increasing their residual strength.

Therefore, the search and development of alternative, resource-saving designs of anchors and methods for their fastening in the rock mass, as well as the technology of using anchor fastening in mine workings, is of great scientific and practical interest.

According to the well-established opinion of specialists in anchoring in the USA, England, and Germany, it is not recommended to use a rock mass in already destroyed rocks, which will be destroyed in the future. However, in most cases, anchor support in the mine workings is recommended to be installed in the mine workings in the radial direction. This practice significantly narrows the scope of the possible use of anchoring, especially in difficult mining and geological conditions (weak rocks and large development depths).

The performed analysis of geomechanical processes in the vicinity of supported workings during the formation and movement of the zone of broken foam from the working path deep into the massif at deep gauging stations showed that the interaction of the moving front of rock destruction with the support occurs due to waves of alternating deformations of the destroyed massif (compression-tension) that move in the radial direction from the boundary of the fracture front to the output contour [2].

Therefore, under such conditions, the installation of anchors in the radial direction is not rational.

114

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

The experience of using a rigid anchor support fixed along its entire length when installing it in the radial direction shows that when the output contour is shifted by 200 mm, it breaks [10]. This is due to the fact that in the conditions of the formation of the zone of destroyed rocks, the area of rocks strengthened by the anchor is much smaller than the distance between them, as a result of which the destroyed rocks beat the anchor support. Therefore, the reinforcing effect of the anchor support installed in the radial direction is carried out only within the rocks directly associated with the anchor. It is possible to increase the area of the anchor influence due to its rational spatial arrangement.

A method for protecting workings has been developed at Donetsk National Technical University, namely, that the anchor support is installed by rosette (4 anchors each), in which the anchors are located along large diagonals of the cube, one side of the base of which coincides with the longitudinal axis of the excavation, and the second linearly approximates the contour of its transverse [ 4].

In order to study the influence of various schemes for installing anchor crepe, the conditionally instantaneous and residual rock strengths at the rock pressure laboratory of Donetsk National Technical University tested the samples made of phosphogypsum with various schemes for their reinforcement. The test results showed that the location of the anchors along the large diagonals of the cube leads to an increase in the conditional instant strength of the sample by 1.6 times. A characteristic feature of sample deformation is the preservation of residual strength after fracture, which is 30% of the conditionally instant strength of a reinforced sample and 50% of the conditional instant strength of a sample without reinforcement. Such a result is explained, in our opinion, by the fact that, with a spatial arrangement, the anchor support plays the role of a rod holder that changes the form of the stress state of the destroyed rocks within itself [3].

In the proposed new concept, the role of the anchor is to provide resistance to the delamination of the roof rocks and their deformation into the cavity of the mine. In the direction of implementing of the proposed new concept at DonNTU, was developed the design of the anchor fastening in the form of a frame of anchor support. Picture 1 shows the design of the anchor support and its composition. The frame of the anchor support consists of two anchors, the tail of which (10 cm long in nature) is bent to 135*, position (1), pickup (2), 2 wedges (for example, from wood). The locking part along the edges of the pickup is made on a 15 cm section in kind in the form of a «swallow tail».

115

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Figure 1 – The construction of the anchor support’s frame (1 – anchor, 2 – grip, 3 – wedge, B – locking part) A distinctive feature of the proposed design of anchor fastening is the simplicity in the execution of components, the absence of threaded connections, as well as the fixing of anchors without the use of binders. It should be noted that this system can work only in a complex structure - two anchors, a pickup with a lock part at its ends and two wedges. Technologically, the installation of the anchor support’s frame is carried out in the following sequence: 1) Drilling of two holes towards each other at an angle of 45 * to the roof of the mine; 2) Pressing with a jack of pickup to the roof of the mine working; 3) The introduction of anchors alternately in the bore holes and their fastening with pickup in the locking parts using wedges.

In this case, the shape of the mine working should be with a flat roof: rectangular; trapezoidal; reverse trapezoid.

References 1. Анкерная крепь: Справочник / А.П.Широков, В.А.Лидер, М.А.Дзауров и др. – М.: Недра, 1990. – 205 с.

2. Зборщик М.П., Касьян Н.Н., Клюев А.П., Азаматов Р.И. Геомеханические процессы в зоне разрушенных пород в окрестности поддерживаемых выработок //Уголь Украины. – 1996 – №4 – С. 7-9.

3. Касьян Н.Н. Повышение эффективности применения анкерной крепи для поддержания выработок глубоких шахт//Известия Донецкого Горного института. – 1996. – №2(4). –с.53-55.

116

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

4. Касьян Н.Н., Клюев, Лысенко В.И. .Влияние анкерной крепи на геомеханические процессы в массиве пород вокруг поддерживаемых выработок//Донецкий горный институт. – 1996 – №1(3). – с.57-60.

5. Кисилев В.Г., Егоров Н.К., Трушин В.С., Рожанская В.И. Исследование анкерной крепи из пластифицированной древесины // Уголь Украины. – 1978. – №8. – с. 17-18.

6. Кравченко Г.И., Венгловский В.Г., Федоренко А.И. Временная технологическая инструкция по применению и установке трубчатых шатнг врзывного закрепления. Новокузнецк. ВостНИГРИ. – 1982. – 26 с.

7. Патент Украины 38093А Е21Д13/02Способы охраны выработок /М.П.Зборщик ,А.П.Клюев,Н.Н.Касьян,И.А.Скидан.-Опуб.15.05.2001.Бюл.№4 8. Петренко Ю.А., Касьян Н.Н. Новиков А.О.,Сахно И.Г. Новый подход к расчету параметров анкерной крепи //Зб.научн.тр. «Физико-технические проблемы горного производства»,вып.7-ИФГП НАН Украины, 2004. – С.167-172 9. Симпсон Р.И. Новый способ закрепления анкерных штанг быстротвердеющей смесью на цементной основе // Глюкауф. – 1980. – №23. с.

41-42.

10. Указания по рациональному расположению охране и поддержанию горных выработок на угольных шахтах СССР.-Л .:ВНИМИ,1986 – 221с.

11. .Широков А.П. Классификация анкерных крепей // Уголь Украины. – 1976. – №12. – с.4-6.

12. Широков А.П. Теория и практика применения анкерной крепи. М.: Недра, 1981. – 391 с.

13. Широков А.П., Давыдов В.В., Дзауров М.А. Армополимерная анкерная крепь // Уголь Украины. – 1977. – №12. – с. 8-12.

14. Шюрман Ф. Прогрессивная анкерная крепь во французской горной промышленности // Глюкауф. – 1979. – №11. – с.9-16. Аннотация. В докладе рассмотрены способы и средства крепления горных выработок анкерной крепью по данным литературных источников. Обоснование новых концепций применения анкерного крепления, его конструктивного изготовления и способов закрепления анкера в горном массиве. Ключевые слова: анкера, металоемкость, подшивка, механизм работы анкера. Сведения об авторах: Русаков Виталий Олегович – аспирант каф.РМПИ, ДонНТУ Петренко Юрий Анатольевич –д.т.н, профессор, зав.кафедры РМПИ Бойко Виктория Николаевна –ст. преподаватель кафедры английского языка, ДонНТУ

117

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

UDC 622.333 ESSENCE AND TYPES OF STRATEGIES FOR THE DEVELOPMENT OF CARBON INDUSTRY ENTERPRISES Savchenko L.A., Borshсh I.V. lejla.savchenko.97@mail.ru

Abstract. The article reveals the content of the concept of «strategy», systematized and generalized numerous interpretations of this concept of modern scientists. The concept of «strategy» is classified according to many different signs, which, if properly organized, will ensure the best functioning of coal mining enterprises.

Keywords: strategy, coal mining activity, perspective, competitiveness, factor, mechanism.

Relevance. A strategy is a combination of tactical actions that helps achieve business goals. It includes setting priorities for tasks based on the assessment of the unique positions and market prospects of a particular enterprise. This is a full-fledged algorithm of actions that makes a business many times more effective, since not only the owners of the company, but also its employees are involved in the decision-making process.

Strategy development is needed in order to think about future steps. As part of this process, a functioning mission is determined, goal maps are developed and plans for the implementation of planned activities for several years are drawn up. A clear understanding of the strategy depends on the degree of transparency of the market environment, the desire of owners to plan their business and the ability to analyze and correctly interpret the information collected and received [1].

In addition, the development of coal enterprises will contribute to improving their financial and economic status, as well as increasing the competitiveness of coal as a commodity in the domestic and foreign markets. It should be noted that all these positive consequences of the development of coal mining enterprises will take place only if appropriate development strategies are formed. That is why the problems of strategic development of coal enterprises are becoming especially relevant for scientific research.

Analysis of publications. Issues of evaluating the effectiveness of the development strategy of coal industry enterprises were studied by Anopchenko T. Yu., Basovsky L.E., Gorbunov S.V., Zabelin P.V., Markova V.D., Shaidulin I.U., Popov S.A., Amosha A.I., Embulaev V.N., Kopytov A. I., Kamenilera S.E., Skryl A.I., Cherevatsky D.Yu. and other scientists.

Goal and tasks. The goal is to study the features of the development strategy of enterprises of a coal organization. Tasks: to study the concepts of «strategy» and 118

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

«development strategy»; to explore the types of strategies and their features; to highlight the factors affecting the choice of development strategy of enterprises of a coal organization; to highlight the features of the development strategy of enterprises of a coal organization [2].

Main text. Current governance practice requires the development of special methodological approaches to the formation of a development strategy mechanism that should take into account the specifics of their industry and the uncertainty of the external and internal environment of enterprises.

For a more complete and deeper disclosure of the concept of «strategy», it is advisable to systematize the numerous interpretations of this concept by modern scientists in Table 1.

Table 1 – Interpretation of the concept «strategy» Anopchenko T. Yu.

Basovskiy L.E.

Gorbunov S.V.

Zabelin P.V. Markova V.D.

Long-term, qualitatively defined direction of the organization's development, relating to the scope, means and form of its activity, the system of relationships within the organization, as well as the organization’s position in the environment The form of manifestation of managerial activity, the relationship of the goal and the way to achieve it; means to achieve certain goals The long-term program of the organization’s activities, which is constantly monitored, evaluated and adjusted in the process of its implementation A generalized model of actions necessary to achieve the goals by coordinating and allocating company resources Identification of key long-term goals together with an appropriate action plan and allocation of resources to achieve them

An analysis of the above interpretations of the concept of “strategy” indicates that most scientists understand the direction of development of an enterprise through the specified goals of the enterprise as a strategy; enterprise activity plan; rules of action or behavior. In modern conditions of economic development, the strategy should be understood through a combination of the three above-mentioned aspects: clearly defined goals for the development of the enterprise, a comprehensive plan and specific measures to achieve them [3].

Based on the analysis of the presented interpretations of the concept of «strategy», one can formulate one’s own definition of this concept: an enterprise’s strategy as an open socio-economic system is a long-term comprehensive plan of enterprise development directions and an effective mechanism for its implementation, 119

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

based on the formed potential of the enterprise, taking into account current environmental conditions.

Having systematized the research of scientists [3], it should be noted that organization strategies can be classified according to many different signs.

1. The level of decision-making is proposed to distinguish between these types of strategies:  a system strategy, which is a model of the behavior of the socio-economic system as a whole, which establishes the order of the distribution of resources between different areas of the system;  a functional strategy, which determines by what actions, resources, competitive advantages the system will achieve success within the framework of a certain direction of activity;  a strategy for improving business processes, which determines how the improvement of the system should be carried out within a specific process.

2. Depending on the direction of action, there are:  stabilization strategy, which offers support for controlled quantities near certain standardized values, despite changes in the external environment;  observation strategy, including compliance with the current state of the controlled quantity to the required level at a given time;  program-oriented strategy, which is expressed in the achievement of set values of controlled quantities that change in time in a known manner;  optimization strategy, which is the fulfillment of specified conditions in the form of extrema of certain functions of the system parameters.

3. According to the style of development distinguish:  planned strategy – a strategy that is formed as a result of a deliberate, fully conscious and controlled mental process; acts as a result of planning;  entrepreneurial type strategy – a strategy that is formed by the leaderentrepreneur on the basis of his personal vision of the situation, trends in the development of the system, occurs semi-consciously);  a strategy based on experience – characterizes adaptive or reactive strategies, is formed in stages, cyclically, according to the dominant influence of impulses from the outside.

4. According to the degree of risk, there are:  high-risk strategy;  medium risk strategy;  low risk strategy.

5. Depending on the interaction with the environment emit:  active strategy;  passive strategy.

6. Depending on the dynamics of the target parameters of the behavior model, there are: 120

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

 development strategy – a strategy that provides for the transition of a system from one state to another, its improvement due to both quantitative growth and qualitative changes;  functioning strategy – a strategy that provides support for the activity of the system at the achieved level, which is divided into restrained (attempts to strengthen and stabilize the position of the system, make it manageable and regulated) and balanced (the transition to different ways of planning and implementing plans depending on the situation and relevance of the decision set task) [4].

The development of coal industry enterprises depends on the influence of external and internal factors and determines the search for ways to ensure their stable operation through innovative and organizational changes in the short and long term.

Given the industry-specific features of the functioning of coal mining enterprises associated with the single-product nature of production, their innovative development, as a rule, is associated with technical re-equipment and the introduction of technical and technological solutions that can provide support and increase the production capacity of coal enterprises. The introduction of new equipment and technology in all subsystems allows some time to increase the efficiency of the coal mining enterprise and improve its technical and economic indicators. However, the constant deterioration of the geological and mining conditions of coal seam mining and the negative impact of this factor increases as mining operations deepen, and it makes it more urgent to solve the problem of developing an appropriate mechanism for forming a long-term development strategy for coal mining enterprises.

Thus, the strategic stability of the entrepreneurial structure of a coal organization will ensure the selection of the right development strategy in accordance with the entrepreneurial potential and the level of competitive advantages. This implies the concept of a coal mining enterprise development strategy as a link in the proper planning and implementation of decisions related to increasing competitiveness, productivity, and mining activities. The development strategy makes it possible to improve the functional elements of the system of coal mining enterprises [4].

References 1. Анопченко Т.Ю. Формирование стратегии развития коммерческой организации на основе критериального выбора [Электронный ресурс] // Фундаментальные исследования. – 2017. – № 10-3 – Электрон. журн. – URL: https://www.fundamentalresearch.ru/ru/article/view?id=41871.

2. Басовский Л.Е Современный стратегический анализ: учебник / Л.Е. Басовский. – ИНФРА-М, 2015. – 255с 3. Горбунов С.В. Стратегический менеджмент: Учебное пособие. – Н. Новгород: НГаС, 2016. – 286 с.

4. Ембулаев В.Н. Формализация задачи управления в угольной промышленности на уровне региона / В.Н. Ембулаев // Фундаментальные исследования. – 2017. – № 10-2. – С. 339-343.

121

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Аннотация. В статье раскрыто содержание понятия «стратегия», систематизированы и обобщены многочисленные трактовки данного понятия современных учёных. Понятие «стратегии» классифицируется по множеству разных признаков, которые при правильной организации обеспечат наилучшее функционирование предприятий угледобывающей деятельности. Ключевые слова: стратегия, угледобывающая деятельность, перспектива, конкурентоспособность, фактор, механизм. Сведения об авторах: Савченко Лейла Ализаминовна – студент группы МОм-19, ДонНТУ Борщ Ирина Владимировна – ст.преподаватель кафедры английского языка. ДонНТУ

UDC 681.586.78 MAGNETIC FIELDS USED IN ELECTROMAGNETIC FLOW METERS Smeshnaya A.V., Kushnirenko Ye.N. linasavinavchsl@gmail.com

Abstract. The article analyzes methods and means of measuring the flow of liquid media in pressure pipelines. After that, the conclusion is made which method is more promising for measuring flow in large-diameter pressure pipelines and its advantages are considered.

Keywords: Flow meter, converter, cross-section, flow rate, measurement, large diameter pipeline All flow measuring tools can be divided into two classes depending on the measurement method: direct flow measurement and indirect flow measurement.

Devices based on the first method measure directly the liquid flow in the pipeline [2]. Devices that implement the indirect method of flow measuring by the «speed-area» method determine the fluid flow rate in the pipe at one point of the cross-section of the pipeline and its area [1,2,4]. In the latter case, the flow measuring is based on the laws of turbulent fluid flow in the pipe, according to which the local flow velocity «V» at a point located at a distance [(0,242 ± 0,013) × R] from the inner surface of the pipe is equal to the average velocity of the fluid in this section – V = Vsr, and therefore is proportional to its flow rate.

Instruments that implement the direct flow measurement method Mechanical flow meters This group of devices includes all variable differential pressure sensors (standard orifices, nozzles, pipes and Venturi nozzles, etc.) [2].

122

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

A typical representative of variable differential pressure flow meters (Fig. 1), which is most widely used in the practice of flow measuring, is a Converter that uses a standard diaphragm as a narrowing device – a flat disk with a concentric hole in the center for the flow of the measured medium. On the inlet side of the flow, the aperture in the diaphragm is cylindrical, and on the outlet side it is conical.

Figure 1 – Flow meter diagram of variable differential pressure The diaphragm installed in the pipeline in the path of the fluid movement creates a pressure drop, by measuring which you can determine the volume flow [2] These flow meters are widely used in industrial practice, due to the simplicity of their design. However, despite the wide variety of narrowing devices and specially designed designs, high accuracy of flow measurement with these measuring tools is not possible. The service life of most diaphragms does not exceed 3 years, since the input edge of the diaphragm is inevitably blunted during operation that reduces the output speed and the measured pressure drop. A negative flow measurement error is formed, which increases as much as the diaphragm is used [2].

In addition, the increased pressure drop that is created by the diaphragm to implement the measuring process leads to energy loss increasing for transporting the fluid through the pipeline. The use of flow meters with variable pressure drop is limited to areas where relatively low measurement accuracy is required. The maximum reduced error of these flow meters is rarely less than (1 ... 2) % [2].

In recent years, vortex flow meters are increasingly used for flow measuring (Fig. 2), based on the dependence of pressure pulsations frequency in the flow on the flow rate. The essence of the method is to create a stable vortex structure in the flow of a moving fluid, the excitation of which is achieved by twisting the flow or by flow around a stationary body of a special shape. There are several varieties of vortex flowmeters [2], but the final criterion of any design is the fluid pressure pulsations, which are characterized by the Reynolds number Re  VD /  , that is the ratio of inertia 123

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

forces to the forces of viscous friction in the flow, and the Strouhal number Sh=fd/V, which determines the periodic processes associated with the fluid moving.

Figure 2 – Diagram of a vortex flow meter Advantages of vortex flow meters are: linear conversion characteristics, high accuracy (error within  0,51,5%), easy conversion, no moving parts, low inertia, frequency output signal. However, such flow meters with large pipeline diameters (over 250-300 mm) operate in the low-frequency region and do not provide stable vortex formation [2]. The disadvantages of vortex flow meters is the sensitivity to contamination of the measured medium.

Tools for measuring flow using the indirect method measurements Of considerable interest for measuring the fluid flow in pressure pipelines of large diameter are devices whose operation is based on determination of volume fluid flow in the pipeline by the indirect method, i.e. the flow rate at one point of the pipeline cross section (the point of average velocity) and its area. As already noted, this measurement method is based on the laws of turbulent flow in pipes, according to which the local flow velocity V at a point located at a distance (0,2420,013)R from the inner surface of the pipeline is equal to the average axial velocity V = Vsr of the medium flow in this section, which means that it is proportional to the volume flow rate [1].

Flow meters with pressure devices Flow meters with pressure devices are flow meters with a variable pressure drop.

The essence of this method is to measure the dynamic and static pressure in the flow, the values of which determine the flow rate [4].

124

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

A classic example of a pressure device is the Pitot-Prandtl tube, which forms a pressure drop depending on the flow rate existing in the place where it is installed.

Currently, Pitot-Prandtl tubes of various modifications are known, but their use for measuring local velocities in fluid flows is mainly limited to laboratory studies. This is primarily due to the requirement that the measured fluid is free of impurities and precipitation. Pitot-Prandtl tubes have a number of disadvantages, the main of which are a) high inertia; b) low sensitivity (about 15-20 cm/s); non-linearity of the calibration characteristic, especially affecting the measurement of low speeds – less than (3-5) m/s [3].

Flow meters with hydrometric turntables Tachometric flow meters for large diameters pipelines (flow meters with hydrometric turntables) differ from conventional turbine flow meters in that their measuring device – turntables – is not affected by the entire flow, but only part of it.

Practically flow meters with hydrometric turntables measure the flow at the crosssection point of the pipeline.

Flow meters with hydrometric turntables for pipelines of large diameters are characterized by insignificant head losses, which are caused by the placement of the primary speed measuring Converter (turntables) in the studied flow. This is their advantage.

The disadvantages of such devices include the presence of moving parts of the flow meter in the pipeline and, as a result, the need for a complex system of their lubrication. In comparison with pressure pipes, hydrometric turntables have large dimensions [2].

Flow meters with MHD converters The need to measure the fluid flow in pipelines of large diameters is great and constantly increasing. Therefore, taking into account the advantages of electromagnetic flow meters when measuring the flow in large diameters pipelines, the interest in them is constantly increasing. However, the disadvantages of flow meters based on the direct measurement method, associated with large dimensions, the consumption of iron and copper, force us to abandon such measurement tools and pay attention to electromagnetic speed converters (MHD – converters) with a magnetic field localized in the zone of installation of sensitive electrodes. Such electromagnetic flow meters with a local magnetic field introduced into the flow are simpler and cheaper, and their advantage increases with the enlargement of the pipeline diameter. The magnetic field in the area of the sensitive electrodes of the Converter is created by a miniature coil fed by alternating current [3, 4]. Besides, to increase the sensitivity of the flow meter, various types of magnetic field concentrators located directly in the electrodes vicinity can be used [1].

The promise of electromagnetic flow meters with a local magnetic field is due to the fact that they have small dimensions, disrupt the flow minimally, are technological, durable, have a low metal content and low power consumption, are easily installed through standard entrances to continuous pipelines, etc. [2, 4]. High 125

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

performance indicators of such flow meters rightfully place them among the most promising devices for measuring water flow in large diameters pipelines.

Conclusions 1. Currently, there are no reliable flow meters that provide reliable information about the fluid flow media in large pipelines and meet all the requirements for modern flow measurement tools.

2. Flow measurement tools can be divided into two classes: direct flow measurement and indirect flow measurement. Devices that implement the direct measuring method directly measure the flow of fluid in the pipeline; devices that implement the indirect measuring method determine the flow rate of fluid in the pipeline at one point of pipeline cross section (the point of average speed) and its area.

3. Measurement of fluid flow in filled pipes of large-diameter flow meters is based on the method of measuring the flow at one point of the pipeline cross-section and its area.

References 1. ГОСТ 8.439-81 Методика выполнения измерений методом площадь скорость. Введ. 23.09.81. – Изд-во стандартов, 1982. – 51 с.

2. Корсунский Л.М. Электромагнитные гидрометрические приборы. – М.: Стандартгиз, 1964. – 180 с.

3. Кремлевский П.П. Расходомеры и счетчики количества вещества: Справочник. - Кн. 2 – СПб.: Политехника, 2004. – 412 с.

4. Шерклиф Дж. Теория электромагнитного измерения расхода: Пер. с. англ. – М.: Мир, 1965. – 268 с. Аннотация. В статье проведен анализ методов и средств измерения расхода жидких сред в напорных трубопроводах. После чего сделан вывод какой метод более перспективен для измерения расхода в напорных трубопроводах большого диаметра, рассмотрены его достоинства. Ключевые слова: расходомер, преобразователь, поперечное сечение, скорость потока, измерение, трубопровод большого диаметра. Сведения об авторах: Смешная Алина Вячеславовна – студент группы ПСм-19, ДонНТУ Кушниренко Елена Николаевна – ст.преподаватель кафедры английского языка, ДонНТУ

126

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

UDC 662.749.33 SMART TRAFFIC LIGHT CONTROL SYSTEM FOR URBAN TRAFFIC FLOWS Strizhko M. A., Sukov S. F., Paniotova L. N. michailstrizhko@gmail.com

Abstract. In this article the goals of road regulation are formulated. The analysis of existing traffic control systems and technical means on which these systems are based is performed. The concept of building a traffic light control system is proposed.

The algorithm for road regulation is developed.

Keywords: traffic light, intersection, control, transport detector.

The growth of the car fleet and traffic volume leads to an increase in traffic intensity that result in a transport problem in cities with historically developed buildings. Traffic delays are increasing, queues and obstructions are forming, which causes a decrease in the speed of communication, unjustified fuel overspending and increased wear on vehicle components and assemblies.

When implementing traffic control actions, a special role is played by the introduction of technical means: road signs and road marking, traffic lights, road barriers and guide devices.

At the same time traffic light regulation is one of the main means of ensuring traffic safety at intersections. The number of intersections equipped with traffic lights in the world's largest cities with a high level of motorization is constantly increasing.

In some cases it reaches the ratio of one traffic light object for 1.5-2 thousand residents of the city.

Traffic is a complex, time-changing system. The system is a set of relationships between moving and stationary vehicles controlled by people and pedestrian flows.

Each object of the system pursues its own specific goals when driving on the road and strives to carry them out with minimal time loss and sufficient degree of safety.

The main point of regulation is to require, prohibit and recommend certain actions to drivers and pedestrians in the interests of speed and safety. It is carried out using the «Traffic regulations», a set of technical means, and administrative actions of traffic police officers.

There are different ways to control traffic depending on the degree of human participation in the control process, the degree of centralization and the availability of feedback.

Automatic control of the traffic light object is carried out without the participation of a person according to a pre-set program. Automated control requires the participation of a human operator. The operator, using a set of technical tools to collect the necessary information and find the optimal solution, can adjust the program of automatic devices. Both in the first and in the second cases computers can be used 127

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

in the control process. Finally, there is manual control, when the operator, assessing the transport situation visually, exerts a controlling influence, which is based on his experience and intuition. The automatic control circuit can be either open or closed.

In a closed circuit, there is feedback between the means and the control object (traffic flow). Automatically it can be carried out by special devices for collecting information — transport detectors. Information enters automation devices. Based on the results of its processing, these devices determine the operation mode of traffic lights or road signs that can change their value on command (controlled signs). This process is called flexible or adaptive control.

When the circuit is open there is no feedback, traffic light control devices (road controllers) switch signals according to a pre-set program. In this case strict software control is performed.

According to the degree of centralization, two types of control can be considered: the local and the systemic one. Both types are implemented in the ways described above.

In local control the signal switching is provided by a road controller located directly at the intersection while in system control the intersection controllers usually act as translators of commands received via special communication channels from the control point. If the controllers are temporarily disconnected from the control center, they can also provide local control. In practice, the terms are used: local controllers and system controllers. The first ones do not have a connection with the control point and work independently, the second ones have this connection and are able to implement local and system management.

With local manual control, the operator is located directly at the intersection watching the movement of vehicles and pedestrians. With the systemic one, he is located in the control point, i.e. away from the control object.

Local control is most used at separate or isolated intersection that has no connection to neighboring intersections either by control or by traffic flow. Changing traffic signals at such intersection is provided by an individual program regardless of traffic conditions at neighboring intersections. The arrival of vehicles to this intersection is random.

The organization of coordinated change of signals at a group of intersections, carried out in order to reduce the time of vehicles movement in the specified area, is called coordinated control (on the principle of «green wave» - GW). In this case, system control is usually used.

This article proposes an adaptive control algorithm that implies coordinated automatic traffic control with feedback. The algorithm is based on analyzing the distance between cars arriving at the intersection. The distance is calculated using the time between the signals received from the transport detector about the passage of two cars. Their speed is assumed equal to the maximum allowed speed for this section. In this case, it is sufficient to use a single recording transport detector. The transport detector is used to determine the decrease in the flow density. It is useful when 128

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

developing a network of connected intersections because the algorithm replaces complex mathematical calculations and provides a variable phase shift.

Interactions between traffic detectors and traffic lights occur as follows: 1. The maximum (Tmax) and the minimum (Tmin) periods of the traffic light cycle and the phase shift time (Tshift) are set.

2. The maximum interval between cars (xint) is set.

3. At each step the distance between arriving vehicles is checked.

4. At the step when the traffic signal switching is coming the following condition is checked:

xi ≤ xint ⇒Ti =Ti + Sshift; xi > xint =Ti = 0.

If Ti = 0 the traffic light switches to red.

This algorithm considers only clusters of cars arriving at the intersection. If the flow turns out to be homogeneous and there are no obvious breaks, the adaptive algorithm will switch to the fixed control mode.

Using the algorithm described above will significantly improve the road situation, namely: - it will minimize traffic jams; - it will increase the average speed of traffic flow; - it will reduce fuel consumption and emissions to the environment.

References 1. Абрамова, Л. С. Способ повышения пропускной способности регулируемых перекрёстков / Л. С. Абрамова, В. В. Ширин // Восточноевропейский журнал передовых технологий. – 2010. – Вып. 4/3 (46). – С. 62-65.

2. Бергер, Г. Автоматизация с помощью программ STEP7 LAD и FBD / Г. Бергер. – Нюрнберг: Siemens AG, 2003. – 605 с.

3. Кретов, А. Ю. Обзор некоторых адаптивных алгоритмов светофорного регулирования перекрестков / А. Ю. Кретов, И. Е. Агуреев, И. Ю. Мацур // Известия ТулГУ. Технические науки. – 2013. – Вып. 7. Ч. 2 – С. 61-66.

4. Нестеров, А. Л. Проектирование АСУ ТП / А. Л. Нестеров – СПб.: ДЕАН, 2006. – 552 с. Аннотация. В докладе сформулированы основные цели дорожного регулирования. Выполнен анализ существующих систем управления движением автотранспорта, а также технических средств, на которых основываются эти системы. Предложена концепция построения САУ дорожным движением. Разработан алгоритм дорожного регулирования. Ключевые слова: светофор, перекрёсток, управление, детектор транспорта.

129

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Сведения об авторах: Стрижко Михаил Александрович – студент группы СУАм-19, ДонНТУ Суков Сергей Феликсович – к.т.н., доцент, профессор кафедры автоматики и телекоммуникаций, ДонНТУ Паниотова Любовь Николаевна – ст. преподаватель кафедры английского языка, ДонНТУ

UDC 007.52 THE ROLE OF AUTOMATION IN THE MODERN WORLD Sukhanov A.A., Sokolova O.V. antonotosha@gmail.com

Abstract: In the article the role of automation in modern world is discussed.

Author gives examples of automation and draws attention to possible consequences.

Advantages and disadvantages are listed in the article.

Keywords: automation, technological, development, progress, production.

Automation is necessary in our life and where it will lead us to. The article is important because the development of robotics and industrial automation can lead to the problems we shall meet in the future.

Automation is one of the areas of scientific and technological progress using self-regulating technical means and mathematical methods to relieve and discharge men in the processes of obtaining, converting, transmitting and using energy, materials, products or information, or to significantly reduce the man’s participation hard operation work. Automation can be compared with a human evolution. People have tried to «simplify» their existence since ancient times. This led to the developments of electricity, water pipes, computers, antibiotics and etc. Automation is the most significant in modern manufacturing [1].

Automation allows partly or completely to relieve a worker in cyclic processes or in the processes being performed according to a strictly specified algorithm.

Nowadays almost all levels of manufacturing or industry are controlled without men notifying him only in the case of a fault or a pre-emergency situation.

Industrial and technological progress caused the development of automation.

Automation in everyday life has taken its rise from industrial production. The desire to accelerate the process and to increase the benefit helped to introduce the latest at that time high-tech automation. Automated manufacturing let release a number of workers, optimize the time and staff employment.

130

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

At the end of the last century automation had the role of a performer of routine cyclic operations, now the entire manufacturing are being operated automatically due to the development of software and computer facilities.

Automation touches upon the fuel and energy industries too. Their maximum efficiency is impossible without co-operative and reliable equipment. automated process control systems based on the most advanced microprocessor equipment have been actively developed recently. They have been improved and implemented. The computing power of different systems and field-level devices, including sensors, converters, flow meters, are constantly being improved.

The automated control system monitors all components of the equipment unlike local automation systems. It can more accurately prevent the development of emergencies that are very expensive for the energy industry. The use of the latest developments in the field of microprocessor technology lets integrate automation systems into an Automated Process Control System of a high level. Apart from everything else such systems collect and diagnose information, they are engaged in group regulation of units operation and other technological schemes. Their coordination and reliability increase the production of energy resources [5].

Industrial programmable logic controllers have achieved tremendous performance, coupled with redundancy they make the work as reliable and fast as possible.

Now scientific and technological progress is advanced. All new automation systems are being tested and implemented. The development and cheapening of the microelectronic components allow developers to experiment and to create more advanced systems.

According to the information we have received, we can conclude that today it is impossible to achieve sustainable success, remaining within the framework of the previous enterprise production management system. It requires periodic improvements as well as the development and application of new technologies suitable for each enterprise at this stage of its activity [3].

It is possible to take out the advantages of automation according to the information mentioned above:  The improvement of the products quality;  The growth of manufacturing;  The improvement of enterprise efficiency;  The increase of safety for employees; Automation has negative ones too. One of the most significant problems of automation is «technological unemployment». Today there is negative attitude towards the introduction of automation. The lack of qualified staff is a big problem now too.

There are a lot of «old» specialists in the plants and young workers often do not have experience and do not know modern standards of work. Table 1 shows advantages and disadvantages of automation for enterprises:

131

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Table 1 – «Advantages» and «disadvantages» of automation Advantages Increase in enterprise profit Product cost reduction Creation of a productive product quality control system Perfect production system Effective product quality control system Decrease in defective products Decrease in defective products Replacing a person in hard work

Disadvantages Hacking threats Staff re-qualification Unemployment growth Complicating the production system

Automate your company or not, it already depends on the management, but how can automation affect the lives of ordinary people? To explain how automation can affect life, we can give a couple of examples from real life [1].

Recently, in some countries, chipping people is gaining popularity. People implant A microchip implant under the skin, which is a device built on an integrated circuit, or using RFID (radio frequency identification) technology. The microchip implant has a glass case and is implanted into the human body by analogy with the technique of animal chipping. Such implant usually contains a unique identification number. If necessary, it can be connected with an external database, which contains information about a person’s personal data, his medical history, contact information, etc. [2].

It sounds like the plot of a science fiction film, but it is reality. Chipping also has its advantages and disadvantages. One of the most important advantages is the complete destruction of paper documents, and the transfer of all important information about a person to a chip. Disadvantage can be seen in the situation when the chip is broken and the user's personal information leaks into the hands of the attacker and can be used by him.

For example, auto manufacturers improve their product every year. They invent new assembling systems and create new software controlling the systems of a smart car.

Automation has two sides in modern society. One of them is uncertainty: factories and plants are afraid of full automation and people had made such a big scientific and technological leap that they didn’t notice how vulnerable they are in their own networks and developments. The flip side of the coin is reduction of the accidents at the enterprises and the automation of routine work and the disposal of paper documents. Automation is necessary but only the future will show where it will lead.

132

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

References 1. Свободная энциклопедия «Википедия».

2001.

URL: https://ru.wikipedia.org/wiki/Автоматизация 2. Свободная энциклопедия «Википедия». 2001. – Режим доступа: https://ru.wikipedia.org/wiki/Микрочип_имплант_(человек) 3. Хлебенских, Л. В. Автоматизация производства в современном мире / Л. В. Хлебенских, М. А. Зубкова, Т. Ю. Саукова. Текст : непосредственный, электронный // Молодой ученый. – 2017. – № 16 (150). – С. 308-311. – Режим доступа: https://moluch.ru/archive/150/42390/ 4. Цветаев С. С., Логачев К. И. Актуальные проблемы автоматизации промышленных предприятий // Вестник Белгородского государственного технологического университета им. В. Г. Шухова. – 2012. – № 1. – С. 87-89.

5. Электротехнический журнал. Роль автоматизации в современном производстве. Режим доступа: https://www.el-info.ru/rol-avtomatizacii-vsovremennom-pro/ Аннотация: В статье рассмотрена роль автоматизации в современном мире. Автор приводит примеры автоматизации и обращает внимание на возможные последствия. Преимущества и недостатки автоматизации перечислены в статье. Ключевые слова: автоматизация, технологии, разработка, прогресс, производство. Сведения об авторах: Суханов Антон Алексеевич – студент группы ПИ-18в, ДонНТУ. Тел.: 071-407-8527 Соколова Ольга Викторовна – ассистент кафедры английского языка ДонНТУ

UDC 620.72 NANOSTRUCTURED CERAMIC FUNCTIONAL MATERIALS Tomilov M. K., Prilipko Y. S., Boyko V. N. super.maksimtomilov@yandex.ua

Abstract. The review discusses methods for the synthesis of bulk nanostructured ceramic materials. Particular attention is paid to the size effect – the effect of reducing the grain size in the nanoscale. Obtaining nanostructured ceramics with a high bulk density and a very small grain size is an extremely difficult task, which is most often solved by sintering previously synthesized nanopowders. The features of the basic 133

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

methods for the synthesis of nanopowders, as well as their consolidation and compaction into compact nanostructured products, are considered.

Keywords: сeramics, nanoceramics, nanomaterials, nanopowder, size, synthesis, properties.

Nanotechnology is considered in relation to obtaining fundamentally new or to multiple improvement of consumer properties of existing materials of structural and functional purpose. Strength is one of the main indicators that determine the very possibility of their use for their intended purpose [1].

Nanoceramics can be defined as a ceramic material in which crystallites are less than 100 nm in size [2]. During the heat treatment of the source material, along with the sintering process (compaction and formation of strong intergranular contacts), the process of ceramic grain growth proceeds. Due to the high chemical activity of nanocrystalline substances, the grain size in the final material can many times exceed the original, which leads to the formation of ordinary coarse-grained ceramics. To effectively inhibit the growth of crystallites and at the same time accelerate the sintering process, special methods are used. For effective compaction of the initial nanopowder, magnetic pulse or ultrasonic pressing is used; in some cases, the sample is additionally heated. Subsequent heat treatment is carried out at lower temperatures than for conventional ceramics, and additives that inhibit grain growth are added.

With a decrease in grain size, the strength of ceramics increases: a product from it becomes not so easy to break as, for example, a porcelain cup and in this case wear resistance increases as well. Plasticity may also occur at higher or even at room temperature. The sintering process itself requires relatively low temperatures, and the use of defect-free nanocrystalline powders leads to the production of a very homogeneous material at both the macro and micro levels. As a result, dielectric, magnetic and optical properties are improved. Dense ceramics become transparent like glass, and nanoporous ceramics turn out to be a better heat insulator than microporous.

The presence of fine uniform pores makes the material also useful for selective filtration and catalysis. For example, heat-protective tiles for modern spacecraft are made from nanoporous ceramics [3].

In samples of small sizes, which include dispersed powders obtained by various methods, a change in the temperature of phase transitions is observed (for example, a decrease in the melting temperature, a shift in the temperature of polymorphic transformations) that depends on the characteristic size (thickness, radius). Under certain conditions, the formation and stabilization of phases occurs in nanoceramic objects that is not observed in bulk samples at all . With a change in the dispersion of substances, the phase transition temperature changes – with a decrease in the particle size, the melting and evaporation temperatures decrease. This means that when a certain dispersion of the powders is achieved, it is possible to obtain the material at such temperatures that it is difficult to achieve with the usual fineness of the material.

The establishment of appropriate dimensional laws opens up the possibility of a transition to a new generation of materials whose properties are changed by adjusting the size and shape of their structural elements. These features can also be used to create 134

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

elements of instrumentation devices. Structural dimensional effects manifest themselves in the form of changes in interatomic distances, rearrangement of the crystal structure up to the transition to an amorphous state, and with chemical dimensional effects, the phase composition of the substance and its reactivity are changed. This means that the implementation of reactions with nanoparticles becomes possible, as it does not occur with substances in a compact state [4].

Functional systems play a particularly important role in practical activities.

These systems of material objects are used to solve practical problems and possess functional properties that determine the field of their practical application as well.

Functional systems are made on the basis of various materials, which are divided into raw materials (materials not previously processed), and semi-finished products (materials subjected to preliminary, partial processing). Functional nanosystems, like nanomaterials, are characterized by a nanometer scale in at least one of three dimensions. The properties of functional nanosystems, as well as the properties of nanomaterials, can manifest themselves in a very unusual way due to their inherent nanometer scale. In practice, solid-state functional nanosystems are used the most widely.

Up to now, a large number of methods and methods for producing nanomaterials have been developed. This is due to the diversity of the composition and properties of nanomaterials, on the one hand, and it allows to expand the assortment this class of substances, creating new and unique samples on the other hand.

The improvement of previously known and the development of new methods for producing nanomaterials has determined the basic requirements, they must meet [5]: - the method should ensure the production of material of controlled composition with reproducible properties; - the method should provide temporary stability of nanomaterials, i.e. first of all to protect, protecting the surface of the particles from spontaneous oxidation and sintering during the manufacturing process; - the method should have long-term efficiency and profitability; - the method should ensure the production of nanomaterials with a specific particle or grain size, and their size distribution should, if necessary, be sufficiently narrow.

Methods for producing nanomaterials are divided into mechanical, physical, chemical and biological. This classification is based on the nature of the process of synthesis of nanomaterials. Chemical methods for producing nanosized materials can be divided into groups, one of which can be attributed to methods where the nanomaterial is obtained by a particular chemical reaction in which certain classes of substances are involved. Different variants of electrochemical reactions can be attributed to another group. Physical methods of preparation are based on physical transformations: evaporation, condensation, sublimation, quenching or heating, atomization of the melt, etc.

Chemical and physical methods of nanoparticle synthesis were being followed over the decades, but their formation was found to be expensive. However, recently 135

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

these methods have become an important discovery for the industrial production of nanoparticle powders [5]. Chemical methods include chemical vapor deposition, sol‐ gel, hydrothermal route, Atomic or Molecular Condensation, and other precipitation processes. The sol gel technique is a long-established industrial process for the generation of nanoparticles from liquid phase. The main advantages of sol-gel techniques for the preparation of materials are low temperature of processing and versatility. They offer unique opportunities for access to organic-inorganic materials.

Sol-gel method is a long established industrial process for generating colloidal nanoparticles from liquid phase. It has been further developed in the last years for the production of advanced nanomaterials and coatings. Sol-gel process is a chemical method, which is based on hydrolysis or condensation reactions (Figure 1). With the correct amount of reactants, nanosized particles precipitate. Sol-gel techniques show many advantages like low temperatures during processing, versatility and easy shaping and embedding. Common precursors, which are used for the production of oxides, are alkoxides, due to their availability and to the high liability of the M-OR bond allowing facile tailoring in situ during processing. This method lucks the risk for nanoparticle release after the drying of the solution [5].

Figure 1 – Main stages of sol-gel method The dispersed state of the powder helps not only to produce of certain components, but also greatly effect on the properties of the obtained sample. The improved properties of nanoceramics are associated with new prospects in the nanomaterials industry: the production of durable, light and heat-resistant parts – turbine blades, nozzles, rocket fairings, cutters; production of biocompatible materials with adjustable porosity to replace bone tissue and heart valve; design of fuel cells, sensors, solar panels; creation of materials for magnetic recording, elements of microelectronics, optoelectronics and micromechanical systems.

Over the years research has provided researchers with a better understanding of the mechanisms involved during the synthesis of nanomaterials, their intrinsic properties affecting from increased surface area and quantum effects, and it still leads to the development of advanced analytical techniques for their characterization and their systematic tailored synthesis. Since many industrial sectors (such as aerospace, energy, transportation, and medicine) are highly benefited and improved by the application of nanomaterials, their demand is increasingly growing. The available 136

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

consumer products containing nanomaterials (as particles or fibres) most commonly carbon, cerium oxide, silver, silica, titanium dioxide, magnesium oxide and zinc oxide now exceeds the number of 1000 and is still raising. However, apart from the advantages that nanotechnology offers to society, it may also have negative impacts on human health and the environment that are not yet understood, since the quantum mechanics which regulate nanomaterials’ interactions with other substances also make their toxicological behavior difficult to predict.

Nowadays, in the aftermath of the nanotechnology breakthrough, we live in the era of nanotechnology application and commercialization. Ceramic nanomaterials (nanoceramics) are about to dominate the respective market due to their numerous and advantageous properties. Several industries have already incorporated within their production processes the manufacture of different types of ceramic nanoparticles, as well as the application of those nanomaterials on conventional products.

The development of products and processes containing ceramic nanoparticles has generated novel and fascinating applications of these materials in the past decades.

In addition to these exciting findings, ceramic nanoparticles tend to be highly stable.

Their routes of synthesis are well known and relatively cheap. The combination of tchnical advantages and profuse investment in research and development increased the number of patents and publications in this area. Even more recently (since 2002), research programs based on toxicology, eco-toxicology, ethics and public perception of nanotechnologies have pointed out potential risks and impacts associated with nanotechnologies. Because of their wide employment, the ceramic nanoparticles extensively have been studied by means of these new approaches and several unexpected hazardous effects such as high toxicity and environmental persistency were observed.

References 1. Витязь, П. А. Наноматериаловедение / П. А. Витязь, Н. А. Свидунович, Д. В. Куис – Минск: Вышэйшая школа, 2015 – 511 с.

2. Жабрев, В. А., Введение в нанотехнологию (общие сведения, понятия и определения) / В. А. Жабрев, В. И. Марголин, В. С. Павельев – Самара: Издательство СГАУ, 2007 – 173 с.

3. Гусев, А.И. Наноматериалы, наноструктуры, нанотехнологии / А. И. Гусев – М.: Физматлит, 2005 – 416 с.

4. Гладких, Н. Т., Поверхностные явления и фазовые превращения в конденсированных пленках / Н. Т. Гладких, С. В. Дукаров, А. П. Крышталь и др.

– Харьков: ХНУ имени В. Н. Каразина, 2004 – 276 с.

5. Колесник, И. В., Химические методы синтеза наноматериала / И. В. Колесник, А. А. Елисеев. – М.: Изд-во Мос. гос. ун-та, 2011. – 41 с. Аннотация. В обзоре рассматриваются методы синтеза объемных наноструктурированных керамических материалов. Особое внимание уделяется размерному эффекту - эффекту уменьшения размера зерна в 137

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

наноразмерном диапазоне. Получение наноструктурированной керамики с высокой объемной плотностью и очень малым размером зерна является чрезвычайно сложной задачей, которая чаще всего решается путем спекания ранее синтезированных нанопорошков. Рассмотрены особенности основных методов синтеза нанопорошков, а также их консолидации и уплотнения в компактные наноструктурированные продукты. Ключевые слова: керамика, нанокерамика, наноматериалы, нанопорошок, размер, синтез, свойства. Сведения об авторах: Томилов Максим Константинович – студент группы ТТМм-19, ДонНТУ Прилипко Юрий Степанович – профессор, к.т.н., ДонНТУ Бойко Виктория Николаевна – ст.преподаватель кафедры английского языка, ДонНТУ UDC 81'33 NEOLOGISMS AS A MEANS OF LANGUAGE DEVELOPMENT Yakhina I.E. yahina_i@kaltschool1.ru

Annotation: The rapid changes in society, the strengthening of international relations led to the rapid development of the language. The lexical composition of the language is overgrown with new words, old tokens fall out of the lexicon of the Russian language, new ones appear in their place. Neologisms are new words in a language that are adapted by society and find their fixation in the language. The appearance of neologisms is associated with the laws of the development of a language that is constantly evolving and improving.

Keywords: neologism, word, meaning, occasionalism, language, society, development.

The vocabulary of the Russian language is constantly updated with new words.

They appear like leaves on trees, the old words «fall off», and the new ones «grow».

Strengthening international relations of the country. There are changes in cultural life.

New sports are emerging. All this is reflected in the language.

A rapidly developing society, constant changes and changes in life are manifested in the language. Every day, society tries to live interesting and favorable, for this they make discoveries and some of them completely rebuild our lives. These words are actively analyzed in Russian linguistics, starting from the 20th century, as evidenced by collections, articles, magazines, and dictionaries [1].

138

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Thanks to such inventions as the satellite, the Internet and telephones have appeared, which allows us to communicate even with other continents. But, many discoveries come to us from other countries, and their names are not always clear to us. These changes are more expressed in the emergence of new words with a temporary sense of innovation. The assimilation of semantic neologisms is also happening rapidly.

Their timeliness for native speakers and its lexical-semantic system shows or refutes time, because only secondary names are stored in the language, which does not mind its laws, models and mechanisms.

They can be divided into: linguistic, lexical, semantic and copyright.

1) Language neologisms are created mainly to define a new subject, phenomenon or concept. They enter a blank vocabulary. Neologisms do not last long as a new word, but only as long as the word keeps updating. Then, the word goes into the category of common vocabulary or in a separate type of terms.

For example, neologism of social networks is considered the word «hype», which means aggressive and intrusive advertising.

2) Lexical neologisms – words newly formed or borrowed. Recently, words such as «Komsomolets», «designer» and so on belonged to this category. The peculiarity of words, the category of lexical neologisms is only a temporary state in the lexical language. Upon the cessation of time, the word enters the active lexical category of the language, and later becomes historicism or archaism, which happens with many words of the Soviet period.

3) Semantic neologisms – represent a transition in understanding reality, about the most active interpretations, and about evaluation by native speakers.

4) Individually-stylistic, author, neologisms – are formed by writers and poets, to give imagery to the literary text. Neologisms of this type are “attached” to the context, have an author. The individual stylistic neologisms are close in their artistic significance to the tropes: the creation and the other are based on the desire to give a sensual description of the subject, the author does not set himself the goal of introducing the words he invented into general use. Occasionalisms are distinguished among speech innovations in the scientific literature.

Occasionalism is used exclusively in the context of this context, as a lexical means of artistic expression or a language game. The author denies the sign of modernity of the word and includes the sign in its unusualness in the definition of occasionalism. In the field of fine art, a process of experimentation is taking place, the entrance of which creates new manners of writing, paintings, and also works of art.

Changes are taking place in theatrical life, new types of theaters appear, and with them new names. To reinforce new words, any language must create lexical units. More often, their creation is a consequence of a certain perception, the need to indicate the opportunity that has arisen. With the introduction of new words, a struggle of two types arises: the first is development and the second is the preservation of the language. As a result of mastering a new word, most native speakers lose their neologism status. As you know, neologisms appear as a result of semantic derivation, as well as borrowing words from other languages [2].

139

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Neologisms should be used in accordance with literary and linguistic norms. On the one hand, excessive use of foreign words is unacceptable, as this results in pollution of the Russian language.

On the other hand, if you do not know the lexical meaning of neologisms, it can lead to a misunderstanding of some words of modern life.

A person creating a new word strives for his individuality and originality. Then, this word is accepted by society and enshrined in the language system. The word is considered by intermediaries who use it among people. As a rule, they are teachers, reporters, and media workers. The word is fixed in the next print. The periodic stage of socialization, the adoption of the word by the masses of the owners of the language.

Next comes the process of inanation, and then the acquisition of skills for the specific use of a new word, that is, the assimilation of practical knowledge of native speakers.

Conclusion, the formation of new words with the help of elements located in the language determines the most important direction in the development of vocabulary and the nature of language acquisition of various borrowed words; the formation of new words by affixing is constantly happening in modern language. With development collocations of new objects, phenomena and concepts appear in technology, culture, science, industry. Neologisms, which become units of the language, eventually enter dictionaries, show the relevance of the state of vocabulary, and lexicography, formed in recent decades, opens up opportunities for understanding the modern history of Russian vocabulary.

In connection with the introduction of neologisms into speech, the language has a huge opportunity to constantly improve and develop, as well as transmit new concepts that have appeared in the language.

References 1. Котелова Н.З. Первый опыт показания русских неологизмов / Котелова Н.З. – М., 1982. – 456с.

2. Новые слова и значения // Под ред. Е.А. Левашова. СПб. – 2007.

3. Социальная сеть [электронный ресурс]. – Режим доступа: http://otherreferats.allbest.ru/languages/00033754_0.html 4. Социальная сеть (электронный) ресурс]. – Режим доступа: http:// kindlebook.ru/referat/dlya-studenta/neologizmy-v-russkom-iazyke/ Аннотация: Быстрое изменения в обществе, усиление международных связей привело к стремительному развитию языка. Лексический состав языка обрастает новыми словами, старые лексемы выпадают из лексикона русского языка, вместо них появляются новые. Неологизмы – новые слова в языке, которые адаптируются обществом и находят свое закрепление в языке. Появление неологизмов связано с законами развития языка, который постоянно развивается и совершенствуется. Ключевые слова: Неологизм, слово, значение, окказионализм, язык, общество, развитие.

140

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Сведения об авторах: Яхина Яна Эльверовна – студент 2 курса факультета философии и социологии БашГУ, г. Уфа

UDC 004.3`144

THE DEVELOPMENT OF COMPUTER TECHNOLOGIES IN RUSSIA Yasinskaya Ya.O., Sokolova O.V. yasinskaya_01@mail.ru

Abstract: The article consider the history of computer technology development in Russia. The most characteristic features of the first computers (electronic computers)are defined. The indicators of the formed computer competence are emphasiidвыделенеы in the artcle. Those touches upon there the necessity for the individual in our informative society.

Keywords: computer competence, computer, information technologies, data,(Random Access Memory)RAM, development.

Computers are widely used nowedays, but even ten years ago it was rare to see any personal computer. They were used but were so exspensive that not every company could afford personal computers for the work in an office. And what about now? Now almost every house has a personal computer without which it is difficult to imagine our everyday.

Computer application is constantly expanding. It is largely facilitated with the spread of personal computers and especially with the use of microcomputers. Digital computer has gone a long way since 50s of XX century [3]. At the very beginning of the development of computer technologies it was a «magical» machine, expensive and unique, having large vacuum tubes. But today it is a device consisting of millions of tiny semiconductors packed in small plastic boxes.

Computer technologies are used everywhere, they control the operation of cash registers, monitor the operation of car ignition systems, keep records of the family budget or just used as an entertainment complex. These are only a small part of the capabilities of modern computers. The rapid progress of semiconductor microelectronics, which is the basics for computer technology, indicates the current level of both computers and their areas of application. Gradually the study of computer technology is being introduced into school curricula as a compulsory subject. So the children know the desing and computer capabilities from an early age. Computers have been used for many years to maintain educational documentation at schools, and now they are used for studying different subjects that are not related to computer technology [2].

141

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

The studey of the development of computer technologies is necessary to improve computer competence.

A common indicator of the formed computer competence is having the followingi feachers: development - an interest to the problems of information technologies; - a conscious attitude to the use of computer technologies in educational and future professional activities; - an ability to think in collapsed forms and formalized structures; - an ability quickly and effectively to become involeved in any virtual situation; - an ability to simulate objects using computer facilities; - knowledge of the etiquette of electronic dialogue communication.

The object of the article is the development of national electronic computers in Russia. The subject of the article is to define characteristics of the national computer that was developed in the Soviet Union in the XX century.

The purpose of thet article is to analyze the history of the development of computer technologies in Russia and to consider the innovative technologies of the future.

Universal computers of the first and the second generations were desinged in the USSR on the original projects of soviet engineers, which were created in simultaneously with the world ones, but withi their own peculiarities. The maino works were carried out at the Institute of Cybernetics (Institute of machiine energy management) in Kiev. In 1950, Mikhail Lavrentiev and Sergey Lebedev founded the Institute laboratory for the development of computing machines. In this laboratory a lot of computers projects were created (BESM, BESM-2, M-20, BESM-6, Elbrus series computers which are similar to the series of American IBM-360 machines) [3]. Under the diraction of Lebedev the first compquter (MESM) was prodused in USSR in 19481951. It was a small electronic calculating machiine of the first generation [1].

The fundamental principles of compquterdesing were implemented by Lebedev in his first machine: - availability of arithmetic devices, memory, in / out devices and controls devices; - storing and encoding of the program in thememory; - binary number system for encoding numbcers and commands; -automatic execution of calculations based on the stored program; -availability of bothi arithmetic and logical opersations; - hierarchical principle of memory architecture; -the use of numerical methods to implement computations.

BESM-6 was the top of scientific and engineering achievements of S.A. Lebedev [4]. It was created in 1967 and implemented new principles and solutions such as parabllel processing of multiple commands, ultra-fast regijster memory, stratification and dynamic distribution of random access memory (RAM), multiprogram operation mode and an advanced interrupt system. BESM-6 is the secondgeneration of computers [1].

142

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

The characteristic features of a second-generation compquters are: -the combination of input/output operations withi the computations in the central procdessing unit (CPU); - storage expansion of RAM and extefrnal memory; -the use of character-output devices for data inpuvt/output; - «closed» modef for users: the programmer is not allowed to enter the machine room, but he hands over the program made in algoprithmic language (a high-level langhuage) to the operator for further passing on the machine.

Russian science has had hard period of time since 1991. Russian government of that time has takefn a course to destroy Russtian science and its techinologies. The investment of many sceintific projects has been stopped, a lott of computer manufacturing planots have suspended their relations due to the ruination of USSR. After its ruination, many computer manufacturing planots occurred in diffgerent countries and it was impossible to go on manufacturing. Many Russian computer engineers and developers had to work in the specialty they has not trained in having lost the time and skills. The onlyz instance of the computer «Elbrsus-3» developed in the Soviet Union was disassembled in 19945. It is worth to admit that «Elbrsus-3» was twice as fast American machine called Cray Y-IMPI [1].

Many of the creabtors of Soviet computers went abroad. Thus, today the leading desing engineer of Intel microprocessor is Vladeimir Pentkovsky. He was educated in the USSRS and worked at the Institute of mechanics and computing science.

S.Pentkovsky took part in the deveflopment of the mentioned above computers «Elbrus-1» and «Elbrus-2». He led the development of the processor for «Elbrsus-3» – El-90[3]. As a result of the purposeful policy of the government of Russian Federation to destroy Russian science the investment of «Elbrus» project stopped.

Vladimir Pentukovsku had to emigrate to the United States of America and got a job at Intel Corporation. Soon he became the leading computer engineer of Intel Corporation. The Pentium processor was developed under his leadership in 1993 [4].

Pentkovsky used the national knowledge and technologies that he knew himself, but thought a lot in the process. And by 1995, Pentkovsky had created a Pentium Pro processor that was similar in capabilities to the Russian El-90 microprocessor.

Pentkovsky is currently devefloping the next generation of Intefl processors [1].

Now, unfoprtunately, in the market of personal compquters, national computers are compqletely absent. Despite the difficulties, the development of the E2C processor («Elbrus-2000»), which is engaged in the company «Elbrus», continues [1].

We concluded that the Soviet school had a high intellectual potential. The ruination of Soviet Union had a huge impact on the further development of native computer technologies. Unfortunately, after 1991 native scientific project were closed or sometimes even were sold and passed abroad. Many specialists emigrated because there was no investment for the project and there were not jobs for them. Russian computer technologies have developed under the great influence of foreign companie, since the late 90's. Only in the mid-2000s native researchers began to be invested again, when the economy of the Russian Federation stabilized. Computer engineering in 143

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Russia is starting to develop its own products and has the opportunities to entering the international computer technology market.

References 1. Ревич Ю. В. Информационные технологии в СССР. Создатели советской вычислительной техники. – BHV, 2014. – 363 с.

2. Электронный ресурс// Электронный научный журнал «Вестник Омского государственного педагогического университета» Выпуск 2006 URL:http://www.omsk.edu/article/vestnik-omgpu-114.pdf 3. Электронный ресурс// Сайт «Виртуальный компьютерный музей». 1997 – 2020. URL: https://computer-museum.ru/histussr/m1rogach.htm 4. Электронный ресурс// Свободная энциклопедия «Википедия». 2001.

URL: https://ru.wikipedia.org/wiki/Список_советских_компьютерных_систем Аннотация: В статье рассматривается история развития компьютерных технологий в России. Определены наиболее характерные особенности первых ЭВМ (электронно-вычислительных машин). В статье акцентируются показатели сформированной компьютерной компетентности. Тем самым затрагивается вопрос о необходимости индивидуума в нашем информационном обществе. Ключевые слова: компьютерная компетентность, ЭВМ, компьютер, информация, технологии, данные, оперативная память, развитие. Сведения об авторах: Ясинская Яна Олеговна – студент группы ПИ-18в, ДонНТУ Соколова Ольга Викторовна – ассистент кафедры английского языка, ДонНТУ

UDC 004.04 WEB-ORIENTED SYSTEM FOR OPTIMUM SELECTION OF TEMPORARY ACCOMMODATION Yasnitsky M.V., Vasyaeva T.A., Revina N.V. max.yasnitsky@gmail.com

Abstract. The paper deals with the approaches used for the system for optimum accommodation selection. The advantages and disadvantages of the approaches under consideration are emphasized. Attention is paid to the comparative analysis of the approaches applied to solving the problem.

Keywords: Collaborative filtering, recommender systems, strategy, objects.

144

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Recent years have been characterized by rapid and vigorous Internet development. Terabytes of information are accumulated and generated in it daily. It becomes quite difficult for a person to select the information he is interested in by mere viewing it, since the relevant information can be most often lost among large amounts of data. For this reason, tools are developed to help the user search, offering more preferable content for him. Such software tools are called recommender systems.

Recommender systems determine content preference for a particular user based on his interaction with the system or data explicitly specified by the user.

Recommender systems should have the following properties: the system should adapt to a specific user, since preferences can vary significantly among different users; the system should take into account the user's current preferences, adapting to them in due course; the system must constantly find new areas of information and offer them to the user.

Recommender systems are also applicable to the tourism sector, namely the selection of temporary accommodation. Nowadays there is a large number of travel services that deal with an even more impressive number of hotels themselves. Thus fulfilling the task of manual searching and analyzing existing reservation proposals requires more time and resources. Therefore, this task is becoming increasingly important: the user of booking services often wants the system itself to offer services appealing to him.

The paper considers the development of a web-oriented system that will select temporary accommodation in optimum way for each user individually. When working the system will have to adapt to the preferences of each user, take into account the current preferences of the user, coordinating with him in due course as well. The system must constantly find new objects and offer them to the user.

There are several basic approaches for constructing recommendations: Based on collaborative filtering, it uses data on the user’s interaction with the system in the past, for example, a list of evaluated objects or orders made on the Internet system site before by the users of the same interest group.

Based on content filtering, in this case, the system stores profiles that contain individual user information, namely age, social status, occupation, place of residence, as well as characteristics that express the user's interest in the object; interest profiles include parameters the user is interested in.

Recommender systems combining the above mentioned approaches are called hybrid. They combine the strong points of these approaches to develop methods that can work more efficiently in narrow-profile systems.

Construction of recommender systems based on collaborative filtering Collaborative filtering produces recommendations based on a model of user’s prior behavior. This model can be constructed exclusively on the basis of the behavior of the user concerned or – more efficiently – taking into account the behavior of other users with the similar characteristics [1].

Collaborative filtering uses the following types of input data (Fig. 1): a great number of objects of interest and a great number of users. The relationship between the 145

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

objects of interest and users is often expressed by using 10-15 ratings given by users, and then used to predict ratings that the user could give to objects that have not been rated yet. When a user interacts with a collaborative recommender system, the system must first identify the nearest users with the same behavior of the current user, and then extrapolate the rating of the current user from the ratings of similar users.

The main two approaches to collaborative filtering are user-based similarity and item-based similarity. The goal of both approaches is to identify similar objects and group them on the basis of evaluation matrices.

User-based filtering finds k nearest users whose ratings are similar to the current user, and uses their ratings to predict the preferences of the current user. The main advantage of this approach is its high accuracy. However, the disadvantage of this approach is the high entry threshold, without any data on the users’ interests, it is almost impossible to find recommendations [5].

The difference between item-based and user-based filtering is not the use of the behavior of user ratings, but the use of similarities between estimates of object models.

As a rule, two elements having the same users’ ratings are similar, which means that users should have similar preferences for selected objects. The advantage of this approach is the ability to calculate the degree of proximity in deferred mode, since the rating of objects is available before the recommendations are formed [5].

Grades

Current User (Ua)

Recommendations

Data

Collaborative filtering Finding users similar to Ua (closest users)

Users

Objects

U1

O1

Using estimates of closest users to predict recommendations to the current user

U2

O2

Un

On

Display generated recommendations to the current user

Figure 1 – Collaborative filtering Construction of recommender systems based on content filtering Content-based filtering assumes that the user's range of interests is constant within some time. The input data are many categories of tagged items of interest and 146

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

many users (Fig. 2). The task of content-based filtering is to compute the set of items nearest to the categories the current user is interested in (Ua).

The main thing in this filtering is the comparison of objects viewed by the user with new ones, which potentially can be recommended to the user. The basic method for determining similarity is the extraction of keywords from the context contained in the object of interest or from the method that annotated the information object [4].

The main advantages of this approach are the possibility of recommendations to be given to new users, as well as recommendations of objects that were not estimated previously. A higher development speed and low accuracy can be considered as the disadvantage.

Categories

Current user (Ua) RecommenDations

Data

Content Based Filtering Finding objects similar to selected Ua Displaygeneratedre commendationsтек ущему пользователю

Users

Categories

Objects

U1

K1

O1

U2

K2

O2

Un

Kn

On

Figure 2 – Content-Based Filtering Hybrid Recommender Systems Constructing A hybrid approach combines the capabilities of basic approaches, thus higher accuracy can be achieved. When constructing hybrid classifiers, the following strategies are used [3]: - Weighted strategy. The predicted estimate for an object is calculated as the weighted arithmetic mean of the estimates predicted by various algorithms.

- Switching strategy. Before constructing a recommendation, the value of a certain criterion is calculated, on the base of which a decision is made on the choice of an algorithm for constructing recommendations. Such a criterion can be expressed by a comparison of the number of the user actions with a certain threshold selected in advance, starting from which a collaborative filtering model can be applied to the user.

- Mixed strategy. It is based on the idea that forecasts of individual recommendations are displayed in one integrated result.

147

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

- A cascade strategy is an iterative method for constructing recommender systems, the first algorithm playing the role of a coarse filter, and the following ones correct estimates [2].

Comparison of basic approaches Basic recommender systems have been compared by the following parameters: - Adaptability. It implies automatic accounting of a user ratings during further launches of the recommendation algorithm; - Fast deployment. No need for sensible information about the properties of the subject; - Intuitiveness. The possibility to display the relevant object at random even without initiating a search by the user; - Transparency. The ability to justify the result of the recommendation algorithm; - Cold start. It implies the need to provide initial estimates before the algorithm starts; While comparing the basic recommender systems, the following table was compiled: Table 1 – Comparison of the characteristics of the main recommender approaches Approach Adaptability Fast Deployment Intuitiveness Transparency Cold start

Collaborative Filtering Yes Yes Yes No Yes

Content-based filtering Yes Yes No No Yes

Conclusion The article considers the issues of constructing a web-oriented system for the optimum selection of temporary accommodation. The existing approaches to solving the problem under consideration are presented and described. The main advantages and disadvantages of the described approaches to constructing recommender systems are emphasized. A comparative analysis of the main approaches is given.

References 1. Джонс Тим. Рекомендательные системы. [Электронный ресурс] // https://www.ibm.com/developerworks/ru/library/os-recommender1/index.html 2. Международный научно-технический журнал «ТЕОРИЯ. ПРАКТИКА. ИННОВАЦИИ» [Электронный ресурс] // https://docplayer.ru/60855201-Mezhdunarodnyy-nauchno-tehnicheskiy-zhurnalteoriya-praktika-innovacii-iyun-2017-avtomatika-vychislitelnaya-tehnika.html 148

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

3. Burke R. Hybrid web recommender systems // The adaptive web. – 2007. – С. 377-408.

4. Jannach D., Zanker M., Felfernig A., Friedrich G. Recommender Systems – An Introduction. Cambridge University Press, 2010. 360 P. 5. Recommender Systems — User-Based and Item-Based Collaborative Filtering [Электронныйресурс] // https://medium.com/@cfpinela/recommendersystems-user-based-and-item-based-collaborative-filtering-5d5f375a127f Аннотация. В работе описаны подходы применимые к системе подбора оптимального жилья. Выделены достоинства и недостатки подходов. Приведен сравнительный анализ подходов решения поставленной проблемы. Ключевые слова: Коллаборативная фильтрация, рекомендательные системы, стратегия, объекты. Сведения об авторах: Ясницкий Максим Валерьевич – студент группы ИСм-19, ДонНТУ Васяева Татьяна Александровна – к.т.н., доц. кафедры «Автоматизированные системы управления», ДонНТУ Ревина Наталья Владимировна – ст. пр. каф. Английского языка, ДонНТУ

UDC 330.873 ANALYSIS OF INDUSTRIAL ENTERPRISES SOCIAL RESPONSIBILITY SYSTEM IN THE DONETSK PEOPLE'S REPUBLIC Zaglada E.A. ekaterynazaglada@gmail.com

Abstract. The problems of social activity and business responsibility, its participation in the development of the territory in which it operates, have recently attracted increasing attention. Industrial enterprises have a significant impact on the development of the economy of the Donetsk People's Republic. The article considers the main directions for the implementation of the social responsibility of industrial enterprises of the DPR. Keywords: social responsibility, concept, theory, principles, directions of social responsibility, industrial enterprises.

The Donetsk People’s Republic is currently on its developing stage and is facing many acute social, economic and environmental problems. Since business has a significant impact on the development of society and in its hands are concentrated significant financial and material resources, the social responsibility of industrial 149

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

enterprises is becoming increasingly important. Due to the multifaceted concept of «social responsibility», a single definition has not yet been developed, however, most scientific interpretations emphasize good business and voluntary participation in improving society. At the same time, social responsibility receives significant development due to incentives from representatives of civil society and the authorities.

The objective value of social responsibility lies in the fact that it allows industrial enterprises to minimize the negative impact of their production activities on society and helps to create predictability, trust, shared values and social capital. Thanks to social responsibility, companies become an important element in developing the institutional structure of society.

The need for industrial enterprises of the Donetsk People's Republic to obtain public recognition and consolidation in the domestic and foreign markets is an incentive for the introduction of a social responsibility system.

Analysis of research and publications. A great contribution to the development of the concept of social responsibility was made by both native and foreign scientists and economists, including: Blagov Y., Goncharov S., Grigoryan E., Korsakova M., Kostin A., Krichevsky N., Litovchenko S., Petrunya Y., Tulchinsky G., Chernov E., Yurasov I., Bowen H., Drucker P., Carroll A., Matten D., Moon J., Porter M., Stanley R., Friedman M. The purpose of the study is to analyze the social responsibility system at industrial enterprises of the Donetsk People's Republic and the degree of its implementation.

The main results of the study. The first mention of social responsibility dates back to the 50’s of the 20th century. Thus, H. Bowen noted the importance of the influence of social issues on the development of economic indicators of the enterprise.

P. Drucker in his scientific works notes that genuine «social responsibility» consists in turning social problems into economic opportunity and economic benefit, into production capacities, staff competence, well-paid work and, finally, wealth [5].

According to N. Krichevsky and S. Goncharova in the West in the early 70’s of XX century another theory has arisen. This theory states that if a company complies with labor safety conditions and does not violate environmental standards, has a sufficient level of employee wages, then such a company can be called socially responsible [2].

Y. Petrunya gives the following definition of the social responsibility of business – it is the responsibility of a business organization for the impact of its activities on various spheres of public life and the natural environment through transparent and ethical behavior [4].

According to Y. Blagov social responsibility, understood as a rational response of the company to the system of conflicting expectations of stakeholders aimed at the sustainable development of the company, can be interpreted as an element of the competitive strategy, considered as part of the industry concept, and as an independent concept of strategic management [1].

150

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

In the post-Soviet space, social responsibility of business appeared in the 1990’s in the form of support of social institutions by large enterprises. At the beginning of the XXI century, the relevance of social responsibility has increased due to globalization and access to the international market, social responsibility programs have been further developed, and corporate social responsibility has become one of the strategic directions for the development of industrial enterprises and society as a whole.

Awareness of the need to attract industrial enterprises in socially responsible activities in the Donetsk People's Republic appeared relatively recently in connection with the development of the economy of DPR. Previously, social responsibility was mainly implemented by corporate structures.

The economy of the Donetsk People's Republic is at the formation stage. The military situation has a significant impact on the economic situation, as well as the transport blockade introduced in early 2017, which resulted in the breaking of traditional socio-economic ties.

The main economic sector of the Donetsk People's Republic is industry (metallurgy, coal mining, mining of raw materials for the metallurgical industry, coke chemistry, mechanical engineering). The main industrial centers are Donetsk, Makeevka, Khartsyzsk, Yasinovataya, Enakievo, Gorlovka, Kirovskoye, Zhdanovka, Torez, Snezhnoye, Shakhtyorsk, as well as the urban village Novyi Svet of Starobeshevo district, where the largest power station of the Starobeshevo district is located.

As a result of the infrastructural destruction caused by the war, the termination of production ties with the territories controlled by Ukraine, the conservation of coal enterprises, which previously formed the basis of the region’s industry, the DPR economy was in decline.

At present, the economy of the DPR is developing, faced with certain difficulties.

New enterprises appear on the territory of the DPR (light industry, agriculture, telecommunications and postal activities), some of the enterprises that previously suspended their work re-launch (food industry, mechanical engineering), but there are also such enterprises that are not able to function in modern conditions or forced to work using only part-time capacities.

Despite the short period of existence of the Donetsk People's Republic, we can confidently say that the social responsibility of industrial enterprises of the DPR has its own developmental features, which were formed due to the history and geographical location of the DPR, the mentality of the population, traditions of governance, the social and political situation in the Republic.

The current state of the DPR is characterized by a general weakening of state influence on the socio-economic activity of enterprises of the republic caused by objective reasons.

The main problems that can be considered obstacles for implementation of social responsibility principles are represented in Table 1.

151

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Table 1 – Obstacles for implementation of social responsibility principles Problem Low living standards and poverty Income polarization Excessive population migration Undermining perfect competition Unemployment

Alcoholism

Others

Explanation The main problem at the moment is the discrepancy between the incomes of citizens and their expenses. High prices for basic consumer goods force most of the population to live from paycheck to paycheck.

Income inequality is one of the acute problems of the DPR. While the majority of the population receives wages close to the cost of living, employees of the administrative apparatus receive wages comparable in level to the salaries of the Russian Federation.

Previously, it was mainly associated with military operations in certain territories of the DPR; with the loss of jobs and the need to find a new job; at the moment – mainly in connection with the issuance of Russian passports.

Due to the non-recognition of this territory, goods manufactured in the territory of the DPR cannot compete in the world market, despite the high quality of the produced goods.

Mainly associated with the closure of enterprises that wished to remain under Ukrainian jurisdiction; with the closure of enterprises whose production facilities suffered during the fighting.

More often observed among people affected by hostilities, people with disabilities; people who have lost their jobs and are unable to find a job again. The situation is aggravated by low prices for alcohol.

Problems that impede the efficient operation of enterprises on the territory of the DPR include: a decrease in the level of social protection of the population; monopolization of economic activity; danger to life and health for people living in territories that are bombarded; social tension.

The presence of existing problems, which at the moment is not possible to solve exclusively by state forces, leads to the realization of the need to share the burdens of these problems with large business. As a result, there is an urgent need to develop and implement a system of social responsibility at industrial enterprises of the DPR, since they have significant financial and material resources concentrated in their hands, which will positively influence the development of society.

With the beginning of transformations in the state and economic structure of the Donetsk People's Republic, it can be concluded that the social responsibility of industrial enterprises is realized mainly in the following areas: 1) production of quality products and services; 152

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

2) job creation, stable salary payments, investments in human development (advanced training courses for employees, retraining of personnel, etc.); 3) safety and labor protection at the enterprise; 4) compliance with legislation (tax, environmental, labor, etc.).

Nevertheless, social responsibility also involves work in the following areas: 1) taking into account public expectations and ethical standards in the work of the enterprise; 2) contribution to the formation of civil society through local community development programs and projects; 3) honest and conscientious business ethics; 4) implementation of measures to prevent environmental pollution; 5) use of energy saving in production; 6) the implementation of staff training in order to increase their awareness and understanding of personal responsibility; 7) the willingness of enterprises to participate in crisis situations.

The implementation of the principles of social responsibility in practice can bring the enterprise the following positive result in the long term: 1) creating a positive image; 2) improving business reputation; 3) creating a stable business environment; 4) growth of investment attractiveness; 5) growth of the company's capitalization in the long term; 6) creating predictability and trust on the part of society; 7) creating mutually beneficial ties with the state.

Now, we can talk about the impossibility to cover fully all areas of social responsibility, since some enterprises are at the level of the struggle for survival.

Nevertheless, there are certain backgrounds for the development of a social responsibility system at industrial enterprises of the DPR. Conclusions. Therefore, the main prerequisites for the development of the social responsibility system in the DPR are the search for effective subjects of ownership, the orientation of industrial enterprises towards the social needs of society, and interaction with the state and society in order to build a strong civil society. This will ensure the implementation in the Republic of the basic principles of social responsibility, ensuring social freedom of a person, the principle of responsibility of all members of the society for the situation in society, for the development of market rules and their implementation, for the creation of appropriate conditions for optimal economic and social life.

References 1. Благов Ю. Е., Журавкова М. В. 2003. Торговая фирма «Шик»: социальная ответственность в малом и среднем бизнесе. Сборник учебных кейсов: Опыт российских компаний (Северо-Западный регион) Под ред. С. П. Куща. СПб.: Издат. Дом С.Петербургского гос. Ун-та. С. 87-112 153

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

2. Инвестиционный портал Министерства экономического развития Донецкой Народной Республики. Режим электронного доступа: https://invest.govdnr.ru/#Catalog 3. Кричевский Н.А., Гончаров С.Ф. Корпоративная социальная ответственность // Москва: Дашков и Ко, 2006. – С. 7.

4. Петруня Ю.Е. Социальная ответственность бизнеса: украинские измерения/ Ю.Е. Петруня // Экономический вестник Национального горного университета. 2012. № 1. С. 7.

5. Drucker Р. F. The new meaning of corporate social responsibility // California Management Review. 1984. 26 (2): P. 53-63. Аннотация. Проблемы социальной активности и ответственности бизнеса, его участие в развитии территории, на которой он функционирует, в последнее время привлекают все большее внимание. Промышленные предприятия оказывают существенное влияние на развитие экономики Донецкой Народной Республики. В статье рассмотрены основные направления реализации социальной ответственности промышленных предприятий ДНР. Ключевые слова: социальная ответственность, концепция, теория, принципы, направления социальной ответственности, промышленные предприятия. Сведения об авторах: Заглада Екатерина Александровна – аспирант, АДИ ДонНТУ

UDC 004.942 (004.912) CLASSIFICATION OF TEXTS Zemlianskiy D.A., Kolomoytseva I.A., Gilmanova R.R. Velheor19375@mail.ru

Abstract. This article considers the process of text classification, the relevance of the task, the steps and algorithms that can be used to implement the classification.

The article presents the main stages of classification and their description. It gives some examples and description of indexing methods, and provides methods to determine the weight of features. Methods of text classification and training the classifier are described.

Key words: classification, text, indexing, machine learning.

Text classification relates to the tasks of computational linguistics. It includes such tasks as identifying the subject matter of the text, authorship of the text, as well 154

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

as the emotional colouring of the text. There are a number of algorithms and methods for text classification.

Most modern approaches to the problem considered relate to machine learning methods, in particular with regard to supervised learning. The method involves training the system on the basis of examples where the resulting class is already known.

Texts classification allows to make quick analysis of large amount of text information, which can be used in various systems in the future. These systems include web search engines that use text classification to select the most relevant search results, to identify authorship of texts, and to determine illegal information on the Internet.

Stages of constructing classification process During the process of text classification, the initial data pass through the steps shown in Figure 1. The right side of Figure 1 shows the possible algorithms for some of them.

Preprocessing includes separation of the text into components, removal of words that do not contain information about the class (commonly used) and morphological analysis, which reduces the feature space.

Text indexing means transforming a text into the form convenient for further processing.

The choice of features reduces the dimensional feature space in order to accelerate the classification process and permits to avoid the effect of retraining.

Preprocessing

Indexing

Bag-of-words Word2vec Doc2Vec

Feature Selection

TF-IDF LSA PMI CRF IG

Construction and training the classifier

NB KNN DT SVM ANN

Quality control

Figure 1 – Stages of text classification [3] 155

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Indexing Methods Indexing methods include the following methods: - Bag-of-words; - Word2vec; - Doc2Vec. Bag-of-words is a model that is trained on the dictionary composed of the words of all documents [4].

The algorithm of model building: Step 1. A dictionary is compiled from all the words that appear in the text, punctuation marks, numbers and «stop words» (interjections, introductory words, definitions, adverbs, pronouns) having been previously omitted from the text.

Step 2. For each document, a vector is defined, each component of which corresponds to the term from the dictionary, and its value is determined by the number of times this word occurs in the text. The dimension of the vector corresponds to the word dictionary size.

The vector representation of a word in Word2vec contains a list of neighbouring words and their frequency as coordinates.

Word2vec means machine learning. Initially, the dimension of vectors is set with random variables. During training, the component values of the vectors will change, while the vector of each word will resemble the vector of neighbouring words as much as possible and differ from the vectors of words that are not neighbours of the word.

The algorithm of building the model: Step 1. The dictionary of terms or words that are found in all texts is compiled; Step 2. The frequency of word occurrence in the documents is set for each one; Step 3. A Huffman tree is built to encode the dictionary; Step 4. The importance of frequently occurring words is reduced; Step 5. One of the algorithms CBOW (Continuous Bag-of-Words) or Skip-gram is used for the words; Step 6. A neural network of direct distribution with hierarchical softmax activation function or negative sampling is applied.

Algorithm Doc2Vec relates to classification algorithm of unsupervised learning with the task to learn how to allocate distributed vectors for parts of the text, the texts being of different lengths. This model involves learning vector representations of text to predict words in a document. Combining the text vector with several words from it, the model tries to predict the next word, he text vectors being unique and the vectors of the same words in different texts coinciding.

Methods for determining the weight of text attributes Methods for determining the weight of document features include: - TF-IDF; - LSA; - PMI; - CRF; - IG. 156

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

TF-IDF is a statistical measure that is used to estimate the weight of a word in a text, which in turn is a part of the collection [6]. The weight of a word in a document is directly proportional to the word frequency, i.e. how frequently the word is used in the current text, and is inversely proportional to the number of its uses in the entire set of documents.

The formula TF is shown in Figure 2. In this formula, nt is the number of occurrences of the word t in the text, and the denominator shows the total number of words in the text.

Figure 2 – Formula of the TF component in the TF-IDF measure The IDF formula is shown in Figure 3. In the formula, |D| – is the number of texts in the collection, and |{ di D | t  di }| – the number of texts in the collection where the word t occurs.

Figure 3 – Formula of the IDF component in the TF-IDF measure The TF-IDF measure itself is a multiplication of the TF and IDF components.

TF-IDF is used for text analysis and information retrieval, as one of the criteria for document relevance to a search query, or for calculating the document proximity for a clustering task.

Latent-semantic analysis is a method that involves analysing the relationships between the collection of documents and the terms in the documents and identifying the characteristic factors (topics) that are inherent in all documents and terms [1]. This method is used in the classification and clustering documents to extract contextsensitive values of lexical units using the statistical processing of large collections of texts.

Conditional random fields are a class of statistical modelling methods applied in pattern recognition in machine learning and used for structured prediction [5]. In this class of methods, prediction is modelled as a graphical model that implements the dependencies between the predictions.

Methods of constructing and training the classifier Classification methods include: - probabilistic methods, an example of which is the Bayesian method; - metric methods, an example of which is the method of nearest neighbours; - logical methods, an example of which is the decision tree method; - linear methods, an example of which is a method of support vectors;

157

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

- methods based on artificial neural networks, an example of which is feedforward neural networks, recurrent neural networks, dynamic neural networks and convolutional neural networks.

The Bayesian method is based on calculating the conditional probability of the event occurrence such as the diagnosis Di with the appearance of specific implementation of the complex of signs K. KNN (K nearest neighbours) is one of the widely used classification methods.

The main principle of the K nearest neighbours method is that the object belongs to the class to which most of its nearest neighbours belong. In other words, for an object, let us consider its K-nearest neighbours, then for each class, we consider the quantity of these neighbours belonging to the class, the object belonging to the class with the largest number of neighbours [2].

The decision tree method consists in dividing the source data into groups until homogeneous sets are obtained. The resulting set of rules allows us to predict new data in the future.

Conclusions In this paper, classical methods and approaches for classifying texts have been analysed, as well as the sequence of procedures for classifying texts. The main algorithms and methods of indexing, determining the weight of text attributes, methods for constructing and classifying texts have been considered.

References 1. Латентно-семантический анализ // Wikipedia.

URL: https://ru.wikipedia.org/wiki/Латентно-семантический_анализ (дата обращения: 29.02.2020) 2. Ле Мань Ха / Оптимизация алгоритма KNN для классификации текстов – М., МФТИ, 2016. – 92 с.

3. Методы автоматической классификации текстов // ResearchGate. URL: https://www.researchgate.net/publication/315328102_Metody_avtomaticeskoj_klassi fikacii_tekstov(дата обращения: 23.02.2020) 4. Мишенин А.Н. /Анализ тональности текстов с использованием нейросетевых моделей– Спб, СПГУ, 2016. – 67 с.

5. Conditional random field // Wikipedi.

URL: https://en.wikipedia.org/wiki/Conditional_random_field (дата обращения: 01.03.2020) 6. TF-IDF // Wikipedia. URL: https://ru.wikipedia.org/wiki/TF-IDF (дата обращения: 24.02.2020) Аннотация. В данной статье рассматривается процесс классификации текстов, актуальность данной задачи, этапы и алгоритмы, которые можно использовать для реализации. В статье представлены основные этапы классификации и их описание. В ней приведены некоторые примеры и описание 158

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

методов индексации, а также методы для определения веса признаков. Описаны методы классификации текста и обучения классификатора. Ключевые слова: классификация, текст, индексация, машинное обучение. Сведения об авторах: Землянский Дмитрий Андреевич – студент группы ПИм-19, ДонНТУ Коломойцева Ирина Александровна – ст. преподаватель кафедры «Программная инжененрия», ДонНТУ Гильманова Роза Разимовна – ст. преподаватель английского языка, ДонНТУ

159

DONETSK NATIONAL TECHNICAL UNIVERSITY YOUNG SCIENTISTS’ RESEARCHES AND ACHIEVEMENTS IN SCIENCE (APRILE 16, 2020)

Научное издание

МАТЕРИАЛЫ научно-технической конференции для молодых ученых «Young scientists’ researches and achievements in science» 16 апреля 2020 г. г. Донецк Под редакцией Кушниренко Е.Н. Адрес редакции: ДНР, 83001, г. Донецк, ул. Артема, 131, Донецкий национальный технический университет, 11-й учебный корпус, факультет компьютерных наук и технологий, кафедра английского языка, ком. 248, тел.: (062) 301-03-74 e-mail: kaf_engl-2017@mail.ru

Web-сайт кафедры английского языка: http://www.ka.fknt.donntu.org

160