М. Хаджийски. Тенденции в развитието на индустриалната автоматика в ерата на изкуствения интелект
- П. Петков. Жордановата канонична форма – митове и реалност
- В. Стефанова-Стоянова, К. Стоянов. Същност, свойства и предимства на интелигентните разпределителни енeргийни мрежи със съхранение на електрическа енергия
- Н. Делийски, Н. Тричков, Д. Ангелски, Ж. Гочев, Н. Тумбаркова. Изчисляване на температурното поле в обли дървени материали съхранявани на открит склад
- В. Димитров, К. Спасов, С. Сярова. Анализ на обхвата на концептуален модел за споделен център за операции по киберсигурността на индустриални управляващи системи
- Р. Хрисчев. Информационната сигурност в системите за планиране на корпоративни ресурси (ERP)
- Н. Николов. 50-годишен юбилей на катедра „Автоматизация на производството“ при Техническия университет – Варна
- Доц. инж. Веселин Спиридонов
Key Words: Industrial automation; artificial intelligence; trends; distributed control systems; cyber security.
Abstract. In the paper is presented analysis of the main now days challenges in the field of business, basic industrial technologies and ecology and the potential of the advanced control technologies as an important component for solving them. As dominant are considered the Industrial Automation (IA) methods and implementation of Artificial Intelligence (AI) achievements in industrial automation in order to meet automatic and operational management as well equipment reliability and cybersecurity. The historical development of industrial automation at different levels in modern Distributed Control Systems (DCS) is considered. Special attention is paid to the rapid development of the basic control level through PLC, PAC and EPIC controllers and expansion of their technological capabilities for control and communication with the higher hierarchical levels of DCS. The reasons why AI is becoming a leading paradigm in modern times are analyzed. The historically formed connections and mutual influence between the control theory and artificial intelligence are discussed. The main directions in which the fastest and most effective ways of introduction the AI’s methods and techniques in industrial automation are under consideration. The problems of suitability for solving the tasks of industrial automation with the methods of AI depending on the amount of available data are treated. It is specifically focused on one of the key points of advanced industrial automation – creation of mathematical models and their maintenance with the necessary accuracy due to the evolution of the environment and the elements of the control system itself. The integration of classical control methods and AI-based approaches are considered in two case studies: (i) process control of cement production with emphasis on the clinker kiln and (ii) control of the regime of heat treatment of wood in an autoclave with a focus on combining analytical modeling of heat transfer processes and data-driven sub-optimal control under conditions of parametric uncertainties. The study examines the effectiveness of the application of artificial intelligence methods to expand the scope of traditional industrial automation to include subsystems for reliability and cybersecurity. The reliability of the technological equipment is ensured by modules for achieving fault tolerance model-based diagnostics and technical maintenance based on the assessment of the state of the system. Cyber security is guaranteed by elements that provide protection against cyberattacks and reduce operational uncertainty. As an example for condition-based maintenance is considered an integrated control system of Peirce-Smith converter from the metallurgical industry. It is concluded that the methods of artificial intelligence give a new inspiration to the future development of industrial automation. These methods allow to achieve new functional capabilities for technological and operational control, reliability and cyber security compared to traditional means of industrial automation. The integration of artificial intelligence in industrial control systems can be successful only if the combination of domain knowledge with the achievements of advanced industrial automation and the new methods, techniques and tools of artificial intelligence will be realized in full degree.