Научная статья на тему 'ЗАДАЧИ УПРАВЛЕНИЯ ТЕХНОЛОГИЧЕСКИМИ ПРОЦЕССАМИ'

ЗАДАЧИ УПРАВЛЕНИЯ ТЕХНОЛОГИЧЕСКИМИ ПРОЦЕССАМИ Текст научной статьи по специальности «Электротехника, электронная техника, информационные технологии»

CC BY
178
21
i Надоели баннеры? Вы всегда можете отключить рекламу.
Ключевые слова
ТЕХНОЛОГИЧЕСКИЙ ПРОЦЕСС / НАДЕЖНОСТЬ / ЦИФРОВИЗАЦИЯ ПРОИЗВОДСТВА / СИСТЕМЫ УПРАВЛЕНИЯ / ПЕЧАТНЫЕ ПЛАТЫ / ТРАВЛЕНИЕ ПЕЧАТНЫХ ПЛАТ / ПРОИЗВОДСТВО / ИНДУСТРИЯ 4.0

Аннотация научной статьи по электротехнике, электронной технике, информационным технологиям, автор научной работы — Ванцов С. В., Васильев Ф. В., Хомутская Ольга Владиславовна, Коробков М. А.

Получение качественного конечного продукта является целью управления любым технологическим процессом. При формальном описании задач управления это положение позволяет сформировать краевые условия решения задачи управления. Дополнительным условием выбора варианта решения является обеспечение надежности технологического процесса. Начальными условиями при решении задачи управления являются компоненты вектора начальных параметров состояния, определяемого степенью настройки исходных параметров и режимов процесса. Понимание начальных и граничных условий существования позволяет сформулировать обобщенную формальную задачу управления технологическим процессом, заключающуюся в нахождении вектора управляющих переменных, обеспечивающих минимум ошибки с учетом влияния вектора случайных внешних переменных. Решение задачи осуществляется методом последовательных приближений. Рассмотрены этапы первичной и вторичной оптимизации построения алгоритмов управления. Показана возможность снижения сложности задач управления путем отказа управления по каждому из компонентов вектора состояния и сведения управления к обеспечению минимума обобщенной меры ошибки. В такой постановке задача управления технологическим процессом может быть сведена к задаче регулирования процесса. Приводится схема регулирования технологического процесса, учитывающая разделимость минимизации ошибки по каждому компоненту вектора состояния. Высказанные положения иллюстрируются на примере задачи управления участком цифрового производства для выполнения технологической операции травления печатных плат. Эта операция наиболее полно отражает исходные положения о непрерывности процедур, производимых при ее осуществлении, и деградации исходного вектора состояния.

i Надоели баннеры? Вы всегда можете отключить рекламу.
iНе можете найти то, что вам нужно? Попробуйте сервис подбора литературы.
i Надоели баннеры? Вы всегда можете отключить рекламу.

MANUFACTURING PROCESS CONTROL TASKS

Obtaining a quality end product is the goal of any manufacturing process control. In the formal description of control problems, this goal allows us to form the boundary conditions for solving the control task. An additional condition for choosing a solution is to ensure the reliability of the process. The starting conditions for solving the control problem are the components of the vector of initial state parameters, determined by the degree of adjustment of the initial parameters and process modes. Understanding the initial and boundary conditions of existence allows us to formulate a generalized formal problem of process control, which consists in finding a vector of control variables that provide a minimum of error, taking into account the influence of a vector of random external variables. The problem is solved by the method of successive approximations. The stages of primary and secondary optimization in the construction of control algorithms are considered. The possibility of reducing the complexity of control tasks is shown by refusing to control individual components of the state vector and transforming control to ensure a minimum of the generalized error measure. In this formulation, the problem of process control can be reduced to the problem of process regulation. A process control scheme is presented. It considers the separability of error minimization for each component of the state vector. The assertions are illustrated by the example of the problem of managing a digital production site for performing the technological operation of etching printed circuit boards. This operation most fully reflects the initial provisions on the continuity of the procedures performed during its implementation and the degradation of the original state vector.

Текст научной работы на тему «ЗАДАЧИ УПРАВЛЕНИЯ ТЕХНОЛОГИЧЕСКИМИ ПРОЦЕССАМИ»

ISSN 0868-5886

НАУЧНОЕ ПРИБОРОСТРОЕНИЕ, 2022, том 32, № 4, c. 124-137

МАТЕМАТИЧЕСКИЕ МЕТОДЫ И МОДЕЛИРОВАНИЕ -В ПРИБОРОСТРОЕНИИ

УДК 658.512.26

© С. В. Ванцов, Ф. В. Васильев, О. В. Хомутская, М. А. Коробков, 2022

ЗАДАЧИ УПРАВЛЕНИЯ ТЕХНОЛОГИЧЕСКИМИ ПРОЦЕССАМИ

Получение качественного конечного продукта является целью управления любым технологическим процессом. При формальном описании задач управления это положение позволяет сформировать краевые условия решения задачи управления. Дополнительным условием выбора варианта решения является обеспечение надежности технологического процесса. Начальными условиями при решении задачи управления являются компоненты вектора начальных параметров состояния, определяемого степенью настройки исходных параметров и режимов процесса. Понимание начальных и граничных условий существования позволяет сформулировать обобщенную формальную задачу управления технологическим процессом, заключающуюся в нахождении вектора управляющих переменных, обеспечивающих минимум ошибки с учетом влияния вектора случайных внешних переменных. Решение задачи осуществляется методом последовательных приближений. Рассмотрены этапы первичной и вторичной оптимизации построения алгоритмов управления. Показана возможность снижения сложности задач управления путем отказа управления по каждому из компонентов вектора состояния и сведения управления к обеспечению минимума обобщенной меры ошибки. В такой постановке задача управления технологическим процессом может быть сведена к задаче регулирования процесса. Приводится схема регулирования технологического процесса, учитывающая разделимость минимизации ошибки по каждому компоненту вектора состояния. Высказанные положения иллюстрируются на примере задачи управления участком цифрового производства для выполнения технологической операции травления печатных плат. Эта операция наиболее полно отражает исходные положения о непрерывности процедур, производимых при ее осуществлении, и деградации исходного вектора состояния.

Кл. сл.: технологический процесс, надежность, цифровизация производства, системы управления, печатные платы, травление печатных плат, производство, индустрия 4.0

ВВЕДЕНИЕ

С физической точки зрения основной задачей, решаемой при управлении технологическими процессами, является получение качественного конечного продукта. Решение этой задачи является многовариантным, зачастую зависящим от состава оборудования на конкретном предприятии, возможностей технологического оборудования и уровня подготовки исполнителей.

Определим, что под качеством конечного продукта понимается его состояние, при котором он полностью удовлетворяет требованиям пользователя этого продукта. Для технических объектов состав эти требований сформулирован в технических условиях (ТУ) на изделие. Отметим, что определение понятия качества не требует от изделия расширения требований более тех, что зафиксированы в ТУ. Тогда при формальном описании задачи управления технологическими процессами состав этих требований позволяет сформировать краевые условия решения.

Дополнительным условием выбора варианта решения этой задачи из допустимого множества является обеспечение надежности технологиче-

ского процесса. Согласно стандарту ГОСТ 27.0022009 [1], надежность понимается как "свойство готовности и влияющие на него свойства безотказности и ремонтопригодности, и поддержка технического обслуживания". При этом под готовностью понимается "способность изделия выполнить требуемую функцию при данных условиях в предположении, что необходимые внешние ресурсы обеспечены", а под безотказностью понимается "способность изделия выполнить требуемую функцию в заданном интервале времени при данных условиях".

Однако представленные в стандарте определения ориентированы на изделие как конечный продукт, свойства функций которого не изменяются по времени, по крайней мере до момента, определенного, например, назначенным ресурсом [2, 3]. Особенностью технологического процесса как сложной технической системы является наличие в составе вектора наблюдаемых, но не управляемых параметров системы функции времени [4]. Практика осуществления технологических процессов показывает, что неотъемлемым его атрибутом является деградация по времени, т.е. явление ухудшения параметров и режимов их протекания.

Пожалуй, первое и наиболее полное определение надежности технологического процесса дано А.С. Прониковым [5]: "Надежность технологического процесса — свойство обеспечивать изготовление продукции в заданном объеме, сохраняя во времени установленные требования к ее качеству". Надежность технологического процесса связана со структурой процесса, применяемыми методами контроля и надежностью выполнения отдельных технологических операций.

Ключевым моментом для определения количественных характеристик надежности является понятие отказа технологической операции. Под отказом технологического процесса или технологической операции будем понимать потерю способности обеспечивать качество или объемы изготавливаемой продукции в соответствии с требованиями технической документации. Отказ — это случайное событие, возникающее под влиянием ряда случайных факторов.

ОСНОВНАЯ ЧАСТЬ Методы

Параметры качества конечного объекта формируют граничные условия при решении задачи управления технологическим процессом, а требования по надежности процесса накладывают дополнительные ограничения на существование граничных условий.

Начальным условием при решении задачи управления технологическими процессами является вектор параметров состояния x(t0), определяемый степенью настройки исходных параметров и режимов процесса. Понимание начальных и граничных условий существования позволяет сформулировать обобщенную формальную задачу управления технологическим процессом [6].

С учетом ранее рассмотренного атрибута деградации технологического процесса, определяющего вектор текущего состояния процесса x(t), наличия вектора случайных внешних переменных f(t), а также при условии обеспечения вектором

начальных параметров состояния x(t0) идеальных параметров процесса задача управления может быть сведена к минимизации вектора разности исходного и текущего состояний процесса, т.е. компенсации влияния деградации на качество конечного объекта:

g(t) = {x(to) - x(t min. (1)

Целью решения этой задачи является нахождение вектора управляющих переменных r(t), обеспечивающих минимум ошибки s(t), с учетом вли-

яния вектора случайных внешних переменных

f со.

При решении этой экстремальной задачи ввиду ее очевидной сложности следует использовать метод последовательных приближений, содержащий:

- этап идеальной, или первичной, оптимизации, состоящий в нахождении идеального, но, как правило, нереализуемого на практике вектора

управления г„д (0;

- этап вторичной оптимизации, или оптимизации качества управления, состоящий в нахождении оптимального и в то же время реализуемого на практике имеющимися техническими средствами вектора управления гопт (£).

Этап первичной оптимизации заключается в том, что экстремальная задача (1) решается в идеализированной постановке, учитывающей лишь важнейшие свойства системы управления технологическим процессом, основные его параметры и режимы.

Поясним решение задачи первичной оптимизации процесса управления технологическим процессом на примере работы гибкого производственного модуля (ГПМ) механической обработки. В качестве целевой функции примем выполнение производственного задания в минимальный срок [7]. Вектор исходного состояния х(£0) характеризуется идеальной настройкой всех рабочих органов ГПМ и обеспечением всех запасов (заготовок, инструмента, смазочно-охлаждающих жидкостей и т.п.), необходимых для выполнения производственного задания. Вектор управления гид (¿) сформирован на основании общих представлений об алгоритме работы ГПМ и предыдущего производственного опыта, но не учитывает вектор случайных внешних переменных к которым относятся, например, сбои в выполнении кинематических и динамических процессов функционирования рабочих органов ГПМ, а также явления износа обрабатывающего инструмента, ведущие к увеличению погрешностей характеристик конечного продукта технологического процесса [8, 9]. Таким образом, на этапе первичной оптимизации предполагаем вектор внешних возмущений f (£) = 0.

Зависимость вектора идеального управления гид (¿) от вектора состояния

Гид Ц) = Rид {*('), '}

назовем идеальным алгоритмом управления.

Идеальный алгоритм управления, как отмечалось ранее, можно рассматривать как первое приближение к решению проблемы управления, причем он практически нереализуем. Однако его построение является базовым для дальнейших действий

и позволяет упростить нахождение решения, сведя задачу управления к задаче регулирования, т.е. удержания вектора управляющих переменных г(7) в некоторых допустимых пределах относительно вектора гид (7) . Пределы допустимости отклонений вектора г(7) определяются пределами величины деградации параметров и режимов технологического процесса, а также требованиями к состоянию конечного продукта.

Этап вторичной оптимизации заключается в том, чтобы получить решение, наименее отличающееся от идеального и в то же время учитывающее максимально полно ограничения и связи, налагаемые свойствами системы управления, т.е. в том, чтобы получить не только оптимальное, но и реализуемое решение с учетом ненулевого вектора внешних возмущений f (7).

При этом следует учесть особенность управления технологическими процессами, которая заключается в том, что в векторе f (7) наличествуют в основном два типа переменных, а именно:

- Переменные, представляющие собой характеристики случайных событий, не зависящих от времени. Примером таких событий являются сбои в работе рабочих органов технологического оборудования и сбои в выполнении функций вектора г(7).

- Переменные, случайные значения которых существенно зависят от времени и образуют временные ряды. Примером таких переменных являются характеристики, связанные, например, с износом обрабатывающего инструмента или старением химических растворов, применяемых в ходе технологического процесса. Общая тенденция изменения этих характеристик по времени, как правило, известна из практического опыта предыдущих реализаций технологического процесса, однако остается существенно случайной в каждый момент действия вектора г (7).

В целом решение этих оптимизационных задач в общем виде представляется достаточно сложным в силу многокомпонентности векторов, участвующих в процессе. Однако существует возможность сокращения множественности компонентов, если воспользоваться критерием ошибки (1), т.е. осуществлять управление не по каждому компоненту индивидуально, а по обобщенному компоненту ошибки, который позволит отслеживать в ходе управления не сами величины компонентов вектора состояния х(7), а их отклонения от идеального, определенного вектором хид (7). В такой постановке задача управления технологическим процессом может быть сведена к задаче регулирования процесса.

Условимся называть проблемой регулирования процесса частный случай проблемы управления,

к которому последняя сводится при следующих упрощающих предположениях:

- идеальный вектор состояния хид (7) предполагается известным, совпадающим с вектором

х(70) и отвечающим идеальному алгоритму управления ^(7);

- при решении задачи управления решается задача отыскания экстремума обобщенного показателя эффективности управления

с=ср},

где Q = Q{Euд -Е}, Еид — экстремум цели управления, соответствующий идеальному течению технологического процесса; Е — значение показателя цели управления, соответствующее реальному состоянию процесса; Р — основная характеристика условия реализуемости управляющей системы, в качестве которой может выступать ее стоимость, сложность как в создании, так и в эксплуатации, степень надежности и т.п. В целом характеристика Р может служить обобщенной мерой совокупности всех этих факторов.

Эта задача для целей регулирования распадается на п раздельных или самостоятельных задач отыскания экстремума п показателей эффективности

е, = Ф {хжт(7) -х(7)} = ф {а,}, I = \п, (2)

каждый из которых зависит только от одной составляющей 8, (7) вектора ошибки г(7).

Вектор управления г(7) в явной форме зависит не от вектора состояния х(7), а от вектора ошибки г(7).

С учетом высказанных предположений проблема регулирования технологического процесса может быть сформулирована следующим образом: предполагая вектор идеального управления гид (7) известным, найти вектор управления г(7) как функцию вектора ошибки г(7), а в общем случае и вектора возмущения f (7), который обеспечивает экстремум п частных показателей эффективности (2) и удовлетворяет имеющимся связям, образованным математической моделью управляемой системы и ее ограничениями. В этом случае зависимость можно называть алгоритмом, или законом, регулирования, а вектор г(7) — вектором регулирования:

г(0 = R {е(7 ), 7}. (3)

Таким образом, глобальная задача управления технологическим процессом сводится к задаче регулирования отдельных факторов, определяющих состояние процесса и удовлетворяющих условиям (1).

Рис. 1. Схема регулирования по минимуму ошибки

Однако следует иметь в виду, что эти факторы не являются полностью независимыми друг от друга. Зависимость между факторами обладает рядом особенностей:

- Зависимость не должна являться функциональной. В противном случае такие факторы для целей управления процессом должны быть объединены.

- Возможно существование корреляционной связи между факторами, что приводит к возникновению случайных связей между составляющими вектора г(?) . Факт возникновения такой связи определяется принципиально случайным характером составляющих вектора f (1).

Следует отметить, что вектор состояния х(?) может содержать не только параметры, но и признаки, которые косвенно отражают качество технологического процесса. В качестве таких признаков могут служить в том числе и параметры изготавливаемых изделий. Для оценки влияния признаков необходимо внедрение в контур управления технологическим процессом дополнительной интеллектуальной системы обработки, которая позволит установить зависимость (4) между признаками и подлежащими контролю параметрами:

хир (1) = X {х(0,(4)

Для формирования таких систем целесообразно использовать методы машинного обучения и искусственного интеллекта [10, 11], а также эмпирические и имитационные модели процессов [12, 13].

Схема регулирования технологического процесса, учитывающая разделимость минимизации ошибки по каждой составляющей вектора, представлена на рис. 1.

Механизм разделимости ошибки наиболее эффективно работает именно для технологических процессов, в которых целый ряд составляющих вектора Г(1) оказывает влияние в течение не всего процесса, а только его частей — технологических операций.

Эксперименты, новые закономерности

Рассмотрим высказанные выше положения на примере задачи управления участком цифрового производства для выполнения технологической операции травления печатных плат (ПП). Эта операция, как и большинство химических операций [14], наиболее полно отражает исходные положения об "аналоговости" процессов, производимых при ее осуществлении, и деградации исходного вектора состояния х(10).

Основным параметром, влияющим на результат технологической операции, является время травления, которое в данном случае определяет вектор регулирования г(1). Однако в настоящее время отсутствуют строгие аналитические зависимости, связывающие время выполнения операции и состояние травильного раствора, особенно с учетом его постоянной и зачастую случайной деградации. На основе анализа операции [15] выделены основные параметры, влияющие на время травления и время использования подготовленного травильного раствора с исходным состоянием х(?0) (рис. 2).

Всего можно выделить три категории параметров:

1) параметры травильного раствора, которые в большей степени определяются его составом и объемом;

2) технологические параметры, такие как: расположение заготовки в ванне травления, температура раствора, температура окружающей среды,

Рис. 2. Параметры, определяющие время травления заготовки и время использования подготовленного травильного раствора

способ переноса раствора на заготовку, перемешивание раствора;

3) параметры заготовки, а именно: количество стравливаемой меди, т.е. толщина слоев медной фольги, площадь стравливаемой меди и класс точности изготавливаемой 1111.

Оценить прямое влияние некоторых параметров, например времени перемешивания, на качество процесса затруднительно, или же оценка потребует большого количества трудоемких ручных операций, что относится к исследованию состава раствора. Поэтому вектор состояния х(?) для управления технологической операцией травления содержит в себе как параметры, напрямую влияющие на качество процесса, так и признаки, которые косвенно отражают качество операции.

В качестве контролируемых параметров выбраны:

- температура раствора;

- объем раствора;

- толщина медной фольги на заготовке;

- площадь стравливаемой меди;

- класс точности изготавливаемой 1111.

Также в контролируемый вектор состояния х(1) включены следующие признаки:

- время жизни раствора, т.е. время, прошедшее с момента приготовления раствора до текущего момента времени;

- цвет травильного раствора;

- кислотность травильного раствора;

- содержание стравленной меди в растворе.

Стоит отметить, что такие параметры, как режим перемешивания, ориентация заготовки и метод переноса раствора, не включены в вектор состояния х(?), т.к. они в большей степени обусловлены конструкцией установки травления и не имеют возможности изменения. В частности, эксперименты проводились на лабораторной установке погружного травления с вертикальным расположением заготовки без перемешивания, а в качестве травителя использован раствор, который состоит из 1.5 л дистиллированной воды, 150 мл персульфата аммония, 35 мл поваренной соли.

«

я и

Рис. 3. Зависимость времени травления Т заготовки от количества N протравленных заготовок.

Сплошной линией показаны результаты эксперимента; пунктирной — эмпирическая зависимость Т(Щ = 1.65 N - 16.08 N + 70.33

4 5 в 7 8 9 10 11 12 13 14

Количество протравленных заготовок N

Проведенные эксперименты с последовательным травлением 15 однотипных заготовок двустороннего фольгированного стеклотекстолита FR4 с размерами 80 х 80 мм с толщиной меди 35 мкм

позволили построить зависимость времени Т травления заготовки от общего количества N протравленных заготовок, по которой можно оценить степень деградации травильного раствора (рис. 3).

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Количество протравленных заготовок N

С помощью метода наименьших квадратов, ограничившись уравнением второго порядка, получена эмпирическая зависимость (5) с коэффициентом детерминации 0.97:

Т^) = 1.65^-16.08N + 70.33. (5)

Зная емкость по меди раствора [16], определим предельное содержание меди Меи [г] в растворе по формуле (6):

Меи {V, ) = V = 58.98, (6)

где: у — емкость по меди раствора, 35 г/л; V, — объем раствора, 1.88 л.

iНе можете найти то, что вам нужно? Попробуйте сервис подбора литературы.

Поскольку в ходе эксперимента со всех заготовок, кроме последней, произошло полное удаление меди, то количество меди в растворе тСи [г ] в зависимости от количества протравленных заготовок N можно вычислить по формуле (7):

теи {N ) = 2Sh pN х 10-9 = 4N, (7)

где: , — площадь стравливаемой поверхности платы, 6400 мм2; h — толщина медной фольги, 35 мкм; р — плотность меди, 8930 кг/м3. Сопоставив значение предельного содержания меди в растворе Меи с полученной зависимостью (рис. 4), можно сделать вывод о том, что в процессе травления 15-й заготовки произошло

насыщение раствора медью, что привело к остановке процесса травления.

Сопоставив выражения (5), (6) и (7), получим зависимость времени травления от содержания меди в растворе (8):

Т(тСи ) = 0.1тСи - 4.02тСи + 70.33, (8)

тси ^ Мси (V, ).

Дополнительно эксперименты показали, что:

- цвет травильного раствора изменяется в значительной мере на начальных стадиях использования подготовленного травильного раствора;

- кислотность травильного раствора снижается в ходе выполнения технологической операции с зависимостью (9), которая имеет коэффициент детерминации 0.92:

рН(г) = 3.33 X ехр{-4.79х10-5 хг), (9)

где рН — водородный показатель раствора; г — время жизни раствора, мин.

ВЫВОДЫ

В целом решение задач управления ходом технологического процесса требует предварительных экспериментальных исследований, связанных с формированием вектора исходного состояния процесса х(г0) и определением как вектора управления г(г), так и значимости влияния отдельных его составляющих на величину ошибки уклонения состояния процесса от идеального.

Формализация задач управления технологическими процессами позволяет унифицировать как ход формирования набора исходных параметров, так и определение структуры алгоритмов управления и регулирования, что позволит организовать переход предприятий от превентивной к предик-тивной модели обслуживания, тем самым снизив вероятность отказа оборудования.

В ходе экспериментального исследования для разработанного опытного образца цифрового участка травления определены составляющие вектора состояния х(г), который содержит как параметры технологического процесса, так и косвенные признаки качества операции. Для учета информации от признаков в контуре управления технологическим процессом получены эмпирические зависимости, связывающие время травления заготовки с содержанием меди в растворе (8),

а также время жизни раствора с водородным показателем (9).

Работа выполнена в рамках гранта Президента Российской Федерации для государственной поддержки молодых российских ученых - кандидатов наук (МК-582.2022.4).

СПИСОК ЛИТЕРАТУРЫ

1. ГОСТ 27.002-2009 "Надежность в технике. Термины и определения".

2. Васильев Ф.В., Ванцов С.В., Медведев А.М., Степанова М.А., Хомутская О.В. Оценка надежности непаяных прессовых соединений по омическому сопротивлению // Надежность и качество сложных систем. 2016. № 3 (15). С. 85-91. DOI: 10.21685/2307-42052016-3-13

3. Vantsov S.V., Vasil'ev F.V., Medvedev A.M., Khomuts-kaya O.V. Influence of nonfunctional contact pads on printed-circuit performance // Russian Engineering Research. 2020. Vol. 40, no. 5. P. 442-445. DOI: 10.3103/S1068798X20050202

4. Leng J., Wang D., Shen W, Li X., Liu Q., Chen X. Digital twins-based smart manufacturing system design in Industry 4.0: A review // Journal of Manufacturing Systems. 2021. Vol. 60. P. 119-137.

DOI: 10.1016/j.jmsy.2021.05.011

5. Проников A.C. Надежность машин. М.: Машиностроение, 1978. 592 с.

6. Жикеев А.А., Алдашева Д.Т., Жаксылыкова А. Автоматизированные системы управления технологическими процессами (АСУТП) // Вестник СевероКазахстанского Университета им. М. Козыбаева. 2022. № 1 (53). С. 157-160. DOI: 10.54596/2309-6977-2022-1157-160

7. Коробков М.А. Исследование аспектов конструирования и изготовления печатных плат со встроенными компонентами // Сборник тезисов работ XLVII Международной молодежной научной конференции. М., 2021. C. 218-219.

URL: https://gagarin.mai.ru/files/2022/abstracts2022.pdf

8. Ванцов С.В., Медведев А.М., Маунг-Маунг З., Хомутская О.В. Анализ процесса сверления отверстий в композиционных материалах оснований печатных плат // Надежность и качество сложных систем. 2016. № 2 (14). С. 37-44.

URL: https://www.elibrary.ru/item.asp?id=26024122

9. Khomutskaya O.V., Medvedev A.M., Korobkov M.A., Van-cov S. V. The method of automated evaluation of the deformation of the printed circuit board // ICOECS 2021: 2021 International Conference on Electrotechnical Complexes and Systems, Ufa. P. 510-512. DOI: 10.1109/IC0ECS52783.2021.9657420

ISSN 0868-5886

NAUCHNOE PRIBOROSTROENIE, 2022, Vol. 32, No. 4, pp. 124-137

10. Коробков М.А., Васильев Ф.В. Применение искусственного интеллекта в управлении технологическими процессами // Тезисы 20-й Международной конференции "Авиация и космонавтика". М., 2021. C. 233234. URL: https://aik.mai.ru/files/abstracts2021.pdf

11. Хомутская О.В. Управление качеством технологического процесса с использованием интеллектуального анализа данных // 14-я Международная конференция "Авиация и космонавтика - 2015", Тезисы. М., 2015. C. 465-467. URL:

https://files.mai.ru/site/conf/aik/2015/sbornik_aik.pdf7refe rer=https%3A%2F%2Fwww.google.com%2F

12. Korobkov M., Vasilyev F., Mozharov V. A comparative analysis of printed circuit boards with surface-mounted and embedded components under natural and forced convection // Micromachines. 2022. Vol. 13, no. 4. Id. 634. DOI: 10.3390/mi13040634

13. Лийн Е.А. Анализ методов оптимизации производственных процессов // Сборник тезисов работ международной молодежной научной конференции XLVIII. М., 2022. C. 261-262.

URL: https://gagarin.mai.ru/files/2022/abstracts2022.pdf

14. Бараковский Ф., Ванцов С., Васильев Ф. Струйный метод получения проводящего рисунка печатной платы // Электроника: Наука, технология, бизнес. 2020. № 3 (194). С. 108-113. DOI: 10.22184/19924178.2020.194.3.108.112

15. Исаев В.В., Коробков М.А. Влияние параметров проектирования и технологических процессов на вероятность появления дефектов на печатных платах // 19-я Международная конференция "Авиация и космонавтика", Тезисы. М., 2020. С. 265-267. URL: https://aik.mai.ru/files/abstracts2020.pdf

16. Маркин С. Как травить платы? // Химия и жизнь. 1990. № 7. С. 76-77. URL: ftp://nozdr.ru/biblio/j/hj/1990/

17. Дзюбаненко А.А. Внедрение M2M на технологическую линию автоматического монтажа печатных плат в условиях цифровизации производства // Сборник материалов V Международной научно-практической конференции. Махачкала, 2021. С. 104-107. URL: https://www.elibrary.ru/item.asp?id=46415984

Московский авиационный институт (национальный исследовательский университет), Москва

Контакты: Хомутская Ольга Владиславовна, khomutskayaov@gmail.com

Материал поступил в редакцию 22.09.2022

MANUFACTURING PROCESS CONTROL TASKS S. V. Vantsov, F. V. Vasiliev, O. V. Khomutskaya, M. A. Korobkov

Moscow Aviation Institute (National Research University), Russia

Obtaining a quality end product is the goal of any manufacturing process control. In the formal description of control problems, this goal allows us to form the boundary conditions for solving the control task. An additional condition for choosing a solution is to ensure the reliability of the process. The starting conditions for solving the control problem are the components of the vector of initial state parameters, determined by the degree of adjustment of the initial parameters and process modes. Understanding the initial and boundary conditions of existence allows us to formulate a generalized formal problem of process control, which consists in finding a vector of control variables that provide a minimum of error, taking into account the influence of a vector of random external variables. The problem is solved by the method of successive approximations. The stages of primary and secondary optimization in the construction of control algorithms are considered. The possibility of reducing the complexity of control tasks is shown by refusing to control individual components of the state vector and transforming control to ensure a minimum of the generalized error measure. In this formulation, the problem of process control can be reduced to the problem of process regulation. A process control scheme is presented. It considers the separability of error minimization for each component of the state vector. The assertions are illustrated by the example of the problem of managing a digital production site for performing the technological operation of etching printed circuit boards. This operation most fully reflects the initial provisions on the continuity of the procedures performed during its implementation and the degradation of the original state vector.

Keywords: manufacturing process, reliability, digitalization of production, control systems, printed circuit boards, printed circuit board etching, manufacturing, industry 4.0

INTRODUCTION

From a physical point of view, the main task solved in technological process control is to obtain a high-quality final product. The solution to this problem is multivariate, often depending on the composition of the equipment at a particular enterprise, the capabilities of technological equipment, and the level of training of staff.

We define the quality of the final product as the state in which it fully satisfies the requirements of the user of this product. For technical objects, the composition of these requirements is formulated in the technical specifications (TSs) for the product. Note that the definition of the concept of quality does not require the product to expand the requirements beyond those fixed in the TSs. Then, in the formal description of the process control problem, the composition of these requirements allows one to form the boundary conditions for the solution.

An additional condition for choosing a variant of solving this problem from an admissible set is to ensure the reliability of the technological process. According to the GOST 27.002-2009 [1] standard, reliability is understood as "the property of availability, affected by the properties of failure-free operation, maintainability, and maintenance support". At the same time, readiness is understood as "the ability of a product to perform the required function under given conditions, assuming that the necessary external resources are provided", and reliability is understood as "the ability of a product to perform the required function in a given time interval under given conditions".

However, the definitions presented in the GOST are focused on the product, as the end item, the properties of which do not change over time, at least until the moment determined, for example, by its assigned resource [2, 3]. A feature of the technological process as a complex technical system, is the presence in the vector of observable but not uncontrollable parameters of the time function system [4]. The practice of implementing technological processes shows that its attribute is degradation over time, i.e., the phenomenon of deterioration of parameters and modes of their flow.

Perhaps the first and most complete definition of the reliability of the technological process was given by A.S. Pronikov [5]: "The reliability of the technological process is the property of ensuring the manufacture of products in a given volume while maintaining the established requirements for its quality over time". The reliability of the technological process is related to the structure of the process, the methods of control used, and the reliability of the execution of individual technological operations.

The key point for determining the quantitative characteristics of reliability is the concept of failure of a technological operation. By the failure of a technological process or technological operation, we mean the loss of the ability to ensure the quality or volume of manufactured products in accordance with the requirements of technical documentation. Failure is a random event that occurs under the influence of a number of random factors.

MAIN PART Methods

The quality parameters of the final object form the boundary conditions when solving the problem of technological process control, and the requirements for process reliability impose additional restrictions on the existence of boundary conditions.

The initial condition in solving the problem of controlling technological processes is the vector of state parameters x(t0), determined by the degree of adjustment of the initial parameters and modes of the process. Understanding the initial and boundary conditions of existence allows us to formulate a generalized formal problem of controlling the technological process [6].

Taking into account the previously considered degradation attribute of the technological process, which determines the vector of the current state of the process x(t), the presence of a vector of random external variables f(t), and that the vector of initial state parameters x(t0) provides ideal process parameters, the control problem can be reduced to minimizing the vector of the difference between the initial and current state of the process, i.e., compensation for the impact of degradation on the quality of the final object:

<0 = {x(0 - x(t)}^ min. (1)

The purpose of solving this problem is to find the vector of control variables r(t), providing the minimum error s(t) while taking the influence of the vector of random external variables f(t) into account.

Due to its obvious complexity, when solving this extremal problem, the method of successive approximations should be used, containing:

- the stage of ideal or primary optimization, which consists in finding the ideal, but, as a rule, unrealizable control vector rud (t);

- the stage of secondary optimization, or optimization of the quality of control, which consists in finding the control vector ronm (t), that is optimal and can

be implemented in practice by the available technical means.

The primary optimization stage highlights the fact that the extremal problem (1) is solved in an idealized formulation, taking into account only the most important properties of the process control system, its main parameters and modes.

Let us explain the solution to the problem of primary optimization of the process control process using the example of the work of a flexible production module (FPM) in machining. As an objective function, we take the performance of the production task in the shortest possible time [7]. The initial state vector x(t0) is characterized by the ideal setting of all the working bodies of the FPM and the availability of all supplies (blanks, tools, cutting fluids, etc.) necessary to complete the production task. The control vector rud (t) is formed on the basis of general ideas about the algorithm of the FPM operation and previous production experience, but does not take into account the vector of random external variables f(t), which include, for example, failures in the performance of kinematic and dynamic processes of the functioning of the working bodies of the FPM, as well as wear phenomena of the processing tool, leading to an increase in the errors in the characteristics of the final product of the technological process [8, 9]. Thus, at the stage of primary optimization, we assume the vector of external disturbances f (t) = 0.

Dependence of the ideal control vector rud (t) on the state vector

rud (t) = Rud {*(0, t}

is what we call the ideal control algorithm.

The ideal control algorithm, as noted earlier, can be considered as the first step to solving the control problem, and it is practically unrealizable. However, its construction is basic for further actions and makes it possible to simplify finding a solution by reducing the control problem to a regulation problem, i.e., keeping the vector of control variables r(t) within

certain allowable limits relative to the vector rud (t). The tolerance limits of the vector r(t) deviations are determined by the limits of the magnitude of the degradation of the parameters and modes of the technological process as well as the requirements for the state of the final product.

The purpose of the secondary optimization step is to obtain a solution that is the least different from the ideal one while also taking into account as fully as possible the restrictions and connections imposed by the properties of the control system. That is, to obtain not only an optimal but also a feasible solution, based

on a given non-zero vector of external disturbances f(t).

In this case, one should take into account the peculiarity of process control, which is that there are mainly two types of variables in the vector, namely:

- Variables represent the characteristics of random events that do not depend on time. An example of such events is failures in the operation of the working bodies of technological equipment and failures in the performance of vector functions r(t).

- Variables, random values of which significantly depend on time and form time series. An example of such variables are characteristics associated, for example, with the wear of the machining tool or the aging of the chemical solutions used during the technological process. The general trend of changing these characteristics over time, as a rule, is known from the practical experience of previous implementations of the technological process. However, it remains essentially random at each moment of the vector action

r(t).

In general, the solution of these optimization problems seems to be rather complicated due to the multi-component nature of the vectors involved in the process. However, it is possible to reduce the multiplicity of components if we use the error criterion (1), that is, to exercise control not for each component individually but to take advantage of the possibility of controlling for the generalized error component, which will allow us to track not the values x(t) of the state vector components during control but their deviations from the ideal defined by the vector xud (t). In this formulation, the problem of process control can be reduced to the problem of regulating the process.

Let us consider the process regulation problem as a specified case of the control problem, to which the latter is reduced under the following simplifying assumptions:

- the ideal state vector xud (t) is assumed to be known, coinciding with the vector x(t0) and corresponding to the ideal control algorithm Rud (t);

- when solving the control problem, the problem of finding the extremum of the generalized indicator of control efficiency is solved:

C = C {Q, P},

where Q = Q{Eud - E}, Eud is the extremum of the

control goal corresponding to the ideal flow of the technological process; E is the value of the control goal indicator corresponding to the actual state of the process; P is the main characteristic of the conditions for the implementation of the control system, which can be its cost, complexity, both in creation and op-

eration, degree of reliability, etc. In general, the characteristic P can serve as a generalized measure of the totality of all these factors.

For the purposes of regulation, this task is divided into n separate or independent tasks of finding the ex-tremum of n performance indicators:

^ = ®t {Xonm (t) - X(t)} = {s } , 1 = \n, (2)

each of which depends on only one component s, (t) of the error vector s(t).

The control vector r(t) does not explicitly depend on the state vector x(t) but on the error vector s(t).

Taking into account the above assumptions, the problem of controlling the technological process can be formulated as follows: assuming that the ideal control vector rud (t) is known, find the control vector r(t) as a function of the error vector s(t), and in the general case also of the perturbation vector f(t), which provides an extremum of n partial performance indicators (2) and satisfies the existing relations formed by the mathematical model of the controlled system and its limitations. In this case, the dependence can be called an algorithm or regulation law, and the vector r(t) can be called a regulation vector:

r(t) = R{e(t), t}. (3)

Thus, the global task of controlling the technological process is reduced to the task of regulating individual factors that determine the state of the process and satisfy the conditions (1).

However, it should be kept in mind that these factors are not completely independent of each other. The relationship between the factors has a number of features:

- The dependency must not be functional. Otherwise, such factors should be combined for process control purposes.

- There may be a correlation between the factors, which leads to the emergence of random relationships between the components of the vector r(t) . The fact of the emergence of such a connection is determined by the fundamentally random nature of the components of the vector f (t).

It should be noted that the state vector x(t) can contain not only parameters but also signs that indirectly reflect the quality of the technological process. The parameters of manufactured products can serve as such signs, among other things. To assess the influence of signs, it is necessary to introduce an additional intelligent processing system into the process control loop, which will allow establishing the relationship between the signs and the parameters to be controlled

(4):

iНе можете найти то, что вам нужно? Попробуйте сервис подбора литературы.

Xnp (t) = X {x(t), t}. (4)

To form such systems, it is advisable to use the methods of machine learning and artificial intelligence [10, 11], as well as empirical and simulation models of processes [12, 13].

The regulation scheme of the technological process, taking into account the separability of minimizing the error for each component of the vector, is shown in Fig. 1.

Fig. 1. Regulation scheme for the minimum error

The error separability mechanism works most effectively for technological processes in which a number of vector f(t) components affect not the entire process but only its parts — technological operations.

Experiments, new patterns

Let's consider the above statements on the example of the problem of managing a digital production site to perform the technological operation of printed circuit board (PCB) etching. This operation, like most chemical operations [14], fully represents the initial assumptions about the analogity of the processes produced during its implementation as well as the degradation of the initial state vector x(t0).

The main parameter affecting the result of the technological operation is the etching time, which in this case determines the regulation vector r(t). However, at present, there are no strict analytical dependences that link the operation time and the state of the pickling solution, especially taking into account its constant and often random degradation. Based on the analysis of the operation [15], the main parameters that affect the pickling time and the time of using the prepared pickling solution in the initial state x(t0) are identified (Fig. 2).

Fig. 2. Parameters that determine the pickling time of the workpiece and the time of use of the prepared pickling solution

Three categories of parameters can be distinguished:

1) the parameters of the pickling solution, which are largely determined by its composition and volume;

2) technological parameters, such as: the position of the workpiece in the pickling bath, the temperature of the solution, the ambient temperature, the method

of transferring the solution to the workpiece, mixing the solution;

3) workpiece parameters, namely: the amount of etched copper, which is the thickness of the copper foil layers, the area of the etched copper and the accuracy class of the manufactured PCB.

It is difficult to assess the direct impact of some parameters, such as mixing time, on the quality of the process, or the assessment will require a large number of time-consuming, labour-intensive operations related to the study of the composition of the solution. Therefore, the state vector x(t) for controlling the etching process contains both parameters that directly affect the quality of the process and features that indirectly reflect the quality of the operation.

The following parameters were selected for control:

- solution temperature;

- the volume of the solution;

- thickness of the workpiece' copper foil;

- copper etched area;

- manufactured PCB accuracy class.

The following features are also included in the controlled state vector:

- solution lifetime, that is, the time elapsed from the moment the solution was prepared to the current moment;

- hue of pickling solution;

- pickling solution acidity;

- portion of etched copper in solution.

It should be noted that parameters such as mixing mode, workpiece orientation, and solution transfer method are not included in the state vector x(t), since they are largely determined by the design of the etching unit and cannot be changed. In particular, the experiments were carried out on a laboratory submersible etching setup with a vertical arrangement of the workpiece without stirring. As an etchant, a solution was used, consisting of 1.5 l of distilled water, 150 ml of ammonium persulfate, and 35 ml of common salt.

The conducted experiments with successive etching of 15 identical workpieces of double-sided foiled fiberglass FR4 with dimensions of 80 x 80 mm and a copper thickness of 35 ^m made it possible to plot the dependence of the workpiece etching time T on the total number of etched workpieces N, which can be used to estimate the degree of degradation of the etching solution (Fig. 3).

Fig. 3. Dependence of the workpiece etching time T on the number of etched workpieces N. The solid line shows the results of the experiment, the dotted line shows the empirical dependence (5)

Using the method of least squares, limiting ourselves to the second-order equation, an empirical dependence was obtained with a determination coefficient of 0.97 (5):

T(N) = 1.65N2 -16.08N + 70.33. (5)

Knowing the copper capacity of the solution [16], we calculate the limiting copper MCu [g] content in the solution using formula (6):

Mcu V ) = jVs = 58.98, (6)

where: j — the solution' copper capacity, 35 g/l; VS — solution volume, 1.88 l.

Since, during the experiment, copper was completely removed from all workpieces except the last one, the amount of copper in the solution mCu [g], depending on the number of pickled workpieces, can be calculated by formula (7)

mCu (N) = 2Sh pN x 10-9 = 4N, (7)

where: S — area of the etched board surface, 6400 mm2; h — copper foil thickness, 35 ^m; p — copper density, 8930 kg/m3. By comparing the value of the limiting copper content in the solution MCu with the obtained dependence (Fig. 4), we can conclude that in the process of pickling the 15th workpiece, the solution was saturated with copper, which led to the stop of the pickling process.

Fig. 4. Dependence of the copper content in the solution mCu on the number of etched workpieces N

By comparing expressions (5), (6) and (7), we obtain the dependence of the etching time on the copper content in solution (8):

T(mcu) = 0.1mCu -4.02mcu + 70.33,

(8)

mcu ^Mcu (Vs).

Additionally, experiments have shown that:

- the color of the pickling solution changes to a large extent at the initial stages of using the prepared pickling solution;

- the acidity of the pickling solution decreases in the course of the technological operation with dependence (9), which has a determination coefficient of 0.92.

pH(t) = 3.33xexp{-4.79x10-5 xt], (9)

where pH is the hydrogen index of the solution; t — solution life time, min.

CONCLUSIONS

In general, solving the problems of controlling the course of a technological process requires preliminary experimental studies related to the formation of the initial state vector of the process x(t0) and the determination of both the control vector r(t) and the significance of the influence of its individual components on the value of the deviation error of the process state from the ideal one.

The formalization of process control tasks allows unification of both the generation of a set of initial parameters and the definition of the structure of control and regulation algorithms, which makes it possible to organize the transition of enterprises from a preventive to a predictive maintenance model, thereby reducing the likelihood of equipment failure.

In the course of an experimental study, for the developed prototype of a digital etching site, the components of the state vector x(t) were determined, containing both the parameters of the technological process and indirect signs of the quality of the operation. To take into account the signs in the process control loop, empirical dependences were obtained that relate the pickling time of the workpiece to the copper content in the solution (8) and also the lifetime of the solution with the hydrogen index (9).

The work was carried out within the framework of the grant of the President of the Russian Federation for state support of young Russian scientists — candidates of sciences (MK-582.2022.4).

REFERENCES

1. GOST 27.002-2009 "Nadezhnost' v technike. Terminy i opredeleniya" [GOST 27.002-2009 Reliability in technology. Terms and Definitions].

URL: https://meganorm.ru/Data/524/52431.pdf (In Russ.).

2. Vasil'ev F.V., Vantsov S.V., Medvedev A.M., Stepano-va M.A., Khomutskaya O.V. [Evaluation of reliability of non-soldered press joints by ohmic resistance]. Nadezhnost' i kachestvo slozhnykh sistem [Reliability & quality of complex systems], 2016, no. 3 (15), pp. 85-91. DOI: 10.21685/2307-4205-2016-3-13 (In Russ.).

3. Vantsov S.V., Vasil'ev F.V., Medvedev A.M., Khomuts-kaya O.V. Influence of nonfunctional contact pads on printed-circuit performance. Russian Engineering Research, 2020, vol. 40, no. 5, pp. 442-445. DOI: 10.3103/S1068798X20050202

4. Leng J., Wang D., Shen W., Li X., Liu Q., Chen X. Digital twins-based smart manufacturing system design in Industry 4.0: A review. Journal of Manufacturing Systems, 2021, vol. 60, pp. 119-137.

DOI: 10.1016/j.jmsy.2021.05.011

5. Pronikov A.S. Nadezhnost' mashin [Machine reliability]. Moscow, Mashinostroenie Publ., 1978. 592 p. (In Russ.).

6. Zhikeev A.A., Aldasheva D.T., Zhaksylykova A. [Automated control system of technological process (ACS TP)]. Vestnik of M. Kozybayev North Kazakhstan University [Bulletin of the North Kazakhstan University named after M. Kozybaev], 2022, no. 1 (53), pp. 157-160. DOI: 10.54596/2309-6977-2022-1-157-160 (In Kazakh.).

7. Korobkov M.A. [Explore design and manufacturing aspects of printed circuit boards with embedded components]. Sbornik tezisov rabot XLVII Mezhdunarodnoi mo-lodezhnoi nauchnoi konferentsii [Proc. XLVII International Youth Scientific Conference], Moscow, 2021, pp. 218-219. (In Russ.).

URL: https://gagarin.mai.ru/files/2022/abstracts2022.pdf

8. Vantsov S.V., Medvedev A.M., Maung-Maung Z., Khomutskaya O.V. [Analysis of hole drilling process in composite materials of printed circuit board bases]. Nadezhnost' i kachestvo slozhnykh sistem [Reliability & quality of complex systems], 2016, no. 2 (14), pp. 37-44.

URL: https://www.elibrary.ru/item.asp?id=26024122 (In Russ.).

9. Khomutskaya O.V., Medvedev A.M., Korobkov M.A., Vancov S.V. The method of automated evaluation of the deformation of the printed circuit board. ICOECS 2021: 2021 International Conference on Electrotechnical Complexes and Systems, Ufa, pp. 510-512. DOI: 10.1109/IC0ECS52783.2021.9657420

10. Korobkov M.A., Vasilyev F.V. [Application of artificial intelligence in manufacturing process control]. Tezisy 20-i Mezhdunarodnoi konferentsii "Aviatsiya i kosmonavtika" [20th International Conference "Aviation and Cosmonautics" (AviaSpace-2021)], Moscow, 2021, pp. 233-234. URL: https://aik.mai.ru/files/abstracts2021.pdf (In Russ.).

11. Khomutskaya O.V. [Technical process quality management using data mining]. 14-ya Mezhdunarodnaya konfe-rentsiya "Aviatsiya i kosmonavtika - 2015", Tezisy [14 th International Conference "Aviation and Cosmonautics -2015"], Moscow, 2015, pp. 465-467. URL: https://files.mai.ru/site/conf/aik/2015/sbornik_aik.pdf7refe rer=https%3A%2F%2Fwww.google.com%2F (In Russ.).

12. Korobkov M., Vasilyev F., Mozharov V. A comparative analysis of printed circuit boards with surface-mounted and embedded components under natural and forced convection. Micromachines, 2022, vol. 13, no. 4, Id. 634. DOI: 10.3390/mi13040634

13. Liyn E.A. [Process optimization analysis]. Sbornik tezisov rabot mezhdunarodnoi molodezhnoi nauchnoi konferentsii XLVIII [Proc. of international youth scientific conference XLVIII], Moscow, 2022, pp. 261-262.

URL: https://gagarin.mai.ru/files/2022/abstracts2022.pdf (In Russ.).

14. Barakovsky F., Vantsov S., Vasiliev F. [Spray method for PCB conductive pattern formation]. Ehlektronika: Nauka, tekhnologiya, biznes [Electronics: Science, Technology, Business], 2020, no. 3 (194), pp. 108-113. DOI: 10.22184/1992-4178.2020.194.3.108.112 (In Russ.).

15. Isaev V.V., Korobkov M.A. [Influence of design parameters and technological processes on probability of defects on PCB]. 19-ya Mezhdunarodnaya konferentsiya "Aviatsiya i kosmonavtika", Tezisy [19th International Conference "Aviation and Cosmonautics" (AviaSpace-2020)],

Moscow, 2020, pp. 265-267. (In Russ.). URL: https://aik.mai.ru/files/abstracts2020.pdf

16. Markin S. [How to poison boards?]. Khimiya i zhizn' [Chemistry and Life], 1990, no. 7, pp. 76-77. URL: ftp://nozdr.ru/biblio/j/hj/1990/ (In Russ.).

17. Dzyubanenko A.A. [Implementation of M2M on the process line of automatic installation of printed circuit

boards in conditions of digitalization of production]. Cbornik materialov V Mezhdunarodnoi nauchno-prakticheskoi konferentsii [Proc. of V international scientific and practical conference], Makhachkala, 2021, pp. 104-107. (In Russ.).

URL : https://www.elibrary.ru/item.asp?id=46415984

Contacts: Khomutskaya Ol'ga Vladislavovna, khomutskayaov@gmail.com

Article received by the editorial office on 22.09.2022

i Надоели баннеры? Вы всегда можете отключить рекламу.